Ann Surg Oncol (2015) 22:1392–1400 DOI 10.1245/s10434-014-4103-x
ORIGINAL ARTICLE – UROLOGIC ONCOLOGY
Prognostic Significance of Lymphovascular Invasion in Upper Urinary Tract Urothelial Carcinoma is Influenced by Tumor Location Hsiang-Ying Lee, MD, MS1,2, Ching-Chia Li, MD1,2,3,4, Chun-Nung Huang, MD, PhD1,3, Hung-Lung Ke, MD, MS1,3,4, Wei-Ming Li, MD, MS1,3,4,5, Peir-In Liang, MD6,7, Sheau-Fang Yang, MD6,7,8, Hung-Pin Tu, PhD9, Wen-Jeng Wu, MD, PhD1,3,4,10,11, and Hsin-Chih Yeh, MD, MS1,2,3,4 Department of Urology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan; 2Department of Urology, Kaohsiung Municipal Ta-Tung Hospital, Kaohsiung, Taiwan; 3Department of Urology, School of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 4Graduate Institute of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 5Department of Urology, Ministry of Health and Welfare, Pingtung Hospital, Pingtung, Taiwan; 6Department of Pathology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan; 7 Department of Pathology, School of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 8 Department of Pathology, Kaohsiung Municipal Ta-Tung Hospital, Kaohsiung, Taiwan; 9Department of Public Health and Environmental Medicine, School of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 10 Center for Infectious Disease and Cancer Research, Kaohsiung Medical University, Kaohsiung, Taiwan; 11Department of Urology, Kaohsiung Municipal Hsiao-Kang Hospital, Kaohsiung, Taiwan 1
ABSTRACT Background. To examine the prognostic value of lymphovascular invasion (LVI) in different tumor locations (i.e., renal pelvis and ureter) of upper urinary tract urothelial carcinoma (UTUC). Methods. Data from a total of 250 patients with nonmetastatic UTUC who received radical nephroureterectomy between 2004 and 2010 at our institution were analyzed retrospectively. The significance of LVI and other relevant factors on cancer-specific survival (CSS), metastasis-free survival (MFS), and intraluminal recurrence-free survival (IRFS) were evaluated. Results. Lymphovascular invasion was present in 60 patients (24 %) and was related to advanced pathological T Hsiang-Ying Lee and Ching-Chia Li contributed equally to this article, and both should be considered first author. Ó Society of Surgical Oncology 2014 First Received: 23 May 2014; Published Online: 20 September 2014 Wen-JengWu, MD, PhD e-mail: [email protected] H.-C. Yeh, MD, MS e-mail: [email protected]
stage (P \ 0.001), higher tumor grade (P \ 0.001), lymph node metastasis (P = 0.005), and pyelocaliceal tumor location (P = 0.002). By Kaplan–Meier analysis, LVI was found to be significantly correlated with worse CSS and MFS but not with IRFS. Multivariate analysis showed that high pathological T stage and regional lymph node involvement were significant prognostic factors for CSS and MFS, and LVI was an independent predictor for MFS (hazard ratio 1.71, 95 % confidence interval 1.00–2.93, P = 0.049). In patients with ureteral tumors, LVI represented the only significant prognosticator for both CSS and MFS in multivariate analysis. The prognostic value of LVI was not observed in pyelocaliceal tumors. Conclusions. The implication of LVI on prognosis, particularly in ureteral tumors but not in pyelocaliceal tumors, may imply diverse disease characteristics between different tumor locations among UTUC. LVI is essential to identify patients at high risk for metastasis/mortality and can facilitate treatment planning and surveillance strategies, especially in patients with ureteral tumors. Upper urinary tract urothelial carcinomas (UTUC) are defined as tumors arising from urothelium along the renal calyx, pelvis, and ureter. UTUC is a relatively uncommon cancer among urinary malignancies, accounting for approximately 5–10 % of patients with urothelial
Effect of LVI on UTUC by Location
1393
carcinomas, but the incidence is rising.1 In Taiwan, an endemically high incidence of UTUC has been reported and is gaining increased attention.2,3 Although patients with UTUC generally receive standard treatment—that is, radical nephroureterectomy (RNU) with excision of the bladder cuff—they have a high risk of local or distal recurrence and poorer prognosis compared to patients with bladder cancer, especially in advanced disease.4 As a result of the aggressive characteristic of UTUC, it is important to recognize risk factors to plan proper adjuvant treatment and follow-up strategies. Therefore, many researchers are dedicated to discovering prognostic indicators for predicting disease outcomes after surgery in an attempt to help clinicians’ decision making. In most studies, tumor stage and grade were considered the most important prognostic factors for UTUC. Other variables, including age, tumor location, lymph node (LN) involvement, multifocality, and architecture, have also been reported as significantly relevant factors.1,5 Lymphovascular invasion (LVI)—either blood vessel or lymphatic channel invasion by tumor cells—is considered to be the initial and crucial step of tumor dissemination, according to the metastatic cascade theory. LVI has been demonstrated to be an essential prognosticator in many kinds of malignancies, such as bladder cancer, prostate cancer, and rectal cancer, and several studies have examined the role of LVI in UTUC.6–9 In previous studies, LVI has been related to established features of aggressive UTUC, such as high-tumor grade, tumor stage, and LN metastasis. However, the prevalence of LVI in UTUC has been reported to be significantly different between white and ethnic Japanese populations.10 Moreover, previous studies from Taiwan discussing the effect of LVI on UTUC had small case numbers.11,12 The aim of the present study was to confirm the impact of LVI and other associated factors on survival, metastatic progression, and intraluminal recurrence of UTUC after RNU in Taiwan. In addition, because different properties or even outcomes have been reported according to tumor location, we also investigated the respective influence of LVI in pyelocaliceal and ureteral tumors.
stage, tumor grade, LN involvement, and LVI status were recorded. Patients who received neoadjuvant chemotherapy, who had previous or concurrent invasive bladder cancer, who had evidence of metastasis at diagnosis, and who had incomplete clinical information were excluded. Tumor location was defined as either renal pelvis or ureter based on dominant tumor features, in sequential order of stage, grade, and size. Tumor multifocality was defined as the concomitant presence of two or more pathologically confirmed tumors in any location of the upper urinary tract.
METHODS
Demographic and clinicopathologic factors between those with and without LVI were compared by Student’s t test for continuous variables and the chi-square test for categorical variables. The characteristics between pyelocaliceal and ureteral tumors were also analyzed. The Kaplan–Meier method was used to estimate the impact of LVI on survival, progression, and intraluminal recurrence. Survival rates were calculated from the date of RNU to the date of cancer-specific death, metastatic progression, and intraluminal recurrence or last visit. Survival curves were compared using the log-rank test. Significant prognostic factors in univariate analysis were included in the
Patient Collection We retrospectively analyzed the data of 250 patients with nonmetastatic UTUC who received RNU by either open or laparoscopic approach between 2004 and 2010 at our institution. The study protocol was approved for data collection by our institutional review board (KMUH-IRB20120138). Parameters including age, gender, smoking, tumor location, type of operation, presence with synchronous bladder tumor, tumor multifocality, pathological T
Pathological Evaluation All surgical specimens were assessed in detail after surgery. Tumor stage was evaluated according to the tumor, node, metastasis classification system, and tumor grade was decided by the 2004 World Health Organization/ International Society of Urologic Pathology consensus classification. Using standard hematoxylin and eosin staining, the presence of LVI in all cases was examined by two independent pathologists (PIL and SFY) who were unaware of the patients’ characteristics. The concordance rate for intervention of LVI was [90 %, and discrepant cases were reviewed to achieve consensus. Postoperative Follow-Up The patients were followed up every 3 months in the first 2 years after RNU, then every 6 months during the next 2 years and annually thereafter. During surveillance, physical examination, cystoscopy, urine cytology, and periodic imaging studies were performed following institutional guidelines. Intraluminal recurrence was defined as recurrence of tumor in the bladder or contralateral upper urinary tract. The definition of metastatic progression was tumor recurrence in the tumor bed or regional LNs and distant metastasis. During follow-up, 42 patients received adjuvant chemotherapy. Statistical Analysis
1394
H.-Y. Lee et al.
