Upper tract urothelial carcinoma (UTUC) arising from the urothelium that lines the ureter and renal pelvis is a rare malignancy and accounts for only 5% of all urothelial carcinomas [1,2]. Although kidney-sparing surgery can be carried out in selected patients with low-risk UTUC, radical nephroureterectomy (RNU) with bladder cuff excision is considered the current standard of management for the majority of non-metastatic UTUCs [3]. The conventional surgical approach for RNU has been open nephroureterectomy (ONU). Recently, a shift toward minimally invasive treatments has emerged, and laparoscopic nephroureterectomy (LNU) has increasingly been used as an accepted alternative to ONU. LNU has been associated with reduced blood loss, faster recovery, and shorter hospital stay compared with ONU [4]. However, there remain some concerns about oncologic safety following LNU because of a higher risk of recurrence due to the high-pressure environment of the pneumoperitoneum, and the oncologic outcomes between ONU and LNU remain controversial.

In recent years, numerous studies have been conducted to compare the oncologic efficacy of ONU and LNU in patients with UTUC [5-13]. The comparative outcomes of these studies were various, without a definitely accepted conclusion on which surgical approach was more beneficial for patients with UTUC. Although a large number of studies in patients with UTUC undergoing RNU have shown no difference in recurrence-free survival (RFS), cancer-specific survival (CSS), or overall survival (OS) based on the type of surgical approach [5-10], some studies have shown an association of LNU with inferior CSS and OS in locally advanced UTUC [5,13]. In contrast, other studies reported that LNU could improve the CSS [11,12,14]. Thus, it is not fully established whether LNU is an effective and safe substitute for ONU in the treatment of UTUC. Furthermore, the results of previous retrospective studies were limited by the small number of patients and single-institution experience.

Knowledge of the efficacy of the two different surgical approaches is essential not only to guide patient counseling and clinical decision making, but also to develop clinical practice guidelines. The aim of this study was to compare the oncologic outcomes between ONU and LNU approaches in a large population obtained from a multicenter collaborative group. We also evaluated predictive factors associated with oncologic outcomes.

Overall, 1,521 patients were included in the present study. Of these, 906 (59.6%) underwent ONU and 615 (40.4%) underwent LNU. Median age was 65 years (IQR, 57 to 72) and 74.1% (1,127/1,521) of the patients were male. Median follow-up duration was 54.9 months (IQR, 32.7 to 89.7). The clinical and pathological details for each of the groups are described in Table 1. Relative to the ONU group, patients in the LNU group had significantly higher body mass index (p=0.003), higher American Society of Anesthesiologists score (p=0.017), less LVI (p=0.010), were less likely to have lymph node metastases (p < 0.001), and had a shorter followup duration (p < 0.001).

Survival outcomes are summarized in Table 2. During follow-up, there were 631 (41.5%) intravesical recurrences, including 396 (43.7%) in the ONU group and 235 (38.2%) in the LNU group. The 5-year IVRFS estimates were 51.0% and 57.7% for patients treated with ONU or LNU, respectively, and this difference was statistically significant (p=0.010) (Fig. 1A). The total number of patients showing progression in ONU and LNU groups was 272 (30.0%) and 155 (25.2%), respectively. The 5-year PFS estimates for ONU and LNU were 68.9% and 73.9% respectively, which was not significantly different (p=0.079) (Fig. 1B). Overall, 453 (29.8%) patients died during the study period, including 307 (33.9%) in the ONU group and 146 (23.7%) in the LNU group, and 342 UTUC-related deaths occurred (229 in ONU group and 113 in LNU group). The 5-year CSS estimates and the 5-year OS were 76.4% and 71.4% respectively for patients treated with ONU versus 80.4% and 75.8% for patients treated with LNU. The LNU group showed better 5-year CSS (p=0.032) (Fig. 1C) and OS (p=0.026) (Fig. 1D) than the ONU group.

