Prognostic significance of sarcopenia in upper tract urothelial carcinoma patients treated with radical nephroureterectomy

Abstract We investigated the prognostic significance of sarcopenia in upper tract urothelial carcinoma (UTUC) patients treated with radical nephroureterectomy (RNU). The skeletal muscle index (SMI) was calculated from computed tomography images. Sarcopenia was defined as SMI <43 cm2/m2 for males with body mass index <25 kg/m2, <53 cm2/m2 for males with BMI ≥25 kg/m2, and <41 cm2/m2 for females. Associations of sarcopenia with cancer‐specific survival (CSS) and overall survival (OS) were evaluated in 81 consecutive UTUC patients who underwent RNU. Forty‐seven (58%) out of 81 patients had sarcopenia. Multivariate analyses identified sarcopenia as a significant and independent poor prognostic factor for both CSS (hazard ratio [HR] 8.58, 95% confidence interval [CI]: 1.63–158.1, P = 0.008) and OS (HR 6.05, 95%CI 2.00–26.21, P < 0.001). In patients with locally advanced disease (pT3/4 or pN+), those with sarcopenia showed significantly worse CSS and OS than those without (5‐year CSS rate 55% vs. 100%, P = 0.014; 5‐year OS rate 40% vs. 86%, P = 0.007). In contrast, no prognostic difference was observed between patients with and without sarcopenia in those with organ‐confined disease (pTa‐2pN0/x). Sarcopenia is an independent poor prognostic factor for UTUC patients treated with RNU, particularly for those with locally advanced disease.


Introduction
The incidence of upper urinary tract carcinoma (UTUC) remains low, but has increased to two cases per 100,000 person-years in the United States [1]. Although radical nephroureterectomy (RNU) is the standard of care for nonmetastatic UTUC patients, their prognosis after RNU is relatively poor; 20-30% develop metastatic recurrence and eventually die of the disease [2,3]. Thus, the identification of high-risk patients as candidates for adjunctive therapies is essential in order to improve oncological outcomes. Previous studies investigated various prognostic factors in an attempt to accurately stratify the risk for mortality after RNU. Several pathological parameters, including the pT and pN stages, tumor grade, and lymphovascular invasion (LVI) have been established as prognostic factors in UTUC patients treated with RNU [4,5]. However, a relatively small number of studies have examined clinical parameters associated with the prognosis of these patients.
Regarding urological cancers, we recently reported the prognostic significance of sarcopenia in metastatic renal cell carcinoma patients [9] and advanced urothelial carcinoma patients for whom curative surgery was not

ORIGINAL RESEARCH
Prognostic significance of sarcopenia in upper tract urothelial carcinoma patients treated with radical nephroureterectomy

Cancer Medicine
Open Access H. Fukushima et al. Prognostic Impact of Sarcopenia in UTUC indicated [10]. The prognostic impact of sarcopenia was also demonstrated in bladder cancer patients treated with radical cystectomy [12]. In this study, we investigated the prognostic role of sarcopenia in operable UTUC patients treated with RNU.

Patients
This study protocol (approval number 1420) was approved by the Institutional Ethical Committee. This retrospective study consisted of 85 consecutive Japanese UTUC patients who underwent RNU at a single cancer center between November 2001 and February 2015. Four patients were excluded (one due to the lack of pathological findings and three due to less than 3 months follow-up) and the remaining 81 patients were subjects for analysis. All patients underwent open or gasless retroperitoneoscopic RNU [14] for nonmetastatic UTUC. A bladder cuff was excised by the extravesical approach. The regional lymph nodes were dissected in patients with swollen nodes detected radiologically or intraoperatively. All patients were preoperatively assessed by computed tomography (CT) within 1 month prior to RNU. Variables included age at RNU, gender, body mass index (BMI), Eastern Cooperative Oncology Group performance status (ECOG PS), tumor location, preoperative C-reactive protein (CRP), tumor grade, pT and pN stages, LVI, and neoadjuvant and/or adjuvant therapy. BMI was calculated as follows: BMI (kg/ m 2 ) = ((weight)/(height) 2 ). BMI was categorized using a cut-off value of 25 kg/m 2 according to the WHO criteria of overweight/obesity specific for Asia-Pacific populations [15]. The cut-off point of CRP was determined according to a previous study [16]. Tumors were staged and graded in accordance with the 2002 TNM classification and 1973 WHO grading system, respectively. LVI was defined as the existence of tumor cells within an endothelium-lined space with no underlying muscular walls.

