Abstract
Aim: To investigate the impact of preoperative platelet count on pathological findings at the time of Radical Cystectomy for Bladder Cancer and postoperative cancer-specific and overall survival. Patients and Methods: A total of 906 consecutive patients treated with Radical Cystectomy for Bladder Cancer between 1995 and 2012 at a tertiary referral Center were included in the study. Thrombocytosis was defined as >400,000 platelets/μl, in agreement with the standard assumed by the central laboratory of our Institution. Univariable and multivariable logistic regression analyses were used to investigate the impact of preoperative platelet count on pathological stage. Univariate and multivariate Cox regression analyses were also adopted to predict both cancer-specific and overall survival. Results: The mean age at cystectomy was 67.25 years. The mean and median platelet counts were 242,100/μl and 227,500/μl. At a mean follow-up time of 41 months, the 2- and 5-year cancer-specific and overall survival were found to be 83.1% and 75.2% and 68.3 and 59.8%, respectively. At Univariable analysis, thrombocytosis count was significantly associated with adverse pathological disease stage (p≤0.007) and lymph node invasion (p=0.05). Platelet count was significantly associated to patient survival at univariable analysis (Hazard Ratio=1.76 and 1.39 for overall survival and cancer specific survival, respectively; all p<0.05). At multivariate Cox regression analysis, platelet count was documented to be significantly related only to overall survival (Hazard Ratio=64,1.03-2.81; p=0.05). Conclusion: Preoperative platelet count should be taken into account as a factor predictive of postoperative oncological outcomes after radical cystectomy for bladder cancer and patients should be counseled accordingly.
Bladder cancer represents the fifth most commonly diagnosed tumor in the Western world and the second genitourinary cancer in incidence (1). At first diagnosis, 70% of cases are assessed as non-muscle-invasive bladder cancer and 30% as muscle-invasive bladder cancer (2). Although radical cystectomy (RC) with pelvic lymphadenectomy is still recognized as being the most effective and recommended treatment for both high-risk non-muscle invasive and muscle invasive bladder cancer, at least 50% of patients overall die within 5 years after radical surgery (3). Although the pathological disease stage according to the TNM system remains a reliable predictor, novel preoperative prognostic factors are urgently advocated in order to stratify at once the subsets of patients to be proposed for dedicated counseling and additional treatment to RC. In this direction, many biological markers have been investigated during the last 10 years. (4) However, most of them require a tumor tissue specimen and sophisticated and costly procedures for their identification, thus reducing their use in clinical practice. (5). There is increasing evidence that some preoperative modifications of hematological parameters could be related to oncological outcomes after radical surgery for many human solid tumors (6). Currently, these parameters may be quickly and easily obtained with low cost laboratory procedures. The association between preoperative thrombocytosis and tumor aggressiveness has been recognized for non-urological malignancies (7-11), however, so far, only a single report on this association in patients suffering from bladder cancer is available in literature. (12) In the present study, the role of pre-operative thrombocytosis, as predictive factor of cancer-specific (CSS) and overall survival (OS) was investigated in a large series of patients treated with RC for bladder cancer.
Patients and Methods
Study population. Between 1995 and 2012, 906 consecutive patients found with both non-muscle and muscle invasive bladder cancer who had never been treated with neo-adjuvant chemotherapy underwent RC and pelvic lymphadenectomy at our institution. For each patient, comprehensive clinical and pathological data was prospectively collected in a dedicated database approved by the Institutional Ethics Committee, with code: “2012/Vescica”.
Clinical and pathological evaluation. Preoperative clinical and pathological patient characteristics assessed before radical cystectomy included: age, gender, body mass index, American Society of Anesthesiologists score, as well as tumor stage and grade. Preoperative blood samples obtained three to seven days before RC included platelet count, leukocyte count and hemoglobin level. Elevated platelet count (thrombocytosis) was defined as >400,000 platelets/μl and normal hemoglobin level as >14 g/dl and >12 g/dl in men and in women, respectively. Pathological tumor stage was assessed according to the TNM sixth edition (13) by a dedicated team of expert uropathologists.
