Prognostic value of systemic inflammatory responses ...

2 downloads 0 Views 934KB Size Report
Prognostic value of systemic inflammatory responses in patients with upper urinary tract urothelial carcinoma. Myong Kim · Kyung Chul Moon · Woo Suk Choi ·.
World J Urol DOI 10.1007/s00345-015-1484-9

ORIGINAL ARTICLE

Prognostic value of systemic inflammatory responses in patients with upper urinary tract urothelial carcinoma Myong Kim · Kyung Chul Moon · Woo Suk Choi · Chang Wook Jeong · Cheol Kwak · Hyeon Hoe Kim · Ja Hyeon Ku 

Received: 20 November 2014 / Accepted: 3 January 2015 © Springer-Verlag Berlin Heidelberg 2015

Abstract  Purpose  To determine the utility of systemic inflammatory response and develop a model based on serum inflammatory indices to aid prognostication in patients with upper urinary tract urothelial carcinoma (UTUC). Methods  The study population comprised 277 patients with non-metastatic UTUC who underwent nephroureterectomy between 1999 and 2010. Multivariate survival analysis was performed using Cox’s proportional hazards model, and a new model was developed to predict recurrence-free (RFS) and disease-specific survival (DSS). The factors considered, in relation to systemic inflammatory responses, were as follows: albumin, white blood cell count, neutrophil count, lymphocyte count, red blood cell count, serum creatinine level, platelet count, prognostic nutritional index, platelet/lymphocyte ratio, neutrophil/lymphocyte ratio (NLR), and derived NLR. Results  The final model consisted of bladder cuffing, pathologic T stage, lymphovascular invasion, and derived NLR. The bootstrap-concordance indices of the model were 0.778 [95 confidence interval (CI) 0.730–0.826] for RFS and 0.802 (95 % CI 0.752–0.851) for DSS, respectively, and integrated area under the curve value for time M. Kim · W. S. Choi · C. W. Jeong · C. Kwak · H. H. Kim · J. H. Ku  Department of Urology, Seoul National University College of Medicine, Seoul, South Korea K. C. Moon  Department of Pathology, Seoul National University College of Medicine, Seoul, South Korea J. H. Ku (*)  Department of Urology, Seoul National University Hospital, 101 Daehak‑ro, Jongno‑Gu, Seoul 110‑744, South Korea e-mail: [email protected]

to RFS and DSS for patients was 0.738 and 0.760, respectively. When we generated calibration curves, the nomograms were reasonably calibrated. Conclusions  Derived NLR may be used in combination with conventional staging techniques and other clinicopathological parameters to improve the prediction of survival in patients with UTUC. Keywords  Ureter · Urothelial carcinoma · Nephroureterectomy · Inflammation · Survival · Prognosis

Introduction Upper urinary tract urothelial carcinoma (UTUC) is an uncommon disease and accounts for 5 % of urothelial carcinoma (UC) overall [1]. Nephroureterectomy is the surgical standard of care for patients with invasive, nonmetastatic, UTUC [2]. However, despite excellent local control achieved by surgery, systemic recurrences are frequent and lethal [2, 3]. The pathological T and N categories and pathologically defined tumor grade are powerful indicators of prognosis [4]. However, these variables are not enough to predict outcome. Thus, there is a need to identify new prognostic markers to provide additional prognostic information. Improved risk stratification and accurate individualized prediction of postoperative recurrence and survival outcomes can help guide patient counseling, follow-up scheduling, administration of adjuvant therapies, and design of clinical trials. Chronic inflammation plays an essential role in cancer development and progression [5]. Laboratory makers of systemic inflammation have been investigated as prognostic biomarkers in several cancers. The host systemic inflammatory response can be assessed by examining the changes

13



in the circulating concentrations of acute-phase proteins including C-reactive protein (CRP) and albumin [6]. However, acute-phase proteins are only one aspect of the systemic inflammatory response [7]. Cellular components of the systemic inflammatory response, such as neutrophils, lymphocytes, and platelets, have been reported to have prognostic value in patients with a variety of cancers [8–11]. Since neutrophils inhibit the immune system and lymphocytes have a role in host cellmediated immunity, these cellular components may reflect the host inflammatory and immune response. Neutrophil/ lymphocyte ratio (NLR) was first reported as an independent prognostic factor in colon cancer by Walsh et al. [12]. Gondo et al. [13] reported that NLR is an independent prognostic factor in bladder cancer patients treated with radical cystectomy. Moreover, Proctor et al. [14] suggested derived NLR (neutrophil count/white cell count-neutrophil count) as another independent prognostic factor, with similar predictive value to NLR. Preoperative thrombocytosis is an independent risk factor for decreased cancer-specific survival after radical treatment of invasive bladder cancer [15]. Preoperative neutrophil count is an independent predictor of recurrence-free survival in patients with non-metastatic UTUC [16]. High white blood cell (WBC) counts are of paramount prognostic significance besides tumor stage, grade, and multifocality in patients with UC of the ureter [17]. We noticed that CRP-based prognostic scores (CRP [6], Glasgow Prognostic Score [GPS] [18], modified GPS [19], and Prognostic Index [20]) could not predict outcome in patients with UTUC in our cohort [21]. The aim of this study was to determine the utility of non-CRP-based prognostic scores and develop a nomogram based on inflammatory indices to aid prognostication in patients with UTUC.

