Ryan S. Chiang, Arya Ashok, Elizabeth Mauer, Alex Barrett, Christian R Hoerner, Osama A Khan, Chia-Sui Kao, Sumit Shah, Sandy Srinivas, Alice C. Fan, Ali Raza Khaki
Background: RCC may metastasize to diverse sites, which confers variable prognosis and treatment response. Pancreatic metastases in RCC have been associated with indolent biology, low genomic instability and decreased inflammation, which may lead to worse responses to immunotherapy. To further investigate this phenomenon, we compared the genomic landscape and immunotherapy biomarkers of metastatic RCC patients with and without pancreatic metastases.
Methods: We identified a cohort from the de-identified Tempus database, consisting of patients with metastatic RCC who received next generation sequencing via the Tempus|xT assay (DNA-seq of 648 genes at 500x coverage; whole-exome capture RNA-seq). Records were included if patients had RCC diagnosis and metastases to at least one of: pancreas, liver, brain, or lung (regardless of the presence of other metastases). RCC diagnosis and metastatic sites were determined from abstracted clinical notes received at the time of sequencing. If individuals had multiple xT results, the results from the most recent sample were included (irrespective of tissue site). We compared the prevalence of somatic gene alterations (using false discovery rate corrected q-values) and immunotherapy markers among records with vs without pancreatic metastases. Statistical significance was determined via Wilcoxon rank-sum tests or Chi-squared/Fisher’s exact tests.
Results: 654 patients (median age 61, 30% female, 62% clear cell) with metastatic RCC were included. 58 harbored pancreatic metastases and 596 had metastases to sites other than pancreas (36 brain, 88 liver, 358 lung, 114 brain/liver/lung combinations). Females had a significantly higher representation in the group with pancreatic metastases (43% vs 29%, p = 0.02). The group with pancreatic metastases had significantly more PBRM1 alterations (52% vs 25%, q < 0.001), and trended towards more prevalent copy number loss of ALK (22% vs 10%, q = 0.06) and NTRK3 (16% vs 6%, q = 0.08) and less prevalent alterations in CDKN2A (3% vs 13%, q = 0.08), BAP1 (2% vs 13%, q = 0.06) and MTAP (0% vs 9%, q = 0.06). Individuals harboring pancreatic metastases had similar MSI status and TMB, but had lower rates of PD-L1 positivity, lower proportions of B-cells and higher proportion of NK cells in immune cell composition (Table).
Conclusions: RCC patients with pancreatic metastases have a distinct molecular profile and lower rates of PD-L1 positivity, which suggests a unique disease biology and potential non-immunogenic phenotype. Further investigation and prospective validation is needed to determine immunotherapy outcomes in this population.
VIEW THE POSTER
VIEW THE PUBLICATION