TABLE 1 Demographics and clinicopathologic characteristics of 250 patients with UTUC according to presence of LVI
TABLE 2 Demographics and clinicopathologic characteristics of 250 patients with UTUC according to tumor location
Variable
Variable
LVI present No (n = 190)
P Yes (n = 60)
C68 Gender
64 (50)
75 (61.5)
Male
54 (42.2)
54 (44.3)
28 (53.3)
Female
74 (57.8)
68 (55.7)
32 (46.7)
Smoking
0.849 26 (43.3)
C68
105 (55.3)
34 (56.7)
Gender
0.534 110 (57.9)
Smoking Yes
35 (18.4)
18 (30)
No
155 (81.6)
42 (70)
Location Pyelocaliceal Ureteral
103 (54.2)
41 (68.3) 19 (31.7) 0.053
120 (63.2) 70 (36.8)
46 (76.7) 14 (23.3)
Presence with synchronous bladder tumor No Yes
0.756 120 (62.8) 70 (36.6)
Multifocal
46 (24.2)
pT1
48 (25.3)
pT2 pT3
58 (30.5) 40 (21.1)
pT4
4 (2.1)
95 (77.9)
83 (64.8)
83 (68)
Laparoscopic
45 (35.2)
39 (32)
Presence with synchronous bladder tumor No Yes
87 (68.5) 40 (31.5)
72 (59) 50 (41) 0.121
25 (19.5)
34 (27.9)
pTis/pTa
22 (17.2)
18 (14.8)
47 (78.3)
pT1
29 (22.7)
24 (19.7)
13 (21.7)
pT2
29 (22.7)
44 (36.1)
pT3
37 (28.9)
33 (27)
pT4
11 (8.6)
3 (2.5)
20 (33.3)
0 (0)
0.071
5 (8.3)
Grade
15 (25) 30 (50)
Low
30 (23.4)
25 (20.5)
High
98 (76.6)
97 (79.5)
0.574
10 (16.7)
pN stage pN0 pNx
24 (18.8) 94 (73.4)
28 (23) 86 (70.5)
pN1/2
10 (7.8)
8 (6.6)
41 (32.0)
19 (15.6)
87 (68.0)
103 (84.4)
\0.001*
Grade Low
54 (28.4)
3 (5.0)
High
136 (71.6)
57 (95)
pN stage
0.005*
pN0
42 (22.1)
10 (16.7)
pNx
140 (73.7)
40 (66.7)
Negative
8 (4.2)
10 (16.7)
UTUC upper tract urothelial carcinoma, LVI lymphovascular invasion
0.690
Presence of LVI Positive
pN1/2
0.119
Tumor multifocality Solitary
0.725 0.594
Open
Pathological T stage
\0.001* 40 (21.1)
27 (22.1)
102 (79.7)
Multifocal
40 (66.7)
Pathological T stage pTis/pTa
26 (20.3)
No
88 (72.1)
0.686 144 (75.8)
Yes
103 (80.5)
Tumor multifocality Solitary
0.741
Type of operation 0.002*
87 (45.8)
Type of operation Open Laparoscopic
0.056
P
0.068 47 (38.5)
85 (44.7)
Female
Age (years) 64 (50)
\68
80 (42.1)
Ureter (n = 122)
\68
Age (years)
Male
Renal pelvis (n = 128)
0.002*
UTUC upper tract urothelial carcinoma, LVI lymphovascular invasion * P \ 0.05
* P \ 0.05
multivariate Cox proportional hazard model to identify independent predictors. In the multivariate analysis, we used several variable selection schemes, including stepwise selection, backward elimination, and entering all significant variables in the model. Because the results were similar, we only presented results that comprised every significant covariate. The prognostic significance of LVI and other clinicopathologic variables were further evaluated on the
basis of tumor location. In all analyses, P \ 0.05 was considered statistically significant, and power tests achieved [80 % from our results. Statistical analyses were performed with SPSS software, version 19 (IBM, Armonk, NY). RESULTS Median follow-up was 41 months. Our study population, consisting of 108 men and 142 women, were divided
Effect of LVI on UTUC by Location
A
1395
tumor, multifocality, and operation method. All clinicopathologic factors between patients with tumors in renal pelvis and ureter were not significantly different except the presence of LVI (Table 2).
1.0
Cancer-Specific Survival Estimates, %
LVI-
0.8
Cancer-Specific Survival (CSS) 0.6
LVI+
0.4
LVILVI+ LVI-censored LVI+censored
0.2 P = 0.003
0.0 0
12
24
36
48
60
72
84
Time after surgery, months
B
1.0
Metastasis-Free Survival (MFS)
Metastasis-Free Survival Estimates, %
LVI-
0.8
0.6
LVI+
0.4
LVILVI+ LVI-censored LVI+censored
0.2 P < 0.001
0.0 0
12
24
36
48
During the follow-up, the 3-year CSS rate (SD) was 90.6 (2.4) % in the LVI-negative group and 74.6 (6.4) % in the LVI-positive group. Patients with LVI showed an apparently worse CSS than those without LVI (P = 0.003, Fig. 1a). In addition to LVI status, advanced T stage, higher tumor grade, LN involvement, multifocality, and open RNU were also significantly related to lower CSS rates (Table 3). Multivariate analysis showed that advanced tumor T stage and LN involvement were independent prognostic factors for CSS [Cox regression hazard ratio (HR) 3.38, 95 % confidence interval (CI) 1.01–11.31, P = 0.048; and HR 4.06, 95 % CI 1.33–12.41, P = 0.014, respectively; Table 3].
60
72
Time after surgery, months
FIG. 1 Kaplan–Meier curves for CSS (a) and MFS (b) according to LVI status
into two groups: LVI positive (n = 60, 24 %) and LVI negative (n = 190, 76 %). The comparison of clinicopathologic characteristics between these two groups are shown in Table 1. The prevalence of LVI was significantly higher in patients with advanced pathological T stage (P \ 0.001), higher tumor grade (P \ 0.001), LN metastasis (P = 0.005), and tumor located in renal pelvis (P = 0.002). The distribution of non-organ-confined disease, determined by pathological stage, was clearly different in LVI-positive (n = 40, 66.7 %) and LVI-negative (n = 44, 24.2 %) groups. There were no significant difference in terms of age, gender, concomitant bladder
Sixty-one patients (24 %) experienced disease progression in this cohort. Figure 1b indicates that patients with LVI had a significant tendency to develop metastatic progression (P \ 0.001). The MFS rates (SD) after 3 years were 83.2 (3.0) % in the LVI-negative group and 52.2 (7.0) % in the LVI-positive group. In univariate analysis, LVI, pathological T stage, tumor grade, LN involvement, and multifocality were significant predictors of MFS. Advanced T stage, LN involvement, and the presence of LVI were associated with a significantly lower MFS in multivariate analysis (respectively, HR 2.99, 95 % CI 1.17–7.63, P = 0.022; HR 2.68, 95 % CI 1.11–6.51, P = 0.029; and HR 1.71, 95 % CI 1.00–2.93, P = 0.049; Table 3). Intraluminal Recurrence-Free Survival (IRFS) Overall, 79 patients (31.6 %) had an intraluminal recurrence in bladder or contralateral upper urinary tract. No significant difference between the LVI-positive and LVInegative groups was detected in terms of IRFS (P = 0.511). In both univariate and multivariate analyses, concomitant bladder tumor was the only independent factor to predict intraluminal recurrence (HR 8.89, 95 % CI 5.25–15.06, P \ 0.001; Table 3). Effect of LVI on Survival in Different Tumor Locations In the subgroup of patients with ureteral tumors, LVI positivity had a significantly negative impact on CSS and MFS (Fig. 2a, b) but similarly not on IRFS. In univariate
2.43 (1.12–5.29)
0.87 (0.46–1.64)
1.12 (0.52–2.44)
1.29 (0.67–2.47)
0.94 (0.50–1.76)
0.175
1.85 (1.08–3.16)
1.25 (0.75–2.08)
0.024* 1.47 (0.83–2.63) 0.190
0.394
0.083
P
0.020* 1.47 (0.37–5.80)
1.29 (0.77–2.14)
0.582
6.96 (2.18–22.20)
0.007*
0.001* 2.65 (0.75–9.44) 0.132 \0.001*
0.68 (0.42–1.10)
2.56 (1.35–4.85)
0.004* 1.38 (0.69–2.75)
0.97 (0.41–2.32) 0.361
0.946
3.02 (1.82–5.00)
0.82 (0.44–1.55)
0.80 (0.42–1.50) 0.480
0.75 (0.45–1.27) \0.001* 1.71 (1.00–2.93) 0.049* 0.83 (0.48–1.44)
0.545
0.218
0.431
0.662
0.790
0.607
Multivariate analysis, HR (95 % CI)
P
0.515
0.284
0.759
0.562
0.117
0.297
0.331
* P \ 0.05
CSS cancer-specific survival, MFS metastasis-free survival, IRFS intraluminal recurrence-free survival, UTUC upper tract urothelial carcinoma, RNU radical nephroureterectomy, HR hazard ratio, CI confidence interval, LVI lymphovascular invasion
LVI positive versus LVI negative
LVI
0.819
0.91 (0.39–2.11)
0.007*
7.94 (2.83–22.25) \0.001* 4.06 (1.33–12.41) 0.014* 5.81 (2.57–13.16) \0.001* 2.68 (1.11–6.51) 0.029* 0.84 (0.29–2.49)
\0.001*
P
8.89 (5.25–15.06) \0.001* 8.89 (5.25–15.06) \0.001*
1.35 (0.84–2.19)
1.12 (0.77–1.86)
0.88 (0.49–1.57)
0.94 (0.60–1.47)
1.12 (0.72–1.76)
Univariate analysis, HR (95 % CI)
0.001* 3.38 (1.01–11.31) 0.048* 5.92 (2.55–13.74) \0.001* 2.99 (1.17–7.63) 0.022* 0.79 (0.50–1.24)
0.020* 1.63 (0.80–3.33)
0.669
1.66 (0.94–2.93)
0.882
0.737
0.562
0.918
Multivariate analysis, HR (95 % CI)
pNx versus pN0
4.06 (1.25–13.20)
5.58 (1.98–15.70)
2.18 (1.13–4.21)
1.15 (0.61–2.18)
0.126
0.96 (0.58–1.59)
1.11 (0.60–2.05)
1.16 (0.70–1.95)
1.03 (0.62–1.70)
P
IRFS
pN1/2 versus pN0
N stage
High versus low
Grade
pT2–4 versus pTa/Tis/T1
Pathological stage
Multifocal versus solitary
Tumor multifocality
Yes versus no
0.025* 1.90 (0.84–4.29)
0.664
0.775
0.453
0.846
P
Univariate analysis, HR (95 % CI)
Multivariate analysis, HR (95 % CI)
Univariate analysis, HR (95 % CI)
P
MFS
CSS
Presence with synchronous bladder tumor
Open versus laparoscopic
Type of operation
Ureteral versus pyelocaliceal
Location
Yes versus no
Smoking
Female versus male
Gender
C68 versus \68 years
Age
Characteristic
TABLE 3 Univariate and multivariate analyses predicting CSS, MFS, and IRFS in patients with UTUC after RNU
1396 H.-Y. Lee et al.
Effect of LVI on UTUC by Location
A
1397
LVI-negative groups in patients with pyelocaliceal tumors (data not shown).
