Image Analysis of Early Lung Adenocarcinoma and Its Significance in Pathological Typing
0 likes88 views
This study analyzed histopathological features of 94 early lung adenocarcinoma samples using image analysis. Samples were classified as adenocarcinoma in situ, microinvasive adenocarcinoma, or lepidic predominant adenocarcinoma. Several parameters were measured from images of gland structures and nuclei, including generalized line reference number density of cancer cells, adhesion rate of nuclei, nucleolus occurrence rate, maximum distance between glands, and nuclear morphology parameters. Statistical analysis found several parameters had significant differences among the three pathological types, while others did not. These parameters may help with quantitative analysis and differential diagnosis of early lung adenocarcinoma types.
![118 Analytical and Quantitative Cytopathology and Histopathology®
Luo et al
the gland interstitial distance is 0, the glands are
merged with each other and there is a back-to-
back phenomenon, which is considered to be one
of the growth characteristics of the cancerous
glands. Our results showed that the maximum dis-
tance between glands is significantly different in
different pathological types of early lung adeno-
carcinoma, while the difference in the minimum
distance between glands is not significant. The
maximum distance between glands of LPA was
the smallest, the interstitial width was the small-
est, and the gland was relatively dense. There was
no significant difference in the maximum distance
between AIS and MIA, indicating that the glandu-
lar interstitial width was not significantly differ-
ent, and the gland interstitial structure was simi-
lar. Therefore, in early lung adenocarcinoma the
glands of AIS and MIA are relatively sparse and
the interstitial fibrosis is more obvious.
The morphological parameters of the nucleus
(long axis, short axis, perimeter, and area of the
nucleus) were significantly different in different
pathological types of early lung adenocarcinoma.
The long axis, short axis, perimeter, and area of
the AIS cells were the smallest. There was no sig
nificant difference in the long axis, short axis, pe-
rimeter, and area of MIS and LPA cells, indicating
that the size of the nucleus was equivalent. The
nuclear roundness and nuclear regulatory shape
factor were not significantly different in AIS, MIA,
and LPA. The nuclear regulatory shape factor of
cancer cells was 0.96, and the nuclear roundness
ranged from 1.10–1.17, which was close to 1. There-
fore, it is considered that the long axis, short axis,
perimeter, and area of the nucleus have applica-
tion value in the quantitative analysis of early lung
adenocarcinoma, which can be used as a quantita-
tive parameter for differential diagnosis between
AIS and MIA. In different pathological types of
early lung adenocarcinoma, the nucleus of cancer
cells in the gland is relatively regular and relatively
close to ellipse or circle, and the degree of variation
between nuclei is not large.
References
1. Lara-Guerra H, Waddell TK, Salvarrey MA, AM Joshua,
Chung CT, N Paul, Boerner S, Sakurada A, Ludkovski O,
Ma C, Squire J, Liu G, Shepherd FA, Tsao MS, Leighl NB:
Phase II study of preoperative gefitinib in clinical stage I
non-small-cell lung cancer. J Clin Oncol 2009;27:6229-6236
2. Li YM, Qi WJ, Shen H: Quantitative analysis of Tiam1 expres-
sion in lung cancer and its clinical significance. Nan Fang Yi
Ke Da Xue Xue Bao 2011;31:1774-1777
3. Webster JD, Dunstan RW: Whole-slide imaging and auto-
mated image analysis: Considerations and opportunities in
the practice of pathology. Vet Pathol 2014;51:211-223
4. Travis WD, Brambilla E, Burke AP, Marx A, Nicholson AG:
Introduction to the 2015 World Health Organization Classi-
fication of Tumors of the Lung, Pleura, Thymus, and Heart.
Thorac Oncol 2015;10(9):1240-1242
5. Aru A, Nielsen K: Stereological estimates of nuclear volume
in primary lung cancer. Pathol Res Pract 1989;185:735-739
6. Di Donato G, Laufer-Amorim R, Palmieri C: Nuclear mor-
phometry in histological specimens of canine prostate can-
cer: Correlation with histological subtypes, Gleason score,
methods of collection and survival time. Res Vet Sci 2017;114:
212-217
7. Ladekarl M, Hansen TB, Nielsen RH, Mouritzen C, Hen-
riques U, Sørensen FB: [Objective histological grading of
squamous and adenosquamous lung cancer]. [Article in
Danish] Ugeskr Laeger 1996;158:5611-5615
8. Sagol O, Kargi A, Ozkal S: Stereologically estimated mean
nuclear volume and histopathologic malignancy grading as
predictors of disease extent in non-small cell lung carcinoma.
