AI in Ophthalmology | Startup Landscape
28 likes9,746 views
The document provides an overview of the business and startup landscape of artificial intelligence in the ophthalmology sector, emphasizing the complexities of integrating AI into healthcare. It discusses advancements in ophthalmic imaging, particularly the effectiveness of SD-OCT over traditional methods for diagnosing retinal diseases and highlights emerging imaging technologies like handheld OCT and new biomarkers. Additionally, it notes the growth potential of the optical coherence tomography (OCT) market, driven by technological innovation and increasing demand in developing countries.
![Typical volumetric medical formats
DICOM NIFTI .NII
brainder.org
http://nipy.org/nibabel/gettingstarted.html
http://people.cas.sc.edu/rorden/dicom/index.html
http://dicom.nema.org/
NEMA standard PS3, and as ISO standard 12052:2006
Practically outdated but still used
The Nifti format has rapidly replaced the Analyze in neuroimaging
research, being adopted as the default format by some of the most
widespread public domain software packages, as, FSL [12], SPM [13],
and AFNI [14]. The format is supported by many viewers and image
analysis software like 3D Slicer [15], ImageJ [16], and OsiriX, as well as
other emerging software like R [17] and Nibabel [18], besides various
conversion utilities.
An update version of the standard, the Nifti-2, developed to manage
larger data set has been defined in the 2011. This new version encode
each of the dimensions of an image matrix with a 64-bit integer instead
of a 16-bit as in the Nifti-1, eliminating the restriction of having a size
limit of 32,767. This updated version maintains almost all the
characteristics of the Nifti-1 but, as reserve for some header fields the
double precision, comes with a header of 544 bytes [19].
doi:10.1007/s10278-013-9657-9
doi:10.1007/s10278-013-9657-9
This project aims to offer easy access to Deep Learning for
segmentation of structures of interest in biomedical 3D scans. It
is a system that allows the easy creation of a 3D Convolutional
Neural Network, which can be trained to detect and segment
structures if corresponding ground truth labels are provided for
training. The system processes NIFTI images, making its use
straightforward for many biomedical tasks.
https://github.com/Kamnitsask/deepmedic
Transition
to NIFTI](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-12-320.jpg)
![Picture Archiving and Communications System (PACS)
PACS exit lessons learned report (PDF, 320Kb)
Picture archiving and communication system
A picture archiving and communication system (PACS)
is a medical imaging technology which provides
economical storage and convenient access to images
from multiple modalities (source machine types).[1]
Electronic images and reports are transmitted digitally
via PACS; this eliminates the need to manually file,
retrieve, or transport film jackets. The universal format
for PACS image storage and transfer is DICOM. Non-
image data, such as scanned documents, may be
incorporated using consumer industry standard
formats like PDF (Portable Document Format), once
encapsulated in DICOM.
A PACS consists of four major components: The
imaging modalities such as X-rayplain film (PF),
computed tomography (CT) and MRI, a secured
network for the transmission of patient information,
workstations for interpreting and reviewing images,
and archives for the storage and retrieval of images
and reports. Combined with available and emerging
web technology, PACS has the ability to deliver timely
and efficient access to images, interpretations, and
related data. PACS breaks down the physical and time
barriers associated with traditional film-based image
retrieval, distribution, and display.
wikipedia.org](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-13-320.jpg)
![Innovative OCTs Horizon 2020 project
New €4.9 million European project aiming to create a tiny and low-cost system for
optical coherence tomography (OCT). Now a team led by Wolfgang Drexler from
the Medical University of Vienna is aiming to shrink the core technology to no
more than the size of a coin, primarily to diagnose eye diseases including
diabetic retinopathy and glaucoma.
According to the official abstract describing the four-year project, the plan is to exploit recent advances in silicon photonics and
CMOS electronics for a healthcare application that is expected to grow in significance as the global population ages and diabetes
becomes more prevalent. The key difference compared with datacoms applications is that in biophotonics a visible-range light source
is typically required and most silicon-on-insulator (SOI) waveguides are not compatible with wavelengths shorter than 1.1 µm.
“To this end, a novel CMOS-compatible, low-loss silicon nitride waveguide based [on a] photonic integrated circuit (PIC)
technology platform will be developed in “OCTCHIP” (short for ophthalmic OCT on a chip, project began at the start of the year
2016) and directly applied in the field of OCT for ophthalmology,” states the team, adding: “The PIC technology developed in
OCTCHIP will make a new generation of OCT systems possible with step-changes in size and cost beyond state-of-the-art. The
monolithic integration of silicon nitride optical waveguides, silicon photodiodes and electronics combined with the hybrid
integration of a III-V laser source will enable a compact, low-cost and maintenance-free solution.”
http://optics.org/news/7/6/19
cordis.europa.eu/project/rcn/199593
Supermarket OCT?
Drexler himself even envisages a time in the future when the ultra-compact technology could feature on
supermarket shelves and be purchased by consumers for self-diagnosis. “State-of-the-art OCT technology has
its limitations: it is bulky, the size of a desktop and quite expensive, costing anything in the region of €100,000 per
unit,” he said in an announcement about the project from Photonics21. “It can detect abnormalities but at the
present moment, compact, cost-effective versions that can be used outside of hospitals and in private practice in a
hand-held mode do not exist.”
However, with diabetic retinopathy now thought to be the cause of 200 million cases of blindness worldwide,
including 60 million people in Europe, there is a clear case for a lower-cost, point-of-care diagnostic platform. That
said, the retina is an extremely complex part of the body, composed of more than ten layers of tissue despite being
only 0.25 mm thick, and is very difficult to access. Drexler said that the core component under development via the
project will be no larger than the size of a 1 cent coin. “It will reduce costs and is maintenance-free,” he added.
“OCTCHIP fosters widespread use to visualize and quantify the retina in more definition, so we can
diagnose diseases better, quicker, and cheaper.”
The researcher even thinks that the core technology may extend to use in battery-powered capsules that patients
could swallow for gastrointestinal diagnosis. “Perhaps in the future this will be available in supermarkets, for
self-diagnosis" he said.
jeppix.eu
Photonic Integration in Healthcare: ‘Towards optical
coherence tomography on a chip’ by Jeroen Kalkman](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-25-320.jpg)
![Proprietary formatS Vendor-specific OCT files
zeiss.com
FromHuang etal.(2013):
“Scans were obtained with certified photographers to
minimize the OCT data acquisition artifacts [15], [20].
The data samples were saved in the Heidelberg
proprietary .e2e format. They were exported from a
Heidelberg Heyex review software (version 5.1) in .vol
format and converted to the DICOM (Digital Imaging
and Communication in Medicine) [21] OPT (ophthalmic
tomography) format using a custom application built in
MATLAB. “
These plugins interpret raw binary files exported from Heidelberg
Spectralis Viewing Software. They successfully import both 8-bit SLO
and 32-bit SD-OCT images, retaining pixel scale (optical and SD-
OCT), segmentation data, and B-scan position relative to the SLO
image (included in v1.1+). In addition to single B-scan SD-OCT
images, the plug-in also opens multiple B-scan SD-OCT images as a
stack, enabling 3-D reconstruction, analysis, and modeling. The plug-
in is compatible with Spectralis Viewing Module exporting raw data in
HSF-OCT-### format. Compatability has been tested with HSF-
OCT-101, 102, and 103
http://dx.doi.org/10.1016/j.exer.2010.10.009
Heidelberg Engineering Spectralis OCT RAW
data (.vol ending): Circular scans and Optic
Nerve Head centered volumes are supported
www5.cs.fau.de .. octseg/
Problem:
Vendors have their own
proprietary image management
platforms and file formats for
lock-in purposes that might
pose problems for automated
AI pipelines.
uocte Reverse-engineered file
readers in C++ by Paul Rosenthal et al.
UOCTML, Eyetec, Heidelberg, NIDEK,
Topcon, Zeiss](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-37-320.jpg)
![Healthcare cutting EHR costs
By Peter Groen, and Roger A. Maduro | July 3, 2013: “Cost transparency is obviously a big issue in the healthcare industry. … , or the costs associated with
acquiring and implementing an Electronic Health Record (EHR) system for a hospital - Why are all these costs often carefully hidden? Is there something
special about the healthcare industry that says – "Let's not talk about how much things really cost."
In a Forbes Magazine article written by Zina Moukheiber and titled "
The Staggering Cost Of An Epic Electronic Health Record Might Not Be Worth It," Moukheiber
states that "not-for-profit hospitals or healthcare provider organizations will need every penny
of those taxpayers’ dollars [in meaninful use reimbursements], but they won’t cover anywhere
near the staggering cost of an Epic EHR." Moukheiber points out "Duke University Health
System will shell out $700 million, so will Boston's Partners Healthcare; University of
California in San Francisco will pay $150 million," adding that "Customers, such as New
Hampshire’s Dartmouth-Hitchcock Medical Center are feeling the pinch. DHMC which
implemented Epic last year at a cost of $80 million, expects a weak operating performance
in 2012, partly because of expenses related to Epic."
According to a recent article from KTOO News, "the owners of Juneau, Alaska-based
Bartlett Regional Hospital voted recently to break their contract with Cerner for a new
system, rather than incur the expense of implementation." Bartlett is one of the latest U.S.
hospital to run into financial trouble while installing a new EHR system.
Implementation of the new EHR would have cost $7.37 million, plus an annual
maintenance fee of $1.155 million. Bartlett Regional Hospital is a small 48 bed acute
care hospital, with a 12-bed Adult Behavioral Health Facility, and a 16-bed Chemical
Dependency Recovery Center.
A recent FierceEMR article, titled "EHR transition may be financially risky for hospitals"
reports that "Henry Ford Health System's investment in an Epic system was a major factor
in its 15 percent decrease in net income from 2011 to 2012. Wake Forest Baptist Medical
Center also reported that its adoption of a new Epic system caused it to suffer
unanticipated losses and business cycle disruptions."
In an article titled "The Costly Darkside Of EMR Implementations," Dr. Edmund Billings
points out that the West Virginia Department of Health & Human Resources spent $9
million to acquire and implement the VistA open source EHR system across its 7
hospitals that operate 776 beds in total. Contrast this with the $92 million spent by West
Virginia University (WVU) to acquire and implement a proprietary EHR system in its 7
hospitals that operate 526 beds.
In a recent article in Open Health News, Dr. Billings details the catastrophic financial
situation facing the 9-hospital Maine Health system in the State of Maine as a result of
their Epic Implementation. While the often cited cost figure for this Epic Implementation is
$160 million, Dr. Billing does a financial analysis that shows that the real cost is going to
be almost $370 million over the next five years. In a comment to the article, our own
Roger A. Maduro points that even this figure is a vast underestimate. Maduro points out
that Billings missed the cost of licensing upgrades imposed by proprietary EHR vendors
every 18 to 36 months. Adding those "licence upgrade" costs and projecting over six years
(to account for two "upgrade" cycles), the projected cost for MaineHealth's EHR is $600
million. It should be noted that the Maine Health hospital system is very similar to West
Virginia's, which, as noted in the note above, implemented the open source VistA
system at a cost of only $9 million.
When it comes to health data interoperability, there are two schools of thought. The first believes
that since there are so many systems already in place, we should agree on the exchange format (and
recently the APIs) and convert the proprietary data into that format as needed. This is the preferred
approach today when faced with a large installed base of megasuite solutions. The focus here is on
the applications.
The second approach is standardizing the health data first, building the new systems on top and
avoiding the interoperability issues altogether. Sure, sometimes this is not possible. But when looking
at projects spending billions of dollars, Euros or pounds on new systems, I firmly believe it is a much
wiser path to take. For a detailed discussion on the two approaches, see Wolandscat blog.
It would also prevent vendor lock-in by making applications easier to replace. Due to the costs of
data migration, introducing a new application often means starting fresh with a nearly empty
database. This is true even of well funded projects like the ongoing US DoD EHR implementation. I
tried to show why in healthcare, data is more important than applications. Now let’s see what
defines the data layer (More details are available here):
●
First, a set of common, open application programming interfaces (APIs). They need to cover
the breadth of health data required by health applications, including systems as complex as EHRs.
They need to be vendor-neutral and technology neutral as well.
●
Second, common, open data models that are multi-lingual, comprehensive, semantically correct
and modeled by clinicians. They need to be governed, enable versioning and querying. This is best
achieved if they support multi-level modeling, where an underlying maximal data set can be
constrained and customized for a specific use case at a higher level of the model.
●
Third, support for clinical terminologies ensuring the computability of health data.
●
Finally, all of the above specifications must be published and freely available for use.
http://www.openhealthnews.com/story/2016-05-04/postmodern-ehr-data-layer
http://www.openhealthnews.com/articles/2013/ehr-systems-cost-transparency-healthcare-industry](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-42-320.jpg)
![Healthcare general model
The NHS Structure (UK)
The NHS is divided into two sections: primary and
secondary care. Primary care is the first point of contact
for most people and is delivered by a wide range of
independent contractors, including GPs, dentists,
pharmacists and optometrists.
