Hereditary-like urothelial carcinomas of the upper urinary tract benefit more from adjuvant cisplatin-based chemotherapy after radical nephroureterectomy than do sporadic tumours

Authors


Abstract

Objective

  • To evaluate the impact of ‘hereditary-like’ status in upper tract urothelial carcinoma (UTUC) on the survival of patients who have undergone radical nephroureterectomy (RNU) and adjuvant chemotherapy.

Patients and Methods

  • A multicentre retrospective study was performed on all patients with high-risk UTUC who underwent RNU and adjuvant cisplatin-based chemotherapy.
  • Using a patient risk identification tool, we distinguished tumours suspected to be hereditary from sporadic tumours and compared survival rates.

Results

  • A total of 112 patients with a median age of 67 years were included. Hereditary-like tumour status was detected in 35 patients (31.3%), while 77 patients (68.7%) had sporadic tumours.
  • The median age was significantly younger in the hereditary-like tumour group (56.0 vs 69.8 years, P < 0.001). Overall survival (OS) after chemotherapy was significantly better in the group with hereditary-like tumours than in the group with sporadic tumours (5-year OS: 48.2 vs 32%; P = 0.008).
  • The cancer-specific survival (CSS) rate was significantly better in the group with ‘hereditary-like’ tumours than in the group with sporadic tumours (5-year CSS: 58 vs 35%; P = 0.006).
  • Although there was a trend in favour of the hereditary-like tumours, we observed no significant difference regarding progression-free survival (PFS) between the two groups (5-year PFS: 71 vs 52%; P = 0.07).

Conclusion

  • Adjuvant chemotherapy after RNU improves survival outcomes in patients with hereditary-like UTUC compared with those with sporadic tumours.
Abbreviations
UTUC

upper tract urothelial carcinoma

RNU

radical nephroureterectomy

OS

overall survival

CSS

cancer-specific survival

PFS

progression-free survival

HNPCC

hereditary nonpolyposis colorectal cancer

MSI

microsatellite instability

Introduction

Urological malignancies are an integral part of the hereditary nonpolyposis colorectal cancer (HNPCC)-associated tumour spectrum, particularly upper tract urothelial carcinomas (UTUCs) [1, 2]. The relationship of UTUC with HNPCC is profound, as individuals with HNPCC have an approximately twenty-twofold increased risk of developing a UTUC compared with the general population [1]. HNPCC is the most common monogenetic predisposition to colorectal cancer [3]. HNPCC, previously known as Lynch syndrome, is a form of colorectal cancer with an autosomal dominant mode of inheritance. Previous analysis has shown that individuals with HNPCC and their relatives have significantly higher rates of extracolonic cancers, which is also a genetically mediated phenomenon [3]. There is a lack of appreciation of the association of HNPCC with a wide variety of extracolonic tumours, and therefore it is presumed that some hereditary cancers are misclassified as sporadic and their incidence is underestimated [3, 4]; however, most urologists are not aware of this distinct association even though their daily practice involves the management of patients with this potential. In colorectal cancer, several studies have shown that patients linked to HNPCC who received adjuvant chemotherapy after surgery had a better prognosis, lower recurrence rates and better survival outcomes compared with patients with sporadic tumours [3, 5-7]. To date, there is no strong evidence that adjuvant chemotherapy is effective in patients with UTUC after radical nephroureterectomy (RNU) [8-10]. In a recent study, we proposed a simple, clinical and user-friendly patient-specific risk identification tool that can be used to distinguish ‘hereditary-like’ tumours from sporadic tumours [11]. The aim of the present study was to assess whether patients with suspected hereditary UTUC had better survival outcomes after adjuvant chemotherapy than patients with sporadic tumours.

