Cancer-specific mortality in high-risk or very high-risk prostate cancer patients treated with focal therapy vs....

Focal therapy (FT) in high-risk or very high-risk prostate cancer (CaP) is not guideline-recommended. However, it may be occasionally used, but its effect on cancer-specific mortality (CSM) relative to standard treatment options such as radical prostatectomy (RP) is unknown and was addressed in the present study.

In the Surveillance, Epidemiology, and End Results (SEER) database (2010-2022), high-risk or very high-risk CaP patients according to the National Comprehensive Cancer Network (NCCN) guidelines, treated with either FT or RP, were identified. A multivariable logistic regression (MLR) model assessed the associations between sociodemographic characteristics and receipt of FT. Multivariable competing risks regression (CRR) models, after 1:1 propensity score matching (PSM), tested for CSM differences after adjustment for other cause mortality (OCM).

Overall, 20,192 high-risk or very high-risk CaP patients treated with FT or RP were identified. Of those, 416 (2.0%) underwent FT and 19,776 (98.0%) RP. In the MLR model, rural area of residence and unmarried status predicted a 2.4-fold (P < 0.001) and 1.8-fold (P < 0.001) higher likelihood of receiving FT, respectively. In PSM analysis, 416 FT patients could only be matched with one RP control from 19,776 RP patients. After 1:1 PSM, at 120 months of follow-up, CSM after RP was 9.2 vs. 16.0% after FT. The corresponding OCM rates were 19.5 after RP vs. 31.5% after FT. In multivariable CRR models adjusted for OCM, FT increased CSM in a 2.0-fold fashion (mHR: 1.98; 95% CI: 1.17-3.34; P = 0.011).

Despite lack of guidelines endorsement, occasional high-risk or very high-risk CaP patients may be treated with FT. Specifically, rural and unmarried patients with high-risk or very high-risk CaP are more likely to receive guideline-discordant FT. Such practice doubles the risk of CSM relative to RP, and its use should therefore be discouraged in this patient population.