1.0
Cancer-specific Survival Estimates, %
LV-
DISCUSSION
0.8
0.6 LV+
0.4
0.2 LVLV+ LV-censored LV+censored
P < 0.001
0.0 0
12
24
36
48
60
Time after surgery, months
B
1.0
Metastasis-Free Survival Estimates, %
LV-
0.8
0.6
0.4
LV+
0.2 LVLV+ LV-censored LV+censored
P < 0.001
0.0 0
12
24
36
48
60
Time after surgery, months
FIG. 2 Kaplan–Meier curves for CSS (a) and MFS (b) in patients with ureteral tumors according to LVI status
analysis, T stage, LN involvement, and the presence of LVI were significant factors for both MFS and CSS (Table 4). LVI positivity was an independent prognostic factor for both CSS and MFS in multivariate analysis (HR 7.87, 95 % CI 2.78–22.26, P \ 0.001; and HR 4.87, 95 % CI 2.10– 11.31, P \ 0.001, respectively). Concerning IRFS, UTUC with concomitant bladder tumor was the most important factor to predict intraluminal recurrence (P \ 0.001; Table 4). However, there was no significant difference in CSS, MFS, and IRFS between LVI-positive and
The value of LVI to predict poor prognosis has been discussed in a number of solid organ malignancies, including prostate, bladder, testis, penis, liver, and rectum.13 Cheng et al. reported that LVI is a substantial risk factor to predict disease progression and CSS in prostate cancer, and they recommended that LVI should be included in the pathological staging system after radical prostatectomy.14 LVI was also found to be associated with survival rates in patients with invasive bladder cancer after radical cystectomy.15 UTUC are rare carcinomas with a poor prognosis and aggressive behavior among genitourinary cancers. Identifying prognostic factors to predict a group at high risk of UTUC is crucial to facilitate individualized therapy and proper surveillance protocol. Pathological T stage and tumor grade were established as major prognostic factors, and the role of LVI in UTUC has been extensively discussed in recent years.16 In this study, we not only confirmed the prognostic value of LVI in UTUC but also discovered different significance of LVI according to tumor location. We demonstrated for the first time that LVI is a more essential factor in ureteral tumors than in pyelocaliceal tumors. On the basis of previous reports, the incidence of LVI positivity is approximately 15–30 % in patients with UTUC after surgery.9,17,18 In this study, 24 % of cases were LVI positive. LVI was believed to be an important stage for tumor cell dissemination and LN metastasis, and it was considered to be a poor prognostic factor in UTUC.13,19,20 Consistent with previous studies, LVI was associated with higher pathological T stage, grade, and LN involvement in our analysis. Also in accordance with preceding reports, we confirmed LVI is associated with a significant decline in CSS and MFS but not in IRFS by Kaplan–Meier analysis.17,21,22 As a result, LVI represented a strong predictor for disease progression and cancer-specific death, but it did not correlate with bladder and contralateral recurrence in UTUC. In the present study, pathological T stage, tumor grade, LN involvement, tumor multifocality, type of operative approach, and LVI were significant factors in influencing CSS in univariate analysis. In multivariate analysis, pathological T stage and LN involvement were independent predictors for CSS. A retrospective study compared efficacy between open and laparoscopic RNU and illustrated no significant difference in CSS.23 In our series, the poorer prognosis in open RNU group was assumed to be related to higher T stage compared with laparoscopic RNU group,
1.52 (0.53–4.31)
2.09 (0.77–5.66)
0.77 (0.30–2.00)
0.58 (0.22–1.51)
Yes versus no
11.97 (4.39–32.65)
3.05 (0.63–14.66) 0.91 (0.28–2.94)
\0.001* 7.87 (2.78–22.26)
0.648
0.070
0.046*
0.102
0.023* 5.60 (0.69–45.63)
0.362
0.611
0.79 (0.33–1.91)
3.77 (1.09–13.08)
9.23 (1.25–68.04)
8.23 (1.95–34.79)
\0.001* 8.79 (4.03–19.18)
0.869
0.165
0.210
0.107
1.21 (0.55–2.68)
1.31 (0.62–2.76)
0.98 (0.45–2.13)
2.51 (1.17–5.36)
0.70 (0.33–1.48)
0.61 (0.29–1.28)
1.11 (0.45–2.75) \0.001* 4.87 (2.10–11.31)
0.604
0.037* 2.09 (0.58–7.54)
0.022*
0.029* 3.23 (0.37–27.90)
0.004* 3.39 (0.71–16.20)
0.633
0.477
0.957
0.018* 2.05 (0.88–4.79)
0.352
0.192
0.819
0.261
0.504
0.287
0.126
0.097
\0.001*
Multivariate analysis, P HR (95 % CI)
IRFS
0.90 (0.35–2.31)
0.59 (0.30–1.18)
0.88 (0.25–3.13)
0.50 (0.26–0.97)
0.66 (0.35–1.23)
1.25 (0.64–2.42)
13.78 (5.72–33.22)
1.54 (0.77–3.09)
1.61 (0.80–3.22)
1.00 (0.54–1.86)
1.03 (0.55–1.94)
0.227
0.184
0.998
0.924
0.827
0.134
0.846
0.303
0.040*
0.191
0.514
0.61 (0.31–1.18)
* P \ 0.05
0.141
\0.001*
Multivariate analysis, P HR (95 % CI)
\0.001* 13.42 (5.55–32.45)
Univariate analysis, P HR (95 % CI)
CSS cancer-specific survival, MFS metastasis-free survival, IRFS intraluminal recurrence-free survival, RNU radical nephroureterectomy, HR hazard ratio, CI confidence interval, LVI lymphovascular invasion
LVI positive versus LVI negative
LVI
4.07 (0.89–18.61)
0.76 (0.24–2.44)
pNx versus pN0
5.42 (0.72–40.96)
10.50 (1.39–79.55)
1.59 (0.59–4.32)
1.28 (0.49–3.33)
pN1/2 versus pN0
N stage
High versus low
Grade
pT2–4 versus pTa/Tis/T1
Pathologic stage
Multifocal versus solitary
Tumor multifocality
0.436
0.146
0.588
0.265
MFS Univariate analysis, P HR (95 % CI)
Multivariate analysis, P HR (95 % CI)
CSS
Univariate analysis, P HR (95 % CI)
Presence with synchronous bladder tumor
Open versus laparoscopic
Type of operation
Yes versus no
Smoking
Female versus male
Gender
C68 years versus \68 years
Age
Characteristic
TABLE 4 Univariate and multivariate analyses predicting CSS, MFS, and IRFS in patients with ureteral tumor after RNU
1398 H.-Y. Lee et al.
Effect of LVI on UTUC by Location
and consequently the effect of operation method diminished after adjusting other variables. Similar to the findings of Chromecki et al., tumor multifocality was associated with increased risk of CSS in univariate analysis, but no significant difference was observed in multivariate analysis.24 For MFS, pathological T stage, tumor grade, LN involvement, tumor multifocality, and LVI were significant factors in univariate analysis. In multivariate analysis, pathological T stage, LN involvement, and LVI represented important factors for disease progression. Both our univariate and multivariate analyses showed that UTUC with synchronous bladder tumor is the only independent risk factor for IRFS, which was also verified in a recent study.25 Among UTUC, tumors in the renal pelvis and ureter may have different characteristics. Chung et al. found that the thicker anatomic barrier of renal pelvis than ureter can lead to different consequences, and they suggested that it may be more reasonable to individually evaluate tumors in different locations.12 We found that the rate of prevalence of LVI is higher in pyelocaliceal tumors than in ureteral tumors. All previous studies have shown this feature, although some detected a significant difference in LVI according to location.9,22,26 Others, however, did not.17,27,28 Pyelocaliceal tumors did not result in poorer prognosis. Instead, they had similar or even better outcomes than ureteral tumors.26,28,29 There are limited data concerning prognostic factors of ureteral tumors in particular, and none of them incorporated LVI for survival analysis.30,31 In this cohort, pT stage, LN metastasis, and LVI were significant factors for both CSS and MFS of ureteral tumors in univariate analysis, and LVI represented an independent predictor in multivariate analysis. In contrast, LVI failed to be independently associated with CSS and MFS in pyelocaliceal tumors. We hypothesized that the diverse effect of LVI in different location is related to the thickness of adjacent barrier. As mentioned before, LVI is an important step for tumor cells to disseminate. If LVI is present in ureteral tumors, its thinner muscular wall is easier for invasion, and thinner adventitia, which is rich in blood vessels and lymphatic plexus, can facilitate tumor spread. Therefore, the prognostic value of LVI is further highlighted with respect to ureteral tumors specifically. There are several limitations to the present study. First, the study design is retrospective and the number of cases is small. Second, there is no initial centralized pathological review of the LVI status. However, all cases were reanalyzed by two independent pathologists, with high consistency in the interpretation of LVI. Third, most subjects did not undergo LN dissection. Patients who received LN dissection can have more accurate N stage, which may dilute the prognostic value of LVI. Nevertheless, the percentage of retrieved LNs was similar between different
1399
tumor locations (27 % in pyelocaliceal and 30 % in ureteral tumors; P = 0.604). Furthermore, in multivariate analysis, we demonstrated that LVI is a powerful and practical parameter for prediction of prognosis, especially in ureteral tumors. One of our study’s strengths is its separate evaluation of MFS and IRFS; another is our detailed subgroup analysis. The patient population we studied was different from those described from Western countries because the incidence of UTUC is far higher in Taiwan than in any other area. In addition, in both Taiwan and China, female predominance in UTUC is another unique feature that is dissimilar to other countries.2,3,32,33 This study provided information about identifying prognostic factors to predict outcome of UTUC in Taiwan. More importantly, this is the first study to demonstrate the enhanced effect of LVI in ureteral tumors specifically. The diverse effect of LVI in different tumor locations in UTUC will require larger prospective studies to validate. CONCLUSIONS We confirmed the imperative role of LVI in predicting disease progression and survival of patients with UTUC after receiving RNU in Taiwan. The remarkable importance of LVI on prognosis, particularly in ureteral tumors rather than pyelocaliceal tumors, may explain the different characteristics between these two locations of UTUC. We think that LVI status can be a part of the histopathological assessment to identify patients at high risk for receiving individualized treatment and surveillance strategies, especially in ureteral tumors. ACKNOWLEDGMENT This study was supported by grants from the Kaohsiung Medical University Hospital (KMUH 97-7G06), the Center for Infectious Disease and Cancer Research, Kaohsiung Medical University (KMU-TP103E19), and the Health and Welfare Surcharge of tobacco products, Ministry of Health and Welfare (MOHW103-TD-B-111-05). DISCLOSURE
The authors declare no conflict of interest.