Secondary care is known as acute healthcare and can be
either elective care or emergency care. Elective care
means planned specialist medical care or surgery,
usually following referral from a primary or community
health professional such as a GP.
http://dx.doi.org/10.1007%2Fs11606-010-1464-0
Primary healthcare results in better health outcomes, reduced health disparities and
lower spending, including on avoidable emergency room visits and hospital care. With
that being said, primary care physicians are an important component in ensuring that the
healthcare system as a whole is sustainable. However, despite their importance to the
healthcare system, the primary care position has suffered in terms of its prestige in
part due to the differences in salary when compared to doctors that decide to specialize.
In a 2010 national study of physician wages conducted by the UC Davis Health System
found that specialists are paid as much as 52 percent more than primary care physicians,
even though primary care physicians see far more patients.[10]
Primary care physicians earn $60.48 per hour; specialists on average earn $88.34.
[10] A follow up study conducted by the UC Davis Health System found that earnings
over the course of the careers of primary care physicians averaged as much as $2.8
million less than the earnings of their specialist colleagues.[11]This discrepancy in pay has
potentially made primary care a less attractive choice for medical school graduates.
https://en.wikipedia.org/wiki/Primary_care#United_States
The Continuum of Healthcare
OJIN: The Online Journal of Issues in Nursing
Rising to the Challenge of Health Care Reform With Entrepreneurial and Intrapreneurial Nursing Initiatives
Anne Wilson, PhD, MN, BN, FRCNA; Nancy Whitaker, BPsych (Hon); Deirdre Whitford, PhD](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-52-320.jpg)
![RETINAL AI Companies DeepMind #1
Say you succeed and create a super intelligence. What happens next? Do you
donate the technology to the United Nations?
I think it should be. We’ve talked about this a lot. Actually Eric Schmidt [executive
chairman of Alphabet, Google’s parent] has mentioned this. We’ve talked to him. We
think that AI has to be used for the benefit of everyone.
https://www.bloomberg.com/features/2016-demis-hassabis-interview-issue/](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-73-320.jpg)
![Retinal Parkinson's Disease prediction
http://dx.doi.org/10.1186/s40478-016-0346-z
The occurrence of ocular manifestations in many neurodegenerative
diseases, including PD, emphasizes the strong connection between the
brain and the retina. Indeed, PD patients often suffer from visual
symptoms such as reduced visual acuity, low contrast sensitivity and
disturbed colour vision [9, 10]. In particular, recent findings have
highlighted the retina as a potential biomarker for PD. Thinning of the
inner retinal layer has shown to be an early event in patients with early
PD, where early disease was defined as diagnosis within 2.5 years with
an average disease stage of 2 (Hoehn–Yahr scale) [11, 12]. The
severity of PD symptoms correlates with RNFL thickness [13, 14].
Functional changes in the retina have also been recorded in early PD
(grade 1–1.5 Hoehn–Yahr scale, diagnosed within 3 years) [15] with
some suggestion that RGCs are involved early [16].
The aim of this study is to determine, using a rotenone-induced rodent
model of PD, whether firstly, retinal imaging can identify quantifiable
changes including retinal layer thickness (OCT) and RGC apoptotic
counts (DARC) in the natural history of the disease; and secondly, if
these changes can be used as surrogate biomarkers, and applied to
assessing a potential PD treatment strategy.
A modified Heidelberg HRA-OCT Spectralis (Heidelberg Engineering) was used for OCT imaging
[33]. A posterior pole scanning protocol, centred on the rat optic nerve head, was obtained in
each treatment group with TruTrack® software (Heidelberg Engineering), to ensure pixel to pixel
correspondence over time. Mean retinal thickness was determined using the HEYEX® thickness
map analysis (Heidelberg Engineering).
Affiliated with: UCL Institute of Ophthalmology, University College London
http://dx.doi.org/10.1371/journal.pone.0085718
http://dx.doi.org/10.1007%2Fs10633-015-9503-0](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-81-320.jpg)
![Regulatory affairs FDA Status of AI
dispatch.com
itnonline.com
Simply put, the FDA wants to keep computers on the periphery of diagnosis – and
even then is not comfortable with the use of diagnostic software. The agency has
cautiously defended the right of physicians – and only physicians – to practice
medicine. The FDAabsolutely will not step over that line, which pretty much negates
the chance that deep learning algorithms with even modest autonomy will pass FDA
review.
http://www.cio.com/article/3024715/government/luddites-clobber-ai-says-advocate.html
How will FDA regulate predictive clinical software as predictive analytics
pushes medical frontiers?
FDA & FTC Law, Healthcare Law
michaelhcohen.com/2015/06
Predictive analytics will push the frontiers of clinical care; the question is whether
FDA regulation will promote or stifle innovation in the name of consumer
protection. FDA regulation in this area is a moving target. Let’s see what we know
so far.
Fortune: Carl, why did you decide to move from real estate into healthcare and has it panned out
like you thought it would?
Carl Berg: I have been in the venture capital business for 40 years but I never once touched biotech
because I was concerned about the risk associated with government approval – it’s bad enough when
you’re doing venture capital but adding one more equation, like getting approval from the FDA [Food and
Drug Administration] makes it a lot harder. But about eight years ago I said, instead of getting into a whole
bunch of small companies, I am in a position now where I can do something really big in a hope that it
changes the world. So that’s what motivated me, and then I met with Niven, and that’s what got it started.
http://fortune.com/2015/04/16/cancer-cure-artificial-intelligence/](https://image.slidesharecdn.com/retinalailandscape-160819021000/85/AI-in-Ophthalmology-Startup-Landscape-84-320.jpg)
AI in Ophthalmology | Startup Landscape
- 1. Artificial intelligence in ophthalmology Business and startup landscape Petteri Teikari, PhD http://petteri-teikari.com/ version Mon 12 September 2016
- 2. Introduction ● Shallow introduction for ophthalmologic / healthcare market in a dense format – Mainly meant for people with machine learning background with little knowledge to the healthcare sector – Best to be read from a tablet (or similar device with easy zoom in/out), do not work very well for project despite the slide format. ● Purpose of the presentation to illustrate the “non-technical” complexities related to healthcare ventures which not be that obvious from start. – Not sufficient necessarily to have technically sophisticated AI solutions if no one is willing to pay for it, and you don't know understand how to bring value to the clinicians.
- 5. What modality is the best for diagnosis? From “old school methods” (visual field, fundus photograph and SD-OCT), the SD-OCT seems to offer clearly the best diagnostic capability Results:” Among the four specialists, the interobserver agreement across the three diagnostic tests was poor for VF and photos, with kappa ( ) values of 0.13 and 0.16, respectively, and moderate for OCT,κ with value of 0.40. Using panel consensus as reference standard, OCTκ had the highest discriminative ability, with an area under the curve (AUC) of 0.99 (95% 0.96–1.0) compared to photograph AUC 0.85 (95% 0.73–0.96) and VF AUC 0.86 (95% 0.76–0.96), suggestive of closer performance to that of a group of glaucoma specialists.” Blumberg et al. (2016) Retinal Diseases Signs In One Picture ophthnotes.com/retinal-diseases-signs-in-one-picture + Bone spicule pigments (BSP) in Retinitis pigmentosa (RP), Chorioretinal Atrophy, Congenital hypertrophy of the retinal pigment epithelium (CHRPE), Asteroid hyalosis, Haemangioma, Choroidal neovascularization (CNV), Retinoschisis, etc. For more detailed analysis see →
- 6. Future of OCT and retinal biomarkers ● From Schmidt-Erfurth et al. (2016): “The therapeutic efficacy of VEGF inhibition in combination with the potential of OCT-based quantitative biomarkers to guide individualized treatment may shift the medical need from CNV treatment towards other and/or additional treatment modalities. Future therapeutic approaches will likely focus on early and/or disease-modifying interventions aiming to protect the functional and structural integrity of the morphologic complex that is primarily affected in AMD, i.e. the choriocapillary - RPE – photoreceptor unit. Obviously, new biomarkers tailored towards early detection of the specific changes in this functional unit will be required as well as follow-up features defining the optimal therapeutic goal during extended therapy, i.e. life-long in neovascular AMD. Three novel additions to the OCT armamentarium are particularly promising in their capability to identify the biomarkers of the future:” Polarization-sensitive OCT OCT angiography Adaptiveopticsimaging “this modality is particularly appropriate to highlight early features during the pathophysiological development of neovascular AMD Findings from studies using adaptive optics implied that decreased photoreceptor function in early AMD may be possible, suggesting that eyes with pseudodrusen appearance may experience decreased retinal (particularly scotopic) function in AMD independent of CNV or RPE atrophy.” “...the specific patterns of RPE plasticity including RPE atrophy, hypertrophy, and migration can be assessed and quantified). Moreover, polarization-sensitiv e OCT allows precise quantification of RPE-driven disease at the early stage of drusen”, “Angiographic OCT with its potential to capture choriocapillary, RPE, and neuroretinal fetures provides novel types of biomarkers identifying disease pathophysiology rather than late consecutive features during advanced neovascular AMD.”” Schlanitz et al. (2011) zmpbmt.meduniwien.ac.at See also Leitgeb et al. (2014) Zayit-Soudry et al. (2013)
- 7. Future Multispectral Imaging http://dx.doi.org/10.1038/eye.2011.202 Absorption spectra for the major absorbing elements of the eye. Note that some of the spectra change with relatively small changes in wavelength. Maximizing the differential visibility requires utilizing small spectral slices. Melanin is the dominant absorber beyond 600 nm. Zimmer et al. (2014) Zimmer et al. (2014) The Annidis RHA™ system combines advanced multispectral imaging (MSI) technology with multi-image software processing for early detection of ocular pathologies such as age related macular degeneration, diabetic retinopathy and glaucoma. http://www.annidis.com/page/technology
- 8. OCT towards handheld devices http://dx.doi.org/10.1364/BOE.5.000293 Cited by 5410.1038/nphoton.2016.141 http://dx.doi.org/10.1364/OE.24.013365 Here, we report the design and operation of a handheld probe that can perform both scanning laser ophthalmoscopy and optical coherence tomography of the parafoveal photoreceptor structure in infants and children without the need for adaptive optics. The probe, featuring a compact optical design weighing only 94 g, was able to quantify packing densities of parafoveal cone photoreceptors and visualize cross-sectional photoreceptor substructure in children with ages ranging from 14 months to 12 years. https://aran.library.nuigalway.ie/handle/10379/5481
- 9. Fundus handheld https://www.jedmed.com/products/hd-digital-scope-system http://www.epipole.com/ http://dx.doi.org/10.1167%2Fiovs.12-10449, Cited by 22 articles http://dx.doi.org/10.1089/tmj.2015.0068
- 10. OCT Example report from Zeiss Cirrus
- 11. OCT Device Comparison Comparison of images obtained with 3 different spectral-domain OCT devices (Topcon 3D OCT- 1000, Zeiss Cirrus, Heidelberg Spectralis) of both eyes of the same patient with early AMD changes taken just minutes apart. Comparison of images obtained with 3 different spectral-domain OCTs (Heidelberg Spectralis, Optovue RTVue, Topcon 3D OCT-1000) and with 1 time-domain OCT (Zeiss Stratus) of both eyes of the same patient with a history of central serous chorioretinopathy in both eyes. The same set of images as shown above in pseudocolor. Comparison of horizontal B-scan images and 3D images of a patient with neovascular age-related macular degeneration obtained with Heidelberg Spectralis, Zeiss Cirrus, Topcon 3D OCT-1000. Spectral-domain Optical Coherence Tomography: A Real-world Comparison IRENE A. BARBAZETTO, MD · SANDRINE A. ZWEIFEL, MD · MICHAEL ENGELBERT, MD, PhD · K. BAILEY FREUND, MD · JASON S. SLAKTER, MD retinalphysician.com reviewofophthalmology.com
- 12. Typical volumetric medical formats DICOM NIFTI .NII brainder.org http://nipy.org/nibabel/gettingstarted.html http://people.cas.sc.edu/rorden/dicom/index.html http://dicom.nema.org/ NEMA standard PS3, and as ISO standard 12052:2006 Practically outdated but still used The Nifti format has rapidly replaced the Analyze in neuroimaging research, being adopted as the default format by some of the most widespread public domain software packages, as, FSL [12], SPM [13], and AFNI [14]. The format is supported by many viewers and image analysis software like 3D Slicer [15], ImageJ [16], and OsiriX, as well as other emerging software like R [17] and Nibabel [18], besides various conversion utilities. An update version of the standard, the Nifti-2, developed to manage larger data set has been defined in the 2011. This new version encode each of the dimensions of an image matrix with a 64-bit integer instead of a 16-bit as in the Nifti-1, eliminating the restriction of having a size limit of 32,767. This updated version maintains almost all the characteristics of the Nifti-1 but, as reserve for some header fields the double precision, comes with a header of 544 bytes [19]. doi:10.1007/s10278-013-9657-9 doi:10.1007/s10278-013-9657-9 This project aims to offer easy access to Deep Learning for segmentation of structures of interest in biomedical 3D scans. It is a system that allows the easy creation of a 3D Convolutional Neural Network, which can be trained to detect and segment structures if corresponding ground truth labels are provided for training. The system processes NIFTI images, making its use straightforward for many biomedical tasks. https://github.com/Kamnitsask/deepmedic Transition to NIFTI