Materials and Methods

Study Population

A multi-institutional, international retrospective study was conducted that included patients from our French national database on UTUC [10-12]. A total of 598 patients with high-risk/locally advanced UTUC (pT3N0, pT4N0 and/or lymph node-positive disease and/or metastasis) who underwent RNU as a first step were selected from our multicentre database. Their medical files were reviewed to collect the following data: localization of the primary tumour, tumour stage at diagnosis (TNM 2009), tumour grade (WHO classification), surgical margin status and vascular invasion, Eastern Cooperative Oncology Group performance status, chemotherapy regimen administered, date of diagnosis, date of relapse or disease progression, date of last visit or death and cause of death.

Clinical Criteria for Hereditary-Like UTUC

To identify patients newly presenting with UTUC, who may have an underlying hereditary aetiology, we used the risk assessment tool that we reported recently [11]. Briefly, patients were considered at risk for hereditary UTUC if they presented with at least one of the following clinical criteria: age at diagnosis of <60 years with no previous history of bladder cancer; personal history of any other HNPCC-related cancer regardless of age; and one first-degree relative with HNPCC-related cancer diagnosed before the age of 50 years or two first-degree relatives with HNPCC-related cancer regardless of age. Endometrial or ovarian cancers, cancer of the small bowel, stomach, pancreas, biliary tract, skin, brain tumours or UTUC all belong to the HNPCC-associated cancer spectrum.

Chemotherapy Regimen

Only patients who underwent adjuvant chemotherapy within 6 months of surgery were included in the current study. Cisplatin-based chemotherapy in various combinations was the most common regimen, depending on the patients' eligibility and renal function, as described previously [10]. Briefly, cisplatin was administered i.v. in a 1-h infusion, at a median dose of 75 mg/m2 of body surface area (calculated using appropriate formulas from the height and weight of the patient) every 3 weeks if renal function and serum biochemistry were acceptable (creatinine clearance ≥60 mL/min, platelets ≥100 000/mm2 and polymorphonuclear leukocytes ≥1000/mm2). Mannitol and pretreatment hydration with 1–2 L of fluid, infused for 8–12 h before the cisplatin dose, were used to reduce nephrotoxicity. Adequate hydration was maintained, and urinary output was monitored over the following 24 h. Whenever it was used in combination with paclitaxel, cisplatin was administered after paclitaxel to avoid increased haematological toxicity resulting from decreased clearance. Patients unfit for cisplatin (e.g. those with impaired renal function or poor performance status) were excluded from the current study.

Statistical Analysis

All patients were followed from diagnosis until death or until the data were censored (with the patient still considered to be alive). Overall survival (OS) and cancer-specific survival (CSS) rates were evaluated from the date of surgery to the last follow-up visit or death. Progression-free survival (PFS) was defined as the period between surgery and the subsequent appearance of recurrence or metastasis. Kaplan–Meier survival curves were calculated using censored data and compared using the log-rank test. Univariate analyses were used to assess associations between potential prognostic factors (i.e. age, gender, tumour stage, grade, type of surgery and adjuvant chemotherapy regimen) and survival. A chi-squared test was used for categorical variables, and an unpaired Student's t-test was used for age. Significant variables in the univariate analyses were entered into a multivariate analysis (Cox step-wise regression). A P value of <0.05 was considered to indicate statistical significance. All tests were conducted using spss software (version 17.0).

Results

Population

Overall, 112 patients, with a median age of 67 years, received adjuvant chemotherapy after RNU for high-risk/locally advanced UTUC. The male-to-female ratio was 1.7. According to the criteria selected, 35 patients (31.3%) were considered at risk of hereditary UTUC, while 77 patients (68.7%) were assumed to have a sporadic tumour. Patients' clinical characteristics and demographic data are shown in Table 1. Apart from median age, which was significantly younger in the hereditary ‘at-risk’ group compared with the sporadic tumour group (56 vs 69.8 years; P < 0.001), the two groups were similar.

