This article is mentioned in:

Introduction

Urothelial carcinoma is the second most common urological malignancy worldwide, after prostate cancer (1). A distinguishing feature of urothelial carcinomas is their multiple foci, which cause the tumors to appear synchronously or sequentially throughout the urinary tract, including the upper urinary tracts (renal pelvis or ureter), bladder and urethra (2). Upper tract urothelial carcinomas (UTUC) are uncommon and account for only 5–10% of urothelial carcinomas (3). At present, tobacco exposure is considered the most important risk factor for urothelial carcinoma (4,5). Gross or microscopic hematuria is the presenting symptom in 70–80% of UTUC patients (6). Furthermore, at the time of diagnosis 40–50% of patients exhibit in situ [pTa to pT1 (7)] disease, 50–60% of patients exhibit invasive or advanced disease [p≥T2 (7)], and ~25% patients already exhibit regional metastasis (8,9). Radical nephroureterectomy (RNU) with excision of the bladder cuff is the gold-standard treatment for UTUC (3). However, alternative treatments include ureteroscopic ablation, percutaneous resection and segmental resection (10). The oncological outcomes for patients with high-grade or non-organ-confined disease remain poor, with 5-year cancer-specific survival rates of <60%; while for patients with non-muscle-invasive lesions, the 5-year recurrence-free survival rate is 88.0–91.8% (11). For patients with low-grade carcinomas, conservative strategies, including segmental ureterectomy or endoscopic management, provide cancer-specific survival (CCS) and overall survival (OS) rates equivalent to that achieved using RNU (12,13), with a 5-year cancer-specific survival rate of >93% (14), whereas patients at high-risk (pT3 or N+) may benefit from neoadjuvant chemotherapy (15,16). The ability to accurately predict pathological outcomes prior to initiating therapy may aid in clinical risk stratification and the selection of therapeutic strategies.

Ipsilateral hydronephrosis (HN) is common in UTUC patients, and may be attributed to one of several factors, including luminal obstruction, intramural invasion or extrinsic compression (17). The presence of ipsilateral HN in patients with bladder cancer is a predictive factor for poor pathological outcome and poor prognosis (18–20); however, at present, no consensus has been reached regarding the predictive role of the presence of HN in UTUC patients. Although HN has been reported to be associated with advanced disease (17,21–24), only two studies demonstrated a correlation between HN and poor prognosis based on small sample (17,25).

Our previous work has revealed associations between HN and muscle-invasive and grade 3 diseases (21). In the present study, after revising our database to include the follow-up information of patients treated between 2000 and 2010, we sought to validate the predictive value of preoperative HN on clinicopathological outcome and prognosis with the aim of improving clinical risk stratification and, thus, the ability to provide more optimal and personalized risk-informed therapeutic options.

Materials and methods

Patient selection

The clinicopathological data of consecutive UTUC patients treated between 2000 and 2010 at Peking University First Hospital (Beijing, China) were collected. Among the 631 patients with complete follow-up, 111 were excluded from the analysis: 25 with bilateral synchronous UTUCs, 54 who underwent alternative surgeries rather than RNU, 28 with a follow-up period of <12 months, 2 with metastatic disease and 2 with positive surgical margins. A total of 520 patients were finally enrolled for evaluation. All patients were diagnosed using computed tomography (CT) or magnetic resonance imaging (MRI), urological ultrasound and ureteroscopy with or without biopsy. None of these patients received neoadjuvant chemotherapy, however, for certain patients, adjuvant chemotherapy or radiotherapy was administered when evidence of distant metastasis or retroperitoneal recurrence was documented. All patients underwent surgery within two months after the occurrence of symptoms. Ethical approval was obtained from Peking University Institutional Review Board (IRB00001052-13057).

Ipsilateral HN status

Ipsilateral HN was assessed by upper urinary tract imaging, including CT with or without intravenous contrast in 510 patients, and MRI with or without intravenous contrast in 10 patients. Only imaging studies performed within 6 weeks of RNU and which were evaluated for HN by radiologists blinded to clinical outcomes were considered. As ~100 CT films were not available for re-evaluation, two authors (Dr Xuesong Li and Dr Gengyan Xiong) blinded to the radiology reports reviewed the 10 MRI films and 100 CT films independently. The concordance between the two observers in assessing presence or absence of HN was 95.5% and, by consensus decision, the concordance between re-evaluation and primary reports was 97.3%.