Pathol Res Pract 2000;196:683-689
9. Zhang DS, Liu L, Jin LQ, Wan ML, Li QH: Effect of
phosphorus-32 glass microspheres on the human hepatocel-
lular carcinoma in nude mice. World J Gastroenterol 2004;
16:1551-1554
10. Yang YJ, Tang Y, Ye XT, Lei CH: Medical Biological Electron
Microscope Technology. Shanghai Second Military Medical
University Press 2012, pp 1-204
11. Luo SQ, Mao SZ, Zhuo FS: Medical Cell Biology. Bei Jing,
Military Medical Science Publishing House, 1998, pp 1-198](https://image.slidesharecdn.com/15-210801195934/85/Image-Analysis-of-Early-Lung-Adenocarcinoma-and-Its-Significance-in-Pathological-Typing-7-320.jpg)
Image Analysis of Early Lung Adenocarcinoma and Its Significance in Pathological Typing
- 1. 112 Analytical and Quantitative Cytopathology and Histopathology® 0884-6812/20/4204-0112/$18.00/0 © Science Printers and Publishers, Inc. Analytical and Quantitative Cytopathology and Histopathology® OBJECTIVE: To quantitatively analyze the histopatho logical features and explore the application value of im age analysis parameters in different pathological types of early lung adenocarcinoma. STUDY DESIGN: A total of 94 surgical specimens of early lung adenocarcinoma were collected and classified into adenocarcinoma in situ (AIS), microinvasive ade nocarcinoma (MIA), and lepidic predominant adenocar cinoma (LPA). In each case, 10 relatively intact glands were photographed under 20-fold objective lens, and 100–150 cells in the corresponding gland cavity were photographed under 40-fold objective lens. The relation ship between AIS, MIA, LPA, and related parameters were statistically analyzed. RESULTS: There were significant differences in gener alized line reference number density of cancer cells in glands, the adhesion rate of nucleus in glands and the nucleolus occurrence rate in glands, the maximum dis tance between glands, nuclear area, long axis, short axis, and nuclear perimeter in the AIS, MIA, and LPA (p<0.05). However, nuclear roundness, nuclear regu latory shape factor, and the minimum distance between glands were not significantly different among AIS, MIA, and LPA (p>0.05). CONCLUSION: The generalized line reference num ber density of cancer cells in glands, the adhesion rate of nucleus in glands, the nucleolus occurrence rate in glands, the maximum distance between glands, and nu cleus morphometric parameters have important appli- cation value in the quantitative analysis of early lung adenocarcinoma and can be used as quantitative param eters for the differential diagnosis of AIS, MIA, and LPA. (Anal Quant Cytopathol Histpathol 2020;42:112– 118) Keywords: early lung adenocarcinoma; image analysis; image processing, computer-assisted; lung cancer quantitative study; pathological typing. In the quantitative study of lung cancer, scholars have explored the significance of many parameters in normal lung tissue and lung cancer tissue. For example, nucleus morphological parameters (area, long axis, short axis, perimeter, roundness, and Image Analysis of Early Lung Adenocarcinoma and Its Significance in Pathological Typing Jiao Luo, M.D., Rong Duan, M.D., Hui Li, M.D., and Hong Shen, Ph.D. From the Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou, and the Department of Pathology, School of Basic Medical Sciences, Southern Medical University, Guangzhou, People’s Republic of China. Jiao Luo is Resident Physician, Department of Pathology, Nanfang Hospital, Southern Medical University. Rong Duan is Resident Physician, Department of Pathology, Nanfang Hospital, Southern Medical University. Hui Li is Resident Physician, Department of Pathology, Nanfang Hospital, Southern Medical University. Hong Shen is Professor, Department of Pathology, Nanfang Hospital, Southern Medical University, and Department of Pathology, School of Basic Medical Sciences, Southern Medical University. Jiao Luo and Rong Duan contributed equally. Address correspondence to: Hong Shen, Ph.D., Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, People’s Republic of China, or Department of Pathology, School of Basic Medical Sciences, Southern Medical University 1023, Sha-Tai South Road, Baiyun District, Guangzhou 510515, People’s Republic of China (shenhong2010168@163.com). Financial Disclosure: The authors have no connection to any companies or products mentioned in this article.
