Detection of BRCA1, BRCA2, and ATM Alterations in Matched Tumor Tissue and Circulating Tumor DNA in Patients with Prostate Cancer Screened in PROfound.

PURPOSE: Not all patients with metastatic castration-resistant prostate cancer (mCRPC) have sufficient tumor tissue available for multigene molecular testing. Furthermore, samples may fail because of difficulties within the testing procedure. Optimization of screening techniques may reduce failure rates; however, a need remains for additional testing methods to detect cancers with alterations in homologous recombination repair genes. We evaluated the utility of plasma-derived circulating tumor DNA (ctDNA) in identifying deleterious BRCA1, BRCA2 (BRCA), and ATM alterations in screened patients with mCRPC from the phase III PROfound study. PATIENTS AND METHODS: Tumor tissue samples were sequenced prospectively at Foundation Medicine, Inc. (FMI) using an investigational next-generation sequencing (NGS) assay based on FoundationOne®CDx to inform trial eligibility. Matched ctDNA samples were retrospectively sequenced at FMI, using an investigational assay based on FoundationOne®Liquid CDx. RESULTS: 81% (503/619) of ctDNA samples yielded an NGS result, of which 491 had a tumor tissue result. BRCA and ATM status in tissue compared with ctDNA showed 81% positive percentage agreement and 92% negative percentage agreement, using tissue as reference. At variant-subtype level, using tissue as reference, concordance was high for nonsense (93%), splice (87%), and frameshift (86%) alterations but lower for large rearrangements (63%) and homozygous deletions (27%), with low ctDNA fraction being a limiting factor. CONCLUSIONS: We demonstrate that ctDNA can greatly complement tissue testing in identifying patients with mCRPC and BRCA or ATM alterations who are potentially suitable for receiving targeted PARP inhibitor treatments, particularly patients with no or insufficient tissue for genomic analyses.

Disparities in Adverse Event Reporting for Hospitalized Children.

OBJECTIVES: Hospitals rely on voluntary event reporting (VER) for adverse event (AE) identification, although it captures fewer events than a trigger tool, such as Global Assessment of Pediatric Patient Safety (GAPPS). Medical providers exhibit bias based on patient weight status, race, and English proficiency. We compared the AE rate identified by VER with that identified using the GAPPS between hospitalized children by weight category, race, and English proficiency. METHODS: We identified a cohort of patients 2 years to younger than 18 years consecutively discharged from an academic children's hospital between June and October 2018. We collected data on patient weight status from age, sex, height, and weight, race/ethnicity by self-report, and limited English proficiency by record of interpreter use. We reviewed each chart with the GAPPS to identify AEs and reviewed VER entries for each encounter. We calculated an AE rate per 1000 patient-days using each method and compared these using analysis of variance. RESULTS: We reviewed 834 encounters in 680 subjects; 262 (38.5%) had overweight or obesity, 144 (21.2%) identified as Black, and 112 (16.5%) identified as Hispanic; 82 (9.8%) of encounters involved an interpreter. We identified 288 total AEs, 270 (93.8%) by the GAPPS and 18 (6.3%) by VER. A disparity in AE reporting was found for children with limited English proficiency, with fewer AEs by VER ( P = 0.03) compared with no difference in AEs by GAPPS. No disparities were found by weight category or race. CONCLUSIONS: Voluntary event reporting may systematically underreport AEs in hospitalized children with limited English proficiency.

Body Mass Index Category and Adverse Events in Hospitalized Children.

OBJECTIVE: To identify associations between patient body mass index (BMI) category and adverse event (AE) rate, severity, and preventability in a cohort of children discharged from an academic children's hospital. METHODS: We identified patients 2 to 17 years old consecutively discharged between June and October 2018. Patient age, sex, height, and weight were used to categorize patients as having underweight, normal weight, overweight, or obesity. We used the Global Assessment of Pediatrics Patient Safety trigger tool to identify AEs, which were scored for harm and preventability. The primary outcome was the rate of AEs; these were compared with Poisson regression. We used multivariable logistic regression to model event preventability. RESULTS: We reviewed 834 encounters in 680 subjects; 51 (7.5%) had underweight, 367 (54.0%) had normal weight, 112 (16.5%) had overweight, and 150 (22.1%) had obesity. Our cohort experienced 270 AEs, with an overall rate of 69.7 (61.8-78.5) AEs per 1000 patient-days: 67.7 (46.4-98.7) in underweight, 70.0 (59.4-82.4) in normal weight, 58.6 (42.5-79.7) in overweight, and 80.4 (62.5-103.6) in obesity, P = .46. No associations were seen between BMI category and AE severity. Children with obesity had an increased rate of preventable AEs (P < .01), but this association did not persist in the multivariable model. CONCLUSIONS: In this single-center study, we did not find associations between BMI category and rate, severity, or preventability of AEs.