Integrated radiomic and liquid biopsy monitoring in SWOG S1802: A phase 3 therapeutic trial for metastatic prostate cancer

Abstract Prostate cancer (PC) afflicts more men in the U.S. than any other malignancy and is the second leading cause of cancer death in this population. Recently, new therapies have improved survival in some men but have offered little benefit to others. Tissue profiling has identified genes and pathways associated with resistance and

progression to advanced disease, but little is known about the dynamics of when and how such changes arise during therapy. Such insights can only be gained through minimally invasive monitoring that allows repeated disease profiling over time. Our multidisciplinary team has developed and tested two such minimally invasive

monitoring capabilities: The first is multi-parametric liquid biopsy: streamlined analysis of blood samples that simultaneously measures PC-relevant cellular and molecular features from single circulating tumor cells and matched plasma cell-free nucleic acids. The second is radiomic analysis: high-throughput extraction of

quantitative metrics to identify PC-relevant phenomic imaging features. By synergizing the strengths of the two methods, our goal is to test the hypothesis that liquid biopsy and radiomic techniques can be applied jointly to monitor metastatic PC noninvasively over time, culminating in new prognostic and predictive tools to guide

therapy. To maximize the impact of our findings, liquid biopsy profiling and radiomic analysis will be integrated into Southwest Oncology (SWOG) S1802, a phase 3 prospective therapeutic trial for over 1200 men with newly diagnosed metastatic castrate sensitive prostate cancer (mCSPC) treated with standard systemic therapy alone

or in combination with definitive treatment of the primary tumor. Notably, SWOG and the other participating cooperative groups have reviewed and approved this proposal (see Letters of Support), and the requisite blood samples and CT scans already are being collected as part of the active protocol. In this unique setting, we will

pursue three Specific Aims. We will: 1. Use multi-parametric liquid biopsies to monitor the evolving cellular and molecular landscape of mCSPC during treatment in S1802; 2. Use radiomic analysis of CT scans to monitor the evolving radiographic landscape of mCSPC during treatment in S1802; and 3. Integrate cellular, molecular, and

radiomic data to develop composite disease monitoring models predictive of PFS and OS. Importantly, liquid biopsies and CT images will be matched within patients and analyzed at key inflection points over the course of the trial: at diagnosis, after initiation of systemic therapy, after definitive therapy, and at disease progression. In

this way, the proposed work will illuminate when and how PC-relevant phenotypes arise during therapy, and how they relate to clinical outcomes. In particular, we will better understand which components of the liquid biopsy – circulating cells, plasma, or both – most accurately reflect the tumor's evolving molecular profile, which radiomic

metrics most closely correlate with these molecular features and can serve as early cost-effective indicators of emerging resistance, and which composite models combining liquid biopsy and radiomic features best predict PFS and OS and can serve as powerful minimally-invasive tools to monitor and adjust therapy.