CDISC journey in solid tumor using recist 1.1 (Paper)
5 likes3,779 views
This document summarizes the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 guidelines for evaluating tumor response in clinical trials. It introduces the three types of oncology studies, defines target and non-target lesions, and describes how lesion measurements are used to determine complete response, partial response, stable disease, progression disease, and not evaluable responses. It also discusses how RECIST 1.1 data are organized in CDISC SDTM and ADaM domains, and provides an example of how these domains can be used to evaluate tumor response and analyze outcomes like objective response rate, progression-free survival, and time to progression.
CDISC journey in solid tumor using recist 1.1 (Paper)
- 1. PhUSE 2013 Paper RG02 CDISC Journey in Solid Tumor using RECIST 1.1 Kevin Lee, Cytel, Inc., Chesterbrook, PA ABSTRACT The presentation is intended for Clinical Trial programmers or statisticians who are working on the solid tumor studies in oncology. There are three types of studies in oncology: Solid Tumor, Lymphoma and Leukemia. The solid tumor study usually follow RECIST (Response Evaluation Criteria in Solid Tumor) while Lymphoma follows Cheson and Leukemia follows study-specific criteria. The presentation will provide the brief introduction of RECIST 1.1 such as lesions (target, non target and new) and their selection criteria (size, number and etc). It will also discuss how the changes in tumor measurements will lead to responses (Complete Response, Partial Response, Stable Disease, Progression Disease and Not Evaluable). Then, the presentation will introduce how RECIST 1.1 data are streamlined in CDISC – mainly in SDTM and ADaM. The presentation will introduce the new oncology SDTM domains - TU (Tumor Identification), TR (Tumor Results) and RS (Response) according to RECIST 1.1. The presentation will also show how ADaM datasets can be created for the tumor response evaluation and analysis in ORR (Objective Response Rate), PFS (Progression Free Survival) and TTP (Time to Progression). INTRODUCTION OF ONCOLOGY CLINICAL TRIAL STUDIES There are three types of oncology clinical trial studies. • Solid Tumor • Lymphoma • Leukemia Response The solid tumor studies usually follow RECIST 1.0 or 1.1 on tumor response evaluation criteria, Lymphoma studies usually follow Cheson 1997 or 2007, and Leukemia studies follow the study-specific criteria. RECIST 1.1 INTRODUCTION OF RECIST RECIST is Response Evaluation Criteria in Solid Tumor and there are two versions- 1.0 and 1.1. Most recent studies follow the recent version, RECIST 1.1, released on October 2008. The paper will follow RECIST 1.1. LESION A lesion is any abnormality in the tissue of an organism and can be described as a cut, an injury, an infected area or a tumor. In oncology study, lesions are tumors. They are divided into three types for the purpose of their measurements. • Target lesions • Non-target lesions • New lesions In clinical trials, lesions are categorized as measurable or non-measurable at baseline. A lesion is considered measurable if its length of longest diameter is longer than 10 mm by CT scan, 10 mm caliper measurement by clinical exam or 20 mm by Chest X-ray. A lesion is considered non-measurable if its length is less than 10 mm by CT scan or truly non-measurable. The measurement of lesions helps to determine whether lesions are target or non-target at baseline in oncology studies. . TARGET LESIONS In clinical trials following RECITS 1.1, target lesions selection at baseline follows as • Measurable • 5 lesions total • Maximum of 2 lesions per organ • Representing all involved organs • Quantitative measurements o Longest diameter of lesions o Short axis of Lymph nodes o Sum of diameters (both lesions and lymph nodes) 1
