High-grade nonmuscle-invasive bladder cancer (NMIBC) invading the lamina propria (HG T1) presents a challenge in determining the most appropriate treatment strategy due to its variable clinical behavior. While patients with HG T1 NMIBC can respond to intravesical bacillus Calmette-Guérin (BCG), those who progress are at higher risk of cancer specific mortality—these patients are best served with early radical cystectomy (RC) or enrollment in clinical trials. Existing risk stratification methods based solely on clinical and pathological factors have limitations in predicting the patients who will benefit most from immediate RC vs intravesical BCG. The focus of our current study is to summarize the contemporary outcomes of patients managed with BCG or immediate RC, with a focus on the mutational landscape across the spectrum of HG T1 disease.

We included all patients newly diagnosed with HG T1 NMIBC between 2014 and 2020 at our institution. All patients underwent restaging transurethral resection of bladder tumor with examination under anesthesia and cross-sectional imaging to confirm clinical stage. The outcomes of interest were pathological upstaging at RC, as well as bladder cancer–specific and overall survival. Furthermore, targeted exome sequencing using Memorial Sloan Kettering Cancer Center IMPACT (Integrated Mutation Profiling of Actionable Cancer Targets) was performed on 128 HG T1 tumors and compared with 191 BCG-naïve T2 tumors.

A total of 390 patients met the inclusion criteria, with a median follow-up time of 35 months. Among them, 308 patients received intravesical BCG as initial management, while 82 patients underwent immediate RC. Patients selected for immediate RC were younger and exhibited higher rates of concurrent carcinoma in situ (CIS), deep lamina propria invasion, lymphovascular invasion, and larger tumor size. Additionally, immediate RC patients were more likely to have persistent pT1 disease on re-resection and harbor variant histological subtypes. Notably, 40% of patients undergoing immediate RC were upstaged to pT2 and/or N+ disease at RC. Persistent pT1 disease at re-resection and the presence of variant histological subtypes were associated with pathological upstaging, while other factors such as tumor size, concurrent CIS, bladder neck involvement, or prostatic urethra involvement were not predictive. There were no differences in cancer-specific survival between these 2 groups, reflecting baseline differences in patient selection.

We then defined “very-high–risk T1” NMIBC as HG T1 with CIS and at least 1 additional risk factor (deep lamina propria invasion, lymphovascular invasion, any variant histology, prostatic urethra involvement, and size >3 cm). These tumors displayed a genomic alteration profile more similar to T2 tumors than NMIBC. Specifically, they exhibited more TP53, ERBB2 (Her2), and RB1 mutations, along with fewer FGFR3 mutations (see Figure). The presence of targetable alterations, including ERBB2/Her2, suggests potential therapeutic opportunities such as Her-2 antibody-drug conjugates, which could be explored in future clinical trials. The molecular similarity of very-high-risk T1 tumors and T2 tumors also may indicate a role for systemic therapy in combination with surgery or intravesical therapy, although this hypothesis remains to be tested.

Figure. Comparison of genomic alterations in standard-risk vs very-high–risk high-grade nonmuscle-invasive bladder cancer (A), and very-high–risk high-grade nonmuscle-invasive bladder cancer and T2 bladder cancer (B). *Standard risk = high-grade nonmuscle-invasive disease ± carcinoma in situ. **Very high risk = high-grade nonmuscle-invasive disease + carcinoma in situ + at least one of: T1b, lymphovascular invasion, variant histology, prostatic urethra, or size >3 cm. MIBC indicates muscle-invasive bladder cancer.

In conclusion, the study findings demonstrate that very-high–risk T1 tumors share a genomic similarity with muscle-invasive tumors and contain targetable alterations including Her-2. Integrating genomic sequencing with clinical and pathological features of HG T1 disease has the potential to enhance risk stratification and aid in treatment selection. This research underscores the importance of considering genomic information to optimize management decisions for patients with HG T1 NMIBC.