Pretreatment parameters obtained from peripheral blood sample predicts invasiveness of bladder carcinoma

Abstract

Purpose: To predict the invasiveness of urothelial bladder carcinoma using a logistic regression model on preoperative peripheral blood samples. Patients and Methods: Hospital data of patients operated for urothelial carcinoma were reviewed retrospectively. Preoperative blood samples were collected before the first cystoscopic examination. Any kind of infection or inflammation was an exclusion criterion. Patients were grouped as having a non-muscle-invasive or muscle-invasive urothelial carcinoma. The mean age was 69 years and was determined as the cut-off value. According to receiver operating characteristic curves, threshold points were determined for lymphocytes, neutrophil count, neutrophil-to-lymphocyte ratio (NLR), thrombocytes and mean platelet volume. Demographic specialties, parameters obtained from blood samples, tumor size and multiplicity were evaluated and significant parameters were put into a logistic regression model. Results: The study group consisted of 80 non-muscle-invasive and 102 muscle-invasive patients. Age (≤69 vs. >69), female gender, NLR (2.57), mean platelet volume (7.9/fl) and platelet count (400,000/μl) were significant parameters and put in a model. Using odds ratios, the probability of tumor invasiveness was calculated by a formula. Conclusion: Age, female gender, NLR and platelet count were found to be the predictors of invasiveness of urothelial carcinoma. Copyright © 2012 S. Karger AG, Basel.