REFERENCES 1. Roupret M, Babjuk M, Comperat E, et al. European guidelines on upper tract urothelial carcinomas: 2013 update. Eur Urol. 2013; 63:1059–71. 2. Li CC, Chang TH, Wu WJ, et al. Significant predictive factors for prognosis of primary upper urinary tract cancer after radical nephroureterectomy in Taiwanese patients. Eur Urol. 2008;54: 1127–34. 3. Li WM, Shen JT, Li CC, et al. Oncologic outcomes following three different approaches to the distal ureter and bladder cuff in nephroureterectomy for primary upper urinary tract urothelial carcinoma. Eur Urol. 2010;57:963–9.
1400 4. Verhoest G, Shariat SF, Chromecki TF, et al. Predictive factors of recurrence and survival of upper tract urothelial carcinomas. World J Urol. 2011;29:495–501. 5. Saito K, Kawakami S, Fujii Y, Sakura M, Masuda H, Kihara K. Lymphovascular invasion is independently associated with poor prognosis in patients with localized upper urinary tract urothelial carcinoma treated surgically. J Urol. 2007;178:2291–6. 6. Eisenberg MS, Boorjian SA, Cheville JC, et al. The SPARC score: a multifactorial outcome prediction model for patients undergoing radical cystectomy for bladder cancer. J Urol. 2013;190:2005–10. 7. Lee JH, Jang HS, Kim JG, et al. Lymphovascular invasion is a significant prognosticator in rectal cancer patients who receive preoperative chemoradiotherapy followed by total mesorectal excision. Ann Surg Oncol. 2012;19:1213–21. 8. Baydar DE, Baseskioglu B, Ozen H, Geyik PO. Prognostic significance of lymphovascular invasion in clinically localized prostate cancer after radical prostatectomy. Sci World J. 2008;17:303–12. 9. Godfrey MS, Badalato GM, Hruby GW, Razmjoo M, McKiernan JM. Prognostic indicators for upper tract urothelial carcinoma after radical nephroureterectomy: the impact of lymphovascular invasion. BJU Int. 2012;110:798–803. 10. Matsumoto K, Novara G, Gupta A, et al. Racial differences in the outcome of patients with urothelial carcinoma of the upper urinary tract: an international study. BJU Int. 2011;108(8 Pt 2):E304–9. 11. Lin WC, Hu FC, Chung SD, Chueh SC, Pu YS, Huang KH. The role of lymphovascular invasion in predicting the prognosis of clinically localized upper tract urothelial carcinoma (pT1-3cN0M0). J Urol. 2008;180:879–84. 12. Chung SD, Wang SM, Lai MK, et al. Lymphovascular invasion predicts poor outcome of urothelial carcinoma of renal pelvis after nephroureterectomy. BJU Int. 2009;103:1047–51. 13. Ku JH, Byun SS, Jeong H, Kwak C, Kim HH, Lee SE. Lymphovascular invasion as a prognostic factor in the upper urinary tract urothelial carcinoma: a systematic review and meta-analysis. Eur J Cancer. 2013;49:2665–80. 14. Cheng L, Jones TD, Lin H, et al. Lymphovascular invasion is an independent prognostic factor in prostatic adenocarcinoma. J Urol. 2005;174:2181–5. 15. Lotan Y, Gupta A, Shariat SF, et al. Lymphovascular invasion is independently associated with overall survival, cause-specific survival, and local and distant recurrence in patients with negative lymph nodes at radical cystectomy. J Clin Oncol. 2005;23:6533–9. 16. Hong B, Park S, Hong JH, Kim CS, Ro JY, Ahn H. Prognostic value of lymphovascular invasion in transitional cell carcinoma of upper urinary tract. Urology. 2005;65:692–6. 17. Hurel S, Roupreˆt M, Ouzzane A, et al. Impact of lymphovascular invasion on oncological outcomes in patients with upper tract urothelial carcinoma after radical nephroureterectomy. BJU Int. 2013;111:1199–207. 18. Margulis V, Shariat SF, Matin SF, et al. Outcomes of radical nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration. Cancer. 2009;115:1224–33.
H.-Y. Lee et al. 19. Akao J, Matsuyama H, Yamamoto Y, et al. Clinical significance of lymphovascular invasion in upper urinary tract urothelial cancer. BJU Int. 2008;102:572–5. 20. Ouzzane A, Colin P, Ghoneim TP, et al. The impact of lymph node status and features on oncological outcomes in urothelial carcinoma of the upper urinary tract (UTUC) treated by nephroureterectomy. World J Urol. 2013;31:189–97. 21. Cha EK, Shariat SF, Kormaksson M, et al. Predicting clinical outcomes after radical nephroureterectomy for upper tract urothelial carcinoma. Eur Urol. 2012;61:818–25. 22. Novara G, Matsumoto K, Kassouf W, et al. Prognostic role of lymphovascular invasion in patients with urothelical carcinoma of the upper urinary tract: an international validation study. Eur Urol. 2010;57:1064–71. 23. Zou L, Zhang L, Zhang H, Jiang H, Ding Q. Comparison of postoperative intravesical recurrence and oncological outcomes after open versus laparoscopic nephroureterectomy for upper urinary tract urothelial carcinoma. World J Urol. 2014;32:565–70. 24. Chromecki TF, Cha EK, Fajkovic H, et al. The impact of tumor multifocality on outcomes in patients treated with radical nephroureterectomy. Eur Urol. 2012;61:245–53. 25. Pignot G, Colin P, Zerbib M, et al. Influence of previous or synchronous bladder cancer on oncologic outcomes after radical nephroureterectomy for upper urinary tract urothelial carcinoma. Urol Oncol. 2014;32:23.e1–8. 26. Yafi FA, Novara G, Shariat SF, et al. Impact of tumour location versus multifocality in patients with upper tract urothelial carcinoma treated with nephroureterectomy and bladder cuff excision: a homogeneous series without perioperative chemotherapy. BJU Int. 2012;110(2 Pt 2):E7–13. 27. Kikuchi E, Margulis V, Karakiewicz PI, et al. Lymphovascular invasion predicts clinical outcomes in patients with node-negative upper tract urothelial carcinoma. J Clin Oncol. 2009;27:612–8. 28. Ouzzane A, Colin P, Xylinas E, et al. Ureteral and multifocal tumours have worse prognosis than renal pelvic tumours in urothelial carcinoma of the upper urinary tract treated by nephroureterectomy. Eur Urol. 2011;60:1258–65. 29. Williams AK, Kassouf W, Chin J, et al. Multifocality rather than tumor location is a prognostic factor in upper tract urothelial carcinoma. Urol Oncol. 2013;31:1161–5. 30. Li WM, Li CC, Ke HL, Wu WJ, Huang CN, Huang CH. The prognostic predictors of primary ureteral transitional cell carcinoma after radical nephroureterectomy. J Urol. 2009;182:451–8. 31. Raman JD, Messer J, Sielatycki JA, Hollenbeak CS. Incidence and survival of patients with carcinoma of the ureter and renal pelvis in the USA, 1973–2005. BJU Int. 2011;107:1059–64. 32. Chen XP, Xiong GY, Li XS, et al. Predictive factors for worse pathological outcomes of upper tract urothelial carcinoma: experience from a nationwide high-volume centre in China. BJU Int. 2013;112:917–24. 33. Matin SF, Shariat SF, Milowsky MI, et al. Highlights from the first symposium on upper tract urothelial carcinoma. Urol Oncol. 2014;32:309–16.