Image analysis
Computed Tomograpy images taken within 1 month prior to RNU were analyzed as described previously [9,10]. Briefly, the third lumbar vertebra (L3) was set as a landmark, and the cross-sectional areas of skeletal muscle were measured in two consecutive slices of axial CT images using Hounsfield unit thresholds of −29 to +150. The mean value was calculated for each patient and normalized for stature: skeletal muscle index (SMI) (cm 2 / m 2 ) = ((skeletal muscle cross-sectional area at L3)/ (height) 2 ). SMI was considered as an indicator of wholebody muscle mass because the total lumbar-skeletal muscle cross-sectional area linearly correlates with whole-body muscle mass [17]. OsiriX Imaging Software (Pixmeo, Geneva, Switzerland) was used for analyses. All image analyses were performed by one investigator (H.F.) who was blinded to other clinicopathological parameters and outcomes.
The cut-off values of SMI to best discriminate cancerspecific survival (CSS) were explored for the definition of sarcopenia as described previously [9,10]. Among the various cut-off values reported, a univariate Cox proportional model yielded the highest hazard ratio (HR) (9.18, P = 0.005) when the cut-off values of SMI incorporating BMI, proposed by Martin et al. [7], were utilized. Thus, sarcopenia was defined as SMI <43 cm 2 /m 2 for males with BMI <25 kg/m 2 , <53 cm 2 /m 2 for males with BMI ≥25 kg/m 2 , and <41 cm 2 /m 2 for females in this study [7].

Statistical analysis
Differences in the distribution of variables between groups were evaluated using the chi-squared test for categorical variables and Wilcoxon rank-sum test for continuous variables. CSS was calculated from the date of RNU to death due to UTUC or the last follow-up. overall survival (OS) was calculated from the date of RNU to death or the last follow-up. Survival curves were generated using the Kaplan-Meier method and compared using the log-rank test. The Cox proportional hazards model was used to evaluate the relationships between variables and survival. Variables with P < 0.10 in the univariate analysis were included in the multivariate analysis. A reduced multivariate model was developed using the stepwise backward method, in which the variable with the highest P-value was eliminated from each iteration of the multivariate analysis. All statistical analyses were performed using JMP 9.0.2 (SAS Institute Inc., Cary, NC). Significance required two-tailed P < 0.05.

Relationships between sarcopenia and clinicopathological variables
Relationships between sarcopenia and clinicopathological variables are also shown in Table 1. Sarcopenia was associated with higher pT stage significantly (P = 0.043) and positive LVI with marginal significance (P = 0.063) but not with ECOG PS (P = 0. 84) or CRP (P = 0.26).

