Data sharing in clinical trials: An experience with two large cancer screening trials.

Paul Pinsky of the US National Cancer Institute and colleagues describe the implementation and outcomes of web-based data sharing from the PLCO and NLST cancer screening trials.

The Prostate, Lung, Colorectal and Ovarian Cancer (PLCO) Screening Trial Pathology Tissue Resource.

BACKGROUND: Pathology tissue specimens with associated epidemiologic and clinical data are valuable for cancer research. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial undertook a large-scale effort to create a public resource of pathology tissues from PLCO participants who developed a cancer during the trial. METHODS: Formalin-fixed paraffin-embedded tissue blocks were obtained from pathology laboratories on a loan basis for central processing of tissue microarrays, with additional free-standing tissue cores collected for nucleic acid extraction. RESULTS: Pathology tissue specimens were obtained for prostate cancer (n = 1,052), lung cancer (n = 434), colorectal cancer (n = 675) and adenoma (n = 658), ovarian cancer and borderline tumors (n = 212), breast cancer (n = 870), and bladder cancer (n = 204). The process of creating this resource was complex, involving multidisciplinary teams with expertise in pathology, epidemiology, information technology, project management, and specialized laboratories. CONCLUSIONS: Creating the PLCO tissue resource required a multistep process, including obtaining medical records and contacting pathology departments where pathology materials were stored after obtaining necessary patient consent and authorization. The potential to link tissue biomarkers to prospectively collected epidemiologic information, screening and clinical data, and matched blood or buccal samples offers valuable opportunities to study etiologic heterogeneity, mechanisms of carcinogenesis, and biomarkers for early detection and prognosis. IMPACT: The methods and protocols developed for this effort, and the detailed description of this resource provided here, will be useful for those seeking to use PLCO pathology tissue specimens for their research and may also inform future tissue collection efforts in other settings. Cancer Epidemiol Biomarkers Prev; 25(12); 1635-42. ©2016 AACR.

Lung cancer risk by years since quitting in 30+ pack year smokers.

OBJECTIVE: Current United States recommendations for low-dose computed tomography (LDCT) lung cancer screening limit eligibility to ever-smokers with 30+ pack-years, with former smokers eligible only within 15 years of quitting. The 15 year limit is partly based on perceived decreases in lung cancer risk as years since quitting (YSQ) increase. We examine the relationship between lung cancer risk and YSQ among 30+ pack-year former smokers. METHODS: In the Prostate, Lung, Colorectal, and Ovarian trial, participants aged 55-74 were randomized to screening or usual care; screened subjects received annual chest-radiographs for lung cancer screening. Subjects completed a baseline questionnaire; smoking history included average cigarettes per day and age at starting and stopping smoking. Subjects were followed 13 years. Cox proportional hazards models were utilized to estimate hazard ratios (HRs) associated with YSQ, with YSQ treated as a time-varying covariate. The models adjusted for age and sex. RESULTS: Of 154899 subjects randomized, 27101 were former smokers with 30+ pack-years, and 69182 were never smokers. HRs relative to never smokers ranged from 30.8 (95% CI:23.4-40.5) for YSQ ≤ 5 to 6.4 (95% CI:5.1-8.0) for YSQ > 30. For YSQ of > 10-15, > 15-20, and > 20-25, HRs were 14.8 (95% CI:11.9-18.2), 13.5 (95% CI:11.3-16.2), and 9.9 (95% CI: 8.1-12.0), respectively. CONCLUSIONS: Lung cancer risk decreases gradually with YSQ in 30+ pack year former smokers. A range of upper limits on YSQ may be supportable for LDCT screening.

The prostate, lung, colorectal, and ovarian cancer screening trial and its associated research resource.

The Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial is a large-scale research effort conducted by the National Cancer Institute. PLCO offers an example of coordinated research by both the extramural and intramural communities of the National Institutes of Health. The purpose of this article is to describe the PLCO research resource and how it is managed and to assess the productivity and the costs associated with this resource. Such an in-depth analysis of a single large-scale project can shed light on questions such as how large-scale projects should be managed, what metrics should be used to assess productivity, and how costs can be compared with productivity metrics. A comprehensive publication analysis identified 335 primary research publications resulting from research using PLCO data and biospecimens from 2000 to 2012. By the end of 2012, a total of 9679 citations (excluding self-citations) have resulted from this body of research publications, with an average of 29.7 citations per article, and an h index of 45, which is comparable with other large-scale studies, such as the Nurses' Health Study. In terms of impact on public health, PLCO trial results have been used by the US Preventive Services Task Force in making recommendations concerning prostate and ovarian cancer screening. The overall cost of PLCO was $454 million over 20 years, adjusted to 2011 dollars, with approximately $37 million for the collection, processing, and storage of biospecimens, including blood samples, buccal cells, and pathology tissues.

Building multi-marker algorithms for disease prediction-the role of correlations among markers.

A widely held viewpoint in the field of predictive biomarkers for disease holds that no single marker can provide high enough discrimination and that a panel of markers, combined in some type of algorithm, will be needed. Motivated by a recent study where 27 additional markers for ovarian cancer, many of which had good predictive value alone, failed to substantially increase the predictive ability of the primary marker of CA125, we explore the effect of additional markers on the area under the ROC curve (AUC). We develop a statistical model based on the multivariate normal distribution and linear algorithms and use it to explore how the magnitude and direction of statistical correlation among the markers (in diseased and in non-diseased) is critical in determining the added predictive value of additional markers. We show mathematically and empirically that if the additional marker(s) is negatively correlated with the primary marker, then it will always be able to provide increased AUC when combined with the primary marker (as compared to that obtained with the primary marker alone), even if it has little predictive ability on its own. In contrast, if the additional marker(s) is positively correlated with the primary marker, then it is unlikely to substantially increase the AUC when combined with the primary marker, even when it has good predictive ability on its own. Thus, univariate analyses alone may not be the best approach in choosing which markers to combine in a predictive panel of markers; patterns of statistical correlation should be considered in ranking top-performing biomarkers.