- 2. PhUSE 2013 NON TARGET LESIONS Non-target lesions selection at baseline follows as • All other lesions beside target lesions • Qualitative measurements – present, absent or unequivocal progression. NEW LESIONS New lesions at post-baseline follows as • any lesions that are newly found at post-baseline. • Either quantitative or qualitative measurements RESPONSE CRITERIA There are six types of responses of tumor lesions to treatments • CR – Complete Response • PR – Partial Response • SD – Stable Disease • PD – Progression Disease • NE – Not Evaluable • NonCR/NonPD - Non Complete Response/Non Progressive Disease And, overall response criteria are determined by the responses of target lesions, non-target lesions and new lesions. RESPONSE CRITERIA OF TARGET LESIONS • Complete Response (CR) – Disappearance of all target lesions • Partial Response (PR) – 30 % decrease in the sum of diameters from baseline • Stable Disease (SD) • Progression Disease (PD) – 20 % increase from nadir, the smallest sum of diameter (at least more than 5 mm) • Not Evaluable (NE) RESPONSE CRITERIA OF NON TARGET LESIONS • Complete Response (CR) – Disappearance of all non-target lesions • Non Complete Response/Non Progressive Disease (NonCR/NonPD) • Progression Disease (PD) - Unequivocal progression (an overall level of substantial worsening in non-target diseases) • Not Evaluable (NE) OVERALL RESPONSE AT GIVEN TIME POINTS According to response criteria of target lesions, non-target lesions and new lesions, overall responses at given time point can be derived. Table 1 - Time point response: patients with target (+/-non-target) disease in RECIST 1.1 Target Lesions Non-target Lesions New Lesions Overall Response CR CR No CR CR NonCR/NonPD No PR CR NE No PR PR NonCR/NonPD or NE No PR SD NonCR/NonPD or NE No SD NE NonCR/NonPD No NE PD Any Yes or No PD Any PD Yes or No PD Any Any Yes PD BEST OVERALL RESPONSE Best Overall Response for each subject can be derived in two different ways. 2
- 3. PhUSE 2013 CR or PR confirmation is not required The best one will be selected from all the overall responses. CR or PR confirmation is required Usually, the confirmation is required for non-randomized trials where response is the primary endpoint so that the responses are not measurement errors. The repeat assessment will be performed after the criteria for response are first met. The interval should be longer than 4 weeks, but it could depend on study. Table 2 – Best overall response when confirmation of CR and PR required in RECIST 1.1 Overall response First time point Overall response Subsequent time point BEST overall response CR CR CR CR PR SD, PD or PR CR SD SD provided minimum criteria for SD duration met, PD CR PD SD provided minimum criteria for SD duration met, PD CR NE SD provided minimum criteria for SD duration met, PD PR CR PR PR PR PR PR SD SD PR PD SD provided minimum criteria for SD duration met, PD PR NE SD provided minimum criteria for SD duration met, NE NE NE NE Stable Disease duration is usually from the start of the treatment (in randomization trials, from date of randomization) to the progression disease. ENDPOINS IN SOLID TUMOR The most common efficacy end points in Solid Tumor studies are • OS (Overall Survival) o Time from randomization until death. o The most common and reliable cancer endpoint. • ORR (Objective Response Rate) o Rate of partial and complete responses to non-responses. o until 1970s, FDA usually approved cancer drugs based on ORR. • DFS (Disease Free Survival) o Time from randomization until death or recurrence of tumor. • PFS (Progression Free Survival) o Time from randomization until objective tumor progression or death o Death is NOT censored. • TTP (Time to Progression) o Time from randomization until objective tumor progression o Death is censored. ORR, DFS, PFS and TTP are based on tumor assessments. CIDSC TUMOR DOMAIN CDISC recently developed Oncology-related domains both in SDTM and ADaM. SDTM • TU : Tumor Identification • TR : Tumor Results • RS : Response ADaM • ADTTE : Time to Event ADaM datasets 3