Table 1. Demographic and clinical characteristics of the study population
VariableTotal, N = 112Sporadic UTUC, n = 77Hereditary-like UTUC, n = 35P
  1. IQR, interquartile range.
Median (IQR) follow-up,19.4 (8.5–32.7)15.1 (7.1–27.4)28.4 (18.0–44.0)0.01
Median (IQR) age67.0 (59.3–74.1)69.8 (65.5–75.5)56.0 (51.5–59.0)<0.001
Gender, n (%)    
Male70 (62.5)49 (63.6)21 (60)0.83
Female42 (37.5)28 (36.4)14 (40)
Previous bladder cancer, n (%)    
No78 (69.6)52 (67.5)26 (57.7)0.51
Yes34 (30.4)25 (32.5)9 (26.9)
Location, n (%)    
Renal pelvis51 (45.6)38 (49.4)13 (37.2)0.14
Ureteric38 (33.9)27 (35)11 (31.4)
Renal pelvis and ureteric23 (20.5)12 (15.6)11 (31.4)
Pathological stage, n (%)    
Organ-confined (≤pT2)21 (18.8)14 (18.2)7 (20)0.80
Locally advanced (>pT2)91 (81.2)63 (81.8)28 (80)
Grade, n (%)    
13 (2.7)1 (1.3)2 (5.7)0.26
27 (6.2)6 (7.8)1 (2.9)
3102 (91.1)70 (90.9)32 (91.4)
pN Stage, n (%)    
pN0/x70 (62.5)48 (62.3)22 (62.9)0.95
pN1/242 (37.5)29 (37.7)13 (37.1)
M stage, n (%)    
M0102 (91.1)69 (89.6)33 (94.3)0.72
M+10 (8.9)8 (10.4)2 (5.7)
Lymphovascular invasion, n (%)    
Negative55 (49.1)35 (45.5)20 (57.1)0.30
Positive57 (50.9)42 (54.5)15 (42.9)
Surgical margin status, n (%)    
Positive31 (27.7)21 (27.3)10 (28.6)0.88
Negative81 (72.3)56 (72.7)25 (71.4)

Survival

In the univariate analysis, lymphovascular invasion status was correlated with PFS (P = 0.04). In the multivariate analysis, locally advanced stage (>pT2) was associated with CSS (P = 0.02). In the univariate and multivariate analyses, hereditary-like tumour status was significantly associated with OS (P = 0.009 and P = 0.005) and CSS (P = 0.009 and P = 0.005) but not with PFS (P = 0.07 and P = 0.11). Full results from the univariate and multivariate analyses are shown in Table 2.