The follow-up was carried out according to European Association of Urology guideline recommendations for muscle-invasive bladder cancer, including complete blood count performed every three months and imaging based on computed tomography every six months for the first two years and subsequently every 12 months. Any presentation of local recurrence, distant metastases as well as CCS and OS were recorded in the patients' personal files and in the database.
Statistical analysis. Descriptive statistics were used to analyze clinical characteristics of the study. Mean, medians and interquartile range (IQR) are reported for continuously coded variables. Chi-square and t-test were used to compare, proportions and means, respectively. Univariate and multivariate logistic regression analyses were used to assess the impact of each clinicopathological parameter on thrombocytosis. Only variables available before surgery were considered. Univariate and multivariate Cox regression models were also adopted to predict OS and CSS according to pathology stage, lymph node invasion (LNI), lymph vascular invasion (LVI), thrombocytosis and remaining available covariates. Statistical significance was considered for p<0.05. Statistical analyses were performed using the SPSS v. 20.0 (IBM Corp., Armonk, NY, USA).
Results
Clinical and pathological characteristics of patients involved in this study are shown in Table I. Out of the 906 patients evaluated, the majority were male (83.2%). The median and mean age of patients were 67.25 and 68 (IQR=62-74) years. At definitive pathology after RC, in 747 (82.5%) and 159 (17.5%) patients, a pure transitional cell carcinoma and a non-urothelial or mixed carcinoma was described by pathologists, respectively. Overall, 63 patients (7.0%) were found to have preoperative thrombocytosis, according to our cut-off definition. As shown in Table II, when matching the presence of thrombocytosis with both pre-operative patient and tumor characteristics, no correlation was found from a statistical point of view.
Patents and pathological characteristcs.
Table III shows the correlation between the preoperative platelet count and both clinical and pathological outcomes after RC. A significant correlation from a statistical point of view, was documented with gender (p<0.001), pT stage (p<0.0001), LNI (p<0.0001), tumor grade (p<0.001), hemoglobin level (p<0.0001) and leukocyte count (p<0.0001). Female gender, higher T stage, LNI, tumor grade, low hemoglobin level and low leukocyte count are associated to higher levels of platelet count.
Pre-operative parameters and thrombocytosis.
As shown in Table IV, at univariate analysis both the pathological stage and LNI were statistically significantly correlated with the presence of preoperative thrombocytosis. At a mean of 41 months follow-up, CSS and OS at 2 and 5 years were 81 and 73%, and 68 and 59%, respectively.
At univariate Cox regression analysis (Table V), CSS was found to be significantly negatively influenced by preoperative thrombocytosis (RR=1.39, CI=1.03-1.92; p=0.05) with an expected 5-year CSS of 59% and 51% for patients without and with thrombocytosis, respectively (p<0.05).
CSS also correlated with hemoglobin level (RR=0.78, CI=0.65-0.93; p<0.007), pT stage (RR=1.25, CI=1.05-1.48; p=0.01) and LNI (RR=1.57, CI=1.29-1.90; p<0.01), at univariate analysis.
At multivariate analysis a significant statistical correlation could only be seen between thrombocytosis and OS (RR=1.64, CI=1.03-2.81; p=0.05), as presented in Table VI.
Figure 1, shows graphically OS according to thrombocytosis status.
Discussion
Previous studies have suggested that different preoperative hematological parameters may be associated with both adverse pathological findings at the time of radical cystectomy and postoperative CSS in patients with bladder cancer. However, the prognostic predictive value of preoperative thrombocytosis in patients submitted to RC has only been sporadically investigated and never analyzed in a large series of patients as far as we are aware. The reason explaining cancer-related hematological alterations are still under debate. So far, the increased secretion of different cytokines and growth factors from tumor cells is advocated as the most accredited mechanism (7).
Univariate analysis of the associaton between clinicopathological parameters and platelet count.