- 4. PhUSE 2013 AN EXAMPLE USING RECIST 1.1 AND NEW TUMOR CDISC DOMAIN Below examples are the randomized and open label Phase II study. It measures the solid tumor using RECIST 1.1 and has 5 cycles. The primary efficacy is objective response rate and secondary efficacies are time to progression and progression free survival. At screening, there are 3 target and 3 non-target lesions • Table 3 will introduce SDTM tumor domains (TU, TR, RS) and how SDTM domains are used to collect tumor results and responses. • Table 4 will show how ADaM datasets are used to determine tumor results for response criteria. • Table 5 will show how ADaM datasets are used to determine responses from tumor results ADaM datasets. • Table 6 will show how objective response parameter is derived using ADaM datasets for Objective Response Rate (ORR) • Table 7 will show how new time to event ADaM dataset model helps to create time to progression (TTP) and progression free survival (PFS) datasets. Table 3.1 - SDTM TU (Tumor identification) dataset USUBJID TULINKID TUTESTCD TUTEST TUORRES TULOC TUMETHOD 001-01-001 T01 TUMIDENT Tumor Identification TARGET ABODOMEN CT SCAN 001-01-001 T02 TUMIDENT Tumor Identification TARGET ABODOMEN CT SCAN 001-01-001 T03 TUMIDENT Tumor Identification TARGET THYROID CT SCAN 001-01-001 NT01 TUMIDENT Tumor Identification NON-TARGET LIVER CT SCAN 001-01-001 NT02 TUMIDENT Tumor Identification NON-TARGET KIDNEY CT SCAN 001-01-001 NT03 TUMIDENT Tumor Identification NON-TARGET SPLEEN CT SCAN Key points to note in the example are: • Subject 001 has 3 target and 3 non-target lesions throughout the studies. Table 3.2 - SDTM RELREC (Related Records ) dataset STUDYID RDOMAIN USUBJID IDVAR IDVARVAL RELTYPE RELID 001 TU TULINKID ONE 001 TR TRLINKID MANY Key points to note in the example are: • TU.TULINKID is connected to TR.TRLINKID Table 3.3 - SDTM TR (Tumor Results) dataset USUBJID TRGRID TRLINKID TRTESTCD TRTEST TRCAT TRORRES TRORRESU VISIT TRDTC TRBLFL 001-01-001 Target T01 LDIAM Longest Diameter Measurement 23 mm Screening 2011-01-01 Y 001-01-001 Target T02 LDIAM Longest Diameter Measurement 22 mm Screening 2011-01-01 Y 001-01-001 Target T03 LDIAM Longest Diameter Measurement 25 mm Screening 2011-01-01 Y 4
- 5. PhUSE 2013 USUBJID TRGRID TRLINKID TRTESTCD TRTEST TRCAT TRORRES TRORRESU VISIT TRDTC TRBLFL 001-01-001 Target SUMDIAM Sum of Diameter Measurement 70 mm Screening 2011-01-01 Y 001-01-001 Non-Target NT01 TUMSTATE Tumor State Qualitative PRESENT Screening 2011-01-01 Y 001-01-001 Non-Target NT02 TUMSTATE Tumor State Qualitative PRESENT Screening 2011-01-01 Y 001-01-001 Non-Target NT03 TUMSTATE Tumor State Qualitative PRESENT Screening 2011-01-01 Y 001-01-001 Target T01 LDIAM Longest Diameter Measurement 10 mm Cycle 1 2011-03-01 001-01-001 Target T02 LDIAM Longest Diameter Measurement 10 mm Cycle 1 2011-03-01 001-01-001 Target T03 LDIAM Longest Diameter Measurement 15 mm Cycle 1 2011-03-01 001-01-001 Target SUMDIAM Sum of Diameter Measurement 35 mm Cycle 1 2011-03-01 001-01-001 Non-Target NT01 TUMSTATE Tumor State Qualitative PRESENT Cycle 1 2011-03-01 001-01-001 Non-Target NT02 TUMSTATE Tumor State Qualitative PRESENT Cycle 1 2011-03-01 001-01-001 Non-Target NT03 TUMSTATE Tumor State Qualitative PRESENT Cycle 1 2011-03-01 Key points to note in the example are: • Target lesions were measured quantitatively and non-target lesions qualitatively. • Row 1 to 3 : There are three target lesions at Baseline. • Row 5 to 7 : There are three non-target lesions at Baseline. • Row 8 to 10 : There are still three target lesions at Cycle 1. • Row 12 to 14 : There are still three non-target lesions at Cycle 1. • Row 4 and 10 : Sum of Diameter was collected, not derived. • Row 11 : Sum of Diameter in target lesions changed from 70 mm at Screening to 35 mm at Cycle 1. • Row 12 to 14 : No changes in non-target lesions. • No new lesions were found at Cycle 1. Table 3.4 - SDTM RS (Response) dataset USUBJID RSTESTCD RSTEST RSCAT RSORRES VISIT RSDTC RSSEQ 001-01-001 TRGRESP Target Response RECIST 1.1 PR Cycle 1 2011-03-01 1 001-01-001 NTRGRESP Non-target Response RECIST 1.1 NonCR/NonPD Cycle 1 2011-03-01 2 001-01-001 OVRLRESP Overall Response RECIST 1.1 PR Cycle 1 2011-03-01 3 001-01-001 TRGRESP Target Response RECIST 1.1 SD Cycle 2 2011-06-01 4 001-01-001 NTRGRESP Non-target Response RECIST 1.1 NonCR/NonPD Cycle 2 2011-06-01 5 5