Materials and methods This was an institutional review board-approved study. The study population comprised 277 patients with non-metastatic UTUC who underwent nephroureterectomy between 1999 and 2010. We excluded patients with a history of muscle-invasive bladder cancer, those who received neoadjuvant chemotherapy or radiotherapy, those who were lost to follow-up within 6 months, those who were diagnosed with hematologic diseases or showed clinical evidence of infection or inflammatory conditions such as a fever >37.5 °C, abnormal urinalysis results with urinary symptoms, haziness on chest radiography, gastrointestinal disturbance, and a sense of local heat with tenderness. Routine laboratory measurements were performed preoperatively. The coefficient of variation for these methods, over the range of measurement, was less than 5 %

13

World J Urol

as established by routine quality control procedures. The prognostic subdivisions used for each described prognostic score are consistent with the previously published literature; a combination of albumin and lymphocyte counts in Onodera’s Prognostic Nutritional Index (PNI) [10], platelet/lymphocyte ratio (PLR) [11], NLR [12], and derived NLR [14] (Appendix 1). All surgical specimens were processed according to standard pathologic procedures. Tumors were staged according to the Union Internationale Contre le Cancer/ American Joint Cancer Committee 2004 classification. Tumor grade was assessed according to the World Health Organization’s 2004 classification. Lymphovascular invasion (LVI) was defined as the presence of tumor cells within an endothelium-lined space without underlying muscular walls. Positive margin status was defined as tumor presence at linked areas of soft tissue on the surgical specimen. Patients were generally followed every 3–4 months for the first 2 years, every 6 months from the third through the fourth year, and annually thereafter. Follow-up consisted of medical history, physical examination, blood tests, urine cytology, cystoscopic evaluation, chest radiography, and abdomino-pelvic computed tomography. Disease recurrence was defined as tumor relapse in the operative field or regional lymph nodes, and/or distant metastasis. UC in the bladder or contralateral upper tract was not considered disease recurrence and was coded as a second (metachronous) primary. Cause of death determination was performed by the clinician based on chart review and the authorized death certificate. Statistical analyses Within the model, we tested 28 predictors: gender, patient age at surgery, body mass index, year of surgery, previous or concomitant bladder cancer, bladder cuff excision, pathologic T stage, tumor grade, concomitant carcinoma in situ (CIS), LVI, tumor necrosis, margin status, tumor location, multifocality, hydronephrosis, lymph node stage, adjuvant chemotherapy, albumin, WBC count, neutrophil count, lymphocyte count, red blood count, serum creatinine level, platelet count, PNI, PLR, NLR, and derived NLR. All variables that were significant on univariate analysis were included in Cox’s proportional hazards model. Multivariate regression coefficients of the predictors were used to generate prognostic nomograms. The Harrell concordance index and time-dependent receiver operating characteristics curve were used to test the ability of the nomogram to predict the individual probability of RFS and DSS at 5 years after nephroureterectomy [22, 23]. A value of 1.0 indicates perfect predictions, whereas 0.5 is equivalent to a toss of a coin. The concordance index is unaffected by the amount of censoring [24].

World J Urol

We used bootstrapping to obtain a relatively unbiased estimate. Confidence intervals were obtained by creating 1,000 bootstrap samples from the entire data set and replicating the estimation process. In the second step of the validation process, the relationship between the model-derived and actuarial outcomes was graphically explored within calibration plots to explore nomogram performance. External validation was performed using 200 bootstrap resamples to decrease overfit bias. All statistical tests were performed with the use of R 2.13.0 (R Development Core Team, Vienna, Austria, http:// www.R-project.org) and SPSS software (SPSS, Chicago, IL). All tests were two-tailed with a significance level at 0.05.

Results The descriptive variables of the 277 evaluable patients are shown in Table 1. Median patient age was 63.7 years (range 29.5–90.0 years), and 78.7 % were men. Non-muscle-invasive UC was recorded in 42 (Ta), 1 (Tis), and 69 (T1) patients, which accounted for 40.5 %. T2–T4 disease was present in 165 (59.5 %) patients. A total of 257 patients (34.7 %) had low-grade disease, while high-grade disease was present in 181 (65.3 %) specimens. LVI within the specimen was present in 52 (18.8 %) cases, and concomitant CIS was identified in 20 (7.2 %). The median follow-up for the entire cohort was 57.2 months (range 6.8–158.3 months). Of 277 patients, 75 (27.1 %) suffered a recurrence of disease, 96 (34.7 %) died of any cause. Seventy-three of 96 deaths were attributable to UTUC. Table  2 shows the univariate and multivariate models that predict RFS after nephroureterectomy. On univariate analysis, bladder cuff, pathologic T stage, tumor grade, LVI, margin status, tumor location, multifocality, hydronephrosis, adjuvant chemotherapy, PNI, and derived NLR represented highly statistically significant predictors of RFS. A nomogram was developed to predict RFS in UTUC using the regression coefficients of the final multivariate model (Fig. 1a). The final model consisted of bladder cuff excision, pathologic T stage, LVI, and derived NLR. Table  3 shows the univariate and multivariate models that predict disease-specific survival (DSS) after nephroureterectomy. A nomogram was developed to predict DSS in UTUC using the regression coefficients of the final multivariate model (Fig. 1b). The final model consisted of bladder cuff excision, pathologic T stage, LVI, and derived NLR. When clinicopathologic parameters were compared between the high and low derived NLR groups, no parameter significantly differed, except body mass index and margin status. The high derived NLR group had lower

body mass index and higher positive surgical margin status than the low derived NLR group (Appendix 2). Kaplan– Meier survival analysis of patients with derived NLR ≥2:1 revealed a highly significant decreased RFS and DSS compared with those with derived NLR