ORIGINAL ARTICLE – UROLOGIC ONCOLOGY
Prognostic Significance of Lymphovascular Invasion in Upper Urinary Tract Urothelial Carcinoma is Influenced by Tumor Location Hsiang-Ying Lee, MD, MS1,2, Ching-Chia Li, MD1,2,3,4, Chun-Nung Huang, MD, PhD1,3, Hung-Lung Ke, MD, MS1,3,4, Wei-Ming Li, MD, MS1,3,4,5, Peir-In Liang, MD6,7, Sheau-Fang Yang, MD6,7,8, Hung-Pin Tu, PhD9, Wen-Jeng Wu, MD, PhD1,3,4,10,11, and Hsin-Chih Yeh, MD, MS1,2,3,4 Department of Urology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan; 2Department of Urology, Kaohsiung Municipal Ta-Tung Hospital, Kaohsiung, Taiwan; 3Department of Urology, School of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 4Graduate Institute of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 5Department of Urology, Ministry of Health and Welfare, Pingtung Hospital, Pingtung, Taiwan; 6Department of Pathology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan; 7 Department of Pathology, School of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 8 Department of Pathology, Kaohsiung Municipal Ta-Tung Hospital, Kaohsiung, Taiwan; 9Department of Public Health and Environmental Medicine, School of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan; 10 Center for Infectious Disease and Cancer Research, Kaohsiung Medical University, Kaohsiung, Taiwan; 11Department of Urology, Kaohsiung Municipal Hsiao-Kang Hospital, Kaohsiung, Taiwan 1
ABSTRACT Background. To examine the prognostic value of lymphovascular invasion (LVI) in different tumor locations (i.e., renal pelvis and ureter) of upper urinary tract urothelial carcinoma (UTUC). Methods. Data from a total of 250 patients with nonmetastatic UTUC who received radical nephroureterectomy between 2004 and 2010 at our institution were analyzed retrospectively. The significance of LVI and other relevant factors on cancer-specific survival (CSS), metastasis-free survival (MFS), and intraluminal recurrence-free survival (IRFS) were evaluated. Results. Lymphovascular invasion was present in 60 patients (24 %) and was related to advanced pathological T Hsiang-Ying Lee and Ching-Chia Li contributed equally to this article, and both should be considered first author. Ó Society of Surgical Oncology 2014 First Received: 23 May 2014; Published Online: 20 September 2014 Wen-JengWu, MD, PhD e-mail: [email protected] H.-C. Yeh, MD, MS e-mail: [email protected]
stage (P \ 0.001), higher tumor grade (P \ 0.001), lymph node metastasis (P = 0.005), and pyelocaliceal tumor location (P = 0.002). By Kaplan–Meier analysis, LVI was found to be significantly correlated with worse CSS and MFS but not with IRFS. Multivariate analysis showed that high pathological T stage and regional lymph node involvement were significant prognostic factors for CSS and MFS, and LVI was an independent predictor for MFS (hazard ratio 1.71, 95 % confidence interval 1.00–2.93, P = 0.049). In patients with ureteral tumors, LVI represented the only significant prognosticator for both CSS and MFS in multivariate analysis. The prognostic value of LVI was not observed in pyelocaliceal tumors. Conclusions. The implication of LVI on prognosis, particularly in ureteral tumors but not in pyelocaliceal tumors, may imply diverse disease characteristics between different tumor locations among UTUC. LVI is essential to identify patients at high risk for metastasis/mortality and can facilitate treatment planning and surveillance strategies, especially in patients with ureteral tumors. Upper urinary tract urothelial carcinomas (UTUC) are defined as tumors arising from urothelium along the renal calyx, pelvis, and ureter. UTUC is a relatively uncommon cancer among urinary malignancies, accounting for approximately 5–10 % of patients with urothelial
Effect of LVI on UTUC by Location
1393
carcinomas, but the incidence is rising.1 In Taiwan, an endemically high incidence of UTUC has been reported and is gaining increased attention.2,3 Although patients with UTUC generally receive standard treatment—that is, radical nephroureterectomy (RNU) with excision of the bladder cuff—they have a high risk of local or distal recurrence and poorer prognosis compared to patients with bladder cancer, especially in advanced disease.4 As a result of the aggressive characteristic of UTUC, it is important to recognize risk factors to plan proper adjuvant treatment and follow-up strategies. Therefore, many researchers are dedicated to discovering prognostic indicators for predicting disease outcomes after surgery in an attempt to help clinicians’ decision making. In most studies, tumor stage and grade were considered the most important prognostic factors for UTUC. Other variables, including age, tumor location, lymph node (LN) involvement, multifocality, and architecture, have also been reported as significantly relevant factors.1,5 Lymphovascular invasion (LVI)—either blood vessel or lymphatic channel invasion by tumor cells—is considered to be the initial and crucial step of tumor dissemination, according to the metastatic cascade theory. LVI has been demonstrated to be an essential prognosticator in many kinds of malignancies, such as bladder cancer, prostate cancer, and rectal cancer, and several studies have examined the role of LVI in UTUC.6–9 In previous studies, LVI has been related to established features of aggressive UTUC, such as high-tumor grade, tumor stage, and LN metastasis. However, the prevalence of LVI in UTUC has been reported to be significantly different between white and ethnic Japanese populations.10 Moreover, previous studies from Taiwan discussing the effect of LVI on UTUC had small case numbers.11,12 The aim of the present study was to confirm the impact of LVI and other associated factors on survival, metastatic progression, and intraluminal recurrence of UTUC after RNU in Taiwan. In addition, because different properties or even outcomes have been reported according to tumor location, we also investigated the respective influence of LVI in pyelocaliceal and ureteral tumors.
stage, tumor grade, LN involvement, and LVI status were recorded. Patients who received neoadjuvant chemotherapy, who had previous or concurrent invasive bladder cancer, who had evidence of metastasis at diagnosis, and who had incomplete clinical information were excluded. Tumor location was defined as either renal pelvis or ureter based on dominant tumor features, in sequential order of stage, grade, and size. Tumor multifocality was defined as the concomitant presence of two or more pathologically confirmed tumors in any location of the upper urinary tract.
METHODS
Demographic and clinicopathologic factors between those with and without LVI were compared by Student’s t test for continuous variables and the chi-square test for categorical variables. The characteristics between pyelocaliceal and ureteral tumors were also analyzed. The Kaplan–Meier method was used to estimate the impact of LVI on survival, progression, and intraluminal recurrence. Survival rates were calculated from the date of RNU to the date of cancer-specific death, metastatic progression, and intraluminal recurrence or last visit. Survival curves were compared using the log-rank test. Significant prognostic factors in univariate analysis were included in the
Patient Collection We retrospectively analyzed the data of 250 patients with nonmetastatic UTUC who received RNU by either open or laparoscopic approach between 2004 and 2010 at our institution. The study protocol was approved for data collection by our institutional review board (KMUH-IRB20120138). Parameters including age, gender, smoking, tumor location, type of operation, presence with synchronous bladder tumor, tumor multifocality, pathological T
Pathological Evaluation All surgical specimens were assessed in detail after surgery. Tumor stage was evaluated according to the tumor, node, metastasis classification system, and tumor grade was decided by the 2004 World Health Organization/ International Society of Urologic Pathology consensus classification. Using standard hematoxylin and eosin staining, the presence of LVI in all cases was examined by two independent pathologists (PIL and SFY) who were unaware of the patients’ characteristics. The concordance rate for intervention of LVI was [90 %, and discrepant cases were reviewed to achieve consensus. Postoperative Follow-Up The patients were followed up every 3 months in the first 2 years after RNU, then every 6 months during the next 2 years and annually thereafter. During surveillance, physical examination, cystoscopy, urine cytology, and periodic imaging studies were performed following institutional guidelines. Intraluminal recurrence was defined as recurrence of tumor in the bladder or contralateral upper urinary tract. The definition of metastatic progression was tumor recurrence in the tumor bed or regional LNs and distant metastasis. During follow-up, 42 patients received adjuvant chemotherapy. Statistical Analysis
1394
H.-Y. Lee et al.