- 13. Picture Archiving and Communications System (PACS) PACS exit lessons learned report (PDF, 320Kb) Picture archiving and communication system A picture archiving and communication system (PACS) is a medical imaging technology which provides economical storage and convenient access to images from multiple modalities (source machine types).[1] Electronic images and reports are transmitted digitally via PACS; this eliminates the need to manually file, retrieve, or transport film jackets. The universal format for PACS image storage and transfer is DICOM. Non- image data, such as scanned documents, may be incorporated using consumer industry standard formats like PDF (Portable Document Format), once encapsulated in DICOM. A PACS consists of four major components: The imaging modalities such as X-rayplain film (PF), computed tomography (CT) and MRI, a secured network for the transmission of patient information, workstations for interpreting and reviewing images, and archives for the storage and retrieval of images and reports. Combined with available and emerging web technology, PACS has the ability to deliver timely and efficient access to images, interpretations, and related data. PACS breaks down the physical and time barriers associated with traditional film-based image retrieval, distribution, and display. wikipedia.org
- 14. Functional marker VISUAL FIELD #1
- 15. Functional marker VISUAL FIELD #2 “Patterns of early glaucomatous visual field loss and their evolution over time” http://iovs.arvojournals.org/article.aspx?articleid=2333021 research-innovation.ed.ac. uk
- 16. Functional marker VISUAL FIELD #3 8 September 2016 Visionary technology Innovative optical concept offers simple, affordable, fast glaucoma screening test The new Viewi optical concept developed by Cambridge Consultants shows how it could be possible for patients to monitor any effect on their vision in the comfort of their own homes. At the moment, patients typically have an annual optometrist or hospital check-up using a visual field analyser. Flashing lights at varying points of the visual field test sensitivity – with the patient pressing a button each time they see a light. The novel Viewi technology performs the same test but at a fraction of the cost – around £20 rather than £20,000 for the clinical device. The innovative Viewi concept has been hailed as an important advance by optics expert Chris Dainty, a professor at University College London Institute of Ophthalmology and Moorfields Eye Hospital. Home > Health & Fitness > http://www.digitaltrends.com/health-fitness/viewi-glaucoma-test/ Chris Dainty, a professor at University College London Institute of Ophthalmology and Moorfields Eye Hospital, expressed interest in wider applications of Viewi, noting that it “could provide a valuable early warning system for people at risk of developing glaucoma, as well as patients who need to monitor the effects of the disease on their vision.” Dainty concluded, “It could also make the static perimetry test accessible to more patients in developing countries, where expensive clinical equipment and trained professionals are often in short supply.” Read more: http://www.digitaltrends.com/health-fitness/viewi-glaucoma-test/#ixzz4K4Z4km4Q Follow us: @digitaltrends on Twitter | digitaltrendsftw on Facebook scholar.google.co.uk
- 17. Functional marker PUPILLOMETRY http://iovs.arvojournals.org/article.aspx?articleid=2467290 doi:10.1016/j.visres.2012.07.019 http://www.medicalexpo.com/cat/ophthalmology/pupillometers-pupillographs-F-144.htm l The 3D model of the pupillometer based on a smartphone (left) and the configuration of the measurement system based on the prototype design (right) .http://dx.doi.org/10.3807/JOSK.2013.17.3.249 businesswire.comdoi:10.1364/AO.53.000H27 BOSTON — A simple, quick test performed with a pupillometer appears to be an effective screening tool for acute mild traumatic brain injury, US Army investigators report.
- 18. Imaging Market
- 19. OCT Market #1 Peer-reviewed publications focused on OCT from 1991 to 2015. The commercial release of products is often a catalyst for publications as clearly seen in ophthalmology and cardiology. (Note that while some specialty-focused OCT products were introduced earlier than indicated here, none made large-scale sales, and some companies that launched those products no longer exist.) Revenue from the sale of OCT systems (including biometry) in 2015 is estimated at ~$750 million/year (www.octnews.org), and ~100 companies now supply OCT systems or components. Cumulatively since the 1996 release of the first commercial product, total OCT system revenue (not including components) has likely exceeded ~$5.2 billion. http://www.laserfocusworld.com/articles/print/volume-52/issue-06/ Despite the saturation of the OCT ophthalmology market in developed countries, the OCT market for healthcare and life science is still expected to grow from around € 500M in 2013 to around € 1b in 2019. Strong demand from new biomedical applications, continuous development of innovative technologies, strong demand from developing countries will drive the OCT market growth. tematys.fr The optical coherence tomography segment has been further divided into three sub-segments, namely, time- domain OCT, Fourier-domain OCT and full-field OCT. In 2012, the optical coherence tomography (OCT) segment held the largest share (76.5%) in the global optical imaging market followed by hyperspectral imaging (HSI). Among the three sub-types of the OCT, Fourier-domain OCT held the largest share. Photoacoustic tomography is expected to be the fastest growing segment in the global optical imaging market during the forecast period 2014 to 2020. Some of the major driving factors for the growth of this market are increasing prevalence of various disorders related to different anatomical areas, shift in lifestyle, aging population, and increasing awareness and acceptance of several optical imaging technologies. Changing lifestyle such as, sedentary sitting work style, longer working durations and refraining from exercises and physical activities are leading to a number of health problems among today’s population all over the world. Several diseases such as diabetes, hypertension etc. are also becoming highly prevalent across the world and even in low to medium income countries. http://www.transparencymarketresearch.com/optical-imaging-market.html
- 20. OCT Market #2 The medical markets (ophthalmic, surgical, dental, and therapeutic) for lasers continue to grow, evidenced by continuing expansion in the BiOS Symposium at SPIE Photonics West 2013 as well as the financial statements of companies like Carl Zeiss Meditec (Dublin, CA), which saw revenues grow 13.6% for the nine-month period up to 3Q12 and celebrated the installation of its 10,000th Cirrus HD-OCToptical coherence tomography system in November 2012 (introduced in 1996). laserfocusworld.com/articles/print/volume-49/issue-01 http://www.sweptlaser.com/OCT-market The OCT market was born with the introduction of the first ophthalmic OCT system in the '90s. Today, OCT products are shipping in opthalmology, dermatology, cardivascular, peripheral artery disease, and pearl inspection. Companies are readying produjcts for dental, esophageal, and OCT-guided surgery. Several other areas are in development. The market is just over $1B in 2012, and it is expected to grow by 18–30% per year for the foreseeable future. Additional general information: Some Historical Statistics of Academic Public ations in the Field of Optical Coherence Tomo graphy Optical Coherence Tomography News (October 13, 2012) Some Historical Statistics on Companies in t he Market of Optical Coherence Tomography Sy stems Optical Coherence Tomography News (November 3, 2012)
- 21. OCT Market #3 Today, commercial OCT systems are used regularly as an imaging standard in ophthalmology, cardiology, dermatology, and general research. NinePoint’s Advanced OCT technology, licensed from the Wellman Center for Photomedicine at Massachusetts General Hospital, initially focuses on esophageal and general imaging of tissue microstructure. http://www.ninepointmedical.com/ 12/03/2014 - Posted by Lee Dubay - Associate Editor, BioOptics World Lux Research (Boston, MA) reports that the clinical optical imaging market—led by optical coherence tomography (OCT) —will rise to $2 billion in 2020, more than doubling from 2012. With a market share of over 60 percent, OCT is expected to retain its dominant position as it builds on ophthalmology applications while expanding into cardiology,oncology, and gastroenterology. Other technologies such as near-infrared spectroscopy (NIRS) and photoacoustic tomography (PAT) also hold high potential for growth, while applications such as real-time optical biopsies and surgical guidance will grow into multi-billion-dollar segments over the long term. Lux Research analysts examined 16 technology developers in the optical imaging industry, and evaluated them on the Lux Innovation Grid. Among its findings: ● Leaders ready to deliver to market. Mela Sciences, Bioptigen, and Infraredx are the leaders in novel clinical optical imaging technologies, rated “Dominant” on the Lux Innovation Grid. Each has achieved FDA and CE clearances, allowing it to market its products in the United States and Europe. ● Cost-cutters show potential. Typical OCT systems, the size of a medical cart, cost on average between $80,000 and $250,000. Compact Imaging, rated "High Potential," is working to make these devices more affordable by turning toward solid-state components instead of discrete optical elements. ● Multimodal systems are long-term targets. While investors looking to enter the OCT market can use acquisitions to gain immediate market share, their longer-term focus should be on firms developing multimodal systems such as PAT. These systems combine the most attractive features of constituent technologies—for instance, the penetration depth of ultrasound with resolution of optical systems—and potentially address much larger markets. The report, titled "Advancing Clinical Imaging Beyond the Existing Standard of Care: Evaluating New Optical Imaging Modalities," is available for download at https://portal.luxresearchinc.com/research/report_excerpt/18129. Optical coherence tomography expected to propel clinical optical imaging market to $2B in 2020
- 22. OCT Market #4 Research laboratories serve as the first proving ground for commercial OCT platforms. In a recent Optics Letters paper, professor of biomedical engineering Xingde Li and colleagues Wu Yuan and Jessica Mavadia-Shukla at Johns Hopkins University (Baltimore, Md., USA) investigated how to optimize the operational conditions for an ultrahigh-resolution (UHR) OCT system in the spectral domain using a supercontinuum source for endoscopic imaging. A supercontinuum source enables a broader spectral bandwidth, which supports higher resolution imaging. In their work, Li’s team used a SuperK Extreme low-noise supercontinuum (SC) source, a compact tunable laser from NKT Photonics (Birkerød, Denmark). They were able to optimize the broad bandwidth (approximately 246 nm) around a central wavelength of 800 nm and use an appropriate power from the SC source to achieve shot-noise- limited operation in an SC-based UHR-OCT system. “Because the cost of OCT hardware has been so high,” says Kemp, “it can be hard to get commercial traction. Ophthalmology has a large installed base that enables a reasonable profit margin. In cardiology, the profit comes from the annual sale of thousands of disposable catheters that go along with the OCT system. But in other markets, the cost must be low at the component level to reduce the initial cost barrier.” Axsun decreases the cost by integrating very small, lower-power laser beams with Ethernet- based data acquisition electronics. Another shifting paradigm is the size of OCT platforms. The typical large, bulky, roll-around console is expensive in terms of infrastructure and workflow complexity needed to obtain an OCT image, according to Kemp. “The system has to be rolled in, plugged in, booted up, and it takes up a large area in the office,” said Kemp. Incorporating their more integrated, low-power, laptop- and tablet- ready laser beam component, Axsun demonstrated a one-off SS- OCT system at the Photonics West 2016 conference in February that is briefcase-sized, battery-powered and pointing to the future. The demo featured a 100-kHz swept laser (customizable to either 1060 or 1310 nm), a hand-held probe, and scan rates of more than 125 frames/s. Just as new types of ultrasound equipment are becoming more shoebox-sized with the same level of image quality, smaller OCT platforms seemed destined to follow. SD-OCT is a solid technology with room for market growth, but it has its challenges. The SD-OCT platform uses a wide wavelength band to reflect off the different layers in the eye. The deeper parts of the eye are imaged with the higher-frequency interference fringes, which are more difficult to detect, producing a lower- signal-to-noise image of the deeper structures. In the past year or two, according to Nate Kemp, market development manager at Axsun Technologies (Billerica, Mass., USA), another OCT technology has begun to make waves, shifting the attention away from SD-OCT: swept-source OCT (SS-OCT). Companies are in a race to launch SS-OCT platforms that can create high-resolution images that are deeper, wider and more detailed than previous OCT platforms. In comparison to SD-OCT, SS-OCT uses a tunable laser to rapidly scan at rates of up to 100,000 A-scans/s sequentially through a range of wavelengths to assemble the B-scan cross-sectional image, often in a fraction of a second. The optic nerve head, sclera, vitreous and choroid structures can be imaged all in one fast, detailed scan. While SD-OCT systems are typically based on a central laser wavelength of 840 or 850 nm, swept-source tunable laser systems are centered around 1 µm (for example, at 1030, 1050 or 1060 nm), which penetrates tissue more deeply than 850 nm. The longer wavelengths can penetrate opaque media such as cataracts more effectively, and provide the most accurate biometry measurements possible for the best selection of an intraocular lens (IOL) before cataract surgery. SWEPT-SOURCE OCTCOST and SIZE of OCTSUPERCONTINUUM SOURCE
