Uncovering the tumor microenvironment (TME): Exploring survival and immunotherapy (IO) response in cancer of unknown primary (CUP).

4198 Background: Cancer of Unknown Primary (CUP) is a group of rare and heterogeneous metastatic cancers with unidentifiable site of origin at time of diagnosis, despite extensive investigation. Characterization of CUP remains challenging, and biomarkers are urgently needed to better tailor therapies. Tumor microenvironment (TME) comprises the dynamic and complex network of extracellular matrix, blood vessels, immune cells, signaling molecules surrounding a tumor. It plays a critical role in tumorigenesis, progression and response to treatment. This study investigates the correlation between clinical factors, TME, IO and overall survival outcomes in patients diagnosed with CUP. Methods: We retrospectively evaluated 49 CUP patients who underwent Boston Gene Tumor Portrait (BGTP) testing between January 2023 and July 2024. BGTP utilizes a combination of Next-Generation Sequencing (NGS) and artificial intelligence to analyze a tumor sample’s genetic profile along with its surrounding microenvironment, thereby providing a more comprehensive view to support clinical decision-making and optimize treatment strategies. Clinico-pathological data, including age, ECOG performance status at diagnosis, baseline laboratory results, tumor histology and grade, number of metastatic sites (NMS) and IO treatment history were collected through retrospective chart review. TME was categorized into 4 subtypes: immune-enriched (IE), immune-enriched/fibrotic (IE/F), fibrotic (F) and immune-depleted (D). Results: Baseline characteristics of 49 CUP patients show median age at diagnosis was 63 years (range 32-80). Most patients were male (59%). Grade was reported as poorly differentiated in 69%, with pathology described as carcinoma (47%), adenocarcinoma (33%), squamous cell carcinoma (10%) and malignant neoplasm (8%). TME subtypes distribution was 33% IE, 14% IE/F, 16% F, 37% D. Univariate analysis revealed that better outcomes were associated with lower NLR ( < 5) (median survival 57 vs 11 months, p = 0.006), lower ECOG (57 vs 16 vs 6 months for ECOG 0, 1 and 2 respectively, p = 0.002), lower NMS ( < 3) (45 vs 14 months, p = 0.046). Survival data of TME and NLR subsets is shown in the table. Conclusions: This is the first characterization of TME profile in CUP. A diverse range of TME subtypes were seen in this population. While classical prognostic factors such as NLR, ECOG, NMS were associated with better survival, there also appears to be a trend toward survival benefit with IO in the IE/IE-F subset. Further studies are needed to prospectively explore the role of TME subtypes in determining clinical outcomes and IO response. Median overall survival (months) in TME and NLR subsets, with and without IO. Subset/Treatment IE/IE-F F/D P-value NLR High NLR Low P-value IO 45.3 15.6 0.14 11.2 22.7 0.04 No IO 14.6 10.8 10.8 57.4