- 2. Volume 42, Number 4/August 2020 113 Image Analysis of Early Lung Adenocarcinoma regulatory shape factor) of lung adenocarcinoma, squamous cell carcinoma, small cell carcinoma, and large cell carcinoma have been tested and applied to establish diagnostic and discrimina- tive functions,1-3 which have an important role for quantitative pathological studies of lung cancer. What are the implications of these parameters in adenocarcinoma in situ (AIS), microinvasive adenocarcinoma (MIA), and lepidic predominant adenocarcinoma (LPA) as defined by the 2015 WHO lung adenocarcinoma classification? In the image analysis of early lung adenocarcinoma, we proposed some new research parameters: the gen- eralized line reference number density of cancer cells in glands, the adhesion rate of nucleolus in glands, the nucleolus occurrence rate in glands, and the maximum and the minimum distance be- tween glands. Are these new structural parameters different in these 3 types of early lung adenocar- cinomas? Is there pathological significance? Are these nucleus morphological parameters previous- ly used in lung cancer research differential signif- icance in the early lung adenocarcinoma defined by the 2015 WHO lung adenocarcinoma classifica- tion? The current research is not yet able to an- swer these questions, for which we had explored these issues in our study. Materials and Methods Case Material Surgical specimens of 94 early lung adenocarci- noma patients with a diameter ≤3 cm were di- agnosed from January 2014 to December 2017 in Nanfang Hospital of Southern Medical Universi- ty, including 31 males and 63 females with an age range of 30–84 years, mean age 58±10.7, and 66 patients with right lung disease and 28 patients with left lung disease. The inclusion criteria were as follows: all cases were surgically resected and histologically diagnosed as early lung adenocarci- noma with a diameter ≤3 cm, the pulmonary nod- ules of all patients were solitary, and no history of systemic radiotherapy and chemotherapy. All pa- tients had complete pathological diagnosis data, imaging data, and general data, and the pathologi- cal diagnosis was interpreted by 2 pathologists. Main Instruments and Software The main instruments and software used in this study were as follows: light microscope column (Olympus; FRAME_MANUAL; Japan), objective lens (Olympus; model no. UPlanFLN; 20×, NA: 0.50; 40×, NA: 0.75), microscopic camera (Olym- pus; model no. DP21; 20×, 219.54 nm/pixel; 40×, 109.33 nm/pixel), and image analysis software (Image-Pro Plus 6.0). Experimental Methods Ninety-four patients with early lung adenocarci- noma were included according to the inclusion criteria. Patients were divided into 3 types refer- ring to the new classification criteria for WHO lung adenocarcinoma published in 20154: adenocarcino- ma in situ (AIS), microinvasive adenocarcinoma (MIA), and lepidic predominant adenocarcinoma (LPA) (Figure 1). AIS is a localized small (≤3 cm) adenocarcinoma with growth restricted to neo plastic cells along preexisting alveolar structures (lepidic growth), lacking stromal, vascular, or pleu- ral invasion. Acinar, papillary, solid, or micropap- illary patterns and intraalveolar tumor cells are absent. MIA is a small (≤3 cm), solitary adenocar- cinoma with a predominantly lepidic pattern and ≤5 mm invasion in greatest dimension in any one focus. Multiple infiltrates are based on the largest diameter, rather than adding multiple infiltrates. Figure 1 Pathological classification of early lung adenocarcinoma. (A) Adenocarcinoma in situ, 4×10 times). (B) Microinvasive adenocarcinoma (4×10 times). (C) Lepidic predominant adenocarcinoma (4×10 times).