The evaluation criteria for assessing the presence or absence of HN were similar to those of a previous study (23). For renal pelvic lesions, patients with hydrocalycosis were included in the cohort of patients considered to have HN. A hydrocalyx was defined as any degree of dilation within a focal calyx, with or without the presence of obvious obstruction at the draining infundibulum. For ureteral tumors, any degree of dilation in any component of the ureter or associated renal unit was classified as HN. To avoid small subgroups and heterogeneity with respect to the grading of HN, the status of HN was evaluated strictly as present or absent in the current analysis.

Patients evaluation

All pathological specimens were re-reviewed by a dedicated genitourinary pathologist (Dr Qun He) to unify the reproducibility of the diagnosis. Tumor stage was assessed according to the 2002 Union for International Cancer Control TNM classification of malignant tumors (10). Tumor grade was assessed according to the World Health Organization classification of 1973 (10). Tumor architecture was defined as papillary or sessile by examination of the final specimen. Tumor location was divided into two areas (renal pelvis and ureter) based on the site of the dominant lesion. Tumor multifocality was defined as the synchronous presence of two or more pathologically confirmed macroscopic tumors in any location. The estimated glomerular filtration rate was calculated using the modified glomerular filtration rate estimating equation for Chinese patients (26).

Follow-up schedule

Of the total cohort (n=631), 520 patients were included in the current analyses. For patients who were followed-up at our institute, the follow-up regimen of the affected patients included cystoscopy every 3 months for the first 3 years; cystoscopy intervals were extended to 1 year thereafter. Chest X-ray, serum creatinine level and abdominal ultrasound or CT were examined concurrently. The impact of preoperative HN on CSS, OS, bladder recurrence-free survival and contralateral carcinoma-free survival times was determined. Bladder recurrence was defined as the detection of a subsequent bladder tumor during cystoscopy and confirmation by pathology, while contralateral carcinoma was defined as urothelial carcinoma in the contralateral upper urinary tract. The causes of patient mortality were determined by the treating physicians.

Statistical analysis

All statistical tests were performed using SPSS software version 20.0 (IBM SPSS, Armonk, NY, USA), and the threshold for statistical significance was set at P<0.05. The Pearson test and χ2 test were used to assess the distribution of categorical variables, and the Mann-Whitney U test was used for continuous variables. Univariate analysis using the log-rank test and multivariate analysis using Cox's proportional hazards regression model were also conducted. Only variables that were indicated to be significant upon univariate analysis were considered for the multivariate analysis.

Results

Patient clinical and pathological characteristics and HN

The clinical and pathological data of the included patients and their association with HN are shown in Table I. Of the 520 patients enrolled, ipsilateral HN was present in 271 patients (52.1%). There were 340 patients with muscle-invasive disease (T stage ≥2), and 191 patients were diagnosed with histological grade 3 disease by final pathology. Preoperative HN was associated with advanced age (P=0.007), sessile tumor architecture (P<0.001), ureteral location (P<0.001), high tumor stage (P<0.001) and higher histological grade (P=0.002). No distribution differences in terms of gender, preoperative kidney function, multifocality, presence of carcinoma in situ (CIS) or tumor size were identified.

Table I. Patient clinical and pathological characteristics and ipsilateral hydronephrosis.

Table I.

Patient clinical and pathological characteristics and ipsilateral hydronephrosis.

