Age-, sex-, and race-related differences in outcomes of patients with muscle-invasive bladder cancer treated with radical cystectomy: The role of biology versus access to standard medical care.

person Akshay Sood The University of Texas MD Anderson Cancer Center, Houston, TX info_outline Akshay Sood, Gottfrid Sjödahl, Zhigang Duan, Hui Zhao, Niyati Lobo, Patrick J. Hensley, Kelly K. Bree, Sharon H. Giordano, Jeffrey S. Ross, Ashish M. Kamat

Authors person Akshay Sood The University of Texas MD Anderson Cancer Center, Houston, TX info_outline Akshay Sood, Gottfrid Sjödahl, Zhigang Duan, Hui Zhao, Niyati Lobo, Patrick J. Hensley, Kelly K. Bree, Sharon H. Giordano, Jeffrey S. Ross, Ashish M. Kamat Organizations The University of Texas MD Anderson Cancer Center, Houston, TX, Lund University Cancer Center, Lund, Sweden, University of Texas MD Anderson Cancer Center, Houston, TX, MD Anderson Cancer Center, Houston, TX, University of Texas at MD Anderson Cancer Center, Houston, TX, SUNY Upstate Medical University, Syracuse, NY Abstract Disclosures Research Funding Other Wayne B. Duddlesten Professorship in Cancer Research to Ashish M. Kamat Background: It remains unclear whether the poor outcomes associated with older age, female sex, and African-American (AA) race in patients with muscle-invasive bladder cancer (MIBC) treated with surgery are due to differential access to standard medical care or intrinsic differences in tumor biology. Methods: We designed an integrative study using two complementary datasets to answer our question. We used the SEER-Medicare MIBC dataset (n=2,327) to evaluate the real-world differences in survival outcomes among patients of female v male sex, AA v Caucasian race, and older v younger age before and after adjusting for demographic, tumor, treatment, and access-to-care factors. Next, we used The Cancer Genome Atlas (TCGA) project’s MIBC dataset (n=410) to evaluate the biological differences in molecular profiles of bladder tumors in these patient groups. Biological endpoints examined included DNA-level genomic alterations, mRNA expression subtypes, and RPPA-derived proteomic profiles, and clinical endpoints assessed included unadjusted and adjusted cancer-specific mortality (CSM) outcomes. Multivariable GLM, multivariate MMLR, and Fine-Gray competing-risk models were used. Results: Population-level analysis demonstrated that patients of female sex (10-y CSM: 51.3% v 43.8%, Gray’s p<0.001), AA race (10-y CSM: 56.2% v 46.1%, Gray’s p=0.022) and older age (10-y CSM: 47.8% >=81 y v 44.1% <=70 y, Gray’s p=0.032) had worse CSM at presentation. In adjusted analyses, however, only female sex remained associated with worse CSM (HR=1.17, p=0.028). The TCGA analysis revealed that advanced age was associated with increased total mutational burden and neoantigen load, both of which have been associated with improved survival. No other age-related differences were seen in other DNA alterations, mRNA expression subtype distribution, or RPPA proteomic cluster profiles. Similarly, no racial differences were observed in molecular alterations either at the DNA-, RNA-, or protein-level. However, females were noted to have adverse molecular profiles across all levels of molecular expression i.e. DNA-, RNA-, and proteomic-level. Specifically, females were noted to have lower total mutational burden (in stage 3-4 cancers, median 170 v 240, p=0.01), lower neoantigen load (in stage 3-4 cancers, median 382 v 551, p<0.01), higher Msig2 DNA mutational signature (in stage 3-4 cancers, 65.3% v 50.5%, p=0.04), higher basal/sq mRNA subtype (in stage 3-4 cancers, 52% v 34.2%, p=0.02), and lower RPPA cluster1 subtype (in stage 2 cancers, 15.2% v 36%, p<0.01). Conclusions: Our findings suggest that the poor outcomes seen among patients of AA race and older age in the real-world are likely driven by multilevel biases in healthcare delivery, while in women they may be a result of a combination of biological and access-to-care factors.