- 6. PhUSE 2013 USUBJID RSTESTCD RSTEST RSCAT RSORRES VISIT RSDTC RSSEQ 001-01-001 OVRLRESP Overall Response RECIST 1.1 SD Cycle 2 2011-06-01 6 001-01-001 TRGRESP Target Response RECIST 1.1 SD Cycle 3 2011-09-01 7 001-01-001 NTRGRESP Non-target Response RECIST 1.1 NonCR/NonPD Cycle 3 2011-09-01 8 001-01-001 OVRLRESP Overall Response RECIST 1.1 SD Cycle 3 2011-09-01 9 001-01-001 TRGRESP Target Response RECIST 1.1 PR Cycle 4 2011-12-01 10 001-01-001 NTRGRESP Non-target Response RECIST 1.1 NonCR/NonPD Cycle 4 2011-12-01 11 001-01-001 OVRLRESP Overall Response RECIST 1.1 PR Cycle 4 2011-12-01 12 001-01-001 TRGRESP Target Response RECIST 1.1 PD Cycle 5 2012-03-01 13 001-01-001 NTRGRESP Non-target Response RECIST 1.1 NonCR/NonPD Cycle 5 2012-03-01 14 001-01-001 OVRLRESP Overall Response RECIST 1.1 PD Cycle 5 2012-03-01 15 Key points to note in the example are: • Row 1 : The sum of diameter of target lesions changed from 70 mm to 35 mm, 50 % decrease in sum of diameter at Cycle 1 from baseline. So, it is PR for target lesions. • Row 2 : At baseline, 3 non-target lesions exist and at Cycle 1, 3 non-target lesions still exist, so it is NonCR/NonPD at Cycle 1. • Row 3 : According to table 1, if target lesion is PR, non-target lesion is NonPD/NonCR and no new lesions, then overall response is PR at Cycle 1. • Row 6, 9, 12 and 15 also follow the same patterns. Overall responses are derived according to table 1. ADaM DATASETS FOR TUMOR RESULTS AND OVERALL RESPONSE All the data in SDTM RS (Response) are collected in eCRF, but sometime the programmers should derive response data or the programmers should check response data that are determined by the investigator. The following table 4.1 to 4.4 will show how the programmers can derive tumor results data using ADaM dataset and the table 5.1 to 5.3 will show how the programmers determine the overall response using ADaM dataset. ADTR (Tumor Results Analysis Dataset) ADTR dataset uses CRIT1 to CRIT6 to indicate the response. Table 4.1 - Analysis Dataset Metadata for ADTR Dataset Name Dataset Description Dataset Location Dataset Structure Key Variables of Dataset Class of Dataset Documentation ADTR Tumor Results Analysis Data adtr.xpt one record per subject per parameter per analysis visit USUBJID, PARAMCD, AVISITN BDS c-adtr.txt Table 4.2 - Analysis Variable Metadata including Analysis Parameter Value-Level Metadata for ADTR 6
- 7. PhUSE 2013 Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADTR *ALL* USUBJID Unique Subject Identifier text $20 ADSL.USUBJID ADTR *ALL* SITEID Site ID text $20 ADSL.SITEID ADTR *ALL* SEX Sex text $20 M, F ADSL.SEX ADTR *ALL* FASFL Full Analysis Set Population Flag text $1 Y, N ADSL.FASFL ADTR *ALL* TRTPN Planned Treatment (N) integer 1.0 1 = Control, 2 = Study Drug ADSL.TRTPN ADTR *ALL* TRTP Planned Treatment text $20 Control, Study Drug ADSL.TRTP ADTR LDIAM1, LDIAM2, LDIAM3, SUMDIA,SD FRSM PARCAT1 Parameter Category 1 text $50 TARGET LESIONS ADTR TUMSTAT1, TUMSTAT2, TUMSTAT3, NUMNTG PARCAT1 Parameter Category 1 text $50 NON-TARGET LESIONS ADTR NEWLES PARCAT1 Parameter Category 1 text $50 NEW LESIONS ADTR PARAMCD PARAMCD Parameter Code text $8 LDIAM1, LDIAM2, LDIAM3, SUMDIA, SDFRSM, TUMSTAT1, TUMSTAT2, TUMSTAT3, NUMNTG, NEWLES 7
- 8. PhUSE 2013 Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADTR *ALL* PARAM Parameter text $50 Longest Diameter of Target 1 (mm), Longest Diameter of Target 2 (mm), Longest Diameter of Target 3 (mm), Sum of Diameter (mm), Sum of Diameter from smallest Sum of Diameter (mm), Tumor State of Non-Target 1, Tumor State of Non-Target 2, Tumor State of Non-Target 3, Number of Present Non-Target Lesion, New Lesion ADTR SDFRSM, NUMNTG, NEWLES PARAMTYP Parameter Type text $20 DERIVED ADTR *ALL* AVISITN Analysis Visit (N) integer 3.0 1 = Baseline, 2 = Cycle 1, 3=Cycle 2, 4=Cycle 3, 5=Cycle 4, 6=Cycle 5 TR.VISITNUM ADTR *ALL* AVISIT Analysis Visit text $20 Baseline, Cycle 1, Cycle 2, Cycle 3, Cycle 4, Cycle 5 TR.VISIT ADTR *ALL* ABLFL Baseline Record Flag text $1 Y TR.TRBLFL ADTR LDIAM1, LDIAM2, LDIAM3, SUMDIA AVAL Analysis Value float 8.2 TR.TRSTRESN ADTR SDFRSM AVAL Analysis Value float 8.2 TR.TRSTRESN at TRTESTCD=’SUMDIA’ ADTR TUMSTAT1, TUMSTAT2, TUMSTAT3 AVALC Analysis Value (C ) text $20 TR.TRSTRESC ADTR NUMNTG AVAL Analysis Value float 8.2 Count(AVALC=’PRESENT’) for non-target lesion ADTR NEWLES AVALC Analysis Value (C ) text $1 Y, N ‘Y’ if any TR.TRGRPID=’NEW’ ‘N’ if no TR.TRGRPID = ‘NEW’ 8