	Ipsilateral hydronephrosis, n (%)
Characteristic	Absent	Present	χa	P-value
Gender			0.004	0.951
Male	110 (21.15)	119 (22.88)
Female	139 (26.73)	152 (29.23)
Age, years			7.320	0.007a
<70	157 (30.19)	139 (26.73)
≥70	92 (17.69)	132 (25.38)
Preoperative renal function			3.221	0.522
CKD1 (eGFR≥90)	26 (5.00)	18 (3.46)
CKD2 (90>eGFR≥60)	88 (16.92)	96 (18.46)
CKD3 (60>eGFR≥30)	106 (20.38)	117 (22.50)
CKD4 (30>eGFR≥15)	13 (2.50)	19 (3.65)
CKD5 (eGFR<15)	16 (3.08)	21 (4.04)
Urinary cytology			0.675	0.714
Negative	57 (10.96)	58 (11.15)
Positive	122 (23.46)	128 (24.62)
Missing data	70 (13.46)	85 (16.35)
Tumor architecture			16.604	0.000a
Papillary	219 (42.12)	200 (38.46)
Sessile	30 (5.77)	71 (13.65)
Multifocality			2.788	0.095
No	205 (39.42)	207 (39.81)
Yes	44 (8.46)	64 (12.31)
Location			145.028	0.000a
Ureter	54 (10.38)	202 (38.85)
Pelvis	195 (37.50)	69 (13.27)
Tumor stage			24.199	0.000a
Ta	22 (4.23)	8 (1.54)
T1	76 (14.62)	74 (14.23)
T2	75 (14.42)	127 (24.42)
T3	76 (14.62)	59 (11.35)
T4	0 (0.00)	3 (0.58)
Node stage			2.066	0.356
N0	20 (3.85)	31 (5.96)
Nx	225 (43.27)	233 (42.88)
N+	4 (0.77)	6 (1.15)
Tumor grade			12.568	0.002a
G1	10 (1.92)	9 (1.73)
G2	167 (32.12)	143 (27.50)
G3	72 (13.85)	119 (22.88)
Tumor necrosis			0.286	0.593
No	225 (43.27)	241 (46.35)
Yes	24 (4.62)	30 (5.77)
Tumor size, cm			1.205	0.272
≤3	156 (30.00)	157 (30.19)
>3	93 (17.88)	114 (21.92)
CIS			0.030	0.863
Absent	241 (46.35)	263 (50.58)
Present	8 (1.54)	8 (1.54)

a Statistically significant. CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; CIS, carcinoma in situ.

Survival outcomes and HN

During a median follow-up duration of 54 months (range, 12–151 months), 120 patients (23.1%) died, including 105 patients (20.2%) who succumbed to urothelial cancer. Of these patients, 78 (65.0%) had preoperative HN. The 5-year CSS and OS rates for patients with HN were 86.9 and 86.2%, respectively: Markedly lower than for patients without HN (93.3 and 91.9%, respectively).

Kaplan-Meier estimated CSS and OS curves are shown in Fig. 1A and B. The presence of HN was a significant risk factor for poorer CSS and OS times according to univariate analysis (P=0.004 and P=0.009, respectively). In the multivariate analysis, the presence of HN remained a significant predictive factor for CSS and OS (P=0.001 and P=0.011, respectively). The multivariate analysis also confirmed male gender, advanced age and higher tumor stage as risk factors for reduced survival (Tables II and III).

Figure 1. (A) Kaplan-Meier estimated cancer-specific survival curves stratified by the presence of HN (P=0.004). (B) Kaplan-Meier estimated overall survival curves stratified by the presence of HN (P=0.009). HN, hydronephrosis.

Table II. Univariate and multivariate analysis of risk factors for cancer-specific survival.

Table II.

Univariate and multivariate analysis of risk factors for cancer-specific survival.

				Multivariate analysis
	Patients	Recurrence	Univariate analysis
Variable	n	n	P-value	Hazard ratio	95% CI	P-value
Presence of hydronephrosis			0.004a	1.932	1.294–2.883	0.001a
Absence	249	39
Presence	271	66
Gender			0.000a	0.491	0.327–0.738	0.001a
Female	291	43
Male	229	62
Age, years			0.028a			0.437
<50	28	7
50–60	92	15
60–70	175	31
70–80	187	40
≥80	38	12
Preoperative renal function			0.940
No CKD (eGFR≥60)	228	44
Early CKD (60>eGFR≥15)	255	52
End-stage CKD (eGFR<15)	37	9
Urinary cytology			0.648
Negative	115	22
Positive	250	57
Missing data	155	26
Surgical approach			0.743
Open	348	76
Laparoscopic	172	29
Tumor architecture			0.008a			0.364
Papillary	419	79
Sessile	101	26
Multifocality			0.560
No	412	84
Yes	108	21
Location			0.060
Ureter	256	58
Pelvis	264	47
Tumor stage			0.000a	1.663	1.289–2.145	0.000a
Ta	30	0
T1	150	13
T2	202	46
T3	135	41
T4	3	2
Node stage			0.284
N0	51	10
Nx	458	91
N+	10	3
Tumor grade			0.010a			0.172
G1	19	0
G2	310	56
G3	191	49
Tumor necrosis			0.021a	2.069	1.162–3.686	0.014a
No	466	90
Yes	54	15
Tumor size, cm			0.165
≤3	313	59
>3	207	46
CIS			0.318
Absent	504	103
Present	16	2
Adjuvant therapy			0.470
No	485	95
Yes	30	8

a Statistically significant. CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; CIS, carcinoma in situ.

