AI/ML advances in non-small cell lung cancer biomarker discovery.

Lung cancer is the leading cause of cancer deaths among both men and women, representing approximately 25% of cancer fatalities each year. The treatment landscape for non-small cell lung cancer (NSCLC) is rapidly evolving due to the progress made in biomarker-driven targeted therapies. While advancements in targeted treatments have improved survival rates for NSCLC patients with actionable biomarkers, long-term survival remains low, with an overall 5-year relative survival rate below 20%. Artificial intelligence/machine learning (AI/ML) algorithms have shown promise in biomarker discovery, yet NSCLC-specific studies capturing the clinical challenges targeted and emerging patterns identified using AI/ML approaches are lacking. Here, we employed a text-mining approach and identified 215 studies that reported potential biomarkers of NSCLC using AI/ML algorithms. We catalogued these studies with respect to BEST (Biomarkers, EndpointS, and other Tools) biomarker sub-types and summarized emerging patterns and trends in AI/ML-driven NSCLC biomarker discovery. We anticipate that our comprehensive review will contribute to the current understanding of AI/ML advances in NSCLC biomarker research and provide an important catalogue that may facilitate clinical adoption of AI/ML-derived biomarkers.

Predictive Gene Signatures Determine Tumor Sensitivity to MDM2 Inhibition.

Early clinical trials using murine double minute 2 (MDM2) inhibitors demonstrated proof-of-concept of p53-induced apoptosis by MDM2 inhibition in cancer cells; however, not all wild-type TP53 tumors are sensitive to MDM2 inhibition. Therefore, more potent inhibitors and biomarkers predictive of tumor sensitivity are needed. The novel MDM2 inhibitor DS-3032b is 10-fold more potent than the first-generation inhibitor nutlin-3a. TP53 mutations were predictive of resistance to DS-3032b, and allele frequencies of TP53 mutations were negatively correlated with sensitivity to DS-3032b. However, sensitivity to DS-3032b of TP53 wild-type tumors varied greatly. We thus used two methods to create predictive gene signatures. First, by comparing sensitivity to MDM2 inhibition with basal mRNA expression profiles in 240 cancer cell lines, a 175-gene signature was defined and validated in patient-derived tumor xenograft models and ex vivo human acute myeloid leukemia (AML) cells. Second, an AML-specific 1,532-gene signature was defined by performing random forest analysis with cross-validation using gene expression profiles of 41 primary AML samples. The combination of TP53 mutation status with the two gene signatures provided the best positive predictive values (81% and 82%, compared with 62% for TP53 mutation status alone). In addition, the top-ranked 50 genes selected from the AML-specific 1,532-gene signature conserved high predictive performance, suggesting that a more feasible size of gene signature can be generated through this method for clinical implementation. Our model is being tested in ongoing clinical trials of MDM2 inhibitors.Significance: This study demonstrates that gene expression profiling combined with TP53 mutational status predicts antitumor effects of MDM2 inhibitors in vitro and in vivoCancer Res; 78(10); 2721-31. ©2018 AACR.