Effect of screening on ovarian cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial.

CONTEXT: Screening for ovarian cancer with cancer antigen 125 (CA-125) and transvaginal ultrasound has an unknown effect on mortality. OBJECTIVE: To evaluate the effect of screening for ovarian cancer on mortality in the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. DESIGN, SETTING, AND PARTICIPANTS: Randomized controlled trial of 78,216 women aged 55 to 74 years assigned to undergo either annual screening (n = 39,105) or usual care (n = 39,111) at 10 screening centers across the United States between November 1993 and July 2001. Intervention The intervention group was offered annual screening with CA-125 for 6 years and transvaginal ultrasound for 4 years. Participants and their health care practitioners received the screening test results and managed evaluation of abnormal results. The usual care group was not offered annual screening with CA-125 for 6 years or transvaginal ultrasound but received their usual medical care. Participants were followed up for a maximum of 13 years (median [range], 12.4 years [10.9-13.0 years]) for cancer diagnoses and death until February 28, 2010. MAIN OUTCOME MEASURES: Mortality from ovarian cancer, including primary peritoneal and fallopian tube cancers. Secondary outcomes included ovarian cancer incidence and complications associated with screening examinations and diagnostic procedures. RESULTS: Ovarian cancer was diagnosed in 212 women (5.7 per 10,000 person-years) in the intervention group and 176 (4.7 per 10,000 person-years) in the usual care group (rate ratio [RR], 1.21; 95% confidence interval [CI], 0.99-1.48). There were 118 deaths caused by ovarian cancer (3.1 per 10,000 person-years) in the intervention group and 100 deaths (2.6 per 10,000 person-years) in the usual care group (mortality RR, 1.18; 95% CI, 0.82-1.71). Of 3285 women with false-positive results, 1080 underwent surgical follow-up; of whom, 163 women experienced at least 1 serious complication (15%). There were 2924 deaths due to other causes (excluding ovarian, colorectal, and lung cancer) (76.6 per 10,000 person-years) in the intervention group and 2914 deaths (76.2 per 10,000 person-years) in the usual care group (RR, 1.01; 95% CI, 0.96-1.06). CONCLUSIONS: Among women in the general US population, simultaneous screening with CA-125 and transvaginal ultrasound compared with usual care did not reduce ovarian cancer mortality. Diagnostic evaluation following a false-positive screening test result was associated with complications. Trial Registration clinicaltrials.gov Identifier: NCT00002540.

Ovarian cancer biomarker performance in prostate, lung, colorectal, and ovarian cancer screening trial specimens.

Establishing a cancer screening biomarker's intended performance requires "phase III" specimens obtained in asymptomatic individuals before clinical diagnosis rather than "phase II" specimens obtained from symptomatic individuals at diagnosis. We used specimens from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial to evaluate ovarian cancer biomarkers previously assessed in phase II sets. Phase II specimens from 180 ovarian cancer cases and 660 benign disease or general population controls were assembled from four Early Detection Research Network or Ovarian Cancer Specialized Program of Research Excellence sites and used to rank 49 biomarkers. Thirty-five markers, including 6 additional markers from a fifth site, were then evaluated in PLCO proximate specimens from 118 women with ovarian cancer and 474 matched controls. Top markers in phase II specimens included CA125, HE4, transthyretin, CA15.3, and CA72.4 with sensitivity at 95% specificity ranging from 0.73 to 0.40. Except for transthyretin, these markers had similar or better sensitivity when moving to phase III specimens that had been drawn within 6 months of the clinical diagnosis. Performance of all markers declined in phase III specimens more remote than 6 months from diagnosis. Despite many promising new markers for ovarian cancer, CA125 remains the single-best biomarker in the phase II and phase III specimens tested in this study.

A framework for evaluating biomarkers for early detection: validation of biomarker panels for ovarian cancer.

A panel of biomarkers may improve predictive performance over individual markers. Although many biomarker panels have been described for ovarian cancer, few studies used prediagnostic samples to assess the potential of the panels for early detection. We conducted a multisite systematic evaluation of biomarker panels using prediagnostic serum samples from the Prostate, Lung, Colorectal, and Ovarian Cancer (PLCO) screening trial. Using a nested case-control design, levels of 28 biomarkers were measured laboratory-blinded in 118 serum samples obtained before cancer diagnosis and 951 serum samples from matched controls. Five predictive models, each containing 6 to 8 biomarkers, were evaluated according to a predetermined analysis plan. Three sequential analyses were conducted: blinded validation of previously established models (step 1); simultaneous split-sample discovery and validation of models (step 2); and exploratory discovery of new models (step 3). Sensitivity, specificity, sensitivity at 98% specificity, and AUC were computed for the models and CA125 alone among 67 cases diagnosed within one year of blood draw and 476 matched controls. In step 1, one model showed comparable performance to CA125, with sensitivity, specificity, and AUC at 69.2%, 96.6%, and 0.892, respectively. Remaining models had poorer performance than CA125 alone. In step 2, we observed a similar pattern. In step 3, a model derived from all 28 markers failed to show improvement over CA125. Thus, biomarker panels discovered in diagnostic samples may not validate in prediagnostic samples; utilizing prediagnostic samples for discovery may be helpful in developing validated early detection panels.