- 9. PhUSE 2013 Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADTR SUMDIA BASE Baseline Value float 8.2 AVAL at ABLFL=’Y’ ADTR SDFRSM BASE Baseline Value float 8.2 Previous smallest AVAL at PARAMCD=’SUMDIA’ ADTR SUMDIA, SDRFSM CHG Change from Baseline float 8.2 BASE - AVAL ADTR SUMDIA, SDFRSM PCHG Percent Change from Baseline float 8.2 (BASE – AVAL)/BASE ADTR SUMDIA CRIT1 Analysis Criteria 1 text $50 AVAL = 0 ADTR SUMDIA CRIT2 Analysis Criteria 2 text $50 PCHG < -30 ADTR SDFRSM CRIT3 Analysis Criteria 3 text $50 PCHG > 120 and CHG > 5 ADTR NUMNTG CRIT4 Analysis Criteria 4 text $50 AVAL = 0 ADTR TUMSTAT1, TUMSTAT2, TUMSTAT3 CRIT5 Analysis Criteria 5 text $50 AVALC = ‘UNEQUIVOCAL’ ADTR NEWLES CRIT6 Analysis Criteria 6 text $50 AVALC = ‘Y’ ADTR SUMDIA CRIT1FL Criteria 1 Evaluation Result Flag text $1 Y or Null ‘Y’ if CRIT1 is satisfied. ADTR SUMDIA CRIT2FL Criteria 2 Evaluation Result Flag text $1 Y or Null ‘Y’ if CRIT2 is satisfied. ADTR SDFRSM CRIT3FL Criteria 3 Evaluation Result Flag text $1 Y or Null ‘Y’ if CRIT3 is satisfied. ADTR TUMSTAT1, TUMSTAT2, TUMSTAT3 CRIT4FL Criteria 4 Evaluation Result Flag text $1 Y or Null ‘Y’ if CRIT4 is satisfied. ADTR NUMNTG CRIT5FL Criteria 5 Evaluation Result Flag text $1 Y or Null ‘Y’ if CRIT5 is satisfied. ADTR NEWLES CRIT6FL Criteria 6 Evaluation Result Flag text $1 Y or Null ‘Y’ if CRIT6 is satisfied. ADTR *ALL* TRSEQ Sequence Number float 8.2 TR.TRSEQ Key points to note in the example are: • Target Lesions o SUMDIA (Sum of Diameter) 9
- 10. PhUSE 2013 if CRIT1FL = ‘Y’, overall response on target lesions is CR. if CRIT2FL = ‘Y’, overall response on target lesions is PR. o SDFRSM (Sum of Diameter from Nadir) : if CRIT3FL = ‘Y’, overall response on target lesions is PD. o If CRIT1FL, CRIT2FL, and CRIT3FL are all null, overall response on target lesions is SD. • Non Target Lesions o NUMNTG : if CRIT4FL = ‘Y’, overall response on non-target lesions is CR. o TUMSTAT1, TUMSTAT2, and TUMSTAT3 : if CRIT5FL=’Y’, overall response on non-target lesions is PD. o if CRIT4FL and CRIT5FL are all null, overall response on non-target lesions is NonCR/NonPD. • New Lesions o NEWLES : if CRIT6FL = ‘Y’, overall response on new lesions is Y ; if not, overall response on new lesions is N. Table 4.3 – ADTR dataset at Cycle 1 USUBJID TRTP PARCAT1 PARAM PARAMT YP AVISIT AVA L BAS E CH G PCH G AVALC BASEC CRIT 1FL CRIT 2FL CRIT 3FL CRIT 4FL CRIT 5FL CRIT 6FL 001-01- 001 Study Drug TARGET LESIONS Longest Diameter of Target 1 (mm) Cycle 1 10 23 001-01- 001 Study Drug TARGET LESIONS Longest Diameter of Target 2 (mm) Cycle 1 10 22 001-01- 001 Study Drug TARGET LESIONS Longest Diameter of Target 3 (mm) Cycle 1 15 25 001-01- 001 Study Drug TARGET LESIONS Sum of Diameter (mm) Cycle 1 35 70 -35 -50 Y 001-01- 001 Study Drug TARGET LESIONS Sum of Diameter from smallest Sum of Diameter (mm) DERIVED Cycle 1 35 70 -35 -50 001-01- 001 Study Drug NON- TARGET LESIONS Tumor State of Non- Target 1 Cycle 1 PRESE NT PRESE NT 001-01- 001 Study Drug NON- TARGET LESIONS Tumor State of Non- Target 2 Cycle 1 PRESE NT PRESE NT 001-01- 001 Study Drug NON- TARGET LESIONS Tumor State of Non- Target 3 Cycle 1 PRESE NT PRESE NT 001-01- 001 Study Drug NON- TARGET LESIONS Number of Non- Target Lesion DERIVED Cycle 1 3 001-01- 001 Study Drug NEW LESIONS New Lesion DERIVED Cycle 1 N 10
- 11. PhUSE 2013 Key points to note in the example are: • Row 4 : CRIT2FL = ‘Y’ since PCHG < -30, so this will lead to PR for target lesion at Cycle 1. ADRS (Tumor Response Analysis Dataset) Table 5.1 - Analysis Dataset Metadata for ADRS Dataset Name Dataset Description Dataset Location Dataset Structure Key Variables of Dataset Class of Dataset Documentation ADRS Response Analysis Data adrs.xpt one record per subject per parameter per analysis visit USUBJID, PARAMCD, AVISITN BDS c-adrs.txt Table 5.2 - Analysis Variable Metadata including Analysis Parameter Value-Level Metadata for ADRS Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADRS *ALL* USUBJID Unique Subject Identifier text $20 ADSL.USUBJID ADRS *ALL* SITEID Site ID text $20 ADSL.SITEID ADRS *ALL* SEX Sex text $20 M, F ADSL.SEX ADRS *ALL* FASFL Full Analysis Set Population Flag text $1 Y, N ADSL.FASFL ADRS *ALL* TRTPN Planned Treatment (N) integer 1.0 1 = Control, 2 = Study Drug ADSL.TRTPN ADRS *ALL* TRTP Planned Treatment text $20 Control, Study Drug ADSL.TRTP ADRS PARAMCD PARAMCD Parameter Code text $8 TRGRESP, NTRGRESP, NEWRESP, OVRLRESP ADRS *ALL* PARAM Parameter text $50 Target Response, Non-target Response, New Lesion Progression, Overall Response ADRS TRGRESP, NTRGRESP, NEWLPROG, OVRLRESP PARAMTYP Parameter Type text $20 DERIVED ADRS *ALL* AVISITN Analysis Visit (N) integer 3.0 1 = Screening, 2 = Cycle 1, 3=Cycle 2, 4=Cycle 3, 5=Cycle 4, 6=Cycle 5 ADTR.VISITNUM ADRS *ALL* AVISIT Analysis Visit text $20 Screening, Cycle 1, Cycle 2, Cycle 3, Cycle 4, Cycle 5 ADTR.VISIT 11
- 12. PhUSE 2013 Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADRS TRGRESP AVALC Analysis Value (C ) text $20 CR, PR, PD, SD, NE CR if ADTR.CRIT1FL=‘Y’, PR if ADTR.CRIT2FL=’Y’, PD if ADTR.CRIT3FL=’Y’, SD if CRIT1FL, CRIT2FL and CRIT3FL are null ADRS NTRGRESP AVALC Analysis Value (C ) text $20 CR, NonCR/NonPD, PD, NE CR if ADTR.CRIT4FL=’Y’, PD if ADTR.CRIT5FL=’Y’, NonCR/NonPD if CRIT4FL and CRIT5FL are all null ADRS NEWLPROG AVALC Analysis Value (C ) text $1 Y, N Y if ADTR.CRIT6FL=’Y’, Otherwise N ADRS OVRLRESP AVALC Analysis Value (C ) text $20 CR, PR, NonCR/NonPD, PD, SD, NE See overall response table 1 and 2 in RECIST 1.1 Key points to note in the example are: • ADTR should be derived before ADRS since ADRS is using ADTR. Table 5.3 – ADRS dataset USUBJID TRTP PARAM PARAMTYP AVISIT AVALC 001-01-001 Study Drug Target Response DERIVED Cycle 1 PR 001-01-001 Study Drug Non-target Response DERIVED Cycle 1 NonCR/NonPD 001-01-001 Study Drug New Lesion Progression DERIVED Cycle 1 N 001-01-001 Study Drug Overall Response DERIVED Cycle 1 PR Key points to note in the example are: • Row 1 : Since ADTR.CRIT2FL = ‘Y’, Target Response is PR (Partial Response). • Row 2 : Since ADTR.CRIT4FL and ADTR.CRIT5FL are null, Non-Target Response is NonCR/NonPD (Non Complete Response/Non Progressive Disease) • Row 3 : since ADTR.CRIT6FL is null, New Lesion Progression is N. • Row 4 : Using RECIST 1.1 Overall Response table, overall response at Cycle 1 is PR (Partial Response). BEST OVERALL RESPONSE FOR ORR (OBJECTIVE RESPONSE RATE) For ORR (Objective Response Rate) analysis as a primary endpoint, objective response should be derived. Usually, the FDA defines PR (Partial Response) or CR (Complete Response) as objective response to the treatment. Best overall response will be used to derive objective response. Best overall response can be selected 12
- 13. PhUSE 2013 as the best response among all responses. The best overall response does not worsen over time – if a subject achieves CR at cycle 3 and PD at cycle 5, the best overall response is still CR. ADORS (Objective Response Analysis Dataset) ADORS will be derived from SDTM RS domain, not from ADRS. Table 6.1 - Analysis Dataset Metadata for ADORS Dataset Name Dataset Description Dataset Location Dataset Structure Key Variables of Dataset Class of Dataset Documentation ADORS Objective Response Analysis Data adors.xpt one record per subject per parameter per analysis visit USUBJID, PARAMCD, AVISITN BDS c-adors.txt Table 6.2 - Analysis Variable Metadata including Analysis Parameter Value-Level Metadata for ADORS Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADORS *ALL* USUBJID Unique Subject Identifier text $20 ADSL.USUBJID ADORS *ALL* SITEID Site ID text $20 ADSL.SITEID ADORS *ALL* SEX Sex text $20 M, F ADSL.SEX ADORS *ALL* FASFL Full Analysis Set Population Flag text $1 Y, N ADSL.FASFL ADORS *ALL* TRTPN Planned Treatment (N) integer 1.0 1 = Control, 2 = Study Drug ADSL.TRTPN ADORS *ALL* TRTP Planned Treatment text $20 Control, Study Drug ADSL.TRTP ADORS PARAMCD PARAMCD Parameter Code text $8 OVRLRESP, BESTRESP, OBJRESP ADORS *ALL* PARAM Parameter text $50 Overall Response, Best Overall Response, Objective Response ADORS BESTRESP, OBJRESP PARAMTYP Parameter Type text $20 DERIVED ADORS *ALL* AVISITN Analysis Visit (N) integer 3.0 1 = Screening, 2 = Cycle 1, 3=Cycle 2, 4=Cycle 3, 5=Cycle 4, 6=Cycle 5 ADORS *ALL* AVISIT Analysis Visit text $20 Screening, Cycle 1, Cycle 2, Cycle 3, Cycle 4, Cycle 5 13
