Purpose: Traditional biomarker investigation schemas for chemotherapy response prediction in patients with bladder cancer relying on the pT0 status at radical surgery are confounded by the therapeutic effect of transurethral resection of bladder tumor (TURBT). Studying cN+ patients and assessing pN0 status presents a unique opportunity for overcoming this bias. Materials and Methods: We studied 26 patients with biopsy-proven cTanyN1-3M0 bladder cancer (2005-2016) who underwent induction chemotherapy and radical cystectomy with pelvic lymphadenectomy. Metastasis and overall survival (OS) outcomes were examined using Kaplan-Meier method and compared using log-rank test. Paired pretherapy primary bladder and nodal tissues were available for 10 patients. In these patients, whole-transcriptome RNA-seq analysis was performed on bladder tumor (for pT0 and pN0 prediction) and lymph-nodal metastasis (for pN0 prediction) tissues to identify differentially expressed genes (DEGs) with a false discovery rate < 0.1. Results: pN0 pathologic responders, but not pT0 responders, had significantly improved freedom from metastasis (5-year: pN0 88.9% vs pN+ 27.4%, log-rank P = .018) and OS (5-year: pN0 76.2% vs pN+ 12.2%, log-rank P = .024). Using RNA-seq data, we identified a significant discordance rate of 87.5% between DEG-based predictive signatures for pT0 and pN0 response to cisplatin-based chemotherapy. This datum, combined with the knowledge that ∼40% of patients who achieve pT0 status at radical surgery, achieve it simply through the therapeutic effect of TURBT (SWOG S8710), underscores a substantial bias in the current biomarker discovery initiatives that use pT0 as an end point. Conclusions: Our findings suggest a need for devising novel study designs to aid in the discovery of reliable biomarkers for preoperative chemo/immunotherapy response in bladder cancer. Clinical node-positive patients may be ideally situated but remain understudied.
Read full abstract