Table III. Univariate and multivariate analysis of risk factors for overall survival.

Table III.

Univariate and multivariate analysis of risk factors for overall survival.

				Multivariate analysis
Variable	Patients, n	Recurrence, n	Univariate analysis P-value	Hazard ratio	95% CI	P-value
Presence of hydronephrosis			0.009a	1.587	1.111–2.265	0.011a
Absence	249	52
Presence	271	78
Gender			0.000a	0.604	0.426–0.858	0.005a
Female	291	57
Male	229	73
Age, years			0.005a	1.284	1.068–1.542	0.008a
<50	28	8
50–60	92	17
60–70	175	38
70–80	187	53
≥80	38	14
Preoperative renal function			0.260
No CKD (eGFR≥60)	228	48
Early CKD (60>eGFR≥15)	255	68
End-stage CKD (eGFR<15)	37	14
Urinary cytology			0.385
Negative	116	26
Positive	250	73
Missing data	155	31
Surgical approach			0.503
Open	348	96
Laparoscopic	172	34
Tumor architecture			0.015a			0.262
Papillary	419	100
Sessile	101	30
Multifocality			0.271
No	412	106
Yes	108	24
Location			0.110
Ureter	256	69
Pelvis	264	61
Tumor stage			0.000a	1.581	1.265–1.976	0.000a
Ta	30	4
T1	150	19
T2	202	57
T3	135	48
T4	3	2
Node stage			0.519
N0	52	11
Nx	458	116
N+	10	3
Tumor grade			0.005a			0.133
G1	19	0
G2	310	70
G3	191	60
Tumor necrosis			0.137
No	466	115
Yes	54	15
Tumor size, cm			0.306
≤3	313	76
>3	207	54
CIS			0.344
Absent	504	127
Present	16	3
Adjuvant therapy			0.648
No	485	119
Yes	30	9

a Statistically significant. CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; CIS, carcinoma in situ.

As higher tumor stage is a well-established predictor of survival (11), the predictive role of HN was analyzed only in the 340 patients with muscle-invasive disease (T≥2) (Table III). Of the 189 patients with HN, there were 68 mortalities, comprising 58 cancer-specific mortalities, during follow-up. By contrast, only 39 mortalities, including 31 cancer-specific mortalities, occurred among the remaining 151 patients without HN. The differences in CSS and OS times between the two groups were statistically significant (P=0.009 and P=0.012, respectively; Fig. 2A and B).

Figure 2. (A) Kaplan-Meier estimated cancer-specific survival curves in patients with muscle-invasive disease stratified by the presence of HN (P=0.009). (B) Kaplan-Meier estimated overall survival curves in patients with muscle-invasive disease stratified by the presence of HN (P=0.012). HN, hydronephrosis.

Pathology confirmed that 178 patients (34.2%) had intravesical recurrence and 35 patients (6.7%) had subsequent contralateral UTUC. Of the patients with HN, 99 experienced intravesical recurrence and 22 contralateral disease during follow-up, and there was no association between the presence of HN and bladder cancer recurrence-free survival time (P=0.552) or the development of contralateral carcinoma-free survival time (P=0.164) (Fig. 3A and B).

Figure 3. (A) Kaplan-Meier estimated bladder recurrence-free survival curves stratified by the presence of HN (P=0.552). (B) Kaplan-Meier estimated contralateral upper tract urothelial carcinoma-free survival curves stratified by the presence of HN (P=0.164). HN, hydronephrosis.

Discussion

Preoperative HN can be present in bladder tumors and UTUCs. The incidence of HN in bladder tumors is reported to be 5.3–22.7%, and the presence of HN has been demonstrated to be associated with poor pathological outcomes, tumor recurrence and progression (18–20). The presence of HN is more prevalent in UTUC than in bladder tumors (52.1% in the current study), which may be because urinary obstruction is more likely to occur in the ureter from a small mass.