- 3. 114 Analytical and Quantitative Cytopathology and Histopathology® Luo et al LPA typically consists of lepidic neoplastic cells growing along the surface of alveolar walls sim- ilar to the morphology defined in the above sec- tion on AIS and MIA. Invasive adenocarcinoma is present in at least 1 focus measuring >5 mm in greatest dimension. Invasion is defined as (1) his- tological subtypes other than a lepidic pattern (acinar, papillary, micropapillary, and/or solid), (2) myofibroblastic stroma associated with inva- sive tumor cells, (3) invades lymphatics, blood ves- sels, or pleura, and (4) airway dissemination and tumor necrosis. Next, the test software was cali- brated using the objective micrometer,5 and then 10 relatively intact glands were photographed in 20× objective lens, and 100–150 cells in the corre- sponding gland cavity were photographed under 40× objective lens. The images were saved in TIF format and analyzed by Image-Pro Plus 6.0 image analysis software (Figure 2). The tested parameters included the circumference of the gland (the linear length of the outer edge of the gland), the number of tumor nuclei in the gland (the total number of nuclei in the gland), the number of nucleoli in the gland (the number of nucleoli in the nucleus of the gland), the number of adhesion nuclei in the gland (the number of adjacent nucleus in the gland that are close together or overlap), the maximum dis- tance between glands (the maximum linear dis- tance between the outer edge of the gland and the peripheral gland of the adjacent gland), the min- imum distance between glands (the minimum lin- ear distance between the outer edge of the gland and the peripheral gland of the adjacent gland), nuclear long axis, nuclear short axis, nuclear pe- rimeter, and nuclear area. Calculation of the gener- alized line reference number density of the cancer cells in glands (Formula 1), the nucleolus occur- rence rate in glands (Formula 2), the adhesion rate of nucleus in glands (Formula 3), the roundness of the nucleus (Formula 4), and the nuclear regu Figure 2 (A) Measurement of the circumference of the gland (the linear length of the outer edge of the gland) and counting of nucleus in glands (200 times high magnification view of the picture). (B) Counting of nucleolus and adhered nucleus in glands (400 times high magnification view of the picture; counting by Image-Pro Plus 6.0 software; red arrow shows adhered nucleus and black arrow shows nucleolus). (C) Measurement of nucleus long axis, short axis, perimeter, area, and roundness (400 times high magnification view of the picture, taking clear and complete nucleus for measurement by Image-Pro Plus 6.0 software). (D) The maximum distance between glands and the minimum distance between glands (the maximum distance between glands in this example is 59.35 µm, and the minimum distance between glands in this example is 19.47 µm).
- 4. Volume 42, Number 4/August 2020 115 Image Analysis of Early Lung Adenocarcinoma lation Shape factor (Formula 5) were carried out. The formulas are as follows: The generalized line reference number density of cancer cells in glands (Number/unit length) = Number of tumor nuclei in the gland __________________________________ (1) Outer edge length of gland The nucleolus occurrence rate in glands (%) = Number of nucleoli in the gland _________________________________ ×100 (2) Number of tumor nuclei in the gland The adhesion rate of nucleus in glands (%) = Number of overlapped nuclei in the gland ______________________________________ ×100 (3) Number of tumor nuclei in the gland Nuclear perimeter2 The roundness of the nucleus = __________________ (4) 4×π×Nuclear area The nuclear regulation Shape factor = Nuclear area×[3×(Nuclear long axis+Nuclear short axis)−2×√Nuclear long axis×Nuclear short axis ___________________________________________________ (5) Nuclear long axis×Nuclear short axis×Nuclear perimeter Statistical Analysis All data of image analysis parameters of differ ent pathological types were analyzed by SPSS 25.0 software statistical package for one-way analysis of variance and comparison between groups. The significance criterion of the test was p<0.05, the dif- ference was significant, p<0.01, and the difference was highly significant. Results The Test Results of the Generalized Line Reference Number Density of Adenocarcinoma Cells in Glands and the Adhesion Rate of Nucleus in Glands Among AIS, MIA, and LPA In our research we obtained the minimum, maxi- mum, mean, and standard deviation of the gener- alized line reference number density of the cancer cells in glands and the adhesion rate of nucleus in glands by using software SPSS 25.0 analysis of variance to analyze 94 patients with early lung adenocarcinoma with different pathological types (Table I). Simultaneously, there were significant differences in