Comparisons of Treatment Outcomes and Patterns of Lymph Node Involvement in T4 Prostate Cancer Patients

Little is known about the clinical behaviors and optimal treatment of T4 prostate cancer. Previous studies on locally advanced prostate cancers have a critical lack of T4 patients to compare. In this study, we analyzed patterns of lymph node (LN) involvement and associations between treatment outcomes in a large cohort of T4 prostate cancer patients. Patients with T4 prostate cancer were identified using Natural Language Processing. Patients who had received previous treatment prior to presentation at our institution were excluded. A dedicated GU radiologist (T.B.) reviewed all existing pretreatment imaging to identify LN involvement and T4 status. Categorical variables were compared utilizing the chi-squared test while overall survival was calculated via the Kaplan Meier method with comparisons via the log rank test. Hazard ratio was calculated using Cox proportional Hazards Regression. 481 patients with T4 prostate cancer were diagnosed between 1996 and 2017; 103 met our inclusion criteria for treatment-naïve T4 condition. T4 status was identified by either clinical examination (colonoscopy, cystoscopy, and digital rectal exam; 56.3%) or imaging (43.7%); 45% of patients were M0, 55% M1, 25% N0, and 75% N1. The median age was 62 (range 44-88) and median pre-treatment PSA was 17.5 (range 0.5-900). The majority of patients presented with an adenocarcinoma histology (78.6%) followed by ductal (8.7%) and small cell (5.8%). The median follow-up time was 103.5 months (mo). Patients with any rectal involvement by the primary tumor (n = 42) exhibited significantly increased rates of perirectal and mesorectal lymph node (LN) involvements than patients without (n = 61) (45% vs 26%, p = 0.046), with no significant links with disease invasion to the bladder or pelvic side wall (p = 0.61, 0.67). Definitive local treatment (LT: surgery, definitive radiation therapy, or both) was given to 56 patients (54%). Patients who received LT demonstrated significantly higher rates of survival (median 80 vs 39 mo; HR = 0.37, p = 0.0002). Subgroup analyses revealed a trend towards improved survival with LT among N1 patients (74 vs 39 mo; p = 0.08) as well as M1 patients (74 vs 37.5 mo, p = 0.14). This analysis identified an increased frequency of perirectal and mesorectal LN involvement among patients with rectal invasion by the primary tumor. LT was associated with significantly improved overall survival, which appears to extend to patients with M1 and N1 disease. These results support the utilization of definitive local therapy among T4 prostate cancer patients with benefits that possibly extend to those with regional or metastatic disease. Further consideration should also be given to include mesorectal and perirectal LNs in pelvic radiation fields in patients with rectal involvement from the primary disease.