PD33-03 TREATMENT OF THE PRIMARY TUMOR WITH VASCULAR-TARGETED PHOTODYNAMIC THERAPY (VTP) OVERCOMES RESISTANCE TO SYSTEMIC PD-1/PD-L1 INHIBITION: EFFECTS ON PRIMARY TUMOR CONTROL AND PREVENTION OF LUNG METASTASIS IN A PRE-CLINICAL RCC MODEL

genomic landscape in clear cell renal cell carcinoma (ccRCC). We propose here a novel strategy of capturing variant alleles present in different tumor regions by sampling multiple tumor sites and combining all DNA to be sequenced as one sample. METHODS: Ex vivo core needle biopsies were obtained from six geographically different regions of five primary renal tumors that were resected via partial or radical nephrectomy at a single institution from 4/2014 to 7/2014. All patients had signed informed consents for tissue utilization, and our institutional review board had approved the study. DNA was extracted using standard techniques. Each tumor region was sequenced individually (R1-R6) and also combined together and sequenced as pooled sample. Tumor DNA and matched normal DNA were profiled for genomic alterations in 341 key cancer-associated genes using our custom, deep sequencing MSK-IMPACT assay (Integrated Mutation Profiling of Actionable Cancer Targets). RESULTS: There were 40 samples sequenced: 5 patients x (6 regions þ 1 pool þ 1 normal), and the average sequence coverage is 905x (range 509x e 1920x). 5/5 tumors were clear cell RCC and 4/5 tumors were stage pT3 or higher. The median tumor size was 6.0 cm (IQR 4.2 e 7.5 cm). All five tumors had a VHL mutation present, with each one being present in all six regions and the pooled sample. Other genes including BAP1, KDM5C, TSC1, TP53, TRAF7, and ERBB2 exhibited various degrees of regional heterogeneity. All variant alleles detected in a single tumor region were also detected in that tumor’s pooled DNA sample (Figure 1). CONCLUSIONS: We present here a feasible, cost effective strategy for accurately profiling a renal tumor’s genetic landscape while overcoming the obstacle of intratumor heterogeneity. By pooling DNA from multiple tumor regions and performing ultra deep sequencing on this combined sample, we were able to detect every mutation that was captured when sequencing each region individually.