Faculty Bibliography

BACKGROUND:Microvascular invasion (MVI) is a significant risk factor for early recurrence after resection or transplantation for hepatocellular carcinoma (HCC). Knowledge of MVI status would help guide treatment recommendations, but is generally identified after operation. This study aims to predict MVI preoperatively using quantitative image analysis. STUDY DESIGN/METHODS:One hundred and twenty patients from 2 institutions underwent resection of HCC from 2003 to 2015 were included. The largest tumor from preoperative CT was subjected to quantitative image analysis, which uses an automated computer algorithm to capture regional variation in CT enhancement patterns. Quantitative imaging features by automatic analysis, qualitative radiographic descriptors by 2 radiologists, and preoperative clinical variables were included in multivariate analysis to predict histologic MVI. RESULTS:Histologic MVI was identified in 19 (37%) patients with tumors ≤5 cm and 34 (49%) patients with tumors >5 cm. Among patients with tumors ≤5 cm, none of the clinical findings or radiographic descriptors were associated with MVI; however, quantitative features based on angle co-occurrence matrix predicted MVI with an area under curve of 0.80, positive predictive value of 63%, and negative predictive value of 85%. In patients with tumors >5 cm, higher α-fetoprotein level, larger tumor size, and viral hepatitis history were associated with MVI, and radiographic descriptors were not. However, a multivariate model combining α-fetoprotein, tumor size, hepatitis status, and quantitative feature based on local binary pattern predicted MVI with area under curve of 0.88, positive predictive value of 72%, and negative predictive value of 96%. CONCLUSIONS:This study reveals the potential importance of quantitative image analysis as a predictor of MVI.

Protein kinases represent the most effective class of therapeutic targets in cancer; therefore, determination of kinase aberrations is a major focus of cancer genomic studies. Here, we analyzed transcriptome sequencing data from a compendium of 482 cancer and benign samples from 25 different tissue types, and defined distinct "outlier kinases" in individual breast and pancreatic cancer samples, based on highest levels of absolute and differential expression. Frequent outlier kinases in breast cancer included therapeutic targets like ERBB2 and FGFR4, distinct from MET, AKT2, and PLK2 in pancreatic cancer. Outlier kinases imparted sample-specific dependencies in various cell lines, as tested by siRNA knockdown and/or pharmacologic inhibition. Outlier expression of polo-like kinases was observed in a subset of KRAS-dependent pancreatic cancer cell lines, and conferred increased sensitivity to the pan-PLK inhibitor BI-6727. Our results suggest that outlier kinases represent effective precision therapeutic targets that are readily identifiable through RNA sequencing of tumors.