Endoscopic Thulium Laser Ablation of Upper Tract Urothelial Carcinoma: a Retrospective Study with Subgroup Analyses

The gold standard treatment for upper tract urothelial carcinomas (UTUC) is radical nephroureterectomy with bladder-cuff excision. However, the role of combination of ureterorenoscopy and laser ablation has recently become more essential. This study aimed to investigate the efficacy of Thulium laser ablation of UTUC and estimate clinical outcomes with subgroup analyses. Methods After obtaining informed consent, we retrospectively reviewed all patients who had undergone endoscopic Thulium laser ablation of UTUC as their primary treatment from Jun 2012, to Nov 2018. Sixty-eight patient were enrolled, and thirty-four patients were analyzed after selection by exclusion criteria. Comparisons with bivariate analyses between patients with and without recurrence were examined. Multivariable cox regression model were further applied. Further, Kaplan-Meier survival estimate were presented with comparison of the survival curves, we also combined two of three tumor characteristics (stage, size, and grade of tumor) for subgroup analyses.


Introduction
The gold standard treatment for upper tract urothelial carcinomas (UTUC) is radical nephroureterectomy with bladder-cuff excision. However, the role of combination of ureterorenoscopy and laser ablation has recently become more essential. This study aimed to investigate the efficacy of Thulium laser ablation of UTUC and estimate clinical outcomes with subgroup analyses.

Methods
After obtaining informed consent, we retrospectively reviewed all patients who had undergone endoscopic Thulium laser ablation of UTUC as their primary treatment from Jun 2012, to Nov 2018.
Sixty-eight patient were enrolled, and thirty-four patients were analyzed after selection by exclusion criteria. Comparisons with bivariate analyses between patients with and without recurrence were examined. Multivariable cox regression model were further applied. Further, Kaplan-Meier survival estimate were presented with comparison of the survival curves, we also combined two of three tumor characteristics (stage, size, and grade of tumor) for subgroup analyses.

Results
Fifteen patients (44%) had local tumor recurrence. Seven patients (21%) had recurrence at the urinary bladder. Two patients (6%) had lymph node or distal metastasis. A total of 4 patients (12%) had cancer-specific death. In bivariate analyses, statistical significance is noted for age, tumor grade and tumor size with p-values of 0.018, 0.047 and 0.014, respectively. In the multivariable analysis, statistical significance is noted for age and tumor size with p-values of 0.017, and 0.042, respectively.
In Kaplan-Meier estimate for subgroup analysis, statistical significance is only noted in the tumor size + tumor grade group but almost achieved in the tumor size + biopsy stage group with p-value of 0.0275 and 0.0518, respectively.

Conclusion
In our experience, tumor size and tumor grade are influential recurrence factors. However, biopsy stage does not achieve statistical significance in both the recurrence analysis and survival estimate.
In the subgroup analyses, tumor size is a more influential factor than tumor grade in the prognosis of Introduction Upper tract urothelial carcinomas (UTUC) are rare malignancies and account for 5%-10% of all urothelial carcinomas (UC) [1]. However, an increased prevalence was noted at high-arsenic exposure areas, such as southwestern and northeastern Taiwan, where residents are at high risk of lung cancer, bladder cancer, and black foot disease. Residents have consumed arsenic-contaminated water from artesian wells for more than 40 years, and the odds ratio of bladder cancer is 4.10 [2].
This group of patients has unique characteristics, (1) arsenic exposure will disproportionally induce a high incidence rate of UTUC (20%-25%) among all UCs [3,4]. (2) Ureteral UCs are twice as common as renal pelvic UCs. (3) The male-to-female ratio among this group of patient is 1:2, in contrast to the male predominance seen in other regions of the world [5].
The incidence of UTUC in Western countries is 2.0/100,000 persons and accounts for only 5%-10% of UCs. Renal pelvic UCs are approximately twice as common as ureteral UCs [6]. However, the incidence of UTUC in Taiwan is 5.56/100,000 persons in males and 7.37/100,000 persons in females [7]. Due to its high prevalence in southwestern Taiwan, we thus managed more patients with UTUC than other regions did. Compared with urinary bladder UC, UTUC tends to have a worse prognosis as a multifocal disease with a high recurrence rate; the 5-year urinary bladder recurrence rate is 28%, and the 5-year mortality rate is 23% [8].
Conventionally, the gold standard treatment for UTUC is radical nephroureterectomy (RNU) with bladder-cuff excision. Because of the continuous progress of the image and endoscopic armamentarium, the role of kidney-sparing surgery (KSS) has recently become more essential.
According to the latest European Association of Urology (EAU) guidelines [6], well-selected patients with UTUC are suitable candidates for KSS. In low-risk patients (unifocal disease, tumor size < 2 cm, low-grade cytology, low-grade URS biopsy, and no invasive aspect on computed tomography urography [CTU]), similar cancer-specific survival (CSS) is noted between the KSS and RNU groups [9].
Due to the advanced development of flexible ureterorenoscopy (URS), we can check the collecting system in detail. URS treatment for UTUC by electrode has shifted to laser ablation with Holmium:YAG laser or Thulium laser [10,11]. Laser ablation has been investigated since the last decade, compared with traditional electrode resection; the advantage is that it provides a continuous wave for sharply defined resection and fulguration of the tumor bed with excellent hemostasis. Thulium laser ablation showed superior recurrence-free survival, reduced bleeding, and less mucosal perforation compared with Holmium laser ablation [12].
This study aimed to investigate the efficacy of laser ablation of UTUC and estimate clinical outcomes with subgroup analyses.