When patients were stratified by pathological T stage, the 5-year IVRFS (p=0.078), PFS (p=0.775), CSS (p=0.994), and OS (p=0.859) of the two groups were similar for patients with organ-confined disease (pTis/pTa/pT1/T2) (Fig. 2). In contrast, 5-year IVRFS (p=0.038) (Fig. 3A), CSS (p=0.022) (Fig. 3C), and OS (p=0.018) (Fig. 3D) were significantly different between the two groups in favor of LNU for patients with locally advanced disease (pT3/pT4). No significant difference in the 5-year PFS (p=0.067) was observed when comparing the two groups for patients with locally advanced disease (pT3/pT4) (Fig. 3B).

The results of the multivariable Cox regression analysis are shown in Table 3. A history of previous bladder tumor and presence of concomitant bladder tumor were independent predictive factors of intravesical tumor recurrence, progression, death from UTUC, and all-cause death. Age, pathological T stage, tumor grade, presence of concomitant LVI, presence of concomitant CIS, and pathological N stage were significantly associated with progression, death from UTUC, and all-cause death. The surgical approach was revealed as an independent prognostic factor for intravesical tumor recurrence, but was not related to progression, death from UTUC, and all-cause death.

The current gold standard treatment for UTUC is RNU with bladder cuff excision [3]. ONU, the conventional surgical approach that supports favorable cancer control, has long been accepted as the standard surgical approach for UTUC, especially for large or locally advanced UTUC. Recently, minimally invasive approaches, such as LNU or robotic nephroureterectomy, have gained wide acceptance as viable alternatives to traditional ONU due to their faster recovery time and a decreased likelihood of perioperative complications. Actually, after the first successful LNU procedure in 1991 by Clayman et al. [19], widespread use of LNU was initially limited by concerns over tumor cell dissemination in a pneumoperitoneal environment and the significant operator learning curve. With increased operative skills and the demonstration of comparable oncologic outcomes, utilization of LNU has gradually increased. In the United States, utilization of LNU or robotic nephroureterectomy increased from 57.7% to 71.5% from 2010 to 2013, whereas use of ONU decreased from 42.3% to 28.6% [20].

Numerous studies have evaluated the oncologic outcomes of ONU versus LNU and demonstrated comparable oncologic results between the two different surgical techniques [5-13]. A randomized trial by Simone et al. [5] compared 40 ONU patients with 40 LNU patients. They found that IVRFS, metastasis-free survival, and CSS were not significantly different between the two groups at a median follow-up of 44 months. The 5-year CSS was 89.9% and 79.8% for the ONU and LNU patients, respectively (p=0.2). Similarly, several multicenter retrospective studies showed oncologic equivalence between ONU and LNU with regard to RFS and CSS [8,9]. Walton et al. [8] retrospectively evaluated 703 ONU patients and 70 LNU patients at nine centers worldwide with median follow-up of 34 months. They reported 5-year RFS of 73.7% and 63.4% (p=0.124) and 5-year CSS of 75.4% and 75.2% (p=0.897) for the ONU and LNU groups, respectively. Likewise, Ariane et al. [9] also demonstrated that LNU could provide equivalent oncologic outcomes compared to ONU with median follow-up of 27 months in a large French multicenter collaborative study. In their study, no significant difference was observed in the 5-year RFS between the ONU (50.7%) and LNU (52.2%) patients (p=0.7). The 5-year CSS was 78% for the ONU patients and 90.7% for the LNU patients, but this difference was not statistically significant (p=0.06). Of note, in subgroup analysis by tumor stage, there were no significant differences between the two procedures in the 5-year RFS or the 5-year CSS for any of the pathological tumor stages. A meta-analysis published by Ni et al. [14] including 21 publications also reported no significant differences in the 5-year RFS and 5-year OS between LNU and ONU. However, a few studies have reported that 5-year CSS and OS rates were lower in LNU patients than in ONU patients with locally advanced disease. Simone et al. [5] reported that CSS was significantly different between the two surgical techniques in favor of ONU for pT3 tumors (p=0.039). A retrospective comparative analysis by Kim et al. [13] demonstrated that LNU patients showed inferior 5-year CSS (p=0.015) and OS (p=0.027) compared with ONU patients. Moreover, the 5-year CSS and OS for pT3/pT4 tumors were significantly lower in the LNU patients than in the ONU patients (p=0.005 and p=0.007, respectively).

