Presentation Authors: Keisuke Shigeta*, Eiji Kikuchi, Takayuki Abe, Masayuki Hagiwara, Koichiro Ogihara, Toshikazu Takeda, Kazuhiro Matsumoto, Ryuichi Mizuno, Masahiro Jinzaki, Mototsugu Oya, Tokyo, Japan
Introduction: The clinical indication of laparoscopic radical nephroureterectomy (LRNU) for locally advanced upper tract urothelial carcinoma (UTUC) patients remain controversial. Our specific aim is to examine whether LRNU could have the same oncological outcome as that of open radical nephroureterectomy (ORNU) for both clinical and pathological T3N0M0 UTUC patients in a propensity adjusted cohort using a multi-institutional collaboration dataset.
Methods: Among the 670 UTUC patients who underwent RNU from 1990 to 2015, we identified 376 pT3N0M0 UTUC patients as cohort 1. Of the 405 UTUC patients who underwent RNU and were preoperatively evaluated by CT urography from 2005 to 2015, we identified 143 cT3N0M0 UTUC patients as cohort 2. We adopted a propensity score (PS) matching method to both cohorts so as to achieve uniform homogeneity with respect to patient background. PS matching-adjusted Cox-regression analysis was performed to evaluate the impact of surgical modality on oncological outcomes.
Results: After PS matching, no significant differences were observed among the patient backgrounds between the LRNU and ORNU patients in each cohort. In cohort 1, the 5-year cancer-specific survival (CSS) was 42.4% in the LRNU group, which was significantly lower than the ORNU group (57.6%, p=0.047) (Figure 1a). Multivariate analysis indicated that the type of surgical approach (LRNU vs. ORNU) (HR=1.84 p=0.028) as well as lympho-vascular invasion (LVI) (HR=2.59, p=0.003) became the independent prognostic factors for poor CSS. Referring to intravesical recurrence (IVR), the 5-year IVR free survival was 46.3% in the LRNU group, which was significantly lower than the ORNU group (64.0%, p=0.042). From the multivariate analysis, LRNU (HR=1.74, p=0.041) and previous history of bladder cancer (HR=2.35, p=0.007) were independent risk factors for developing subsequent IVR. In cohort 2, the 5-year CSS was 44.8% in the LNU group, which was significantly lower than the 60.0% in the ONU group (p=0.049) (Figure 1b). Similar to the result for cohort 1, multivariate Cox regression analysis showed that LRNU remained as one of the independent factors for both poor CSS and developing IVR.
Conclusions: From both pathological and clinical settings, our study has demonstrated that LRNU may have poorer long-term oncological outcomes than ORNU in T3N0M0 UTUC patients whose potential baseline differences were adjusted by PS matching.