PCPT and SELECT Cancer Cohorts Briefly describe PCPT and SELECT trial designs Current follow-up status for each cohort SELECT Centralized Follow-up Data elements and biologic samples collected Events ascertained, verification Process for Proposing Research Using Cohorts Cathy Tangen and Phyllis Goodman, 3/24/14, Survivorship Affinity Group Presentation
PCPT Schema 1993-1997 1994-1997 Enrollment Randomize Finasteride 5 mg Placebo Interim Prostate Cancer Lost to F/U Died Refused Biopsy For-cause Biopsy (PSA or DRE) Interim Prostate Cancer Lost to F/U Died Refused Biopsy 2001-2003 EOS Biopsy 7-year Period Prevalence EOS Biopsy 7-year Period Prevalence
SELECT Schema Calendar Year 2001 2004 (Planned 2001 2006) Pre-Randomization Period Randomized Vitamin E + Vitamin E + Placebo + Placebo + Calendar Year 2001-2013 Selenium Placebo Selenium Placebo Follow - up Prostate cancer, other cancer, death
Current Follow-up Status - PCPT Cancer endpoint assessment and follow-up for entire trial ended 12/03. Tried to conduct a long-term follow-up of men diagnosed with prostate cancer but proved infeasible accrual-wise. Closed 5/09 Recent SSDI search (98% of participants reported SSN) led to recent survival publication of PCPT, 3422 deaths found. Missing piece of puzzle: cause of death. Plans for NDI search, have funding. Plans to link CDMS database to evaluate longterm effects of finasteride and other research aims.
Current Status of Follow-up for SELECT Study site closure 6/10-4/11, close out visit Move subjects into centralized cohort followed by SWOG Statistical Center 17,747 transitioned to cohort (n=35,533 randomized) Challenge getting revised consent from participant locally and direct contact info given to Statistical Center Data collected by mailed booklets initially: cancer diagnosis, prostate cancer treatment, ancillary studies follow-up, source documents and prostate tissue requested June 2011 on-line data submission implemented Phone call primary means of contact now contracted interviewers Wrapping up final calls now, ending summer 2014. NDI search, CDMS merge plans
Baseline Characteristics PCPT (n=18,880) SELECT Trial (n=35,533) SELECT CFU (n=17,747) N % N % N % Age at enrollment (median [IQR],y) 63 [59-67] 62 [58-67] 62 [58-67] 50-54 2 0.0% 1,728 4.9% 633 3.6% 55-59 5,909 31.3% 11,341 31.9% 5,933 33.4% 60-64 5,793 30.7% 9,281 26.1% 4,866 27.4% 65-69 4,331 22.9% 6,917 19.5% 3,514 19.8% 70 2,845 15.1% 6,266 17.6% 2,801 15.8% Current age (median [IQR],y) 79 [76-83] 72 [67-77] 72 [68-77] 58-69 1 0.0% 12,334 35.1% 6,385 36.0% 70-79 7,017 37.2% 14,980 42.2% 8,295 46.7% 80-89 6,292 33.3% 4,726 13.3% 2,650 14.9% 90 721 3.8% 324 0.9% 129 0.7% N/A (Deceased) 4,849 25.7% 3,169 8.9% 288 1.6% Race/Ethnicity White 17,382 92.1% 27,592 77.7% 14,753 83.1% African American 712 3.8% 4,907 13.8% 1,990 11.2% Hispanic (not AA) 481 2.5% 1,946 5.5% 555 3.1% Hispanic (AA) 12 0.1% 356 1.0% 71 0.4% Other 293 1.6% 732 2.1% 378 2.1%
Baseline Characteristics PCPT (n=18,880) SELECT Trial (n=35,533) SELECT CFU (n=17,747) N % N % N % History of prostate cancer in 1 st 2,913 15.4% 6,609 18.6% 3,453 19.5% degree relative PSA at entry (median [IQR], ng/ml) 1.1 [0.7-1.7] 1.1 [0.6-1.8] 1.1 [0.7-1.9] 0-<1.0 9,132 48.4% 17,177 48.3% 8,436 47.5% 1.0-<2.0 6,708 35.5% 10,895 30.7% 5,579 31.4% 2.0-<3.0 3,039 16.1% 4,815 13.6% 2,410 13.6% 3.0-<4.0 1 0.0% 2,646 7.4% 1,322 7.4% Education (highest level) High school graduate or GED 3,569 18.9% 8,100 22.8% 2,854 16.1% Some college/vocational school 5,509 29.2% 9,451 26.6% 4,634 26.1% College graduate 9,790 51.9% 17,613 49.6% 10,130 57.1% Cigarette smoking Never 6,218 32.9% 15,107 42.5% 8,158 46.0% Former 11,171 59.2% 17,333 48.8% 8,524 48.0% Current 1,486 7.9% 2,898 8.2% 1,010 5.7% Body mass index (kg/m 2 ) <25 4,817 25.5% 7,129 20.1% 3,571 20.1% 25-<30 9,499 50.3% 16,991 47.8% 8,692 49.0% 30 4,378 23.2% 11,198 31.5% 5,408 30.5%
PCPT: Review of Other Cancers and Deaths Cancer reported by participant to Study Site Study Site learns of participant death Source documentation obtained and submitted to Statistical Center Additional deaths from Social Security Death Index (SSDI) and planned National Death Index search Endpoint Review Committee reviews documentation and codes cancer site/cause of death Endpoint form submitted
SELECT: Data Collection for Other Cancers and Deaths Other cancers Reported by the participant to study site staff Source documentation submitted to the Statistical Center Upon the interest of an investigator, source documents can be reviewed and coded Deaths Sites collected source documentation for coding cause of death on study forms Source documentation was not submitted to the Statistical Center Social Security Death Index (SSDI) search was performed in 2012 National Death Index search planned 2014-2015
Baseline Data Collected Medical history, limited Physical Exam, height, weight, demographics PCPT SELECT Trial Quality of Life (SF-36) Subset Family history cancer Prostate only Prostate, lung, colon Anthropometric measures PSA Measured at Participant report or central facility done at study site Diet and supplement use At year 1 Physical activity
