The impact of smoking on genomic alterations in muscle-invasive bladder cancer

View Poster


Smoking is the most important risk factor for developing muscle-invasive bladder cancer (MIBC). The incidence of MIBC, as with lung cancer, correlates directly to both the duration (years) of smoking and the number of cigarettes smoked per day. Carcinogenic toxins similar to cigarette smoke are administered to rodents to model the effects of smoking on the development of MIBC. Here we aimed to investigate the correlation between genomic alterations present in MIBC and the patient smoking status. Furthermore, we compared these findings with genomic alterations present in lung adenocarcinoma (LAC) and with the biological characteristics of chemically induced MIBC in rodents.


Clinical and genomic data from MIBC and LAC were downloaded from the TCGA and MSK-IMPACT ( The TCGA dataset contained genomic DNA sequencing, mutation count in Signature 5*, total mutational burden, RNA sequencing, follow-up, pack years and smoking history (current/former/never). By using the gene expression data, molecular subtype calls were generated by our recently published genomic subtyping classifier. From the MSK-IMPACT dataset, mutational burden and smoking history were used. The effects of smoking exposure on patient outcome, mutational burden, Signature 5* and molecular subtypes was analyzed in these datasets.


In the bladder TCGA, smoking history, as well as pack years, were not related to disease-free (p=0.4) or to overall survival (p=0.2). There was no association of smoking history (p=0.8) or pack years (p=0.9) to mutational burden in either the bladder TCGA or the MSK-IMPACT datasets. Mutation count in Signature 5* was higher in former or current smokers, but this did not reach significance (p=0.051). This was in contrast to findings in LAC. Mutational burden was higher in cases with more pack years (p=.005). Moreover, current and former smokers had a higher mutational burden when compared to never smokers. Although others have reported that carcinogen-induced MIBC in rodents has almost exclusively a basal-like phenotype (Kim PloS one 2011; Bivalacqua J Urol 2017), a similar association between smoking history and molecular subtypes of MIBC in TCGA patients was not apparent. We generated subtype calls using other published methods for molecular subtyping (Sjödahl Clin Cancer Res 2012, Choi Cancer Cell 2014, TCGA Nature 2014) but found here also no enrichment of one molecular subtype in current and/or former smokers.


Although smoking is recognized as a major risk factor for developing MIBC, its impact on genomic alterations in MIBC is not obvious. This is in contrast to LAC in which smokers have a higher mutational burden and enrichment of mutations in known oncogenes (e.g. TP53). Cigarette smoke contains over 4000 compounds, which may prevent a direct comparison to the rodent model that is induced by a single toxin.

Funding: none