The role of cytokine activity in many kinds of solid human cancers (including transitional carcinoma) growth and progression is well-known. Many pre-clinical and clinical studies supported the influence of different interleukins (namely IL2 and IL6) and tumor growth factors (mainly vascular endothelial growth factor)(14) on cancer-related hematological disorders (14-16), including the alteration of hemoglobin level (17), elevation of C reactive protein (18) and the stimulation of platelet production (7). There is also evidence that circulating platelet–tumor cell aggregates may favor cancer metastasis formation (19). Although the alterations of thrombopoiesis have been correlated with the prognosis of many types of human solid tumors, such as lung, endometrial, esophageal, breast and colorectal cancer (8-11), surprisingly, the prognostic role of alterations to thrombopoiesis still remains almost totally uninvestigated for bladder cancer. A single report on this topic, supporting the prognostic role of thrombocytosis in muscle-invasive bladder cancer, has recently been published by Todenhöfer et al. (12). In this context, our study aimed to assess the correlation between thrombocytosis and both clinical and oncological outcomes of a large series of patients submitted to RC for bladder cancer at a tertiary referral center.
Clinicopathological parameters and thrombocytosis.
For the purpose of the study, thrombocytosis was defined as >400,000 plateles/μl, in agreement with the standard established by the Central Institutional Laboratory and with many reports in literature defining thrombocytosis as the total count of platelets within a range of 370,000/μl and 450,000/μl (10-12). According to our definition, 63 patients (7.0%) were found with thrombocytosis before RC. Todenhöfer et al., using a cut-off of 450,000/μl to define thrombocytosis, documented a rate of 10.1% in their bladder cancer series. Our reduced rate may be related to the higher proportion (24.2%) of patients suffering from high-risk non-muscle invasive bladder cancer. This rate of bladder cancer-related thrombocytosis is in line with that documented for colorectal cancer (12.1%) (11) but clearly lower when compared to those associated with other kinds of tumors such as ovarian (31%) (20) or renal cancer (20%-57%) (21-22). When the presence of thrombocytosis was correlated with all preoperative parameters (including gender, body mass index, age, clinical tumor stage and number of transurethral resections before RC), a statistical significance was only documented for gender (p=0.04). At univariate analysis, platelet count correlated with postoperative pathological tumor stage, LNI, hemoglobin level and leukocyte count. Differently from that described by Todenhöfer et al., a correlation between thrombocytosis and non-urothelial cancer was not found in our analysis, probably due to our higher rate (17.5%) of rare histotypes or mixed tumor. At univariate Cox regression analysis, CSS and OS were significantly correlated with many predictive factors: T stage, LNI, hemoglobin level and thrombocytosis. However, at multivariate Cox regression analysis, a statistically significant correlation way only seen between thrombocytosis and OS. These results are in agreement with those of the reports in literature concerning non-urothelial tumors (8-11).
Univariate Cox Regression predictng CSS and OS.
Multvariate Cox Regression predictng CSS and OS.
Overall survival following radical cystectomy (RC) for bladder cancer, stratified by presence of preoperative thrombocytosis.
This study suffers from the limitations of every retrospective investigation. In addition, we analyzed only patients who underwent RC for bladder cancer, possibly introducing a bias in patient selection. Moreover, although the series of patients included in the study was large, the proportion of patients actually found to have thrombocytosis (7%) was low. Nevertheless, the predictive value of adverse oncological outcomes after RC of thrombocytosis shown by this study is clear and supports its possible advantageous use in clinical practice for the preoperative discrimination of different risk categories among patients with bladder cancer.
Conclusion
Reduced OS and CSS in patients treated for bladder cancer by radical cystectomy was documented in the presence of a preoperative thrombocytosis status. The preoperative platelet count should be taken into account, together with other predictive factors, when counseling patients who are candidates for RC for bladder cancer. Further studies with larger sample sizes are advocated to validate the current results.
- Received February 8, 2014.
- Revision received April 16, 2014.
- Accepted April 22, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