Materials And Methods
After obtaining informed consent, we retrospectively reviewed all patients who had undergone Three months after the first endoscopic Thulium laser ablation, 2nd -look URS was performed. A routine biopsy of the previous ablation site was performed to determine the recurrence status and could be followed by a second Thulium laser ablation of the previous tumor site. The second procedure of laser ablation might not be performed if no suspicious lesion was found. However, if the biopsy of 2nd -look URS showed recurrence, then the third endoscopic Thulium laser ablation would be scheduled. Patients were followed by a stringent surveillance protocol, including cystoscopy, retrograde pyelography, and URS every 3 months until they were tumor-free. Once the patient was free of tumor, a URS evaluation was scheduled every 6 months for 5 years and then yearly thereafter.
A CTU or MRI with contrast was ordered if the baseline retrograde pyelography suggested local recurrence of UTUC.

Statistical Analysis
Continuous and categorical data in the clinical characteristics of patients and tumors were separately presented by count with percentage and median with range (Table 1 & Table 2). Comparisons with bivariate analysis between patients with and without recurrence were examined by Wilcoxon rank sum test for continuous data and by Chi-square test or Fisher's exact test for categorical data (Table 3). Multivariable cox regression model, which is adjusted by age, biopsy stage, tumor grade and tumor size, was further applied for identifying the association between clinical characteristics and risk factors of recurrence (Table 4). Survival estimates were calculated by Kaplan-Meier method (Table 5). Further, Kaplan-Meier survival plot were presented for each group and the comparison of survival curves were tested by log rank test. We also combined two of three tumor characteristics (stage, size, and grade of tumor) for subgroup analyses. In this study, the difference were considered significant if the p-value was smaller than 0.05.     Table 3).
The incidence of UTUC in women was more than 2 fold higher than that in men, but the male-tofemale hazard ratio of UTUC recurrence was 2.090 with 95% CI (0.704-6.201) ( Table 4).
We treated UTUC of the renal pelvis by Flexible URS, compared with ureteral UTUC treated by rigid URS; no significance is noted between these two groups.
In both bivariate and multivariable analyses, tumor stage was not significant. Only 2 patients had Tis stage; no recurrence was found in these patients. Most patients (n = 20) have small tumors (< 2 cm); after laser ablation, the recurrence rate is 25%. Both bivariate and multivariable analyses achieved statistical significance. The hazard ratio of large tumors (> 2 cm) to small tumors was 3.225 with 95% CI (1.041-9.988).
Five patients received intra-ureteral (or intra-pelvic) chemotherapy with mitomycin or epirubicin, the recurrence rate of patients with or without intraureteral chemotherapy is not statistically significant.
We used the Kaplan-Meier estimate for survival rate analysis (Table 5). Because the mean duration of follow-up was 25 months, we estimated 2 years of local recurrence-free survival, bladder recurrencefree survival, metastasis-free survival, and cancer-specific survival. With regard to the 2 years of local recurrence-free survival, the use of the Flexible URS for laser ablation of UTUC was not inferior to the rigid URS, the p-value of the log-rank test was 0.972.
High-grade UTUC is a higher tendency of recurrence than did a low-grade tumor and had statistical significance with a p-value of 0.0477 (Fig. 1). Biopsy stage was not significant by the log-rank test (Fig. 2). Large tumors (> 2 cm) had a higher tendency of recurrence than did small tumors (< 2 cm) and showed statistical significance with a p-value of 0.0066 (Fig. 3).
We combined two of three tumor characteristics (stage, size, and grade of tumor) for subgroup analyses. In the tumor grade + tumor size group, statistical significance was noted with a p-value = 0.0275. We noticed that small tumor size was still a lower recurrence event than of larger tumors despite that the analysis was combined with tumor grade (Fig. 4). In the tumor size + biopsy stage group, statistical significance was almost achieved as noted by the p-value of 0.0518 (Fig. 5). In the tumor grade + biopsy stage group, no significance was noted (Fig. 6).