the generalized line reference num- ber density of the cancer cells in glands and the adhesion rate of nucleus in glands between AIS and MIA, MIA and LPA, and AIS and LPA (p< 0.005) (Table I). It indicated that the generalized line reference number density of adenocarcinoma cells in glands and the adhesion rate of nucleus in glands were with a continuous increasing trend in AIS, MIA, and LPA (Chart 1). The Test Results of the Nucleolus Occurrence Rate in Glands, the Maximum and Minimum Distance Between Glands, and Nucleus Morphological Parameters (Nuclear Long Axis, Short Axis, Area, Perimeter, Roundness, and Regulatory Shape Factor) Among AIS, MIA, and LPA We obtained the minimum, maximum, mean, and standard deviation of the nucleolus occurrence rate in glands, the maximum and minimum dis- tance between glands, and nucleus morphologi- cal parameters (nuclear long axis, short axis, area, perimeter, roundness, and regulatory shape factor) of the 3 different types of early lung adenocarci- noma through software analysis, and different im- age analysis parameters were compared between groups. Our results showed that the minimum distance between glands, nuclear roundness, and nuclear regulatory shape factor were not signifi cantly different in AIS, MIA, and LPA (p>0.05). The nucleolus occurrence rate in glands and the maximum distance between glands were signifi- cantly different between AIS and LPA and between MIA and LPA (p<0.05), but there was no signi- ficant difference between AIS and MIA (p>0.05), indicating that the nucleolus occurrence rate in glands of LPA was the highest, and the maxi- mum distance between glands was the smallest. The long axis, short axis, perimeter, and area of the nucleus were significantly different between AIS and MIA and between AIS and LPA (p<0.05). There was no significant difference in MIA and LPA (p>0.05). It indicated that the nuclear long axis, short axis, perimeter, and area of the AIS were the smallest in these 3 different types of early lung adenocarcinoma (Table II). Discussion Image analysis technology has been applied to the quantitative study of some tumors. At present, scholars have quantitatively tested the morpho- logical structure of partial tumor cells. It is noted that the quantitative parameters of tumor cell nu- clei have important value in pathological diagnosis and classification.6-9 In our study the generalized line reference number density of cancer cells in glands, the adhesion rate of nucleus in glands, the nucleolus occurrence rate in glands, the maximum
- 5. 116 Analytical and Quantitative Cytopathology and Histopathology® Luo et al distance between glands, and the area, perimeter, long axis, and short axis of the nucleus were sig nificantly different and had a certain application value in the quantitative analysis of pathological types of early lung adenocarcinoma. In the image analysis of early lung adenocarci- noma, the density of the cancer cells in the gland of AIS, MIA, and LPA can be quantitatively de- scribed by the generalized line reference number density of cancer cells in glands and the adhesion rate of nucleus in glands; these parameters pre sented an increasing trend. The density of the can- cer cells in the gland of AIS was the sparsest and had the lowest adhesion rate of nucleolus in the glands. Conversely, the generalized line reference number density of cancer cells in glands of LPA was the largest and the adhesion rate of nucle- us in the glands was the highest. Therefore, when the generalized line reference number density of cancer cells in glands is increased or the adhesion rate of nucleus in glands is high, it should be con- sidered first that these glands are likely to have invasive characteristics, which will help us to iden- tify some different types of early lung adenocarci- noma that are difficult to diagnose. The nucleolus is a site for the synthesis of ribo- Table I Results of Measuring Generalized Line Reference Number Density of Cancer Cells in Glands, the Adhesion Rate of Nucleus in Glands of Various Pathological Subtypes, and the Results of Multiple Comparison Mean± Adeno- Minimum Maximum standard Parameter carcinoma Cases value value deviation p Value Generalized line reference AIS (a) 26 48 137 88±19.21b3,c3 number density of cancer MIA (b) 38 72 167 120±19.35a3,c3 0.000 cells in glands (no./mm) LPA (c) 30 93 208 142±31.23a3,b3 Adhesion rate of nucleus in AIS (a) 26 0.00 0.15 0.08±0.04b3,c3 glands (%) MIA (b) 38 0.08 0.30 0.16±0.06a3,c3 0.000 LPA (c) 30 0.04 0.47 0.25±0.11a3,b3 Superscript letters indicate comparison with this group; 1 indicates p<0.05, 2 indicates p<0.01, and 3 indicates p<0.001. Chart 1 Mean contrast bar chart of (A) generalized line reference number density of cancer cells in glands (number/mm) and (B) the adhesion rate of nucleus in glands (%) of AIS, MIA, and LPA.