- 23. OCT Market #5: What next? Even higher laser wavelengths for better image quality? For example in three-photon microscopy 1,700 nm seems to be quite optimal for brain tissue penetration and image quality The spectral response of oxygenated hemoglobin, deoxygenated hemoglobin, and water as a function of wavelength. The red highlighted area indicates the biological optical window where adsorption due to the body is at a minimum. Figure from Doane and Burda (2012) Wavelength-dependent attenuation length in brain tissue and measured laser characteristics. a, Attenuation spectrum of a tissue model based on Mie scattering and water absorption, showing the absorption length of water (la, blue dashed line), the scattering length of mouse brain cortex (ls, red dashed-dotted line), and the combined effective attenuation length (le, green solid line). The red stars indicate the attenuation lengths reported for mouse cortex in vivo from previous work. The figure hows that the optimum wavelength window (for three-photon microscopy) in terms of tissue penetration is near 1,700 nm when both tissue scattering and absorption are considered. Figure from Horton et al. (2013, Cited by 304 articles).https://github.com/petteriTeikari/spectralSeparability/wiki R.F. Spaide et al. “What’s next in laser and OCT?” Rev. Opthal. (March 2013). Direct comparison of spectral-domain and swept-source OCT in the measurement of choroidal thickness in normal eyes Comparison of choroidal thicknesses using swept source and spectral domain optical coherence tomography in diseased and normal eyes Choroidal thickness maps from spectral domain and swept source optical coherence tomography: algorithmic versus ground truth annotation Yamanaka et al. (2016) "Optical coherence microscopy in 1700 nm spectral band for high-resolution label-free deep-tissue imaging" Ishida and Nishizawa (2012) "Quantitative comparison of contrast and imaging depth of ultrahigh-resolution optical coherence tomography images in 800–1700 nm wavelength region"
- 25. Innovative OCTs Horizon 2020 project New €4.9 million European project aiming to create a tiny and low-cost system for optical coherence tomography (OCT). Now a team led by Wolfgang Drexler from the Medical University of Vienna is aiming to shrink the core technology to no more than the size of a coin, primarily to diagnose eye diseases including diabetic retinopathy and glaucoma. According to the official abstract describing the four-year project, the plan is to exploit recent advances in silicon photonics and CMOS electronics for a healthcare application that is expected to grow in significance as the global population ages and diabetes becomes more prevalent. The key difference compared with datacoms applications is that in biophotonics a visible-range light source is typically required and most silicon-on-insulator (SOI) waveguides are not compatible with wavelengths shorter than 1.1 µm. “To this end, a novel CMOS-compatible, low-loss silicon nitride waveguide based [on a] photonic integrated circuit (PIC) technology platform will be developed in “OCTCHIP” (short for ophthalmic OCT on a chip, project began at the start of the year 2016) and directly applied in the field of OCT for ophthalmology,” states the team, adding: “The PIC technology developed in OCTCHIP will make a new generation of OCT systems possible with step-changes in size and cost beyond state-of-the-art. The monolithic integration of silicon nitride optical waveguides, silicon photodiodes and electronics combined with the hybrid integration of a III-V laser source will enable a compact, low-cost and maintenance-free solution.” http://optics.org/news/7/6/19 cordis.europa.eu/project/rcn/199593 Supermarket OCT? Drexler himself even envisages a time in the future when the ultra-compact technology could feature on supermarket shelves and be purchased by consumers for self-diagnosis. “State-of-the-art OCT technology has its limitations: it is bulky, the size of a desktop and quite expensive, costing anything in the region of €100,000 per unit,” he said in an announcement about the project from Photonics21. “It can detect abnormalities but at the present moment, compact, cost-effective versions that can be used outside of hospitals and in private practice in a hand-held mode do not exist.” However, with diabetic retinopathy now thought to be the cause of 200 million cases of blindness worldwide, including 60 million people in Europe, there is a clear case for a lower-cost, point-of-care diagnostic platform. That said, the retina is an extremely complex part of the body, composed of more than ten layers of tissue despite being only 0.25 mm thick, and is very difficult to access. Drexler said that the core component under development via the project will be no larger than the size of a 1 cent coin. “It will reduce costs and is maintenance-free,” he added. “OCTCHIP fosters widespread use to visualize and quantify the retina in more definition, so we can diagnose diseases better, quicker, and cheaper.” The researcher even thinks that the core technology may extend to use in battery-powered capsules that patients could swallow for gastrointestinal diagnosis. “Perhaps in the future this will be available in supermarkets, for self-diagnosis" he said. jeppix.eu Photonic Integration in Healthcare: ‘Towards optical coherence tomography on a chip’ by Jeroen Kalkman
- 26. Innovative OCTs Compact Imaging "Our lenses and beam splitters are all either consumer level items or have a direct analogue in the consumer domain," commented Bogue. "That means we can make a low-cost system, while the design itself allows that system to fit into an inherently small footprint with low operating power. Potential applications for a genuinely low-cost compact OCT platform could include a number of uses in health monitoring and non- destructive testing." IPIC's involvement has been key to the company's success in miniaturizing the technology from its initial bench-top implementation. The first-generation prototypes now measure about 50 x 50 mm, and Bogue indicated that his road map towards even smaller versions was on course to conclude early in 2016. https://www.crunchbase.com/organization/fp-technology#/entity https://compactimaging.com http://optics.org/news/6/9/56
- 27. “Innovative” OCTs BioOptigen Leica→ http://www.leica-microsystems.com/products/optical-coherence-tomography-oct/
- 28. “Established” OCT Topcon Description Topcon’s 3D imaging and analysis functions provide invaluable pathological confirmation of progression. With the addition of noise reducing algorithms and Infra red/3D tracking technology, the 3D OCT-2000 provides you with extremely detailed OCT images. Fundus Images Unique to the Topcon OCT series is its integrated retinal photography function, which is based upon its highly successful non-mydriatic fundus camera. An interchangeable (future proof) 16.2mp digital camera acquires highly detailed images using a sub one millisecond flash at the point of OCT capture or stand alone fundus photography if required. Glaucoma One of many modules within Topcon’s Fastmap software; The Glaucoma module allows fully automated disc topography, normative database comparison and total progression analysis (trend analysis) through various screening options. Complemented by Ganglion cell analysis and anterior chamber angle measurements the glaucoma module is a comprehensive screening tool. Anterior segment scanning By combining both OCT scan technology with traditional photographic imaging a variety of analysis functions and scan protocols allow for the detection and treatment of many corneal conditions. Full corneal thickness topography and automated central thickness values are complemented by corneal curvature topography along with high resolution imaging. http://www.topcon-medical.co.uk/uk/products/32-3d-oct-2000-optical-coherence-tomography.html#description http://www.topcon-medical.co.uk/uk/products/75-imagenet-i-base.html#description
- 29. TopCon Innovative OCT→ Topcon selected to join IBM Watson Health global medical imaging collaborative IBM (NYSE: IBM) announced on June 22nd that it is forming the Watson Health medical imaging collaborative, naming Japanese ophthalmic device manufacturer Topcon Corporation (TYO: 7732) among the sixteen partners, which includes academic medical centers, health systems, ambulatory radiology providers and imaging technology companies. As part of this global effort, foundational members will engage IBM's "augmented intelligence" platform, called Watson, to extract insight from a variety of structured and unstructured data sources, such as medical imaging data, electronic health records, radiology and pathology reports, lab results, doctors' progress notes, medical journals, clinical care guidelines and published outcomes studies. Watson, a cognitive computing system, understands natural language, reasons and learns over time. The inclusion of Topcon as one of the collaborating members is a testament to the company's commitment to integrating medical devices, big data analytics and cognitive computing into its products and services. It will leverage Topcon's large family of imaging devices to facilitate and optimize the Watson training process and accelerate the development of products and services that improve the understanding, diagnosis, and treatment of eye disease, ultimately improving patient care. Foundational members of the collaborative include Agfa HealthCare, Anne Arundel Medical Center, Baptist Health South Florida, Eastern Virginia Medical School, Hologic, ifa systems AG, inoveon, Radiology Associates of South Florida, Sentara Healthcare, Sheridan Healthcare, Topcon, UC Sand Diego Health, University of Miami Health System, University of Vermont Health Network, vRad, and Merge Healthcare (an IBM company). http://fortune.com/2016/06/22/ibm-watson-health-imaging-collaboration/ diagnosticimaging.com/pacs-and-informatics
- 30. Established OCT Zeiss http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2917043 / Zeiss ForumZeiss OCT
- 31. Established OCT Heidelberg Multi-Modality Diagnostic Imaging of the Eye Precision, Detail, and Versatility Multi-modality imaging with SPECTRALIS® is helping drive the development of novel therapeutics and changing the course of patient management in eye care. Using an upgradeable platform approach, SPECTRALIS has enhanced the role of spectral domain OCT by integrating it with confocal scanning laser ophthalmoscopy (cSLO). The combination of these two technologies has enabled new imaging capabilities, such as TruTrack™ active eye tracking, and BluePeak™ blue laser autofluorescence, providing clinicians with unique views of the structure and function of the eye. https://www.heidelbergengineering.com/us/products/spectralis-models/ HEYEX Networking Solutions Heidelberg Eye Explorer, or “HEYEX™”, is the heart of all Heidelberg Engineering instruments, such as SPECTRALIS®, HEP and HRT. The central patient database and uniform user interface allow the review of images nearly anywhere in the practice without compromising on tools or image quality. HEYEX has been designed to integrate into a variety of workflows. It enables connectivity to EHR and PACS systems via a DICOM Interface or via custom interfaces, providing flexible solutions from a solo practice to a large clinic environment. https://www.heidelbergengineering.com/us/products/heyex-networking-solutions/ Based on the website, Heidelberg indeed seems to be “established” especially compared to sleek visual branding of Zeiss
- 33. Established OCT OptoVue http://www.optovue.com/ http://optovueacademy.com/ bloomberg.com/Research
- 34. Established OCT MOptim http://www.moptim.com/en/index.php e.g. Moptim Mocean3000/3000 Plus Intel i5 quad-core CPU, 4G DDR RAM Windows 7 32-bit (3GB of RAM) NVIDIA Quadro K600, 1 GB, 0.34 TFLOPS 1 TB HDD METHODS Seventy-two normal subjects were included. Every subject underwent CMT measurement twice using one of two SD-OCT (OSE-2000, Moptim, Shenzhen, China & 3-D OCT-1000, Topcon, Tokyo, Japan) instruments. http://dx.doi.org/10.3980%2Fj.issn.2222-3959.2012.03.20 linkedin.com/company/moptim
- 36. Electronic Medical records Depending on what you want to do, you may want to have an API available for the image management / electronic medical record if it is not possible to interface the vendor-specific software directly
- 37. Proprietary formatS Vendor-specific OCT files zeiss.com FromHuang etal.(2013): “Scans were obtained with certified photographers to minimize the OCT data acquisition artifacts [15], [20]. The data samples were saved in the Heidelberg proprietary .e2e format. They were exported from a Heidelberg Heyex review software (version 5.1) in .vol format and converted to the DICOM (Digital Imaging and Communication in Medicine) [21] OPT (ophthalmic tomography) format using a custom application built in MATLAB. “ These plugins interpret raw binary files exported from Heidelberg Spectralis Viewing Software. They successfully import both 8-bit SLO and 32-bit SD-OCT images, retaining pixel scale (optical and SD- OCT), segmentation data, and B-scan position relative to the SLO image (included in v1.1+). In addition to single B-scan SD-OCT images, the plug-in also opens multiple B-scan SD-OCT images as a stack, enabling 3-D reconstruction, analysis, and modeling. The plug- in is compatible with Spectralis Viewing Module exporting raw data in HSF-OCT-### format. Compatability has been tested with HSF- OCT-101, 102, and 103 http://dx.doi.org/10.1016/j.exer.2010.10.009 Heidelberg Engineering Spectralis OCT RAW data (.vol ending): Circular scans and Optic Nerve Head centered volumes are supported www5.cs.fau.de .. octseg/ Problem: Vendors have their own proprietary image management platforms and file formats for lock-in purposes that might pose problems for automated AI pipelines. uocte Reverse-engineered file readers in C++ by Paul Rosenthal et al. UOCTML, Eyetec, Heidelberg, NIDEK, Topcon, Zeiss
- 38. Electronic medical/Health record Market Thanks to the White House's stimulus-era initiative to bring the health care industry into the digital age, her company has grown into the country's leading vendor of software in the $9.3 billion electronic health records (EHR) sector. Epic pulled in $1.8 billion in 2014 and is expanding at a rate of about 1,000 new employees a year. But instead of ushering in a new age of secure and easily accessible medical files, Epic has helped create a fragmented system that leaves doctors unable to trade information across practices or hospitals. Epic is not the only barrier to a seamless medical records system. Thanks to legislative maneuvering by former Rep. Ron Paul (R-Texas) in 1999, the federal government can't fund any sort of system with unique health care identification numbers. There have been other signs that the government has had enough with Epic. A massive contract to overhaul and modernize the Department of Defense's health records, worth up to $9 billion over the next 18 years, had long been expected to go to the company. But in July, the Pentagon announced it had instead chosen Cerner to provide the software to serve 9.5 million DOD beneficiaries. Jonathan Woodson, the assistant secretary of defense for health affairs, stressed that the Pentagon believed it was "very important to have a highly integrated system that is portable. The private sector has to position itself to be more interoperable." motherjones.com/politics/2015/10/ Global EHR Market http://blog.capterra.com/top-5-ehr-trends-for-2016-2/ 1. More EHRs are moving to the cloud 2. More EHRs are providing patient portals 3. Telemedicine will blow up 4. EHRs will (finally) go mobile 5. Big data will reveal more connections in US, forbes.com Room for a lot of disruption here! Some have been openly critical of the tactics used by EHR vendors and their ability to “lock in” customers often in a sequential‒ ‒ fashion as they churn through different EHR vendor solutions. In their eBook ‒ Hacking Healthcare authors Fred Trotter and‒ David Uhlman offered this sober assessment: A vendor creates a system so specific, non standardized and esoteric that one and only one vendor can possibly support or‒ maintain it. The vendor then pursues aggressive sales strategies, often combined with low initial pricing. Once fully established in the customer site, the vendor can drain the customer with ever increasing maintenance and upgrade fees.‒ forbes.com