There have been a number of studies focused on the association between the presence of HN and clinicopathological characteristics and prognosis (17,22–25,27). The majority of these studies have reported a predictive role of HN in poor pathological outcomes, however, there has been no consensus on the association between HN and poor prognosis. Ng et al (17) confirmed that HN was independently associated with cancer metastasis and cancer-specific mortality by preoperative multivariable analysis controlling for preoperative clinical features. However, their research was based on only 106 patients and HN was no longer an independent risk factor upon postoperative multivariable analysis. Hwang et al (25) reported that preoperative HN predicted poor prognosis in 114 patients; data on architecture, multifocality and preoperative renal function were unavailable. Ito et al (22) reported that 67 patients (73.6%) exhibited HN in a retrospective study of 91 cases, and validated the correlation between HN and poor pathological outcomes, whilst a higher HN grade was not associated with disease-specific or metastasis-free survival. Bozzini et al (27) conducted a relatively large-scale study with 401 patients, however, HN was present in only 18.4% of patients and the median follow-up period was 26 months. Furthermore, whilst studies by Messer et al (23) and Brien et al (24) demonstrated that HN was associated with muscle-invasive and non-organ confined disease, these studies were lacking in follow-up data.

The proportions of HN reported in previous studies may differ due to the lack of clear criteria with which to evaluate HN. Based on the current analysis, the presence of HN was associated with a number of poor pathological outcomes, including high tumor stage, high tumor grade and sessile tumor architecture. In addition, a greater number of tumors were located in the ureter in patients with HN, and previous studies have demonstrated that patients with ureteral tumors have a poorer prognosis compared with those with renal pelvis tumors, after adjustment for a number of pathological variables (28,29). A recent study attributed this difference in prognosis to the fact that ureteral tumors are more likely to have HN (30). According to univariate and multivariate survival analyses, the presence of HN was associated with poorer survival, which confirmed the role of HN as an independent risk factor for poor prognosis. Preoperative HN must be carefully evaluated as a significant predictive factor for prognosis as well as higher tumor stage and tumor grade. The present study observed no correlation between HN and bladder recurrence or contralateral UTUC, and no such association has been reported previously (31).

Using conservative surgeries, including segmental ureterectomy or endoscopic management, renal function may be preserved and perioperative complications with RNU avoided (32). Clinical consideration of advanced disease based on the presence of HN could allow physicians to better individually assess treatment options in UTUC; patients with HN may not be suitable candidates for less invasive surgical options. In addition, patients with locally advanced UTUC have significantly higher local recurrence and distant metastasis rates following RNU, compared with patients exhibiting early stage disease (11,33). Such findings call for effective strategies for perioperative systemic therapy to improve survival. The presence of HN also indicates a need for aggressive treatment, including lymphadenectomy and systemic chemotherapy.

Neoadjuvant chemotherapy appears to achieve favorable oncological outcomes in high-risk patients (15,16), while adjuvant chemotherapy confers minimal impact on OS or CSS (34,35). In addition, not all patients are able to receive adjuvant treatment due to comorbidities and impaired renal function following RNU (36). Hoshino et al (37) found that patients with no HN or a lower grade of HN have a higher risk of missing the opportunity to undergo adjuvant chemotherapy for impaired renal function following RNU. Thus, if patients without HN are evaluated as high-risk (based on lymph node metastasis or higher biopsy grade) and systemic chemotherapy is considered, neoadjuvant chemotherapy is recommended before renal function becomes impaired.

The limitations of the current study include the retrospective design and data collection, and the lack of re-evaluation of a number of CT films. Therefore, the study cohort may be subject to selection and recall bias. The incidence of UTUC in the Chinese population is markedly higher compared with that of western populations, and the biology may differ (38). Although the mechanisms related to these ethnic differences are still not fully known, dietary exposure to toxins may play a major role (39). In addition, Chinese UTUC patients are more likely to be female, and females are less likely to be of an advanced pathological disease stage compared with males (40,41). Despite its limitations, the present study is currently the largest to report on the predictive role of HN in UTUC patients, and the first study confirming an association between HN and poor prognosis after controlling for other clinical and pathological characteristics in a large sample.

In conclusion, preoperative HN is prevalent in UTUC. The presence of preoperative HN predicted poorer pathological outcomes and was a significant risk factor affecting survival. The evaluation of HN may therefore be informative for decisions concerning surgical strategies, and the presence of HN should raise the possibility of employing an aggressive treatment strategy.

Acknowledgements

The present study was supported by grants from the Clinical Features Research of Capital (grant nos. Z121107001012154 and Z151100004015173), the Collaborative Research Foundation of Peking University Health Science Center and the College of Medicine, National Taiwan University (grant no. BMU20120318), the Natural Science Foundation of China (grant nos. 81172419 and 81372746), the Natural Science Foundation of Beijing (grant no. 7122183), and the Research Foundation of Peking University First Hospital (grant no. 2015QN026).

References