- 6. Volume 42, Number 4/August 2020 117 Image Analysis of Early Lung Adenocarcinoma somal RNA (rRNA), which is closely related to protein synthesis. The number of nucleoli reflects the physiological state of the cell. In cells with ac- tive protein synthesis, the nucleolus is developed, large and obvious, or with multiple nucleoli, while in cells with inactive protein synthesis the nucleo- lus is small or nonexistent.10,11 In the image anal- ysis of early lung adenocarcinoma, we quantita tively analyzed the nucleolus occurrence rate in glands, and the results showed that LPA had the highest nucleolus occurrence rate, indicating that the intracellular protein synthesis was the most active. There was no significant difference in the nucleolus occurrence rate in glands of the AIS and MIA, which indicated that the level of cellular pro- tein synthesis in the gland was comparable. Two new parameters, the maximum and mini mum distance between glands in early lung ade- nocarcinoma, can quantitatively describe the inter stitial width between the glands. The maximum distance between glands is the quantitative anal- ysis of the maximum width of the interstitial be- tween the glands. The maximum interstitial width refers to the degree of interstitial fibrosis and the degree of density of the glands. The more serious the degree of interstitial fibrosis is, the larger the measurement parameter of maximum distance be- tween glands we measured. The minimum distance between glands is the quantitative analysis of the minimum width of the interstitial fibrosis between the glands. The smaller the minimum distance be- tween glands, the closer the glands are. The min- imum interstitial distance of the gland can reflect the degree of density between the glands. When Table II Results of Measuring Nucleolus Occurrence Rate in Glands, the Maximum and Minimum Distance Between Glands, and Nucleus Morphological Parameters of Various Pathological Subtypes and the Results of Multiple Comparisons Mean± Adeno- Minimum Maximum standard Parameter carcinoma Cases value value deviation p Value Nucleolus occurrence rate AIS (a) 26 0.00 0.52 0.12±0.11c1 in glands (%) MIA (b) 38 0.02 0.46 0.14±0.11c1 0.030 LPA (c) 30 0.04 0.42 0.20±0.12a1,b1 Minimum distance between AIS (a) 26 6.98 24.79 13.94±4.68 glands (µm) MIA (b) 38 7.63 19.45 13.10±2.49 0.132 LPA (c) 30 7.13 23.58 12.07±3.28 Maximum distance between AIS (a) 26 21.03 52.29 35.61±7.76c3 glands (µm) MIA (b) 38 21.86 46.70 32.58±6.15c1 0.002 LPA (c) 30 20.72 39.73 29.60±4.13a3,b1 Nuclear long axis (µm) AIS (a) 26 5.95 8.71 7.09±0.72b3,c3 MIA (b) 38 6.50 10.00 8.14±0.98a3 0.000 LPA (c) 30 5.81 9.85 8.18±0.87a3 Nuclear short axis (µm) AIS (a) 26 4.38 6.38 5.11±0.47b3,c3 MIA (b) 38 4.63 6.77 5.75±0.65a3 0.000 LPA (c) 30 4.88 7.12 5.73±0.64a3 Nuclear perimeter (µm) AIS (a) 26 16.50 23.98 19.33±1.89b3,c3 MIA (b) 38 17.68 26.66 22.06±2.59a3 0.000 LPA (c) 30 17.84 26.61 22.16±2.24a3 Nuclear area (µm2) AIS (a) 26 20.32 42.54 27.86±5.50b3,c3 MIA (b) 38 23.17 68.51 36.70±9.66a3 0.000 LPA (c) 30 24.01 51.96 36.24±7.47a3 Nuclear roundness AIS (a) 26 1.05 1.14 1.10±0.02 MIA (b) 38 1.07 1.18 1.11±0.02 0.214 LPA (c) 30 1.07 2.61 1.17±0.27 Nuclear regulatory shape AIS (a) 26 0.94 1.00 0.96±0.01 factor MIA (b) 38 0.95 0.99 0.96±0.01 0.506 LPA (c) 30 0.95 0.96 0.96±0.01 Superscript letters indicate comparison with this group; 1 indicates p<0.05, 2 indicates p<0.01, and 3 indicates p<0.001.