Table 2. Univariate and multivariable analyses predicting OS, CSS and PFS in patients with UTUC after RNU and adjuvant chemotherapy
CovariatesOSCSSPFS
Univariate analysisMultivariate analysisUnivariate analysisMultivariate analysisUnivariate analysisMultivariate analysis
HR (95% CI)PHR (95% CI)PHR (95% CI)PHR (95% CI)PHR (95% CI)PHR (95% CI)P
  1. Values in bold are significant P values.
  2. HR, hazard ratio.
Age, continuously coded1.02 (0.99–1.04)0.15 1.02 (0.99–1.05)0.18 1.01 (0.97–1.04)0.55 
Gender            
Female vs male0.54 (0.27–1.10)0.09 0.70 (0.33–1.45)0.34 0.56 (0.22–1.42)0.22 
Hereditary-like UTUC 0.009 0.005 0.009 0.005 0.07 0.11
Yes vs no0.36 (0.17–0.77) 0.34 (0.16–0.73) 0.32 (0.14–0.76) 0.29 (0.12–0.68) 0.44 (0.18–1.09) 0.46 (0.17–1.20) 
Previous bladder cancer           
Yes vs no0.71 (0.35–1.45)0.35 0.58 (0.26–1.31)0.19 0.65 (0.26–1.62)0.36  
Location 0.75  0.28  0.55 
Ureteric vs pelvicalyceal0.86 (0.36–2.04)0.74 0.88 (0.32–2.41)0.81 1.23 (0.47–3.20)0.67 
Multifocal (pelvicalyceal and ureteric) vs pelvicalyceal1.18 (0.58–2.41)0.64 1.68 (0.77–3.66)0.19 0.68 (2.24–1.91)0.46 
Pathological stage            
Organ-confined (≤pT2) vs locally advanced (>pT2)0.72 (0.36–1.44)0.360.48 (0.23–1.030)0.060.65 (0.31–1.35)0.250.39 (0.17–0.88)0.021.40 (0.47–4.15)0.530.84 (0.25–2.82)0.78
Grade2.45 (0.25–23.85)0.73  0.89 0.30 
2 vs 11.82 (0.24–13.36)0.43 1.71 (0.15–19.11)0.66 1.32 (0.42–4.06)0.93 
3 vs 1 0.55 1.60 (0.21–11.85)0.64 1.56 (0.56–6.98)0.94 
Surgical margin status            
R+ vs R−0.78 (0.35–1.71)0.54 0.95 (0.43–2.13)0.95  2.10 (0.92–4.82)0.071.88 (0.80–4.41)0.14
Lymphovascular invasion            
Positive vs negative1.31 (0.69–2.49)0.39 1.13 (0.56–2.29)0.71 2.45 (1.03–5.80)0.041.62 (5.66–4.69)0.36
N stage            
pN1/2 vs pN0/x1.48 (0.76–2.87)0.241.57 (0.76–3.24)0.211.52 (0.74–3.13)0.251.66 (0.75–3.69)0.22.22 (0.95–5.17)0.061.94 (0.73–5.16)0.18
M stage at diagnosis            
M+ vs M−4.54 (1.86–11.11)0.0015.25 (2.02–13.62)0.0015.89 (2.34–14.78)<0.0017.52 (2.76–20.49)<0.0010.69 (0.93–5.15)0.710.71 (0.08–6.31)0.76

The OS rate was significantly better in the group with hereditary-like tumours than in the group with sporadic tumours: 5-year OS 48.2 vs 32% (P = 0.008). CSS was also significantly better in the group with hereditary-like tumours than in the group with sporadic tumours: 5-year CSS 58 vs 35% (P = 0.006). Although there was a trend in favour of the hereditary-like tumours (i.e. 5-year PFS 71 vs 52%), the difference in PFS between the two populations was nonsignificant (P = 0.07). The Kaplan–Meier curves are shown in Fig. 1. OS and CSS were significantly better in patients with no metastases at diagnosis (P = 0.001 and P < 0.001, respectively). CSS was better in patients whose tumours had a unique pelvicalyceal location than those with a multifocal location (P = 0.02).

Figure 1.

Kaplan–Meier curves of patients with sporadic (grey) and hereditary-like UTUC (black) who all received cisplatin-based adjuvant chemotherapy after RNU. (A) OS (P<0.008). (B) CSS (P = 0.006). C, PFS (P = 0.07).

Discussion

To our knowledge, this is the first time that adjuvant chemotherapy has been shown to have a significant impact on survival in hereditary-like UTUC compared with sporadic tumours. Invasive cases of UTUC have a very poor prognosis, but patients who presented with suspicious hereditary clinical criteria survived longer than patients whose tumours lacked these features. The benefit of chemotherapy on specific survival rate was substantial, which is particularly important in a population likely to have comorbidities [13]. This is clearly in line with previous important findings in colorectal carcinomas, where tumours receive adjuvant chemotherapy contingent upon their suspected hereditary status [6]. Hereditary colorectal carcinomas are also more chemosensitive than sporadic tumours [5, 6, 14]. Recent data suggest that the impact of microsatellite instability (MSI) on benefit to fluorouracil-based adjuvant chemotherapy is dependent of the molecular mechanism involved in this genetic instability since an improved survival has been reported with adjuvant fluorouracil in MSI colorectal cancers of germline origin but not in sporadic cases [15].

