Abstract 315P
Background
Cancers of Unknown Primary (CUP) are rare and heterogeneous tumors accounting for 2-3% of all cancers. PAX8 stains positive in renal, thyroid and mullerian cancers. CUP w/ Renal Profile (CUP-RP) has been proposed as a distinct subset, aiding in selection between cytotoxic chemotherapy for CUP compared to tyrosine kinase inhibitors (TKI) or checkpoint inhibitors used for renal cell carcinoma (RCC). PAX8 positivity is thus critical for personalizing therapy and improving outcomes.
Methods
369 CUP pts had PAX8 immunohistochemical (IHC) testing as a part of their standard clinical work-up, via a prospective/retrospective institutional CUP database at MD Anderson Cancer Center (Houston, TX). Pts were classified as PAX8+ CUP (cases, N=58, PAX8 IHC+) and PAX8-ve CUP (controls, N=311, PAX8 IHC-). Clinical and pathological data including therapy and survival were obtained. Statistical analyses included Fisher-exact test for contingency, Kaplan-Meier method to estimate overall survival (OS), and univariate and multivariate Cox regression for survival analysis.
Results
Median age for the entire cohort was 58 (range 20-90); 68% were female. For PAX8+ pts, pathology was adenocarcinoma in 33% (vs 44% for PAX8-ve). PAX8+ pts were less likely to have elevated neutrophil-to-lymphocyte ratio (NLR) (20% vs 38%; P=.03) and more often received VEGF TKIs or mTOR inhibitors (13% vs. 2%, P=.005). For the entire cohort, on multivariate analysis, worse OS was associated w/ age >60 years (HR 1.64, P=.01), ECOG PS ≥2 (HR 4.01, P<.001), NLR ≥5 (HR 2.15, P<.001), # of metastatic sites ≥3 (HR 1.83, P=.002), and LDH >ULN (HR 1.57, P=.01). Though PAX8+ was prognostic on univariate analysis, it was not an independent prognostic variable on multivariate analysis (HR 0.91, P=.69). Receipt of VEGF TKI or mTORi vs non-RCC therapy was not associated w/ improved OS (P=.81).
Conclusions
PAX8+ CUP is a distinct entity exhibiting unique clinical characteristics. PAX8+ CUP pts had improved prognosis compared to PAX8-ve CUP pts, driven by favorable clinicopathologic variables. Further efforts are needed to assess molecular underpinnings and if therapeutic outcomes in this subset mirror RCC pts.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.