PCPT SELECT Trial SELECT CFU Frequency Every 3 months Every 6 months Annual Medical events/side effects Medications Limited Extensive Prostate health (annually) AUA symptom score PSA and DRE (annually) Follow-Up Data Collected (annually) PSA measured at central facility; DRE by study site Participant report or done at study site Updated supplement use Adherence assessment Updated smoking history Weight Quality of Life (SF-36) On subset Macular degeneration, diabetes, colorectal screen procedures Finasteride, statins, aspirin, NSAIDs PSA, treatments for men diagnosed with prostate cancer
Epidemiologic Sub-studies Nested Within PCPT and SELECT Case-control study within PCPT: Approx. 1800 prostate cancer cases and 1800 biopsy proven controls at 7 yrs. Frequency matched on treatment arm, family history of CaP and age group. Over-sampled African Americans Case-cohort within SELECT: cases= all prostate cancers, cohort stratified on age group and race (AA vs. not) was selected, sampling weights needed
Samples Collected Plasma Serum RBC WBC Prostate tissue Toenail PCPT Blood sample at baseline Annually during follow-up X x x Blood sample at baseline and Year 5 7.8% subset at 6 months, SELECT Year 1, 2, 4, 6 and 8; post-diagnostic sample for those with prostate cancer. Toenails at baseline. X x x x X WBC = white blood cell/buffy coat, RBC=red blood cell
SELECT: Cancers and Samples Collected Cancer site Total reported Some source documentation Baseline plasma Year 5 plasma Total with Buffy Coat available specimen specimen 2 Prostate 1 2,494 2,012 2,321 1,710 2,370 Bladder 249 229 227 156 230 Colorectal 323 298 293 172 298 Lung 381 352 349 139 351 Melanoma 333 311 308 254 315 Pancreatic 118 100 104 29 104 Renal 110 101 103 76 105 Other cancers 807 730 729 414 743 Non-cancers 30,948 N/A 27,934 19,424 28,406
PCPT: Cancers and Samples Collected Serum Samples # Additional Pre-diagnostic Serum Samples DNA source Cancer site Total Baseline 0 1 2 3+ DNA Serum 1 Prostate 2,401 2,261 3 169 118 2,111 1,420 857 Bladder 146 137 10 39 20 77 71 56 Colorectal 210 196 29 43 34 104 82 101 Lung 255 235 35 57 31 132 62 131 Melanoma 158 148 9 15 9 125 90 55 Pancreatic 38 32 2 14 2 20 6 18 Renal 56 53 7 10 8 31 22 25 Other 474 450 53 86 58 277 153 246 Non-cancers 15,301 14,328 26 1,735 655 12,885 7,845 5,757 1 Among cases with no WBC or DNA available
Data Request Process Brief description of proposed data for epidemiologic or statistical research Approximately 2 pages Specific aims Data elements required PCPT/SELECT Cohorts Executive Committee Availability of data Overlap with other approved projects SWOG Executive Committee Review for regulatory issues Meets weekly Data Usage Agreement Who will have access to the individual participant data Data not to be shared with others or used for other projects
Biologic Specimens Requests SWOG webpage (www.swog.org) Visitors Biospecimens Application not to exceed 5 pages: Specific aims Background and significance Preliminary results (if available) Experimental methods and design Type and volume of samples Why are the PCPT/SELECT specimens and associated clinical data necessary Why is the use of the SELECT/PCPT biorepository needed or advantageous. PCPT/SELECT Cohorts Executive Committee Feasibility (population, study design, sample size, sample and data needs) Availability of biospecimens Overlap with other approved projects SWOG Executive Committee Further scientific review Meets weekly Materials reserved for 3 years Letter of support from SWOG for any applications for funding (i.e. R01s, DoD) SWOG will work with the investigator to obtain external funding Material Usage Agreements and Data Usage Agreements (local IRB approval)
Deaths PCPT SELECT Deaths Total 5,125 3,071 Prostate cancer 18 12 Other cancer 579 723 Cardiovascular 602 746 Other 495 671 Unknown/SSDI* 3,431 919 * Planned NDI search for cause
Summary: Endpoint Assessments Prostate cancer Other cancer PCPT SELECT Trial SELECT CFU Study site report Participant report; confirmed by central Medical records/tissue lab requested End of study biopsy and interim cancers Source docs collected, adjudicated Source docs collected, not adjudicated Participant report; Medical records requested Source docs collected, SSDI, NDI Deaths Site reports; SSDI adjudicated; SSDI Cardiovascular Participant report Participant report None
Unique Features of Cohorts Follow-up: median 9 yrs. SELECT, 17 yrs. PCPT 15% AA participation in SELECT Test and validation with both cohorts Nested case-control and case-cohort studies with measures stored at Statistical Center Longitudinal QOL assessment Careful longitudinal assessment of meds and dietary supplements on SELECT RCT underpinnings with significant treatment results may provide keys to mechanistic pathways
Resources for Future Research Upcoming National Death Index Search for PCPT and SELECT SWOG has access to CDMS database Link procedures and medications, long-term effects to both trials/cohorts. Contribute to NCI Cancer Cohort Consortium approved projects Solicit projects from scientific community, TM and data analyses, collaborate on grants WHI strategic collaborations