Developing novel data-driven biomarkers and treatment strategies: In precision oncology, the goal is to personalize patient care by aligning each tumor with the most relevant treatment strategy based on its unique molecular composition. Advanced techniques capturing molecular profiles at an "omic" scale generate extensive data. Our lab’s ongoing research explores innovative approaches to harness and tailor treatment strategies based on individual patient needs, utilizing data from both experimental systems and patient-derived samples.

1. Incorporating the non-coding genome in drug response prediction models: Dr. Nath’s research extended cancer drug response models to include lncRNAs, using large-scale screens with 800+ agents and 1000+ cell lines. New predictive models integrating genomic and transcriptomic data identified EGFR-AS1 and MIR205HG as potent anti-EGFR drug response biomarkers, surpassing clinical biomarkers (Nath et al. 2019, PNAS). Validation using resistant cell lines highlighted non-coding genes' predictive potential. To address transcriptomic data limitations, he developed ML tools for imputing miRNA and lncRNA levels with external datasets (Nath et al. 2020, Bioinformatics; Nath et al. 2020, Briefings in Bioinformatics).

2. Developing novel prognostic biomarkers and treatment strategies for breast cancer: Breast cancer, the leading cause of cancer-related deaths in women, presents a challenge with over 40,000 annual deaths in the US. Approximately 75% of cases are estrogen receptor-positive (ER+), initially treated with endocrine therapy. However, 30-40% experience relapse, necessitating alternative treatments. Dr. Nath’s work addresses this gap by developing AI/ML-based prognostic models. Using transcriptomic data from 800+ ER+ breast cancer patients, the ENDORSE model identifies those likely to benefit from endocrine therapy (Nath et al. 2022, Molecular Systems Biology). Validation across four independent clinical trials, including advanced Stage IV cases, confirmed ENDORSE's reliability. Additionally, an ML biomarker for mTOR inhibitor response was developed with both biomarkers protected by patents (U.S. Patent Application No. 18/031,855 and International Patent Application No. PCT/US2021/055285).

3. Innovative clinical trials incorporating AI/ML biomarkers: A key area of focus of our lab is designing and initiating clinical trials incorporating novel biomarkers for treatment selection. Through our collaborations across City of Hope National Medical Center as well as Inova Schar Cancer Institute and MD Anderson Cancer Center, we are developing new trials to evaluate the utility of AI/ML biomarkers in stratifying cancer patients. To enable seamless integration of multimodal biomarkers into clinical practice, we are also developing cloud-based computational pipelines that will simplify and automate the process of using complex models.