Combined Scintigraphy and Tumor Marker Analysis Predicts Unfavorable Histopathology of Neuroblastic Tumors with High Accuracy

OBJECTIVES: Our aim was to improve the prediction of unfavorable histopathology (UH) in neuroblastic tumors through combined imaging and biochemical parameters. METHODS: (123)I-MIBG SPECT and MRI was performed before surgical resection or biopsy in 47 consecutive pediatric patients with neuroblastic tumor. Semi-quantitative tumor-to-liver count-rate ratio (TLCRR), MRI tumor size and margins, urine catecholamine and NSE blood levels of neuron specific enolase (NSE) were recorded. Accuracy of single and combined variables for prediction of UH was tested by ROC analysis with Bonferroni correction. RESULTS: 34 of 47 patients had UH based on the International Neuroblastoma Pathology Classification (INPC). TLCRR and serum NSE both predicted UH with moderate accuracy. Optimal cut-off for TLCRR was 2.0, resulting in 68% sensitivity and 100% specificity (AUC-ROC 0.86, p < 0.001). Optimal cut-off for NSE was 25.8 ng/ml, resulting in 74% sensitivity and 85% specificity (AUC-ROC 0.81, p = 0.001). Combination of TLCRR/NSE criteria reduced false negative findings from 11/9 to only five, with improved sensitivity and specificity of 85% (AUC-ROC 0.85, p < 0.001). CONCLUSION: Strong (123)I-MIBG uptake and high serum level of NSE were each predictive of UH. Combined analysis of both parameters improved the prediction of UH in patients with neuroblastic tumor. MRI parameters and urine catecholamine levels did not predict UH.