Eur J Clin Invest 2011; 41 (5): 539–545
Background Temozolomide (TMZ) is commonly used for the therapy of malignant glioma and induces thrombocytopenia in a fraction of patients. Currently, no biomarkers predicting TMZ-induced thrombocytopenia are available. In this study, we investigated whether changes in platelet count (PLT) or the immature platelet fraction (IPF) may serve as predictor of TMZ-induced thrombocytopenia in malignant glioma patients.
Methods We prospectively included 52 malignant glioma patients receiving TMZ-containing therapy regimens in this study. Platelet counts and IPF were determined at each clinical follow-up visit (weekly during concomitant radiochemotherapy or at least monthly during TMZ monotherapy) using the Sysmex XE-2100 system. We explored the diagnostic utility of PLT change/day and IPF change/day from the last to the current follow-up visit for the prediction of clinically relevant thrombocytopenia (PLT < 100·000 μl−1) at the next follow-up visit.
Results Relevant thrombocytopenia was observed in 10 of 234 occasions. The areas under the receiver operating characteristic curves for PLT absolute change/day, PLT relative change/day and IPF relative change/day were 0·675, 0·703 and 0·663, respectively. The Youden indices (maximum sum of sensitivity and specificity minus one) were 0·31, 0·39, and 0·29, respectively. The corresponding positive predictive values were 16%, 57%, and 6·7%, and the negative predictive values were 97%, 97%, and 98%, respectively.
Conclusions The rather moderate diagnostic potential of our data indicate that the time course of PLT counts and IPF measured at routine clinical follow-up are not useful for the prediction of thrombocytopenia in glioma patients treated with TMZ.