Adjuvant chemotherapy yields reported recurrence-free rates of up to 50% but thus far no impact on survival has been clearly shown in UTUC [9, 10]. Further long-term data are awaited from a newly commenced prospective randomized trial, the Perioperative Chemotherapy or Surveillance in Upper Tract Urothelial Cancer (POUT) trial [16]; however, platinum-based chemotherapy is still expected to produce similar results to those observed in bladder cancer because UTUC tumours are urothelial tumours [17]. This is the reason that we focused on a population of patients who received only adjuvant cisplatin-based chemotherapy in the current study; however, the addition of a certain chemotherapy-related toxicity, particularly nephrotoxicity from platinum derivatives, to a population with already impaired post-surgical renal function may also be linked with the reduced survival in these patients [18]. Not all the patients are likely to receive adjuvant treatment because the loss of a renal unit after radical surgery can lead to impaired renal function that may reduce a patient's eligibility for chemotherapy, but one could hypothesize from our findings and from the current literature that other agents should be tested in cases of hereditary-like UTUC because irinotecan has been shown to be more efficient than ‘gold-standard’ 5-fluorouracil chemotherapy in this subpopulation of patients with colorectal carcinomas [14, 19, 20]. Moreover, the use of another type of chemotherapy regimen could be considered because the histological variants observed in UTUC are not recognized factors that would influence chemosensitivity [21].

In the present study, the proportion of patients considered to have hereditary-like tumours was ∼30%. This proportion seems abnormally high as in colorectal cancer only ∼5–10% of cases occur in patients with HNPCC. We may have over-represented young patients in the hereditary-like tumour group, meaning that our criteria for distinguishing patients between those with sporadic and hereditary cases were too broad. UTUC carcinomas usually peak in incidence among people in their 70s and 80s [22], but in cases of hereditary tumours, it appears that the median age at presentation is younger in inherited cases than in sporadic cases, as observed in our population (i.e. 56 vs 70 years). Additionally, application of the stringent clinical criteria for the diagnosis of HNPCC (Amsterdam criteria) shows that 5–10% of colorectal cancers are hereditary. From genetic studies, it has been shown that MSI screening identifies a further 5% of hereditary cancers and the incidence of de novo hMSH2 mutations is not negligible [3, 6, 7]. Consequently, some hereditary cancers, whether cancer of the colon or UTUC, are clearly misclassified as sporadic, and their incidence is underestimated [23].

The main limitation of the current study is that our results relied only on the use of our clinical assessment tool for hereditary status and there was no genetic testing (i.e. MSI status and/or germinal DNA sequencing), which was not possible to perform; however, we have previously reported that MSI status and loss of MSH2 expression were associated with survival in patients with high-risk UTUC, although in a smaller group of patients [24]. Some investigators consider that, in patients with colorectal cancer, even the results of MSH2 immunochemistry are sufficiently well correlated with the MSI phenotype to act as a surrogate for MSI determination, especially as it is quicker and cheaper [25]. In light of our current findings, we will attempt to further confirm such data in a prospective manner with appropriate samples and testing in a larger setting [26]. It appears, however, that age and family history are reliable clinical clues to potential hereditary status that should be sought by physicians when taking a medical history [11, 27]. In addition, the considerable survival difference between patients with sporadic and hereditary-like UTUC after adjuvant chemotherapy supports the hypothesis that these tumours are clearly behaving like colonic hereditary tumours. The era of personalized medicine has already arrived in the field of colorectal cancer treatment, which is best exemplified by the change of chemosensitivity according to hereditary status. It seems henceforth amenable in UTUC in the light of our data.

In conclusion, cisplatin-based adjuvant chemotherapy after RNU benefited patients with hereditary-like UTUC significantly more than those with sporadic tumours. In this very particular field where there is a lack of strong data to provide any recommendations in many clinical desperate situations, this is the first time that results can pinpoint a subset of good candidates for adjuvant chemotherapy, although the results were from retrospective data. Current findings strongly support the need for further prospective research including genetic testing, which has already been proven in colorectal carcinomas; however, the impact of genetic variability on predicting treatment response and toxicity of adjuvant chemotherapy in hereditary-like UTUC requires the clarification of additional factors.

Conflict of Interest

None declared.

Ancillary