In contrast to these results, two comprehensive metaanalyses in the literature reported that LNU was superior to ONU in terms of CSS. Zhang et al. [12] conducted a systemic review and meta-analysis to evaluate the oncologic outcome associated with two different surgical approaches (ONU and LNU) across 21 retrospective studies. They demonstrated that LNU showed better CSS compared with ONU (hazard ratio, 0.79; 95% confidence interval, 0.68 to 0.91). Likewise, Ni et al. reported that the 5-year CSS rate was notably higher for patients who underwent LNU (9%) than for those who underwent ONU (p=0.03) [14]. In this study, IVRFS was significantly lower in LNU patients, at 17% (p=0.02).

The present study from the Urothelial Cancer-Advanced Research and Treatment study group extended previous studies to evaluate the associations between surgical approach and oncologic outcomes in a larger cohort of patients with UTUC. The 5-year CSS of this study was 76.4% and 80.4% for ONU and LNU, respectively. These data were similar to those of previous studies, which reported 5-year CSS of 73-90% and 75-91% for ONU and LNU patients, respectively [5,6,8,9]. A summary of our results shows significant differences in IVRFS, CSS, and OS between ONU and LNU. The 5-year IVRFS, CSS, and OS rates were lower in the ONU group than in the LNU group and the benefit of LNU was especially apparent in the subgroup with locally advanced disease (pT3/pT4). The type of surgical approach for RNU was not a significant predictor of oncologic results; significance was achieved for IVRFS, but not for PFS, CSS, and OS. However, these results should be interpreted with caution. The reason for superiority of oncologic outcomes in the LNU group is uncertain and might be affected by several factors, including patient’s clinical and pathological characteristics, surgical experience, and extent of regional lymph node dissection [21]. Above all, it could be mainly affected by selection bias. The choice of surgical approach was usually determined by the surgeon’s preference in addition to the patient’s baseline characteristics. Although absolute indications for each surgical approach are not clearly defined, it is important to select appropriate patients to ensure optimal oncologic outcomes and safety. Many surgeons performing LNU tend to select patients who generally have a good comorbidity profile and typically offer the ONU procedure to patients with more aggressive and bulky UTUC. Therefore, although there were no differences in pathological T stage and tumor grade between two surgical approaches it is possible that ONU patients had more aggressive tumors, which may have affected the oncologic outcome. Indeed, the LNU group was less likely to have lymph node metastases relative to the ONU group in the present study (Table 1).

The present study has some other limitations. Foremost, due to the retrospective and nonrandomized study design, unidentified confounding variables may have been present. Second, multiple surgeons in five different centers performed ONU and LNU. Thus, the individual learning curve of each surgeon could have been a source of bias. Furthermore, the patients analyzed in ONU and LNU groups are not contemporary; the outcomes from the ONU group are based on less contemporary patient cohorts. Because we did not consider the effect of surgeon or time period, it remains unclear whether these factors influenced the results. Third, although all participating centers in this study usually followed recommendations and institutional protocols [3], there was a lack of standardization of the selection criteria and surgical approaches. In addition, we excluded patients with largely incomplete information from our analysis, which could possibly create selection bias. As previously mentioned, this selection bias might affect evaluation of the real impact of surgical approach on oncologic outcomes.

The present study shows that LNU provided better oncologic control of IVRFS, CSS, and OS compared to ONU for the management of patients with UTUC. The clinical benefit was more pronounced among patients with locally advanced disease. Further multicenter randomized trials are necessary to definitively prove that LNU is a safe alternative surgical approach to ONU in patients with UTUC, especially those with more advanced disease.