- 14. PhUSE 2013 Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADORS OVRLRESP AVALC Analysis Value (C ) text $20 RS.RSORRES when RSTESTCD=’OVRLRESP’ ADORS BESTRESP AVALC Analysis Value (C ) text $20 Best of AVALC at PARAMCD=’OVRLRESP’ ADORS OBJRESP AVALC Analysis Value (C ) text $20 ‘Y’ if AVALC at PARAM=’Best Overall Response’ is ‘CR’ or ‘PR’ ‘N’ otherwise ADORS *ALL* RSSEQ Sequence Number float 8.2 RS.RSSEQ Table 6.3 – ADORS dataset when the confirmation is not needed. USUBJID TRTP PARAM PARAMTYP AVISIT AVALC RSSEQ 001-01-001 Study Drug Overall Response Cycle 1 PR 3 001-01-001 Study Drug Overall Response Cycle 2 SD 6 001-01-001 Study Drug Overall Response Cycle 3 SD 9 001-01-001 Study Drug Overall Response Cycle 4 PR 12 001-01-001 Study Drug Overall Response Cycle 5 PD 15 001-01-001 Study Drug Best Overall Response DERIVED End of Study PR Key points to note in the example are: • Row 6 : At Cycle 1 and 4, the subject have PR, but at cycle 5, he has PD. But since it is best overall response, best overall response for the subject will be PR. Table 6.4 – ADORS dataset when the confirmation of CR and PR is needed. USUBJID TRTP PARAM PARAMTYP AVISIT ADT AVALC RSSEQ _NAVALC _DUR _COR 001-01-001 Study Drug Overall Response Cycle 1 2011-03-01 PR 3 SD 61 SD 001-01-001 Study Drug Overall Response Cycle 2 2011-06-01 SD 6 SD 61 SD 001-01-001 Study Drug Overall Response Cycle 3 2011-09-01 SD 9 PR 61 SD 001-01-001 Study Drug Overall Response Cycle 4 2011-12-01 PR 12 PD 61 SD 001-01-001 Study Drug Overall Response Cycle 5 2012-03-01 PD 15 PD 001-01-001 Study Drug Best Overall Response DERIVED End of Study PR 001-01-001 Study Drug Objective Response DERIVED End of Study N 14
- 15. PhUSE 2013 Key points to note in the example are: • Temporary ADaM plus variables o _NAVALC : Subsequent Analysis Value o _DUR : Duration (Days) from first time point to subsequent time point o _COR : Confirmed Overall Response • Row 1 : According to table 2, since PR at the first time point (AVALC = ‘PR’), SD at the subsequent time point (_NAVALC=’SD’) and duration is bigger than 4 weeks (_DUR=61), the best overall response at cycle 1 is SD (_COR=’SD’) • Row 2 : According to table 2, since SD at the first time point (AVALC = ‘SD’), PD at the subsequent time point (_NAVALC=’SD’) and duration is bigger than 4 weeks (_DUR=61), the best overall response at cycle 2 is SD (_COR=’SD’) • Row 3 : According to table 2, since SD at the first time point (AVALC = ‘SD’), PD at the subsequent time point (_NAVALC=’PR’) and duration is bigger than 4 weeks (_DUR=61), the best overall response at cycle 3 is SD (_COR=’PR’) • Row 4 : According to table 2, since PR at the first time point (AVALC = ‘PR’), PD at the subsequent time point (_NAVALC=’PD’) and duration is bigger than 4 weeks (_DUR=61), the best overall response at cycle 4 is SD (_COR=’SD’) • Row 6 : Best overall response is SD at cycle 1, 2 and 3. So, the best overall response for subject 001 is SD. Table 6.5 – Final ADORS dataset with Objective Response parameter USUBJID TRTP PARAM PARAMTYP AVISIT ADT AVALC RSSEQ 001-01-001 Study Drug Overall Response Cycle 1 2011-03-01 PR 3 001-01-001 Study Drug Overall Response Cycle 2 2011-06-01 SD 6 001-01-001 Study Drug Overall Response Cycle 3 2011-09-01 SD 9 001-01-001 Study Drug Overall Response Cycle 4 2011-12-01 PR 12 001-01-001 Study Drug Overall Response Cycle 5 2012-03-01 PD 15 001-01-001 Study Drug Best Overall Response DERIVED End of Study PR 001-01-001 Study Drug Objective Response DERIVED End of Study N Key points to note in the example are: • Temporary ADaM plus variables (_NAVALC, _DUR and _COR) are dropped. • Row 7 : Objective Response parameter is derived. ADTTE(Time to Event Analysis Dataset) FOR TIME TO PROGRESSION (TTP) AND PROGRESSION FREE SURVIVAL (PFS) Table 7.1 - Analysis Dataset Metadata for ADTTE Dataset Name Dataset Description Dataset Location Dataset Structure Key Variables of Dataset Class of Dataset Documentation ADTTE Time to Event Analysis Data adtte.xpt one record per subject per parameter USUBJID, PARAMCD BDS c-adtte.txt 15