TABLE 1 Demographics and clinicopathologic characteristics of 250 patients with UTUC according to presence of LVI
TABLE 2 Demographics and clinicopathologic characteristics of 250 patients with UTUC according to tumor location
Variable
Variable
LVI present No (n = 190)
P Yes (n = 60)
C68 Gender
64 (50)
75 (61.5)
Male
54 (42.2)
54 (44.3)
28 (53.3)
Female
74 (57.8)
68 (55.7)
32 (46.7)
Smoking
0.849 26 (43.3)
C68
105 (55.3)
34 (56.7)
Gender
0.534 110 (57.9)
Smoking Yes
35 (18.4)
18 (30)
No
155 (81.6)
42 (70)
Location Pyelocaliceal Ureteral
103 (54.2)
41 (68.3) 19 (31.7) 0.053
120 (63.2) 70 (36.8)
46 (76.7) 14 (23.3)
Presence with synchronous bladder tumor No Yes
0.756 120 (62.8) 70 (36.6)
Multifocal
46 (24.2)
pT1
48 (25.3)
pT2 pT3
58 (30.5) 40 (21.1)
pT4
4 (2.1)
95 (77.9)
83 (64.8)
83 (68)
Laparoscopic
45 (35.2)
39 (32)
Presence with synchronous bladder tumor No Yes
87 (68.5) 40 (31.5)
72 (59) 50 (41) 0.121
25 (19.5)
34 (27.9)
pTis/pTa
22 (17.2)
18 (14.8)
47 (78.3)
pT1
29 (22.7)
24 (19.7)
13 (21.7)
pT2
29 (22.7)
44 (36.1)
pT3
37 (28.9)
33 (27)
pT4
11 (8.6)
3 (2.5)
20 (33.3)
0 (0)
0.071
5 (8.3)
Grade
15 (25) 30 (50)
Low
30 (23.4)
25 (20.5)
High
98 (76.6)
97 (79.5)
0.574
10 (16.7)
pN stage pN0 pNx
24 (18.8) 94 (73.4)
28 (23) 86 (70.5)
pN1/2
10 (7.8)
8 (6.6)
41 (32.0)
19 (15.6)
87 (68.0)
103 (84.4)
\0.001*
Grade Low
54 (28.4)
3 (5.0)
High
136 (71.6)
57 (95)
pN stage
0.005*
pN0
42 (22.1)
10 (16.7)
pNx
140 (73.7)
40 (66.7)
Negative
8 (4.2)
10 (16.7)
UTUC upper tract urothelial carcinoma, LVI lymphovascular invasion
0.690
Presence of LVI Positive
pN1/2
0.119
Tumor multifocality Solitary
0.725 0.594
Open
Pathological T stage
\0.001* 40 (21.1)
27 (22.1)
102 (79.7)
Multifocal
40 (66.7)
Pathological T stage pTis/pTa
26 (20.3)
No
88 (72.1)
0.686 144 (75.8)
Yes
103 (80.5)
Tumor multifocality Solitary
0.741
Type of operation 0.002*
87 (45.8)
Type of operation Open Laparoscopic
0.056
P
0.068 47 (38.5)
85 (44.7)
Female
Age (years) 64 (50)
\68
80 (42.1)
Ureter (n = 122)
\68
Age (years)
Male
Renal pelvis (n = 128)
0.002*
UTUC upper tract urothelial carcinoma, LVI lymphovascular invasion * P \ 0.05
* P \ 0.05
multivariate Cox proportional hazard model to identify independent predictors. In the multivariate analysis, we used several variable selection schemes, including stepwise selection, backward elimination, and entering all significant variables in the model. Because the results were similar, we only presented results that comprised every significant covariate. The prognostic significance of LVI and other clinicopathologic variables were further evaluated on the
basis of tumor location. In all analyses, P \ 0.05 was considered statistically significant, and power tests achieved [80 % from our results. Statistical analyses were performed with SPSS software, version 19 (IBM, Armonk, NY). RESULTS Median follow-up was 41 months. Our study population, consisting of 108 men and 142 women, were divided
Effect of LVI on UTUC by Location
A
1395
tumor, multifocality, and operation method. All clinicopathologic factors between patients with tumors in renal pelvis and ureter were not significantly different except the presence of LVI (Table 2).
1.0
Cancer-Specific Survival Estimates, %
LVI-
0.8
Cancer-Specific Survival (CSS) 0.6
LVI+
0.4
LVILVI+ LVI-censored LVI+censored
0.2 P = 0.003
0.0 0
12
24
36
48
60
72
84
Time after surgery, months
B
1.0
Metastasis-Free Survival (MFS)
Metastasis-Free Survival Estimates, %
LVI-
0.8
0.6
LVI+
0.4
LVILVI+ LVI-censored LVI+censored
0.2 P < 0.001
0.0 0
12
24
36
48
During the follow-up, the 3-year CSS rate (SD) was 90.6 (2.4) % in the LVI-negative group and 74.6 (6.4) % in the LVI-positive group. Patients with LVI showed an apparently worse CSS than those without LVI (P = 0.003, Fig. 1a). In addition to LVI status, advanced T stage, higher tumor grade, LN involvement, multifocality, and open RNU were also significantly related to lower CSS rates (Table 3). Multivariate analysis showed that advanced tumor T stage and LN involvement were independent prognostic factors for CSS [Cox regression hazard ratio (HR) 3.38, 95 % confidence interval (CI) 1.01–11.31, P = 0.048; and HR 4.06, 95 % CI 1.33–12.41, P = 0.014, respectively; Table 3].
60
72
Time after surgery, months
FIG. 1 Kaplan–Meier curves for CSS (a) and MFS (b) according to LVI status
into two groups: LVI positive (n = 60, 24 %) and LVI negative (n = 190, 76 %). The comparison of clinicopathologic characteristics between these two groups are shown in Table 1. The prevalence of LVI was significantly higher in patients with advanced pathological T stage (P \ 0.001), higher tumor grade (P \ 0.001), LN metastasis (P = 0.005), and tumor located in renal pelvis (P = 0.002). The distribution of non-organ-confined disease, determined by pathological stage, was clearly different in LVI-positive (n = 40, 66.7 %) and LVI-negative (n = 44, 24.2 %) groups. There were no significant difference in terms of age, gender, concomitant bladder
Sixty-one patients (24 %) experienced disease progression in this cohort. Figure 1b indicates that patients with LVI had a significant tendency to develop metastatic progression (P \ 0.001). The MFS rates (SD) after 3 years were 83.2 (3.0) % in the LVI-negative group and 52.2 (7.0) % in the LVI-positive group. In univariate analysis, LVI, pathological T stage, tumor grade, LN involvement, and multifocality were significant predictors of MFS. Advanced T stage, LN involvement, and the presence of LVI were associated with a significantly lower MFS in multivariate analysis (respectively, HR 2.99, 95 % CI 1.17–7.63, P = 0.022; HR 2.68, 95 % CI 1.11–6.51, P = 0.029; and HR 1.71, 95 % CI 1.00–2.93, P = 0.049; Table 3). Intraluminal Recurrence-Free Survival (IRFS) Overall, 79 patients (31.6 %) had an intraluminal recurrence in bladder or contralateral upper urinary tract. No significant difference between the LVI-positive and LVInegative groups was detected in terms of IRFS (P = 0.511). In both univariate and multivariate analyses, concomitant bladder tumor was the only independent factor to predict intraluminal recurrence (HR 8.89, 95 % CI 5.25–15.06, P \ 0.001; Table 3). Effect of LVI on Survival in Different Tumor Locations In the subgroup of patients with ureteral tumors, LVI positivity had a significantly negative impact on CSS and MFS (Fig. 2a, b) but similarly not on IRFS. In univariate
2.43 (1.12–5.29)
0.87 (0.46–1.64)
1.12 (0.52–2.44)
1.29 (0.67–2.47)
0.94 (0.50–1.76)
0.175
1.85 (1.08–3.16)
1.25 (0.75–2.08)
0.024* 1.47 (0.83–2.63) 0.190
0.394
0.083
P
0.020* 1.47 (0.37–5.80)
1.29 (0.77–2.14)
0.582
6.96 (2.18–22.20)
0.007*
0.001* 2.65 (0.75–9.44) 0.132 \0.001*
0.68 (0.42–1.10)
2.56 (1.35–4.85)
0.004* 1.38 (0.69–2.75)
0.97 (0.41–2.32) 0.361
0.946
3.02 (1.82–5.00)
0.82 (0.44–1.55)
0.80 (0.42–1.50) 0.480
0.75 (0.45–1.27) \0.001* 1.71 (1.00–2.93) 0.049* 0.83 (0.48–1.44)
0.545
0.218
0.431
0.662
0.790
0.607
Multivariate analysis, HR (95 % CI)
P
0.515
0.284
0.759
0.562
0.117
0.297
0.331
* P \ 0.05
CSS cancer-specific survival, MFS metastasis-free survival, IRFS intraluminal recurrence-free survival, UTUC upper tract urothelial carcinoma, RNU radical nephroureterectomy, HR hazard ratio, CI confidence interval, LVI lymphovascular invasion
LVI positive versus LVI negative
LVI
0.819
0.91 (0.39–2.11)
0.007*
7.94 (2.83–22.25) \0.001* 4.06 (1.33–12.41) 0.014* 5.81 (2.57–13.16) \0.001* 2.68 (1.11–6.51) 0.029* 0.84 (0.29–2.49)
\0.001*
P
8.89 (5.25–15.06) \0.001* 8.89 (5.25–15.06) \0.001*
1.35 (0.84–2.19)
1.12 (0.77–1.86)
0.88 (0.49–1.57)
0.94 (0.60–1.47)
1.12 (0.72–1.76)
Univariate analysis, HR (95 % CI)
0.001* 3.38 (1.01–11.31) 0.048* 5.92 (2.55–13.74) \0.001* 2.99 (1.17–7.63) 0.022* 0.79 (0.50–1.24)
0.020* 1.63 (0.80–3.33)
0.669
1.66 (0.94–2.93)
0.882
0.737
0.562
0.918
Multivariate analysis, HR (95 % CI)
pNx versus pN0
4.06 (1.25–13.20)
5.58 (1.98–15.70)
2.18 (1.13–4.21)
1.15 (0.61–2.18)
0.126
0.96 (0.58–1.59)
1.11 (0.60–2.05)
1.16 (0.70–1.95)
1.03 (0.62–1.70)
P
IRFS
pN1/2 versus pN0
N stage
High versus low
Grade
pT2–4 versus pTa/Tis/T1
Pathological stage
Multifocal versus solitary
Tumor multifocality
Yes versus no
0.025* 1.90 (0.84–4.29)
0.664
0.775
0.453
0.846
P
Univariate analysis, HR (95 % CI)
Multivariate analysis, HR (95 % CI)
Univariate analysis, HR (95 % CI)
P
MFS
CSS
Presence with synchronous bladder tumor
Open versus laparoscopic
Type of operation
Ureteral versus pyelocaliceal
Location
Yes versus no
Smoking
Female versus male
Gender
C68 versus \68 years
Age
Characteristic
TABLE 3 Univariate and multivariate analyses predicting CSS, MFS, and IRFS in patients with UTUC after RNU
1396 H.-Y. Lee et al.
Effect of LVI on UTUC by Location
A
1397
LVI-negative groups in patients with pyelocaliceal tumors (data not shown).