Discussion
EAU guidelines verified the minimally invasive and conservative kidney-sparing approach for low-risk UTUC as an alternative treatment for RNU [6]. As EAU and NCCN guidelines suggest, URS treatment for UTUC should be performed in patients with low grade tumors; UTUC in solitary kidney, impaired renal function and bilateral UTUC was also imperatively indicated for KSS in highly selected patients [6,13].
In a systematic review by Seisen et al. [9], only patients with low-grade and non-invasive UTUC had similar survival after URS management for UTUC versus RNU. However, in our study, only the tumor grade group and tumor size group achieved statistical significance. However, the biopsy stage group was not statistically significant. This difference may imply that both tumor size and tumor grade are more influential factors than biopsy stage in local recurrence-free survival. We surmise that the biopsy stage by URS is not consistent with the actual tumor stage because biopsy cannot precisely grasp the actual depth of tumor invasion. However, tumor size may be indicative of the grade of tumor invasion.
Due to the high prevalence of UTUC in southwestern Taiwan, many patients cannot fit in the low-risk group. In the review by Fiuk et al., they extended the indication for KSS to the following populations: (1) UTUC in solitary kidney, (2) renal insufficiency imposing the threat of hemodialysis if any further renal insult, (3) bilateral UTUC, (4) comorbidities enough to be life-limiting or to incur additional risk with RNU, and (5) low-risk tumors with stringent follow-up [14].
Therefore, we selected patients with imperative indications and analyzed the results after laser ablation of UTUC. In our study, female predominance is noted; it is speculated that female patients were highly exposed to arsenic steam during cooking from the boiling water [15]. However, our statistics showed that men had a higher tendency of recurrence with near statistical significance. Age was statistically significant in the multivariate analysis, but Chromecki et al. suggested that age was not an independent predictor of outcomes of UTUCs [16].
We used the Flexible URS combined with the Thulium laser system for treating renal pelvic UTUC. The statistical results showed no significant difference between the renal pelvic tumor (flexible URS) and ureteral tumor (rigid URS), and may result from advances in current technology that allows us to approach all parts of the renal collecting system.
In the NCCN guidelines, low-grade UC of the renal pelvis treated by endoscopic resection is suggested to be followed by postsurgical intra-ureteral (intra-pelvic) chemotherapy or BCG [13]. In our study, a small number of patients (n = 5) received intra-ureteral (intra-pelvic) chemotherapy. The outcome of local recurrence was not statistically significant. Due to the high risk of acute pyelonephritis after intra-ureteral (intra-pelvic) chemotherapy, we terminated the therapy after administering it to a few patients. Repeat acute pyelonephritis will also induce permanent deterioration of renal function.
The recurrence rate of bilateral UTUC is lower than that of unilateral UTUC, but only 5 patients had bilateral disease; the statistical analysis also showed no significance. The tumor characteristics of these 5 patients were small tumor size and earlier tumor stage (Ta and Tis); thus, the lower rate of recurrence.
We combined two of three tumor characteristics (stage, size, and grade of tumor) for subgroup analyses. In the tumor grade + tumor size group, statistical significance was noted with a p-value = 0.0275 (Fig. 4). We noticed that the subgroup of small tumor size (tumor < 2 cm with low/high grade) still had a lower local recurrence than did large tumors (tumor > 2 cm with low/high grade) (Fig. 4).
The subgroup analysis of tumor size + biopsy stage also showed the same pattern as that small tumor size (tumor < 2 cm with Tis/Ta/T1) had a lower local recurrence than did large tumors (tumor > 2 cm with Tis/Ta/T1) (Fig. 5). However, no statistical significance was achieved in the subgroup analysis of tumor grade + biopsy stage (Fig. 6). These results imply that the factor of tumor size was more influential than tumor grade and biopsy stage in the prognosis of local recurrence.
Despite considerable experience treating UTUC by endoscopic management, risks of both understaging and undergrading remain. Up to 25% of UTUC lesions had been missed, and almost 50% of carcinoma in situ lesions were missed [19].
Because we excluded 34 patients due to (1) short follow-up duration (less than 6 months), (2) previous history of urinary bladder UC, and (3) concomitant diagnosis of UTUC and bladder UC at the first visit, it resulted in a small number of patients who were included in our study. Also, laser ablation is a novel technique for the management of UTUC in the recent decade and results in a short followup duration. We need a longer follow-up duration for an accurate survival estimate.
We shared our experience of laser ablation of UTUC to provide direction to regions or countries that have a high incidence of UTUC, such as southwestern and northeastern Taiwan, the Balkan area, and other Chinese herb-consuming regions.

Conclusion
Upper tract urothelial carcinoma is a multifocal disease with a high recurrence rate. In our experience, tumor size and tumor grade are influential recurrence factors in kidney-sparing surgery. However, biopsy stage did not achieve statistical significance in both the recurrence analyses and survival estimate. In the subgroup analysis, tumor size was a more influential factor than tumor grade in the prognosis of local recurrence. High-grade UTUC is a higher tendency of recurrence than did a low-grade tumor and had statistical significance with a p-value of 0.0477.

Figure 2
Biopsy stage was not significant by the log-rank test. Large tumors (>2 cm) had a higher tendency of recurrence than did small tumors (<2 cm) and showed statistical significance with a p-value of 0.0066.

Figure 4
We noticed that small tumor size was still a lower recurrence event than of larger tumors despite that the analysis was combined with tumor grade.

Figure 5
In the tumor size + biopsy stage group, statistical significance was almost achieved as noted by the p-value of 0.0518.

Figure 6
In the tumor grade + biopsy stage group, no significance was noted.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.