- 39. Towards better EMR Open-sourcing? Writing for the New England Journal of Medicine (here), Dr. David Blumenthal (former Director of ONC) cited four significant obstacles to the rapid adoption and effective use of EHR’s by the healthcare community. 1) The fee-for-service payment system in the United States does not financially reward the improved quality and efficiency that health information technology makes possible. 2) EHRs in particular are complex products that are difficult to evaluate and understand. The market’s diverse offerings vary enormously in capability and usability, and new products are burgeoning. 3) The ability to effectively transfer electronic health information between different information systems in various institutions and practices is underdeveloped in the United States at this time. 4) Still a fourth problem inhibiting the adoption and use of health information technology is concern about the privacy and security of digital health information. Entire new industries have arisen using personal health information for purposes that were never anticipated by existing privacy statutes, and these uses are not currently regulated. That first obstacle highlights how the healthcare IT vendor landscape mirrors the very healthcare industry it’s designed to serve. Both have been optimized around revenue and profits ‒ not safety and quality. The other three obstacles highlight the significant difference between simply deploying one of the 500+ EHR solutions (very high across the healthcare industry) and it’s effective use around coordinated and continuous patient care with better outcomes. forbes.com/sites/danmunro/2014/07/27 The Top 5 Free and Open Source EMR Software Products Published July 7th, 2016 by Cathy Reisenwitz in EMR 1) Practice Fusion 2) OpenMRS 3) iSALUS EHR 4) VistA 5) FreeMED softwareadvice.com
- 40. Towards better EMR Efficiency issues Medscape Medical News Survey: EHRs Have Taken Over, Except for Hearts and Minds Robert Lowes, August 30, 2016 EHRs Not Reliable for Legal Cases, Experts Say Designed to maximize billing, EHRs have features that lend themselves to gaming and concealment, making them equivalent to 'hearsay' in court. The paper cites a 2012 letter from Attorney General Eric Holder and then–Secretary of Health and Human Services Kathleen Sebelius warning hospitals not to manipulate electronic records for the purposes of getting improper payments from Medicare. This letter followed press reports about the widespread practice of cutting and pasting past EHR notes into current ones. J Am Med Inform Assoc. 2005 Sep-Oct; 12(5): 505–516. doi: 10.1197/jamia.M1700, Cited by 710 articles Lise Poissant, PhD, Jennifer Pereira, MSc, Robyn Tamblyn, PhD, and Yuko Kawasumi, MSc http://dx.doi.org/10.13063%2F2327-9214.1176 http://dx.doi.org/10.1177/1460458214534091 http://www.ncbi.nlm.nih.gov/pubmed/17299922 The experts also observe that not all EHRs are capable of performing audits that show when entries were made and whether and when they were changed. They cite a US Department of Health and Human Services Office of Inspector General report that stated that 44% of hospitals could delete their EHR audit logs. Neither Dr Leone nor George L. Paul, an Arizona attorney who has written a book on digital evidence, could say whether the reliability of EHRs has been a factor in any malpractice case to date. But Paul noted that the widespread use of EHRs is still new and that it can take several years for an issue such as this to be litigated, adjudicated, and appealed to higher courts Ave Maria Law Rev. 2014;12:257-289. Full text
- 41. US Centers for Medicare and Medicaid Services (CMS) and the Office of the National Coordinator for Health Information Technology (ONC) published the EHR Incentive Program and Health IT Certification Program final rules. The final rules emphasize simplicity and flexibility, encourage greater interoperability, and allow stakeholders additional time to phase in the required changes. Health care providers, developers, and consumers may benefit from the changes included in the final rules: deloitte.com/us/en/pages/life-sciences-and-health-care Towards better EMR US Issues As a part of the American Recovery and Reinvestment Act, all public and private healthcare providers and other eligible professionals (EP) were required to adopt and demonstrate “meaningful use” of electronic medical records (EMR) by January 1, 2014 in order to maintain their existing Medicaid and Medicare reimbursement levels. Since that date, the use of electronic medical and health records has spread worldwide and shown its many benefits to health organizations everywhere. “Meaningful use” of electronic health records (EHR), as defined by HealthIT.gov, consists of using digital medical and health records to achieve the following: ● Improve quality, safety, efficiency, and reduce health disparities ● Engage patients and family ● Improve care coordination, and population and public health ● Maintain privacy and security of patient health information The American Recovery and Reinvestment Act also included financial incentives for healthcare providers who prove meaningful use of electronic health records (EHR). EHR is not only a more comprehensive patient history than electronic medical records (EMR), the latter of which contains a patient’s medical history from just one practice, but was also the end-goal of the federal mandate. Penalties were also issued to those healthcare organizations that were non-compliant. For example, EP’s who didn’t implement EMR/EHR systems and demonstrate their meaningful use by 2015 experienced a 1% reduction in Medicare reimbursements. Not surprisingly, the EMR/EHR mandate spurred significant growth in health informatics, an interdisciplinary field of study that merges information technology and healthcare. Healthcare professionals with the skills and knowledge necessary to develop, implement, and manage IT software and applications in a medical environment are already in high demand, and the field is expected to experience continual growth.
- 42. Healthcare cutting EHR costs By Peter Groen, and Roger A. Maduro | July 3, 2013: “Cost transparency is obviously a big issue in the healthcare industry. … , or the costs associated with acquiring and implementing an Electronic Health Record (EHR) system for a hospital - Why are all these costs often carefully hidden? Is there something special about the healthcare industry that says – "Let's not talk about how much things really cost." In a Forbes Magazine article written by Zina Moukheiber and titled " The Staggering Cost Of An Epic Electronic Health Record Might Not Be Worth It," Moukheiber states that "not-for-profit hospitals or healthcare provider organizations will need every penny of those taxpayers’ dollars [in meaninful use reimbursements], but they won’t cover anywhere near the staggering cost of an Epic EHR." Moukheiber points out "Duke University Health System will shell out $700 million, so will Boston's Partners Healthcare; University of California in San Francisco will pay $150 million," adding that "Customers, such as New Hampshire’s Dartmouth-Hitchcock Medical Center are feeling the pinch. DHMC which implemented Epic last year at a cost of $80 million, expects a weak operating performance in 2012, partly because of expenses related to Epic." According to a recent article from KTOO News, "the owners of Juneau, Alaska-based Bartlett Regional Hospital voted recently to break their contract with Cerner for a new system, rather than incur the expense of implementation." Bartlett is one of the latest U.S. hospital to run into financial trouble while installing a new EHR system. Implementation of the new EHR would have cost $7.37 million, plus an annual maintenance fee of $1.155 million. Bartlett Regional Hospital is a small 48 bed acute care hospital, with a 12-bed Adult Behavioral Health Facility, and a 16-bed Chemical Dependency Recovery Center. A recent FierceEMR article, titled "EHR transition may be financially risky for hospitals" reports that "Henry Ford Health System's investment in an Epic system was a major factor in its 15 percent decrease in net income from 2011 to 2012. Wake Forest Baptist Medical Center also reported that its adoption of a new Epic system caused it to suffer unanticipated losses and business cycle disruptions." In an article titled "The Costly Darkside Of EMR Implementations," Dr. Edmund Billings points out that the West Virginia Department of Health & Human Resources spent $9 million to acquire and implement the VistA open source EHR system across its 7 hospitals that operate 776 beds in total. Contrast this with the $92 million spent by West Virginia University (WVU) to acquire and implement a proprietary EHR system in its 7 hospitals that operate 526 beds. In a recent article in Open Health News, Dr. Billings details the catastrophic financial situation facing the 9-hospital Maine Health system in the State of Maine as a result of their Epic Implementation. While the often cited cost figure for this Epic Implementation is $160 million, Dr. Billing does a financial analysis that shows that the real cost is going to be almost $370 million over the next five years. In a comment to the article, our own Roger A. Maduro points that even this figure is a vast underestimate. Maduro points out that Billings missed the cost of licensing upgrades imposed by proprietary EHR vendors every 18 to 36 months. Adding those "licence upgrade" costs and projecting over six years (to account for two "upgrade" cycles), the projected cost for MaineHealth's EHR is $600 million. It should be noted that the Maine Health hospital system is very similar to West Virginia's, which, as noted in the note above, implemented the open source VistA system at a cost of only $9 million. When it comes to health data interoperability, there are two schools of thought. The first believes that since there are so many systems already in place, we should agree on the exchange format (and recently the APIs) and convert the proprietary data into that format as needed. This is the preferred approach today when faced with a large installed base of megasuite solutions. The focus here is on the applications. The second approach is standardizing the health data first, building the new systems on top and avoiding the interoperability issues altogether. Sure, sometimes this is not possible. But when looking at projects spending billions of dollars, Euros or pounds on new systems, I firmly believe it is a much wiser path to take. For a detailed discussion on the two approaches, see Wolandscat blog. It would also prevent vendor lock-in by making applications easier to replace. Due to the costs of data migration, introducing a new application often means starting fresh with a nearly empty database. This is true even of well funded projects like the ongoing US DoD EHR implementation. I tried to show why in healthcare, data is more important than applications. Now let’s see what defines the data layer (More details are available here): ● First, a set of common, open application programming interfaces (APIs). They need to cover the breadth of health data required by health applications, including systems as complex as EHRs. They need to be vendor-neutral and technology neutral as well. ● Second, common, open data models that are multi-lingual, comprehensive, semantically correct and modeled by clinicians. They need to be governed, enable versioning and querying. This is best achieved if they support multi-level modeling, where an underlying maximal data set can be constrained and customized for a specific use case at a higher level of the model. ● Third, support for clinical terminologies ensuring the computability of health data. ● Finally, all of the above specifications must be published and freely available for use. http://www.openhealthnews.com/story/2016-05-04/postmodern-ehr-data-layer http://www.openhealthnews.com/articles/2013/ehr-systems-cost-transparency-healthcare-industry
- 43. Practice Fusion fortune.com/2015/08/19 http://www.practicefusion.com/electronic-health-record-ehr/ San Francisco-based Practice Fusion, the largest U.S. cloud-based electronic health record (EHR) platform for doctors and patients, on June 18th launched a native version of its EHR optimized for iOS and Android based tablets. Data from the recent AmericanEHR study in partnership with the American College of Physicians (ACP) demonstrated that Practice Fusion is the clear leader for solo and practices with 1-3 medical providers. The study found that for solo providers, Practice Fusion has 40% greater market share than eClinicalWorks, the next largest competitor in this category. According to Healthit.gov, a federal website that provides comprehensive up-to date information regarding EHRs, the average upfront cost of implementation of an EHR is $33,000 per provider, along with an annual maintenance fee of $4,000. For many smaller practices, this may not be feasible—thus making the case for implementing Practice Fusion, a completely free, Meaningful Use certified EHR. forbes.com softwareadvice.com https://www.crunchbase.com/organization/practice-fusion#/entity
- 44. OpenMRS https://github.com/openmrs http://openmrs.org Software is available under the Mozilla Public License 2.0 with Healthcare Disclaimer (MPL 2.0 HD).What technologies is OpenMRS built on? OpenMRS is programmed in Java and the core application works through a web-browser. Hibernate is used as an interface layer to the database. Tomcat is used as the web application server. The back end database is currently in MySQL. The system creates XML schemas for form design. Form design and form data entry is currently done in Microsoft Infopath, HTML, or XForms. When form data entered is submitted, it is converted into a HL7 message before going into the database.
