Receiver operating characteristic (ROC) regression methodology is used to identify factors that affect the accuracy of medical diagnostic tests. In this paper, we consider a ROC model for which the ROC curve is a parametric function of covariates but distributions of the diagnostic test results are not specified. Covariates can be either common to all subjects or specific to those with disease. We propose a new estimation procedure based on binary indicators defined by the test result for a diseased subject exceeding various specified quantiles of the distribution of test results from non‐diseased subjects with the same covariate values. This procedure is conceptually and computationally simplified relative to existing procedures. Simulation study results indicate that the approach has fairly high statistical efficiency. The new ROC regression methodology is used to evaluate childhood measurements of body mass index as a predictive marker of adult obesity.

Lung cancer is the number one cancer killer in the United States. This disease is clinically divided into two sub-types, small cell lung cancer, (10–15% of lung cancer cases), and non-small cell lung cancer (NSCLC; 85–90% of cases). Early detection of NSCLC, which is the more common and less aggressive of the two sub-types, has the highest potential for saving lives. As yet, no routine screening method that enables early detection exists, and this is a key factor in the high mortality rate of this disease. Imaging and cytology-based screening strategies have been employed for early detection, and while some are sensitive, none have been demonstrated to reduce lung cancer mortality. However, mortality might be reduced by developing specific molecular markers that can complement imaging techniques. DNA methylation has emerged as a highly promising biomarker and is being actively studied in multiple cancers. The analysis of DNA methylation-based biomarkers is rapidly advancing, and a large number of potential biomarkers have been identified. Here we present a detailed review of the literature, focusing on DNA methylation-based markers developed using primary NSCLC tissue. Viable markers for clinical diagnosis must be detectable in 'remote media' such as blood, sputum, bronchoalveolar lavage, or even exhaled breath condensate. We discuss progress on their detection in such media and the sensitivity and specificity of the molecular marker panels identified to date. Lastly, we look to future advancements that will be made possible with the interrogation of the epigenome.

Post-relapse therapy remains critical for survival in children with acute myeloid leukemia (AML). We evaluated survival, response and prognostic variables following relapse in independent cooperative group studies conducted by COG and the population-based AML-BFM study group. BFM included 197 patients who relapsed after closure of the last I-BFM relapse trial until 2017, while COG included 852 patients who relapsed on the last Phase 3 trials (AAML0531, AAML1031). Overall survival at 5 years (OS) was 42 ± 4% (BFM) and 35 ± 2% (COG). Initial high-risk features (BFM 32 ± 6%, COG 26 ± 4%) and short time to relapse (BFM 29 ± 4%, COG 25 ± 2%) predicted diminished survival. In the BFM dataset, there was no difference in OS for patients who had a complete remission with full hematopoietic recovery (CR) following post-relapse re-induction compared to those with partial neutrophil and platelet recovery (CRp and CRi) only (52 ± 7% vs. 63 ± 10%, p = 0.39). Among 90 patients alive at last follow-up, 87 had received a post-relapse hematopoietic stem cell transplant (HSCT). OS for patients with post-relapse HSCT was 54 ± 4%. In conclusion, initial high-risk features and early relapse remain prognostic. Response assessment with full hematopoietic recovery following initial relapse therapy does not predict survival. These data indicate the need for post-relapse risk stratification in future studies of relapse therapies.

Key Points: Mutations in SETBP1 can be detected using droplet digital polymerase chain reaction in at least 30% of patients with JMML and are associated with a dismal prognosis. Patients harboring rare cells with mutant SETBP1 at diagnosis should be considered candidates for swift hematopoietic stem cell transplant.

Disease prevalence is ideally estimated using a ‘gold standard’ to ascertain true disease status on all subjects in a population of interest. In practice, however, the gold standard may be too costly or invasive to be applied to all subjects, in which case a two‐phase design is often employed. Phase 1 data consisting of inexpensive and non‐invasive screening tests on all study subjects are used to determine the subjects that receive the gold standard in the second phase. Naive estimates of prevalence in two‐phase studies can be biased (verification bias). Imputation and re‐weighting estimators are often used to avoid this bias. We contrast the forms and attributes of the various prevalence estimators. Distribution theory and simulation studies are used to investigate their bias and efficiency. We conclude that the semiparametric efficient approach is the preferred method for prevalence estimation in two‐phase studies. It is more robust and comparable in its efficiency to imputation and other re‐weighting estimators. It is also easy to implement. We use this approach to examine the prevalence of depression in adolescents with data from the Great Smoky Mountain Study.