A logistic regression model for predicting malignant pheochromocytomas

Abstract

It is well known that no single histological feature is diagnostic for malignant pheochromocytomas (PCCs). So we developed a logistic model based on a series of clinical and pathological features to predict malignance in PCCs, and evaluated its diagnostic performance. In all 130 cases with malignant or benign PCCs, 15 predictive variables were observed and entered in the logistic regression analysis in a backward stepwise way. The diagnostic performance of this logistic model was evaluated by calculating the area under the receiver operating characteristic (ROC) curve. In logistic analysis, of the 15 variables entered in the logistic regression analysis, 9 were retained in the model. High cellularity had the highest odds ratio (OR), followed by spindle cell (>10% of tumor volume), atypical mitotic figure, periadrenal adipose tissue invasion, mitotic figures [>3/10 high-power field (HPF)], cellular monotony, capsular invasion, vascular invasion, and central or confluent tumor necrosis. High cellularity, spindle cell (>10% of tumor volume) and atypical mitotic figure were selected to built a logistic model. This model had the area under the ROC curve of 0.927 (95% confidence interval 0.883–0.971). The application of the model can benefit the clinical management decision for patients with PCCs. We still emphasize, however, that a clinical prospective evaluation is needed to confirm its actual value.