- 45. OpenMRS Build on top of it Bahmni is an easy-to-use EMR & hospital system. It combines and enhances existing open source products into a single solution. What is the difference between Bahmni and other OpenMRS based applications? OpenMRS is a great platform on which many have developed end user EMR applications. Most of these applications are by design specific to a particular disease, one type of hospital in a country or for just one hospital. Bahmni is aimed to being a generic system which can be used for multiple diseases, hospitals in different countries (at different levels) - only via configuration and not via software development. The EMR part of Bahmni complements OpenMRS platform (or backend) to provide an end user system. What is the cost / fees of Bahmni? Bahmni is Open Source Software licensed under AGPL 3.0 so there are no license fees to be paid. It can be downloaded and used at no cost. The details of license can be found here. You can implement Bahmni yourself or engage an implementer. Implementer may charge a fee for implementation services including installation, configuration, data migration, reference data set up, user creation, training, go-live support and support. http://www.bahmni.org/faq/ http://www.bahmni.org/ Using OpenMRS, Odoo, dcm4chee, OpenELIS https://github.com/openmrs/openmrs-core https://www.google-melange.com/archive/gsoc/2014/orgs/openmrs Google Code-In | Promo Video on OpenMRS Project - YouTube https://wiki.openmrs.org/display/docs/Developer+Guide
- 46. OpenVista (VistA) VA’s ComputerizedProvider Record System (CPRS), thecore electronicrecordinthebroader VistA platform In their 2014 EHR Report—a survey of 18,575 physicians on their EHR preferences—Medscape concludes that doctors like using the VA’s Computerized Provider Record System (CPRS), the core electronic record in the broader VistA platform, more than any other solution. The highest-rated EHR, with a score of 3.9, is the Veterans Administration EHR: VA-CPRS. It’s regarded as one of the best overall by our physician respondents. So, why is VistA CPRS the preferred choice? In a word, design. The VA built the system with two design goals: improved patient care and rapid adoptability. Physicians at the VA rotate through services and the system has to be adoptable with minimal (2 hours) training; they learn it as they take care of patients. Maybe you’re wondering how a government-derived software system could be more highly rated than private sector alternatives. As mentioned above, the VA’s goals are to develop a system that improves care for veterans and is easy to learn. Contrast that with the natural overarching goals of proprietary EHR providers, which is to automate the enterprise and make money. So, if VistA is the preferred choice, why is adoption of VistA-derived systems outside the VA so low? One explanation is lack of awareness. How many hospitals and clinics know that VistA code is public domain and available without expensive license fees? That private companies are succeeding by offering development and support for VistA-based solutions? Well, for large academic medical centers and cash-rich nonprofit (cough, cough) healthcare systems, it may be true. I mean, who would ever stand in the way of a hospital’s right to overpay for Epic? Edmund Billings, MD, is chief medical officer of Medsphere Systems Corporation, the solution provider for the OpenVista electronic health record. http://hitconsultant.net/2014/07/24/physicians-prefer-the-vas-ehr/ http://www.medsphere.com/open-vista OpenVista Technology Medsphere's Enterprise Assessment process determines each hospital's unique requirements and presents alternatives for meeting those goals. Our collaborative approach answers questions about hosting, server virtualization and other issues before an implementation plan commences. Hardware and Operating System A client/server solution, OpenVista runs effectively on a single Linux or Windows server, even in large hospitals. Workstation requirements are modest, enabling many OpenVista client sites to save money by using existing Windows workstations. Database OpenVista requires either InterSystems Caché (Linux or Windows) or FIS GT.M (Linux only) database management systems. Caché is a commercial product that supports some of the largest hospitals and EHR implementations in the USA. GT.M is an open source product with professional support that runs both hospitals and some of the largest banking systems in the world. Caché and GT.M each have a long track record of exceptional reliability and performance in healthcare scenarios. Network The bandwidth requirements for OpenVista client operations are modest. A reliable DSL or cable connection is adequate for a small satellite clinic; to accommodate more users and interfaces, a larger site, such as a hospital, will require more bandwidth. http://www.medsphere.com/open-vista/technology healthcareitnews.com
- 47. Open-source Eye EMR OpenEyes is a collaborative, open source, project led by Moorfields Eye Hospital. The goal is to produce a framework which will allow the rapid, and continuous development of electronic patient records (EPR) with contributions from Hospitals, Institutions, Academic departments, Companies, and Individuals. https://github.com/openeyes/OpenEyes ● Development of 3rd party applications would be simplified by standardized open source platforms of which is an example of the OpenEyes initiative led by Moorfields Eye Hospital in London, UK
- 51. Health care So what is wrong with it?
- 52. Healthcare general model The NHS Structure (UK) The NHS is divided into two sections: primary and secondary care. Primary care is the first point of contact for most people and is delivered by a wide range of independent contractors, including GPs, dentists, pharmacists and optometrists. Secondary care is known as acute healthcare and can be either elective care or emergency care. Elective care means planned specialist medical care or surgery, usually following referral from a primary or community health professional such as a GP. http://dx.doi.org/10.1007%2Fs11606-010-1464-0 Primary healthcare results in better health outcomes, reduced health disparities and lower spending, including on avoidable emergency room visits and hospital care. With that being said, primary care physicians are an important component in ensuring that the healthcare system as a whole is sustainable. However, despite their importance to the healthcare system, the primary care position has suffered in terms of its prestige in part due to the differences in salary when compared to doctors that decide to specialize. In a 2010 national study of physician wages conducted by the UC Davis Health System found that specialists are paid as much as 52 percent more than primary care physicians, even though primary care physicians see far more patients.[10] Primary care physicians earn $60.48 per hour; specialists on average earn $88.34. [10] A follow up study conducted by the UC Davis Health System found that earnings over the course of the careers of primary care physicians averaged as much as $2.8 million less than the earnings of their specialist colleagues.[11]This discrepancy in pay has potentially made primary care a less attractive choice for medical school graduates. https://en.wikipedia.org/wiki/Primary_care#United_States The Continuum of Healthcare OJIN: The Online Journal of Issues in Nursing Rising to the Challenge of Health Care Reform With Entrepreneurial and Intrapreneurial Nursing Initiatives Anne Wilson, PhD, MN, BN, FRCNA; Nancy Whitaker, BPsych (Hon); Deirdre Whitford, PhD
- 53. Healthcare in US ● It is well-known that the healthcare in US is the least cost-efficient system in Western countries even though they have the access to cutting-edge technologies and high-level medical universities. By Lillian Thomas, Pittsburgh Post-Gazette, June 14, 2014 3:15 p.m. archive.jsonline.com/news The U.S. health care system is neither a true market system, nor a government managed system. It's complicated and hard to navigate. The same forces that make it a bloated drain on the economy drive it out of poor neighborhoods where it's sorely needed. Princeton University health economist Uwe Reinhardt compares the system to one where employees are told they'd be reimbursed for clothing deemed "necessary" and "appropriate" for the job, but are forced to shop blindfolded, stuffing items into a cart without knowing what they cost or what they look like — and only informed months later whether they'd be reimbursed. https://en.wikipedia.org/wiki/Health_care_in_the_United_States https://en.wikipedia.org/wiki/Health_care_in_Canadaforbes.com/sites/robertpearl/2014/01/09
- 54. US Fee-For-Service model This article is part of a series of blog posts by leaders in health and health care who participated in Spotlight Health from June 25-28, the opening segment of the Aspen Ideas Festival. Current Fee-For-Service Primary care fee-for-service only pays a doctor for a certain set of discrete activities—largely confined to doctor sick visits—which are tiered by means of an arcane coding system counting very discrete micro tasks, such as how many organ systems a doctor examines, or what questions a doctor asks a patient about the quality of their symptoms. This encourages every health care issue or question to become a doctor visit (because that is paid for), and for the doctor to do most things instead of others on the team (because that is what is paid for). It leads to reactive care (since thinking of a patient not in front of you isn’t paid for), and leads to framing the job as taking care of one patient at a time, like a never- ending series of widgets on an assembly line. Electronic health records, not surprisingly, are thus built to optimize this fee-for-service payment, particularly the coding level of each visit, and leads to lots of useless points and clicks and incredibly long notes that, in retrospect, are extremely difficult to comprehend. And practices spend a huge percent of their time and overhead dealing with all of this, which is really just a game, and does not lead to one iota of better patient care. Fee-for-service is simply the wrong model to pay for primary care. It is toxic to good care and to physician and team culture, so we should simply stop using it, not try to supplement it. Primary care should be about continuous healing relationships, and discretely paying for services is antithetical to this. healthaffairs.org/blog/2015/08/17 If fee-for-service is a problem, what's the solution? By Paul Demko | February 25, 2015 http://www.modernhealthcare.com/article/20150225/NEWS/150229939 HEDIS® and Quality Compass® HEDIS is a tool used by more than 90 percent of America's health plans to measure performance on important dimensions of care and service. Because so many plans collect HEDIS data, and because the measures are so specifically defined, HEDIS makes it possible to compare the performance of health plans on an "apples-to-apples" basis. Health plans also use HEDIS results themselves to see where they need to focus their improvement efforts. http://www.ncqa.org/HEDISQualityMeasurement/WhatisHEDIS.aspx Advanced Laboratory Analytics — A Disruptive Solution for Health Systems Eleanor Herriman, MD, MBA Chief Medical Informatics Officer
- 55. US Healthcare Misaligned incentives #1 U.S. Health Care Rewards Quantity Over Quality Imagine you’re planning to remodel your kitchen. You hire a contractor and opt to defer entirely to his judgment on the kitchen’s aesthetics and the source of his materials. Instead of requesting a competitive bid or choosing exactly what you want, you agree to a time and materials contract. By the end of the remodel, the contractor has billed more hours than you expected, marked up the cost of the materials and charged you twice for his construction errors. Chances are you’d never agree to such a lopsided arrangement for your kitchen. But that’s the approach most Americans take when they go for medical care. The U.S. health care system pays physicians based on a fee-for-service (FFS) financial model. In short, it’s that “time and materials” contract you’d never agree to for your kitchen. But it’s not the doctors or hospital administrators who are the fundamental problem. It’s the financial model. forbes.com/sites/robertpearl/2014/01/09 Fee-For-Service Model Pervasive Yet Perverse Economics 101 teaches that as supply goes up, costs should come down. But this tenant doesn’t hold true in medical care – not when the supplier also controls demand. In health care, doctors can stimulate demand because (a) health insurance blinds most patients to the costs of services and (b) patients often don’t know whether a complex procedure is as necessary as a non-invasive one. Over the past 15 years, U.S. medical school enrollment has risen by 30 percent. But while the number of specialty residences – and therefore specialists in a community – has grown substantially, the number of primary care residents has remained flat. The reason is simple: Hospitals receive the same financial reimbursement from the federal government whether they train a primary care physician or an orthopedic surgeon. The orthopedic resident will earn the hospital a lot of money while the primary care physician will bring in little or nothing. As a hospital administrator, which clinical training program would you expand? All too often, patients acquire any number of conditions from a hospital stay, from pressure ulcers to post-admission infections. In fact, about 4 percent of beds in a typical hospital are occupied by patients who couldn’t be discharged because of a hospital-acquired complication. As an example, a recent study found that privately insured surgical patients with one or more complications provided hospitals with a 330 percent higher profit margin than those who had no complications. Free Rein Drives Personal Gain Just like the contractor who’s given free rein over a kitchen remodeling project, little limits doctors from making decisions that boost their bottom line. Oncologists routinely purchase the chemotherapy they administer, mark up the price substantially and keep the difference for themselves. Surgeons often buy into ambulatory surgery centers or “surgicenters,” and earn guaranteed double digit returns provided they commit to bringing their fully insured patients there. Meanwhile, drug and device companies pay physicians to talk up new medications or devices. And until the passage of the Sunshine Act, the U.S. health care system didn’t require any visibility or disclosure. forbes.com/sites/robertpearl/2014/01/09According to the MGMA report, the average primary care physician in a group earned $171,519 last year, up 2 percent. Specialty care physicians earned $322,259, up 1.8 percent. Over the last five years, incomes rose 11.9 percent for all primary care doctors, while specialists’ pay surged 17.3 percent managedcaremag.com/archives/2007/12.
