Prediction models perform poorly if based on small numbers of events and developed with common but suboptimal statistical approaches. Alternative modeling strategies to best exploit available predictive information need wider implementation, with collaborative research to increase sample sizes.

This article has an accompanying continuing medical education activity, also eligible for MOC credit, on page e50. Learning Objective-Upon completion of this activity, successful learners will be able to identify clinical factors that indicate an increased likelihood for gastric cancer among individuals with Lynch syndrome. BACKGROUND & AIMS:Lynch syndrome is the most common inherited cause of gastrointestinal cancer and increases risk for a variety of malignancies, including gastric cancer. We aimed to identify clinical factors associated with gastric cancer in carriers of germline variants causing Lynch syndrome. METHODS:We collected data from 52,758 consecutive individuals tested for genetic variants associated with Lynch syndrome from June 2006 through July 2013 at a commercial laboratory. We obtained clinical and demographic data, as well as information on personal and family histories of cancer (first-and second-degree relatives) from forms completed by ordering providers. We performed multivariate logistic regression to identify clinical factors associated with gastric cancer in carriers of mutations that cause Lynch syndrome (pathogenic mutations). RESULTS:After we excluded individuals with missing clinical data (n [ 1664) or with multiple pathogenic mutations (n [ 8), we analyzed data from 51,086 individuals. Of these, 3828 persons carried pathogenic mutations (1346 with mutations in MLH1, 1639 with mutations in MSH2, 670 with mutations in MSH6, 145 with mutations in PMS2, and 28 with mutations in EPCAM). Of the 3828 carriers of pathogenic mutations, 41 (1.1%) had a previous gastric cancer and 350 (9.1%) had 1 or more firstor second-degree relatives with gastric cancer. In multivariate analysis, male sex (odds ratio [OR], 2.82; 95% CI, 1.48-5.38), older age (OR, 2.07 per 10 years; 95% CI, 1.64-2.61), mutations in MLH1 (OR, 6.53; 95% CI, 1.50-28.42) or MSH2 (OR, 5.23 compared to mutations in MSH6, PMS2, or EPCAM; 95% CI, 1.21-22.71), and number of first-degree relatives with gastric cancer (OR, 2.52; 95% CI, 1.42-4.45), but not second-degree relatives (OR, 1.12; 95% CI, 0.40-3.18) were independently associated with gastric cancer among carriers of pathogenic mutations. CONCLUSIONS:In an analysis of data from almost 4000 carriers of Lynch syndrome-associated mutations, we found history of gastric cancer to be independently associated with male sex, older age, mutations in MLH1 or MSH2, and number of first-degree relatives with gastric cancer. These findings suggest that personalized, risk-stratified approaches to gastric cancer surveillance may be appropriate for individuals with Lynch syndrome-associated mutations.

Prior small reports have postulated a link between gastrointestinal polyposis and childhood and young adulthood cancer (CYAC) treatment (therapy-associated polyposis; TAP), but this remains a poorly understood phenomenon. The aim of this study was to describe the phenotypic spectrum of TAP in a multi-institutional cohort. TAP cases were identified from eight high-risk cancer centers. Cases were defined as patients with 10 gastrointestinal polyps without known causative germline alteration or hereditary colorectal cancer predisposition syndrome who had a history of prior treatment with chemotherapy and/or radiotherapy for CYAC. A total of 34 TAP cases were included (original CYAC: 27 Hodgkin lymphoma, three neuroblastoma, one acute myeloid leukemia, one medulloblastoma, one nephroblastoma, and one non-Hodgkin lymphoma). Gastrointestinal polyposis was first detected at a median of 27 years (interquartile range, 20-33) after CYAC treatment. A total of 12 of 34 (35%) TAP cases had 50 colorectal polyps. A total of 32 of 34 (94%) had >1 histologic polyp type. A total of 25 of 34 (74%) had clinical features suggestive of 1 colorectal cancer predisposition syndrome [e.g., attenuated familial adenomatous polyposis (FAP), serrated polyposis syndrome, extracolonic manifestations of FAP, mismatch repair-deficient colorectal cancer, or hamartomatous polyposis] including 8 of 34 (24%) with features of multiple such syndromes. TAP is an apparently acquired phenomenon that should be considered in patients who develop significant polyposis without known causative germline alteration but who have had prior treatment for a CYAC. Patients with TAP have features that may mimic various hereditary colorectal cancer syndromes, suggesting multiple concurrent biologic mechanisms, and recognition of this diagnosis may have implications for cancer risk and screening.

Background: Lynch syndrome confers markedly increased risks of various malignancies, including urinary tract cancers (UTC; renal pelvis, ureter, bladder, and possibly kidney cancers). It is unknown how to determine which Lynch syndrome carriers are at highest UTC risk. Our aim was to identify clinical factors associated with UTC among Lynch syndrome carriers.Methods: The study population was a cohort of 52,758 consecutively ascertained individuals undergoing Lynch syndrome testing at a commercial laboratory. Clinical data were obtained from test request forms completed by the ordering provider. Univariate analysis and multivariate logistic regression were performed to identify factors associated with UTC among Lynch syndrome carriers.Results: Compared with noncarriers, Lynch syndrome carriers were significantly more likely to have had UTC (4.1% vs. 1.2%; P < 0.0001). Lynch syndrome-associated UTC was independently associated with male sex [OR 1.95; 95% confidence interval (CI), 1.38-2.76], increased age (OR 2.44 per 10 years; 95% CI, 2.11-2.82), familial burden of UTC (OR 2.69 per first-/second-degree relative with UTC; 95% CI, 1.99-3.63), and pathogenic EPCAM/MSH2 variants (OR 4.01; 95% CI, 2.39-6.72) but not MLH1 variants (OR 1.17; 95% CI, 0.63-2.17), race, or history of other Lynch syndrome-associated malignancy. A total of 143 of 158 (90.5%) Lynch syndrome carriers with UTC had 1 of the following characteristics: male sex, EPCAM/MSH2 variants, or family history of UTC; 1,236 of 1,251 (98.8%) Lynch syndrome carriers lacking all of these characteristics had no history of UTC.Conclusions: Specific clinical factors can reliably identify Lynch syndrome carriers most likely to be at risk for UTC.Impact: A predictable subset of Lynch syndrome carriers may be most likely to benefit from UTC surveillance/prevention.

No abstract