- 7. 118 Analytical and Quantitative Cytopathology and Histopathology® Luo et al the gland interstitial distance is 0, the glands are merged with each other and there is a back-to- back phenomenon, which is considered to be one of the growth characteristics of the cancerous glands. Our results showed that the maximum dis- tance between glands is significantly different in different pathological types of early lung adeno- carcinoma, while the difference in the minimum distance between glands is not significant. The maximum distance between glands of LPA was the smallest, the interstitial width was the small- est, and the gland was relatively dense. There was no significant difference in the maximum distance between AIS and MIA, indicating that the glandu- lar interstitial width was not significantly differ- ent, and the gland interstitial structure was simi- lar. Therefore, in early lung adenocarcinoma the glands of AIS and MIA are relatively sparse and the interstitial fibrosis is more obvious. The morphological parameters of the nucleus (long axis, short axis, perimeter, and area of the nucleus) were significantly different in different pathological types of early lung adenocarcinoma. The long axis, short axis, perimeter, and area of the AIS cells were the smallest. There was no sig nificant difference in the long axis, short axis, pe- rimeter, and area of MIS and LPA cells, indicating that the size of the nucleus was equivalent. The nuclear roundness and nuclear regulatory shape factor were not significantly different in AIS, MIA, and LPA. The nuclear regulatory shape factor of cancer cells was 0.96, and the nuclear roundness ranged from 1.10–1.17, which was close to 1. There- fore, it is considered that the long axis, short axis, perimeter, and area of the nucleus have applica- tion value in the quantitative analysis of early lung adenocarcinoma, which can be used as a quantita- tive parameter for differential diagnosis between AIS and MIA. In different pathological types of early lung adenocarcinoma, the nucleus of cancer cells in the gland is relatively regular and relatively close to ellipse or circle, and the degree of variation between nuclei is not large. References 1. Lara-Guerra H, Waddell TK, Salvarrey MA, AM Joshua, Chung CT, N Paul, Boerner S, Sakurada A, Ludkovski O, Ma C, Squire J, Liu G, Shepherd FA, Tsao MS, Leighl NB: Phase II study of preoperative gefitinib in clinical stage I non-small-cell lung cancer. J Clin Oncol 2009;27:6229-6236 2. Li YM, Qi WJ, Shen H: Quantitative analysis of Tiam1 expres- sion in lung cancer and its clinical significance. Nan Fang Yi Ke Da Xue Xue Bao 2011;31:1774-1777 3. Webster JD, Dunstan RW: Whole-slide imaging and auto- mated image analysis: Considerations and opportunities in the practice of pathology. Vet Pathol 2014;51:211-223 4. Travis WD, Brambilla E, Burke AP, Marx A, Nicholson AG: Introduction to the 2015 World Health Organization Classi- fication of Tumors of the Lung, Pleura, Thymus, and Heart. Thorac Oncol 2015;10(9):1240-1242 5. Aru A, Nielsen K: Stereological estimates of nuclear volume in primary lung cancer. Pathol Res Pract 1989;185:735-739 6. Di Donato G, Laufer-Amorim R, Palmieri C: Nuclear mor- phometry in histological specimens of canine prostate can- cer: Correlation with histological subtypes, Gleason score, methods of collection and survival time. Res Vet Sci 2017;114: 212-217 7. Ladekarl M, Hansen TB, Nielsen RH, Mouritzen C, Hen- riques U, Sørensen FB: [Objective histological grading of squamous and adenosquamous lung cancer]. [Article in Danish] Ugeskr Laeger 1996;158:5611-5615 8. Sagol O, Kargi A, Ozkal S: Stereologically estimated mean nuclear volume and histopathologic malignancy grading as predictors of disease extent in non-small cell lung carcinoma. Pathol Res Pract 2000;196:683-689 9. Zhang DS, Liu L, Jin LQ, Wan ML, Li QH: Effect of phosphorus-32 glass microspheres on the human hepatocel- lular carcinoma in nude mice. World J Gastroenterol 2004; 16:1551-1554 10. Yang YJ, Tang Y, Ye XT, Lei CH: Medical Biological Electron Microscope Technology. Shanghai Second Military Medical University Press 2012, pp 1-204 11. Luo SQ, Mao SZ, Zhuo FS: Medical Cell Biology. Bei Jing, Military Medical Science Publishing House, 1998, pp 1-198