1.0
Cancer-specific Survival Estimates, %
LV-
DISCUSSION
0.8
0.6 LV+
0.4
0.2 LVLV+ LV-censored LV+censored
P < 0.001
0.0 0
12
24
36
48
60
Time after surgery, months
B
1.0
Metastasis-Free Survival Estimates, %
LV-
0.8
0.6
0.4
LV+
0.2 LVLV+ LV-censored LV+censored
P < 0.001
0.0 0
12
24
36
48
60
Time after surgery, months
FIG. 2 Kaplan–Meier curves for CSS (a) and MFS (b) in patients with ureteral tumors according to LVI status
analysis, T stage, LN involvement, and the presence of LVI were significant factors for both MFS and CSS (Table 4). LVI positivity was an independent prognostic factor for both CSS and MFS in multivariate analysis (HR 7.87, 95 % CI 2.78–22.26, P \ 0.001; and HR 4.87, 95 % CI 2.10– 11.31, P \ 0.001, respectively). Concerning IRFS, UTUC with concomitant bladder tumor was the most important factor to predict intraluminal recurrence (P \ 0.001; Table 4). However, there was no significant difference in CSS, MFS, and IRFS between LVI-positive and
The value of LVI to predict poor prognosis has been discussed in a number of solid organ malignancies, including prostate, bladder, testis, penis, liver, and rectum.13 Cheng et al. reported that LVI is a substantial risk factor to predict disease progression and CSS in prostate cancer, and they recommended that LVI should be included in the pathological staging system after radical prostatectomy.14 LVI was also found to be associated with survival rates in patients with invasive bladder cancer after radical cystectomy.15 UTUC are rare carcinomas with a poor prognosis and aggressive behavior among genitourinary cancers. Identifying prognostic factors to predict a group at high risk of UTUC is crucial to facilitate individualized therapy and proper surveillance protocol. Pathological T stage and tumor grade were established as major prognostic factors, and the role of LVI in UTUC has been extensively discussed in recent years.16 In this study, we not only confirmed the prognostic value of LVI in UTUC but also discovered different significance of LVI according to tumor location. We demonstrated for the first time that LVI is a more essential factor in ureteral tumors than in pyelocaliceal tumors. On the basis of previous reports, the incidence of LVI positivity is approximately 15–30 % in patients with UTUC after surgery.9,17,18 In this study, 24 % of cases were LVI positive. LVI was believed to be an important stage for tumor cell dissemination and LN metastasis, and it was considered to be a poor prognostic factor in UTUC.13,19,20 Consistent with previous studies, LVI was associated with higher pathological T stage, grade, and LN involvement in our analysis. Also in accordance with preceding reports, we confirmed LVI is associated with a significant decline in CSS and MFS but not in IRFS by Kaplan–Meier analysis.17,21,22 As a result, LVI represented a strong predictor for disease progression and cancer-specific death, but it did not correlate with bladder and contralateral recurrence in UTUC. In the present study, pathological T stage, tumor grade, LN involvement, tumor multifocality, type of operative approach, and LVI were significant factors in influencing CSS in univariate analysis. In multivariate analysis, pathological T stage and LN involvement were independent predictors for CSS. A retrospective study compared efficacy between open and laparoscopic RNU and illustrated no significant difference in CSS.23 In our series, the poorer prognosis in open RNU group was assumed to be related to higher T stage compared with laparoscopic RNU group,
1.52 (0.53–4.31)
2.09 (0.77–5.66)
0.77 (0.30–2.00)
0.58 (0.22–1.51)
Yes versus no
11.97 (4.39–32.65)
3.05 (0.63–14.66) 0.91 (0.28–2.94)
\0.001* 7.87 (2.78–22.26)
0.648
0.070
0.046*
0.102
0.023* 5.60 (0.69–45.63)
0.362
0.611
0.79 (0.33–1.91)
3.77 (1.09–13.08)
9.23 (1.25–68.04)
8.23 (1.95–34.79)
\0.001* 8.79 (4.03–19.18)
0.869
0.165
0.210
0.107
1.21 (0.55–2.68)
1.31 (0.62–2.76)
0.98 (0.45–2.13)
2.51 (1.17–5.36)
0.70 (0.33–1.48)
0.61 (0.29–1.28)
1.11 (0.45–2.75) \0.001* 4.87 (2.10–11.31)
0.604
0.037* 2.09 (0.58–7.54)
0.022*
0.029* 3.23 (0.37–27.90)
0.004* 3.39 (0.71–16.20)
0.633
0.477
0.957
0.018* 2.05 (0.88–4.79)
0.352
0.192
0.819
0.261
0.504
0.287
0.126
0.097
\0.001*
Multivariate analysis, P HR (95 % CI)
IRFS
0.90 (0.35–2.31)
0.59 (0.30–1.18)
0.88 (0.25–3.13)
0.50 (0.26–0.97)
0.66 (0.35–1.23)
1.25 (0.64–2.42)
13.78 (5.72–33.22)
1.54 (0.77–3.09)
1.61 (0.80–3.22)
1.00 (0.54–1.86)
1.03 (0.55–1.94)
0.227
0.184
0.998
0.924
0.827
0.134
0.846
0.303
0.040*
0.191
0.514
0.61 (0.31–1.18)
* P \ 0.05
0.141
\0.001*
Multivariate analysis, P HR (95 % CI)
\0.001* 13.42 (5.55–32.45)
Univariate analysis, P HR (95 % CI)
CSS cancer-specific survival, MFS metastasis-free survival, IRFS intraluminal recurrence-free survival, RNU radical nephroureterectomy, HR hazard ratio, CI confidence interval, LVI lymphovascular invasion
LVI positive versus LVI negative
LVI
4.07 (0.89–18.61)
0.76 (0.24–2.44)
pNx versus pN0
5.42 (0.72–40.96)
10.50 (1.39–79.55)
1.59 (0.59–4.32)
1.28 (0.49–3.33)
pN1/2 versus pN0
N stage
High versus low
Grade
pT2–4 versus pTa/Tis/T1
Pathologic stage
Multifocal versus solitary
Tumor multifocality
0.436
0.146
0.588
0.265
MFS Univariate analysis, P HR (95 % CI)
Multivariate analysis, P HR (95 % CI)
CSS
Univariate analysis, P HR (95 % CI)
Presence with synchronous bladder tumor
Open versus laparoscopic
Type of operation
Yes versus no
Smoking
Female versus male
Gender
C68 years versus \68 years
Age
Characteristic
TABLE 4 Univariate and multivariate analyses predicting CSS, MFS, and IRFS in patients with ureteral tumor after RNU
1398 H.-Y. Lee et al.
Effect of LVI on UTUC by Location
and consequently the effect of operation method diminished after adjusting other variables. Similar to the findings of Chromecki et al., tumor multifocality was associated with increased risk of CSS in univariate analysis, but no significant difference was observed in multivariate analysis.24 For MFS, pathological T stage, tumor grade, LN involvement, tumor multifocality, and LVI were significant factors in univariate analysis. In multivariate analysis, pathological T stage, LN involvement, and LVI represented important factors for disease progression. Both our univariate and multivariate analyses showed that UTUC with synchronous bladder tumor is the only independent risk factor for IRFS, which was also verified in a recent study.25 Among UTUC, tumors in the renal pelvis and ureter may have different characteristics. Chung et al. found that the thicker anatomic barrier of renal pelvis than ureter can lead to different consequences, and they suggested that it may be more reasonable to individually evaluate tumors in different locations.12 We found that the rate of prevalence of LVI is higher in pyelocaliceal tumors than in ureteral tumors. All previous studies have shown this feature, although some detected a significant difference in LVI according to location.9,22,26 Others, however, did not.17,27,28 Pyelocaliceal tumors did not result in poorer prognosis. Instead, they had similar or even better outcomes than ureteral tumors.26,28,29 There are limited data concerning prognostic factors of ureteral tumors in particular, and none of them incorporated LVI for survival analysis.30,31 In this cohort, pT stage, LN metastasis, and LVI were significant factors for both CSS and MFS of ureteral tumors in univariate analysis, and LVI represented an independent predictor in multivariate analysis. In contrast, LVI failed to be independently associated with CSS and MFS in pyelocaliceal tumors. We hypothesized that the diverse effect of LVI in different location is related to the thickness of adjacent barrier. As mentioned before, LVI is an important step for tumor cells to disseminate. If LVI is present in ureteral tumors, its thinner muscular wall is easier for invasion, and thinner adventitia, which is rich in blood vessels and lymphatic plexus, can facilitate tumor spread. Therefore, the prognostic value of LVI is further highlighted with respect to ureteral tumors specifically. There are several limitations to the present study. First, the study design is retrospective and the number of cases is small. Second, there is no initial centralized pathological review of the LVI status. However, all cases were reanalyzed by two independent pathologists, with high consistency in the interpretation of LVI. Third, most subjects did not undergo LN dissection. Patients who received LN dissection can have more accurate N stage, which may dilute the prognostic value of LVI. Nevertheless, the percentage of retrieved LNs was similar between different
1399
tumor locations (27 % in pyelocaliceal and 30 % in ureteral tumors; P = 0.604). Furthermore, in multivariate analysis, we demonstrated that LVI is a powerful and practical parameter for prediction of prognosis, especially in ureteral tumors. One of our study’s strengths is its separate evaluation of MFS and IRFS; another is our detailed subgroup analysis. The patient population we studied was different from those described from Western countries because the incidence of UTUC is far higher in Taiwan than in any other area. In addition, in both Taiwan and China, female predominance in UTUC is another unique feature that is dissimilar to other countries.2,3,32,33 This study provided information about identifying prognostic factors to predict outcome of UTUC in Taiwan. More importantly, this is the first study to demonstrate the enhanced effect of LVI in ureteral tumors specifically. The diverse effect of LVI in different tumor locations in UTUC will require larger prospective studies to validate. CONCLUSIONS We confirmed the imperative role of LVI in predicting disease progression and survival of patients with UTUC after receiving RNU in Taiwan. The remarkable importance of LVI on prognosis, particularly in ureteral tumors rather than pyelocaliceal tumors, may explain the different characteristics between these two locations of UTUC. We think that LVI status can be a part of the histopathological assessment to identify patients at high risk for receiving individualized treatment and surveillance strategies, especially in ureteral tumors. ACKNOWLEDGMENT This study was supported by grants from the Kaohsiung Medical University Hospital (KMUH 97-7G06), the Center for Infectious Disease and Cancer Research, Kaohsiung Medical University (KMU-TP103E19), and the Health and Welfare Surcharge of tobacco products, Ministry of Health and Welfare (MOHW103-TD-B-111-05). DISCLOSURE
The authors declare no conflict of interest.