- 56. US Healthcare Misaligned incentives #2 The principal-agent problems in health care: evidence from prescribing patterns of private providers in Vietnam The principal-agent problem in health care asserts that providers, being the imperfect agents of patients, will act to maximize their profits at the expense of the patients’ interests. This problem applies especially where professional regulations are lacking and incentives exist to directly link providers’ actions to their profits, such as a fee-for- service payment system. http://dx.doi.org/10.1093/heapol/czr028 http://dx.doi.org/10.2307/1912246 Cited by 3325 http://dx.doi.org/10.1016/0167-6296(93)90023-8 Cited by 181 NBER Working Paper No. 21930 Issued in January 2016 NBER Program(s): AG HC HE A longstanding literature has highlighted the tension between the altruism of physicians and their desire for profit. This paper develops new implications for how these competing forces drive pricing and utilization in healthcare markets. Altruism dictates that providers reduce utilization in response to higher prices, but profit-maximization does the opposite. Rational physicians will behave more altruistically when treating poorer patients or those that face higher medical cost burdens, and when foregone profits are lower. These insights help explain the observed heterogeneity in pricing dynamics across different healthcare markets. http://dx.doi.org/10.3386/w21930 http://dx.doi.org/10.1097/HMR.0000000000000042 Physician financial conflict of interest is a concern in the delivery of medicine because of its possible influence on the cost and the quality of patient care. There has been an extensive discussion of the ethical, economic, and legal aspects of this issue but little direct empirical evidence of its magnitude or effects. Results indicate that the vast majority of physicians receive industry gifts in various forms, and the receipt of gifts is associated with lower perceived quality of patient care. There is also an inverse relationship between the frequency of received gifts and the perceived quality of care. Physicians need to be aware of the widespread receipt of industry gifts in medical practice and the potential adverse impact of such receipts on the delivery of care.
- 57. US towards value-based medicine #1 Xerox surveyed 761 U.S. adults and 204 healthcare payers and providers to gauge their attitudes about the current state of health care. To view the full results, go to http://www.xerox.com/healthcaresurvey http://www.thechicagodoctor.com/columns/healthcare-it-2/value-based-medicine-success/ http://blog.academyhealth.org/movingawayfromffs/ Lisa McDonnel, SVP, Network Strategy & Line of Business Support, UnitedHealthcare Networks http://slideplayer.com/slide/4895830/ As part of health reform and confronting a health care system with costs that are unsustainable, policymakers and others have been looking at – and testing – new payment models. These would move away from fee-for-service and payment based on volume to systems that encourage more coordinated care, focusing on the overall health of the population.The graph below shows the incentives under the current payment model and the evolution of the incentives as we shift from fee for service to population health. http://www.dartmouth-hitchcock.org/about_dh/new_reimbursement_models.html
- 58. US towards value-based medicine #2 by COURTNEY BAIRD March 08, 2016 At all stages of life – from infancy to old age – the researchers discovered consistently poorer health in the US for every measure. Some of the more remarkable findings include: • The highest rate of women dying due to complications of pregnancy and childbirth; • The highest chance that a child will die before age 5; • The second-highest rate of death by coronary heart disease; • The second-highest rate of death by lung disease; • Some of the worst rates of heart disease, lung disease, obesity and diabetes. A plethora of other reports by groups such as the World Health Organization and the OECD reach similar conclusions. Todd Hixon , CONTRIBUTOR I blog about entrepreneurs, their world, and the new, new thing. forbes.com/sites/toddhixon/2015/06/11 Healthcare and government leaders talk a lot about value- based payment changing the way U.S. healthcare works. But, a closer look shows that “value” is often measured on the basis of procedures. So providers are still paid for delivering volume. More important, value as experienced by medical customers(Wherever possible, I use the term “customer” or “medical customer” instead of “patient”.) is still not a big factor in the provider incentive equation. In primary care particularly, much of the value of good care emerges over time. Eighty percent of healthcare cost is driven by chronic diseases which often result from lifestyle choices: over-eating, no exercise, sun-bathing, smoking, excessive drinking, bad diet, etc. Fixing this problem broadly is a huge challenge because the root problem is how we pay for healthcare, a gnarly political problem. And short-term oriented behaviors occur at every level of the healthcare payment chain: government, corporations, and individuals. But if we don’t start to fix these problems, then in many cases “value-based healthcare” will turn out to be the same old pig with new lipstick. doctorsbag.net/2015/04/16
- 59. US towards value-based medicine #3 http://dx.doi.org/10.1377/hlthaff.2015.1291 forbes.com/sites/brucejapsen/2015/01/26 http://dx.doi.org/10.1056/NEJMp1500445, GoogleScholar: Cited by 251 times http://dx.doi.org/10.1056/NEJMhpr1503614 http://dx.doi.org/10.1001/jama.2015.18161 President Obama is the only sitting president of the United States in modern history to publish an article inJAMA.1 That seems appropriate since he is also the only recent president to sign comprehensive health reform legislation, the Affordable Care Act (ACA). The president adopted a dual mandate for the ACA: it needed not only to expand coverage but also to contain costs (despite the additional utilization associated with the increased coverage) and improve quality. http://dx.doi.org/10.1001/jama.2016.9876 http://dx.doi.org/10.1016/S1474-4422(16)00064-8
- 60. US towards value-based medicine #4 Emad Rizk, MD, McKesson Health Solutions “Navigating the Complexity of New Value-Based Reimbursement Models” http://www.slideshare.net/ClevelandHeartLab/emad-rizk-md-final "Many of the alternative payment models currently being implemented in Medicare not only fail to solve the problems in the current payment system, they can actually make things worse for physicians who want to improve care and reduce spending,” stated Harold D. Miller, CHQPR’s President and CEO. revcycleintelligence.com hbr.org/2013/10
- 61. US Who pays and how much? medicaleconomics.modernmedicine.com US Health Care Spending: Who Pays? Josh Cothran, Georgia Institute of Technology In the past 50 years, the way health care is financed has changed, with private payers and public insurance paying for more care. This interactive graphic shows who paid for the nation's health care and how much it cost. US patients typically do not know the actual price of their medical operations, as insurance companies typically pay the part exceeding the deductibles. One approach to mitigate this, price transparency have been tried to introduce to US healthcare with varying results. doi:10.1001/jama.2016.4325 citeseerx.ist.psu.edu http://www.ncbi.nlm.nih.gov/books/NBK53906/ pwc.com/us/en/health-industries
- 62. Functional marker PUPILLOMETRY http://iovs.arvojournals.org/article.aspx?articleid=2467290 doi:10.1016/j.visres. 2012.07.019 http://www.medicalexpo.com/cat/ophthalmology/pupillometers-pupillographs-F-144.htm l The 3D model of the pupillometer based on a smartphone (left) and the configuration of the measurement system based on the prototype design (right) .http://dx.doi.org/10.3807/JOSK.2013.17.3.249 businesswire.comdoi:10.1364/AO.53.000H27 BOSTON — A simple, quick test performed with a pupillometer appears to be an effective screening tool for acute mild traumatic brain injury, US Army investigators report.
- 63. Eye care in US 2014 Report Medscape Ophthalmologist Compensation Report 2014 Carol Peckham, April 15, 2014 In Medscape's 2014 Compensation Report, ophthalmologists fall slightly above the middle among all physicians, with average earnings of $291,000. As in previous Medscape reports, orthopedists are the earning leaders, followed by cardiologists. Urologists and gastroenterologists are tied for third place.
- 64. Eye care in US 2016 Report Medscape Ophthalmologist Compensation Report 2016 Carol Peckham | April 1, 2016 Bureaucratic tasks were the prime cause of physician burnout, according to this year's Medscape Lifestyle Report (and in previous ones as well). Second was spending too many hours at work. Among ophthalmologists responding to this year's survey, 37% of those who are self-employed and 32% of their employed peers spend 10 hours or more per week on paperwork and administrative tasks.
- 65. Health care in Europe United Kingdom #1 The UK’s National Health Service is the best healthcare system in the world, and the US system is the worst according to a new study by The Commonwealth Fund (direct link). UK citizens are not necessarily more healthy than other citizens when it comes to healthy life expectancy. The obesity crisis, adiabetes epidemic and widespread smoking have all contributed to this. ampp3d.mirror.co.uk/2014/06/18
- 66. Health care in Europe United Kingdom #2: NHS www.gov.uk ● In Europe, there has been more focus on making the system work more efficiently as a whole instead of the piecewise optimized system of the USA.
- 67. Artificial Intelligence in Healthcare While the biggest problems are in medical financial models, there are a lot of room for innovations In healthcare for artificial intelligence -based solutions.
- 68. Healthcare Artificial Intelligence #1 http://blog.ventureradar.com/2015/09/30/top-5-companies-revolutionizing-healthcare-with-machine-learning/ http://www.enlitic.com/ http://www.lumiata.com/
- 69. Healthcare Artificial Intelligence #2 forbes.com businesskorea.co.kr http://www.journals.elsevier.com/artificial-intelligence-in-medicine/ www.wsj.com http://www.cnbc.com/2016/08/17/ cbinsights.com technologyreview.com Ali Parsa, Babylon forbes.com
- 70. Healthcare AI companies #1 stratifiedmedical.com babylonhealth.com
- 71. Enlitic is a startup that uses deep learning and image analysis to help doctors make diagnoses and spot abnormalities in medical images. For example, Enlitic could analyze medical images such as X-rays, MRIs, or CT scans for trends in the data or anomalies in individual images. Cardiogram, an Apple Watch app, uses an algorithm to detect when changes in a user’s heart rate may signal a serious health disorder. The National Institutes of Health–funded AiCure app uses a smartphone’s webcam and AI to autonomously confirm that patients are adhering to their prescriptions: critical for people with serious ailments and participants in clinical trials. San Francisco–based startup Sense.ly has a slew of customers, from the National Health Service to UC San Francisco, for its virtual nurse, Molly. The interface uses machine learning to support patients with chronic conditions between doctor’s visits. Part of a larger effort to offer individuals targeted diagnoses and treatments, Toronto startup Deep Genomics identifies patterns in huge data sets of genetic information, looking for mutations and linkages to disease. Developing pharmaceuticals can take decades. Silicon Valley’s Atomwise speeds things up with supercomputers that root out therapies from a database of molecular structures. Meanwhile, Berg Health also mines data for clues about why some people survive diseases—insights that can inform new therapies. Imagen is building a world without diagnostic errors. We are beginning by applying the latest advances in computer vision and machine learning to medical imaging. Each year hundreds of millions of people across the world are misdiagnosed. Diagnostic errors are a leading cause of patient harm and unnecessary costs, making it one of the greatest problems in healthcare. http://www.imagentechnologies.com/ Quantified Skin. Powered by Deep Learning. Our platform analyzes the contextual and behavioral signals of users from images and sensor data. Total Equity Funding: $280.75k in 3 Rounds. Most Recent Funding: $60.75 Thousand on January 1, 2014 / Undisclosed Round The summit will showcase the opportunities of advancing methods in deep learning and their impact across healthcare & medicine. Discover the deep learning tools & techniques set to revolutionise healthcare applications, medicine & diagnostics from a global line-up of experts. re-work.co Healthcare AI companies #2
- 72. 3rd party APIs for artificial intelligence http://www.i-nside.com/ blog.clarifai.com http://clarifai.com/
- 73. RETINAL AI Companies DeepMind #1 Say you succeed and create a super intelligence. What happens next? Do you donate the technology to the United Nations? I think it should be. We’ve talked about this a lot. Actually Eric Schmidt [executive chairman of Alphabet, Google’s parent] has mentioned this. We’ve talked to him. We think that AI has to be used for the benefit of everyone. https://www.bloomberg.com/features/2016-demis-hassabis-interview-issue/
- 74. RETINAL AI Companies DeepMind #2 At a Glance ● You may have heard of Google’s DeepMind division – which has artificial intelligence (AI) algorithms that can beat not only Atari videogames from the 1980s, but also the world’s best player of the world’s most complex board game, Go ● But their AI platform has health applications too, which inspired Moorfields Consultant Ophthalmologist, Pearse Keane, to assess its potential to automate and transform retinal image analysis – and disease diagnostics ● Moorfields is now sharing 1 million fundus photograph and OCT images with DeepMind, who will train its AI algorithms to detect even the earliest signs of disease pathology ● Manual retinal image analysis today requires highly-trained, experienced specialists, and takes time. AI could help both speed up the process, and prioritize the patients who need review and treatment the earliest Pearse Keane, a Consultant Ophthalmologist at Moorfields, initiated the collaboration. Below is his story of how Moorfields and DeepMind got working together, what they’re currently working on, and what’s next.: “I sent Mustafa Suleyman a message on LinkedIn, and within a few days I was meeting him for coffee to get the project underway.” The AI can surely come to better conclusions if it knows the patient’s history, demographics, treatment history and so on – how much data will be used? We’ve got two strands of approach to this. The first work is on the anonymized data, which will provide more limited information. For example, we will provide the diagnosis (say, AMD), the patient’s age, and this will be linked to a certain OCT scan, and DeepMind’s algorithms will get to work – but there will be no patient identifiable information. This is what we’re doing our initial analysis from. We’re also planning research on pseudonymized data as well. This will include additional labels, and in particular will involve longitudinal image sets. We’ll be able to see if a patient has had OCT scans performed at multiple visits, meaning that we can then track the progression of the disease. We have taken extra measures to be very confident that this won’t allow you to identify any of the patients. We’ve got ethical approval for our pseudonymized data set, but it’s still pending UK Health Research Authority approvals. Do you hope the AI can eventually move beyond diagnostics? I think in the medium- to long-term we definitely hope that it will provide new scientific insight. In the short-term we want to see if we can get accurate diagnoses, and then we’d like to see if we can get information about disease prognosis and pathophysiology – for example, what is the risk of converting from dry to wet AMD, and what timeline might this occur on? One of the really nice things about this collaboration is that DeepMind wants the research to be clinician-led. We don’t just give them data and leave them alone to come up with something; it’s a two-way process. They’re always looking for guidance on what features we’re interested in, what clinical problems will have the most patient benefit if solved, and so on. I think what would be really interesting as a research question, is whether deep learning could pick up features on the OCT scan that we as humans are oblivious to, even with specialist training. I know that deep learning has been applied to things like breast cancer histopathology, and was able to pick up new features on the microscope slide that correlate with five-year survival of patients with breast cancer. To do something similar would be the Holy Grail in terms of OCT retinal imaging research. https://theophthalmologist.com/issues/0716/on-a-quest-to-find-the-holy-grail-of-imaging/
- 75. RETINAL AI Companies Behold.ai http://behold.ai/ medium.com/aws-activate-startuphttps://blogs.nvidia.com/blog/2016/02/09/deep-learning-3/
- 76. RETINAL AI Companies DreamUp Vision http://dreamupvision.com/ http://www.dreamquark.com/ 2015 © DreamUp Vision. Paris, France. All Rights Reserved.