Discussion
This study is the first to demonstrate the prognostic significance of sarcopenia in UTUC patients treated with RNU. Sarcopenia was a strong and independent prognostic factor among the various clinicopathological parameters examined. The prognostic significance of sarcopenia was evident in patients with locally advanced disease (pT3/4 or pN+) but not in those with organ-confined disease (pTa-2pN0/x). This result is conceivable considering that sarcopenia has been identified as an unfavorable prognostic factor in advanced urothelial carcinoma patients [10].
However, the reason why sarcopenia has a prognostic impact only for locally advanced disease is currently unclear. The development of sarcopenia involves both tumor and host factors. Indeed, sarcopenia was observed more frequently in patients with locally advanced disease than those with organ-confined disease (68% vs. 46%, P = 0.043). Our results suggest that, as a tumor factor, locally advanced disease that induces sarcopenia is aggressive; and, as a host factor, patients with sarcopenia are prone to or those without are unlikely to die from locally advanced UTUC. Sarcopenia was not a prognostic factor for organ-confined disease presumably because of its less aggressive nature.
Sarcopenia clearly stratified the prognosis of locally advanced UTUC patients treated with RNU. Patients without sarcopenia appear to show favorable CSS even though they have locally advanced disease. In contrast, the prognosis of sarcopenic patients with locally advanced disease was extremely poor. These patients might be candidates for adjuvant therapies. However, cautions should be paid because patients with sarcopenia are reportedly susceptible to the adverse events of anticancer therapies [18]. Thus, sarcopenia may be helpful in counseling patients on their personalized risk to determine therapeutic strategies including adjuvant therapies and clinical trials.
Cancer cachexia, characterized by fatigue, loss of appetite, weight loss, and frailty, involves host inflammatory responses to tumors [6,[19][20][21]. Several studies have demonstrated the prognostic roles of cancer cachexia-related factors in UTUC patients treated with RNU. CRP, a nonspecific systemic inflammatory marker, has been identified as a useful biomarker for predicting a poor prognosis in UTUC patients treated with RNU [16,22]. However, the clinical usefulness of CRP as a cancer biomarker is limited by its nonspecific elevation due to noncancerrelated conditions including infection, autoimmune disease, and cardiovascular disease. This study did not detect a significant relationship between CRP and survival, possibly because we did not exclude patients with nonspecific CRP elevations. Based on the prognostic significance of sarcopenia beyond CRP demonstrated in this study and the noncancer-specific nature of CRP, sarcopenia appears to surpass CRP as a prognostic factor associated with host inflammatory responses to tumors in UTUC patients treated with RNU.
Body weight and BMI decrease with the development of cancer cachexia. In this study, BMI was not significantly associated with CSS and OS. Similarly, the prognostic roles of decreased BMI in UTUC patients have been inconclusive in previous studies [23,24]. This may partly be because BMI does not reflect body composition, including the proportion of fat to muscle tissue and the degree of fluid accumulation such as ascites and edema [8]. Moreover, the contemporary trend of increasing obesity obscures the clinical significance of weight loss [7].
Sarcopenia has several strengths in its clinical use, with the most advantageous being that it is easy and simple to evaluate with no additional cost. CT scanning, which is universally available in clinical practice, is required for the diagnosis, staging, and follow-up of UTUC patients. The calculation of SMI is easy and simple using free software and CT images without special training. Furthermore, sarcopenia is a highly objective measure because it is possible to estimate the composition of the human body using CT scans with a reported precision error of 1.4% [25]. Sarcopenia is also stable; rapid fluctuations in skeletal muscle volumes in a short interval are unlikely to occur. The objectivity and stability of sarcopenia appears to be advantageous for correctly and precisely predicting patient prognoses.
This study has several limitations. It is limited by its inherent retrospective nature and relatively small patient cohort at a single institution. Thus, our results need to be validated in larger multicenter cohorts. In this study, established prognostic factors including tumor grade, pT stage, and LVI, were not identified as independent prognostic factors in multivariate analysis probably because of a small sample size. In addition, a significant association between pT stage and sarcopenia attenuated the prognostic significance of pT stage in multivariate analysis including sarcopenia. Indeed, pT stage was an independent predictor for both CSS and OS when excluding sarcopenia from multivariate analysis (data not shown). Furthermore, our cohort is heterogeneous in terms of RNU-related adjunctive therapies. However, sarcopenia was still a significant predictor of worse CSS and OS when analyzed in patients who did not receive neoadjuvant therapy (data not shown). Regional lymph node dissection was not performed in most patients. In survival analyses, patients with pNx and pN0 were grouped together because CSS and OS curves were similar between those with pNx and pN0. In addition, weight loss, one of the symptoms of cancer cachexia, and systemic inflammation-related parameters (e.g., the neutrophil-to-lymphocyte ratio, fibrinogen, and interleukin-6) were not evaluated in this study.
Moreover, the BMI-incorporated cut-off values of SMI based on a large Canadian patient cohort were used to define sarcopenia in this study [7]. Since these cut-off values have not been validated in a Japanese population, future studies may find more appropriate definitions of sarcopenia for Japanese patients. However, the prognostic significance of sarcopenia has been demonstrated in Japanese patients with various urological cancers using BMI-incorporated cut-off values [9,10], indicating that this definition is optimal. Finally, we did not investigate the role of salvage chemotherapy in the management of sarcopenic UTUC patients developing local and/or distant recurrence after RNU. Sarcopenic patients may be unlikely to receive salvage chemotherapy at adequate doses due to poor performance status and severe adverse events, which might lead to worse clinical outcomes. Future studies are required to evaluate this point.

Conclusions
We demonstrated for the first time that sarcopenia is a strong and independent prognostic factor in UTUC patients