REFERENCES 1. Roupret M, Babjuk M, Comperat E, et al. European guidelines on upper tract urothelial carcinomas: 2013 update. Eur Urol. 2013; 63:1059–71. 2. Li CC, Chang TH, Wu WJ, et al. Significant predictive factors for prognosis of primary upper urinary tract cancer after radical nephroureterectomy in Taiwanese patients. Eur Urol. 2008;54: 1127–34. 3. Li WM, Shen JT, Li CC, et al. Oncologic outcomes following three different approaches to the distal ureter and bladder cuff in nephroureterectomy for primary upper urinary tract urothelial carcinoma. Eur Urol. 2010;57:963–9.
1400 4. Verhoest G, Shariat SF, Chromecki TF, et al. Predictive factors of recurrence and survival of upper tract urothelial carcinomas. World J Urol. 2011;29:495–501. 5. Saito K, Kawakami S, Fujii Y, Sakura M, Masuda H, Kihara K. Lymphovascular invasion is independently associated with poor prognosis in patients with localized upper urinary tract urothelial carcinoma treated surgically. J Urol. 2007;178:2291–6. 6. Eisenberg MS, Boorjian SA, Cheville JC, et al. The SPARC score: a multifactorial outcome prediction model for patients undergoing radical cystectomy for bladder cancer. J Urol. 2013;190:2005–10. 7. Lee JH, Jang HS, Kim JG, et al. Lymphovascular invasion is a significant prognosticator in rectal cancer patients who receive preoperative chemoradiotherapy followed by total mesorectal excision. Ann Surg Oncol. 2012;19:1213–21. 8. Baydar DE, Baseskioglu B, Ozen H, Geyik PO. Prognostic significance of lymphovascular invasion in clinically localized prostate cancer after radical prostatectomy. Sci World J. 2008;17:303–12. 9. Godfrey MS, Badalato GM, Hruby GW, Razmjoo M, McKiernan JM. Prognostic indicators for upper tract urothelial carcinoma after radical nephroureterectomy: the impact of lymphovascular invasion. BJU Int. 2012;110:798–803. 10. Matsumoto K, Novara G, Gupta A, et al. Racial differences in the outcome of patients with urothelial carcinoma of the upper urinary tract: an international study. BJU Int. 2011;108(8 Pt 2):E304–9. 11. Lin WC, Hu FC, Chung SD, Chueh SC, Pu YS, Huang KH. The role of lymphovascular invasion in predicting the prognosis of clinically localized upper tract urothelial carcinoma (pT1-3cN0M0). J Urol. 2008;180:879–84. 12. Chung SD, Wang SM, Lai MK, et al. Lymphovascular invasion predicts poor outcome of urothelial carcinoma of renal pelvis after nephroureterectomy. BJU Int. 2009;103:1047–51. 13. Ku JH, Byun SS, Jeong H, Kwak C, Kim HH, Lee SE. Lymphovascular invasion as a prognostic factor in the upper urinary tract urothelial carcinoma: a systematic review and meta-analysis. Eur J Cancer. 2013;49:2665–80. 14. Cheng L, Jones TD, Lin H, et al. Lymphovascular invasion is an independent prognostic factor in prostatic adenocarcinoma. J Urol. 2005;174:2181–5. 15. Lotan Y, Gupta A, Shariat SF, et al. Lymphovascular invasion is independently associated with overall survival, cause-specific survival, and local and distant recurrence in patients with negative lymph nodes at radical cystectomy. J Clin Oncol. 2005;23:6533–9. 16. Hong B, Park S, Hong JH, Kim CS, Ro JY, Ahn H. Prognostic value of lymphovascular invasion in transitional cell carcinoma of upper urinary tract. Urology. 2005;65:692–6. 17. Hurel S, Roupreˆt M, Ouzzane A, et al. Impact of lymphovascular invasion on oncological outcomes in patients with upper tract urothelial carcinoma after radical nephroureterectomy. BJU Int. 2013;111:1199–207. 18. Margulis V, Shariat SF, Matin SF, et al. Outcomes of radical nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration. Cancer. 2009;115:1224–33.
H.-Y. Lee et al. 19. Akao J, Matsuyama H, Yamamoto Y, et al. Clinical significance of lymphovascular invasion in upper urinary tract urothelial cancer. BJU Int. 2008;102:572–5. 20. Ouzzane A, Colin P, Ghoneim TP, et al. The impact of lymph node status and features on oncological outcomes in urothelial carcinoma of the upper urinary tract (UTUC) treated by nephroureterectomy. World J Urol. 2013;31:189–97. 21. Cha EK, Shariat SF, Kormaksson M, et al. Predicting clinical outcomes after radical nephroureterectomy for upper tract urothelial carcinoma. Eur Urol. 2012;61:818–25. 22. Novara G, Matsumoto K, Kassouf W, et al. Prognostic role of lymphovascular invasion in patients with urothelical carcinoma of the upper urinary tract: an international validation study. Eur Urol. 2010;57:1064–71. 23. Zou L, Zhang L, Zhang H, Jiang H, Ding Q. Comparison of postoperative intravesical recurrence and oncological outcomes after open versus laparoscopic nephroureterectomy for upper urinary tract urothelial carcinoma. World J Urol. 2014;32:565–70. 24. Chromecki TF, Cha EK, Fajkovic H, et al. The impact of tumor multifocality on outcomes in patients treated with radical nephroureterectomy. Eur Urol. 2012;61:245–53. 25. Pignot G, Colin P, Zerbib M, et al. Influence of previous or synchronous bladder cancer on oncologic outcomes after radical nephroureterectomy for upper urinary tract urothelial carcinoma. Urol Oncol. 2014;32:23.e1–8. 26. Yafi FA, Novara G, Shariat SF, et al. Impact of tumour location versus multifocality in patients with upper tract urothelial carcinoma treated with nephroureterectomy and bladder cuff excision: a homogeneous series without perioperative chemotherapy. BJU Int. 2012;110(2 Pt 2):E7–13. 27. Kikuchi E, Margulis V, Karakiewicz PI, et al. Lymphovascular invasion predicts clinical outcomes in patients with node-negative upper tract urothelial carcinoma. J Clin Oncol. 2009;27:612–8. 28. Ouzzane A, Colin P, Xylinas E, et al. Ureteral and multifocal tumours have worse prognosis than renal pelvic tumours in urothelial carcinoma of the upper urinary tract treated by nephroureterectomy. Eur Urol. 2011;60:1258–65. 29. Williams AK, Kassouf W, Chin J, et al. Multifocality rather than tumor location is a prognostic factor in upper tract urothelial carcinoma. Urol Oncol. 2013;31:1161–5. 30. Li WM, Li CC, Ke HL, Wu WJ, Huang CN, Huang CH. The prognostic predictors of primary ureteral transitional cell carcinoma after radical nephroureterectomy. J Urol. 2009;182:451–8. 31. Raman JD, Messer J, Sielatycki JA, Hollenbeak CS. Incidence and survival of patients with carcinoma of the ureter and renal pelvis in the USA, 1973–2005. BJU Int. 2011;107:1059–64. 32. Chen XP, Xiong GY, Li XS, et al. Predictive factors for worse pathological outcomes of upper tract urothelial carcinoma: experience from a nationwide high-volume centre in China. BJU Int. 2013;112:917–24. 33. Matin SF, Shariat SF, Milowsky MI, et al. Highlights from the first symposium on upper tract urothelial carcinoma. Urol Oncol. 2014;32:309–16.