- 77. RETINAL AI Companies Orobix http://www.orobix.com/ Copyright Daniele Cortinovis, Orobix Srl https://github.com/orobix/retina-unet Retina blood vessel segmentation with a convolution neural network (U-net) http://vmtklab.orobix.com/ Life science work focused on vessel segmentation and 3D modeling of it, rather than purely on retinal pathologies.
- 78. Academic Approaches S6516 - Automatic Gradingof Eye Diseases throughDeep Learning Apaar Sadhwani Ph.D. Candidate, Stanford University We'll outline the development of state-of-the-art medical imaging system using novel deep architectures that harness GPUs for accelerated training. Trained using data from Stanford Byers Eye Institute and Palo Alto VA Hospital, our model grades the severity of eye diseases and localizes lesions to help screen eye patients at primary care. At the heart of this system lies our hybrid approach to deep learning for high resolution images -- a large convnet with millions of parameters trained with downsized images, fused with a net trained on selected tiles of the high-resolution image. This innovative approach involves use of transfer learning, data augmentation, and multi-GPU systems to identify small-scale features that are critical to detecting eye diseases. Tags: Deep Learning & Artificial Intelligence; Medical Imaging; Supercomputing & HPC; Press-Suggested Sessions: AI & Deep Learning http://arxiv.org/abs/1608.01339 https://arxiv.org/abs/1603.04833
- 79. Retinal Alzheimer Prediction #1 forbes.com http://www.cognoptix.com/ http://dx.doi.org/10.4172/2161-0460.1000223 PhD Project: Multi-spectral imaging for in vivo imaging of oxygen tension and -amyloidβ The aim of this project is to build and clinically test a reliable multi-spectral imaging device, that allows in vivo imaging of oxygen tension and -amyloid in human eyes.β Maps showing the possible existence and distribution of -amyloid plaques will beβ obtained in glaucoma patients and possibly patients with (early) Alzheimers’s disease. A second goal is to develop software for hyperspectral image analysis for early detection and diagnosis of these diseases, based on existing models developed in our laboratory (Berendschot et al., 2010; van de Kraats et al., 1996). Applicants Dr. Tos TJM Berendschot Prof. dr. Carroll AB Webers University Eye Clinic Maastricht
- 80. Retinal Alzheimer Prediction #2 NeuroVision http://www.austinblanco.com/nvi/ Diagnostics player NeuroVision raised $10 million in a Series B round, a portion of which was reserved for strategic investors. Wildcat Capital Management led the round with $5 million and Leonard Potter, the firm's chief investment officer and president, will join NeuroVision's board of directors. https://www.crunchbase.com/organization/neurovision-imaging#/entity
- 81. Retinal Parkinson's Disease prediction http://dx.doi.org/10.1186/s40478-016-0346-z The occurrence of ocular manifestations in many neurodegenerative diseases, including PD, emphasizes the strong connection between the brain and the retina. Indeed, PD patients often suffer from visual symptoms such as reduced visual acuity, low contrast sensitivity and disturbed colour vision [9, 10]. In particular, recent findings have highlighted the retina as a potential biomarker for PD. Thinning of the inner retinal layer has shown to be an early event in patients with early PD, where early disease was defined as diagnosis within 2.5 years with an average disease stage of 2 (Hoehn–Yahr scale) [11, 12]. The severity of PD symptoms correlates with RNFL thickness [13, 14]. Functional changes in the retina have also been recorded in early PD (grade 1–1.5 Hoehn–Yahr scale, diagnosed within 3 years) [15] with some suggestion that RGCs are involved early [16]. The aim of this study is to determine, using a rotenone-induced rodent model of PD, whether firstly, retinal imaging can identify quantifiable changes including retinal layer thickness (OCT) and RGC apoptotic counts (DARC) in the natural history of the disease; and secondly, if these changes can be used as surrogate biomarkers, and applied to assessing a potential PD treatment strategy. A modified Heidelberg HRA-OCT Spectralis (Heidelberg Engineering) was used for OCT imaging [33]. A posterior pole scanning protocol, centred on the rat optic nerve head, was obtained in each treatment group with TruTrack® software (Heidelberg Engineering), to ensure pixel to pixel correspondence over time. Mean retinal thickness was determined using the HEYEX® thickness map analysis (Heidelberg Engineering). Affiliated with: UCL Institute of Ophthalmology, University College London http://dx.doi.org/10.1371/journal.pone.0085718 http://dx.doi.org/10.1007%2Fs10633-015-9503-0
- 82. Other retinal markers for disease Bipolar disorder http://dx.doi.org/10.1016/j.comppsych.2016.02.005 Major depressive disorder http://dx.doi.org/10.1016/j.comppsych.2016.02.005 Neuromyelitis optica and multiple sclerosis http://dx.doi.org/10.1016/j.psychres.2015.07.028 Schizophrenia http://dx.doi.org/10.1016/j.pscychresns.2011.08.011 Neurology in general http://dx.doi.org/10.1097/WCO.0b013e328334e99b Neurology in general #2 http://www.ncbi.nlm.nih.gov/pubmed/20065921 Huntington's disease http://dx.doi.org/10.1002/mds.26486 Optical Coherence Tomography and Visual Field Findings in Patients With Friedreich Ataxia http://dx.doi.org/10.1097/WNO.0000000000000068 Retinal nerve fibre layer loss in hereditary spastic paraplegias is restricted to complex phenotypes http://dx.doi.org/10.1186/1471-2377-12-143 Retinal nerve fiber layer thickness in patients with essential tremor. Retinal markers in various neurodegenerative disease retinalphysician.com http://dx.doi.org/10.1016/S1474-4422(16)00068-5
- 83. Regulatory affairs and patents
- 84. Regulatory affairs FDA Status of AI dispatch.com itnonline.com Simply put, the FDA wants to keep computers on the periphery of diagnosis – and even then is not comfortable with the use of diagnostic software. The agency has cautiously defended the right of physicians – and only physicians – to practice medicine. The FDAabsolutely will not step over that line, which pretty much negates the chance that deep learning algorithms with even modest autonomy will pass FDA review. http://www.cio.com/article/3024715/government/luddites-clobber-ai-says-advocate.html How will FDA regulate predictive clinical software as predictive analytics pushes medical frontiers? FDA & FTC Law, Healthcare Law michaelhcohen.com/2015/06 Predictive analytics will push the frontiers of clinical care; the question is whether FDA regulation will promote or stifle innovation in the name of consumer protection. FDA regulation in this area is a moving target. Let’s see what we know so far. Fortune: Carl, why did you decide to move from real estate into healthcare and has it panned out like you thought it would? Carl Berg: I have been in the venture capital business for 40 years but I never once touched biotech because I was concerned about the risk associated with government approval – it’s bad enough when you’re doing venture capital but adding one more equation, like getting approval from the FDA [Food and Drug Administration] makes it a lot harder. But about eight years ago I said, instead of getting into a whole bunch of small companies, I am in a position now where I can do something really big in a hope that it changes the world. So that’s what motivated me, and then I met with Niven, and that’s what got it started. http://fortune.com/2015/04/16/cancer-cure-artificial-intelligence/
- 85. Patents for AI technology One signal of this change is the dramatic increase in the number of companies making serious investments in Machine Learning. IBM is the most famous example with Watson and their innovative cloud service. But Google, Amazon and Microsoft are also deeply involved. Even Uber just bought a whole department of strong University talent from Carnegie-Mellon. These actions are not those of crazed corporate lemmings pouring their money off the cliff. No, not at all. The investment is increasing because finally, after decades of disappointment, Machine Learning algorithms are solving useful problems and delivering useful results. Patent law was designed to protect the work of inventors. Inventors are the ones “who conceived the invention.” Patents give inventors an exclusionary right to their particular invention. Patent protection provides a mechanism to compensate inventors for their investment in developing the invention, e.g., by permitting others to practice their invention for a fee. But the rising capability and prevalence of AI seem to pose a threat to this system. First, as discussed before, identifying the problem itself is not patentable. In some jurisdictions, problem identification by the inventor does not protect against a claim of obviousness when the solution combines prior known elements. Consider then the case where the researcher presents the AI entity with a problem to solve. The AI entity then crafts (or “deep learns”) the solution. Who is the inventor of that solution? Certainly not the researcher, since current law does not credit problem identification by itself worthy of patent protection. Furthermore, the researcher didn’t actually create or even discover the solution (the AI did). The arrival of IBM’s Watson and his cousins from other companies is a game changer. The new technology affects many things including the world of Intellectual Property. Those who hate patents anyway will likely cheer this coming world. With some predicting that artificial intelligence (AI) will allow a patent to be filed and granted without human intervention within the next 25 years, WIPR assesses the potential impact of AI on the IP landscape. worldipreview.com cipa.org.uk/policy-and-news
- 86. Google & Medical Tech fiercebiotech.com http://www.seobythesea.com/2016/03/verily-life-sciences-patents/ AI AND GOOGLE’S KILL SWITCH Artificial intelligence (AI), Virtual Reality (VR), autonomous vehicles, and battery development are the near future (not investment advice, but there’s definitely money here!). All four fields are seeing a rapid and exciting increase in research and funding, and the technology in these fields will almost certainly improve exponentially from the fledgling systems we are currently seeing to sophisticated devices in the next few years or decades. AI AND GOOGLE Google’s Deepmind (which recently won four out of five games of ‘Go’ against a human counterpart, thus hailing a watershed moment from the science world) in conjunction with Oxford University’s Future of Humanity Institute has recently published a document (Safely Interruptible Agents) investigating disabling or turning off AI if a human operator doesn’t like the AI’s actions. A kill switch or ‘big red button’, in effect. But would that really be necessary? Elon Musk and Stephen Hawking think so. However, a number of others, including Eric Schmidt, the executive chairman of Google, feel that super intelligent computers or robots, where the AI reaches human levels of cognition allowing exponential improvements without human interaction are never actually likely to be released, certainly in any near future involving our children or grand-kids. Eric Schmidt points out rather bluntly that Musk and Hawking are not computer scientists, and as such do not have the facts and knowledge to make such statements with any certainty. GOOGLE AND PATENTS Though, that hasn’t stopped Google filing some interesting sounding patents to such things as ‘ neural network training’, ‘parallel convolutional networks’, ‘reinforcement learning with a neural network’ and ‘ classifying data objects’. The patents help protect Google’s financial investment in this rapidly-developing field of research, and it would be assumed much of this technology may ultimately find its way into such things as autonomous cars and other vehicles, which are currently a hot topic. But from the general overviews above, you can easily appreciate the focus on the development of at least semi-self-aware data networks which allow the forming of at least rudimentary AI. REFLECTIONS Is it a good thing that Google is tying up or trying to tie up areas of this technology? A patent is used to control the use or implementation of your technology or development. If an organisation with less than honest intentions (and I’m not suggesting this is Google’s case – I’m a fan and hugely impressed with what Google has achieved in a relatively short space of time) was to develop an AI technology and patent it to better control or utilise it with less restriction or interference, would that cause a problem? The assumption is that Google, whilst being a hugely successful commercial entity and requiring a revenue stream, is aiming to improve the World for everyone, and as such by protecting its financial interests through the use of patenting seems wholly right and sensible. But who actually has a ‘kill switch’ for Google, just in case? p.s. If you do have a tech- or software based patent, you can submit this to Google for review and possible purchase. 28% of submissions have been made a purchase offer, at an average of around £20,000. https://www.albright-ip.co.uk/2016/06/ai-google-kill-switch/