- 16. PhUSE 2013 Table 7.2 - Analysis Variable Metadata including Analysis Parameter Value-Level Metadata for ADTTE Dataset Name Parameter Identifier Variable Name Variable Label Variable Type Display Format Code list / Controlled Terms Source / Derivation ADTTE *ALL* USUBJID Unique Subject Identifier text $20 ADSL.USUBJID ADTTE *ALL* SITEID Site ID text $20 ADSL.SITEID ADTTE *ALL* SEX Sex text $20 M, F ADSL.SEX ADTTE *ALL* FASFL Full Analysis Set Population Flag text $1 Y, N ADSL.FASFL ADTTE *ALL* TRTPN Planned Treatment (N) integer 1.0 1 = Control, 2 = Study Drug ADSL.TRTPN ADTTE *ALL* TRTP Planned Treatment text $20 Control, Study Drug ADSL.TRTP ADTTE PARAMCD PARAMCD Parameter Code text $8 TTP, PFS ADTTE *ALL* PARAM Parameter text $50 Time to Progression (Days), Progression to Survival (Days) ADTTE *ALL* PARAMTYP Parameter Type text $20 DERIVED ADTTE *ALL* STARTDT Time to Event Origin Date for Subject integer YYYY-MM- DD ADSL.RANDDT ADTTE *ALL* ADT Analysis Date integer YYYY-MM- DD Numeric date of DS.DSSTDTC ADTTE *ALL* AVAL Analysis Value integer 8. ADT – STARTDT + 1 ADTTE TTP CNSR Censor integer 1. 0,1 0 if EVNTDESC = ‘PROGRESSIVE DISEASE’ 1 if otherwise ADTTE PFS CNSR Censor integer 1. 0,1 0 if EVNTDESC in ( ‘PROGRESSIVE DISEASE’, ‘DEATH’) 1 if otherwise ADTTE *ALL* EVNTDESC Event or Censoring Description text $50 PROGRESSIVE DISEASE, LOST TO FOLLOW-UP, DEATH, COMPLETED, ADVERSE EVENT, WITHDRAWAL BY SUBJECT DS.DSDECOD 16
- 17. PhUSE 2013 Table 7.3 - ADTTE dataset of Time to Progression(TTP) parameter USUBJID TRTP PARAM AVAL STARTDT ADT CNSR EVNTDESC 001-01-001 Study Drug Time to Progression (Days) 452 2011-01-01 2011-03-01 0 PROGRESSIVE DISEASE 001-01-002 Control Time to Progression (Days) 338 2011-02-01 2011-01-05 1 LOST TO FOLLOW-UP 001-01-003 Control Time to Progression (Days) 212 2011-02-05 2011-09-05 1 DEATH 001-01-004 Study Drug Time to Progression (Days) 463 2011-03-20 2012-06-25 1 COMPLETED 001-01-005 Study Drug Time to Progression (Days) 67 2011-03-26 2011-06-01 0 PROGRESSIVE DISEASE Key points to note in the example are: • Row 3 : In TTP, DEATH is censored Table 7.4 - ADTTE dataset of Progression Free Survival (PFS) parameter USUBJID TRTP PARAM AVAL STARTDT ADT CNSR EVNTDESC 001-01-001 Study Drug Progression Free Survival (Days) 452 2011-01-01 2011-03-01 0 PROGRESSIVE DISEASE 001-01-002 Control Progression Free Survival (Days) 338 2011-02-01 2011-01-05 1 LOST OT FOLLOW-UP 001-01-003 Control Progression Free Survival (Days) 212 2011-02-05 2011-09-05 0 DEATH 001-01-004 Study Drug Progression Free Survival (Days) 463 2011-03-20 2012-06-25 1 COMPLETED 001-01-005 Study Drug Progression Free Survival (Days) 67 2011-03-26 2011-06-01 0 PROGRESSIVE DISEASE Key points to note in the example are: • Row 3 : In PFS, DEATH is NOT censored 17
- 18. PhUSE 2013 CONCLUSION CDISC introduces the new tumor domains in SDTM and Time to Event datasets in ADaM, but it is also very important for programmers and statisticians to understand therapeutic characteristics and methods in Oncology. RECIST 1.1 introduces how to collect tumor data and analyze responses in solid tumor oncology studies. The combination of both RECIST and CDISC can streamline data collection and analysis in solid tumor oncology studies. REFERENCES New response evaluation criteria in solid tumors: Revised RECIST guideline (version 1.1) ADaM Implementation Guide, Version V 1.0 (ADaMIG v1.0) Analysis Data Model, Version 2.1 (ADaM 2.1) ADaM Basic Data Structure for Time-to-Event Analyses SDTM Version 1.3 SDTM Implementation Guide Version 3.1.3 SDTM Supplement Tumor Domains CDER Guidance for Industry – Clinical Trial Endpoints for the approval of Cancer Drugs and Biologics ACKNOWLEDGMENTS I would like to thank Cytel friends for their feedback, inputs and review. CONTACT INFORMATION (HEADER 1) Your comments and questions are valued and encouraged. Contact the author at: Kevin Lee Cytel, Inc. Chesterbrook, PA (610) 994 - 9840 Email:Kevin.lee@cytel.com Tweet : @kevinlee_pharma @cdisc_adam Linkediin : www.linkedin.com/in/kevinlee1995/ 18