Citations

98 Background: Circulating tumor DNA (ctDNA) monitoring provides an additional clinical tool for assessing treatment response and emerging resistance. We evaluated the real-world (RW) use and results of FoundationOne Monitor (F1Monitor), a tissue-free ctDNA monitoring assay, in patients (pts) with prostate cancer (PCa). Methods: The clinical laboratory developed test, F1Monitor, was available to physicians in a limited pilot with full discretion of test eligibility, timing, and frequency. Therapy type was a non-mandatory field on the order form. CtDNA was detected and quantified by ctDNA tumor fraction (TF); variants were tracked with variant allele frequency (VAF), and the origin [tumor somatic (TS), germline, or clonal hematopoiesis (CH)] was predicted algorithmically. Results: F1Monitor was ordered by 28 physicians for 215 pts with PCa. Therapy data were known for 156 pts: ADT+ androgen receptor pathway inhibitor (ARPI, N = 69/156 pts, 44%), ADT alone (N = 32, 21%), ADT + chemotherapy (chemo, N = 20,13%), ADT + radioligand therapy (RLT, N = 17, 11%), and others (N = 18). Median follow-up was 188 days. Longitudinal testing (≥2 tests) occurred in 64 pts, often (N = 37) when the first test was ctDNA TF positive (+). Median cadence for serial testing (≥3 tests, N = 31) was 42 days (IQR: 35, 59). ctDNA TF was + at any timepoint in 50% (107/215) of pts overall and 53% (82/156) of pts known to be on therapy, trending higher in pts on ADT+chemo (80%) and ADT+RLT (65%). Median and maximum ctDNA TF values were 11% (IQR: 0.9,37.8%) and 89% respectively. Lowest ctDNA TF and reportable TS VAF quantified were 0.13% and 0.04%, respectively. ctDNA TF longitudinal changes were observed in 67% (43/64) of pts: 92% of pts on ADT+chemo, 69% on ADT+RLT, 59% on ADT+ARPI, and 38% of pts on ADT alone. Among pts with ≥2 tests, 39% (25/64) had ctDNA TF increase, including 6/27 (22%) pts that were ctDNA TF negative (-) on the first test. Of pts with ≥2 tests and any ctDNA TF+ result, 35% cleared ctDNA TF in later tests. 33% of pts were persistently ctDNA TF -. Each test reported an average of 5 alterations: ~2 TS short variants, copy number variants, or rearrangements, and ~2 predicted CH variants. HRR variants were detected in 37% (80/215) of pts (31 with ≥2 tests) : persistent in 20 pts, cleared in 7, and newly emerging in 14. At least 50% of alterations in CHEK2 and ATM were predicted CH. Alterations in PI3K pathway genes were detected in 20% (43/215) of pts: (16 with ≥2 tests) persistent in 6 pts, cleared in 5, and newly emerging in 5. Conclusions: In this RW pilot of tissue-free ctDNA monitoring in PCa, F1Monitor demonstrated high sensitivity in clinical practice, detecting ctDNA TF as low as 0.13% and TS VAF as low as 0.04% while distinguishing tumor-derived from predicted CH variants. Importantly, ctDNA TF was + in 53% of pts while on therapy with observed longitudinal changes for 67% of pts, supporting potential utility of ctDNA for monitoring treatment response and emerging resistance.

631 Background: For most clear cell renal cell carcinoma (ccRCC) patients with early stage disease, surgical resection offers definitive cure. However, for the small percentage of tumors that metastasize, analyzing gene expression profiles from the primary site at the time of nephrectomy can serve as a model to understand the molecular aberrations behind a metastatic phenotype. Differences in gene expression profiles between patients with Stage I and II ccRCC who experience metastasis versus patients who maintain cure after surgery may help elucidate significant molecular targets and stratify patients at higher risk for metastasis. Methods: Nineteen Stage I and twenty Stage II ccRCC tumors preserved in FFPE blocks after nephrectomy were included in this study. Patients were matched for age, gender, tumor size and grade. In both stages, approximately half the patients that experienced metastasis within 5 years of surgery were part of the experimental group, whereas the control group had > 5 years of follow-up without evidence of disease. Extracted RNA for Stage I patients was sequenced using Illumina TruSeq RNA Access Library. Gene counts were assessed by ht-seq counts and differential expression using DESeq2. Significant genes found were validated in the Stage II group using RT-qPCR. Results: In the Stage I experimental group, statistically significant upregulation of several genes associated with unfavorable prognosis in RCC were found: COL1A1, NUMBL, and STEAP3. Random forest classification accurately separated Stage I control versus experimental patients based on expression of COL1A1. Affected genes were consistent with molecular changes seen in TCGA analysis. In the Stage II group, a double-blinded analysis correctly identified the clinical outcome for the majority of the patients using qPCR expression of COL1A1, NUMBL, and STEAP3. Conclusions: Differences in gene expression profiles harbored in the primary site of early stage ccRCC may be employed to predict patients at high risk for developing metastasis. Validating these findings in a larger study carries the potential to better understand mechanisms of metastasis and identify an at risk cohort of patients with early stage disease.