Accuracy of diagnostic imaging as determined by delayed operative intervention for advanced neuroblastoma

Current treatment of newly diagnosed widespread neuroblastoma may include chemotherapy, delayed surgical resection, marrow ablative chemoradiotherapy, and bone marrow transplantation. Diagnostic imaging (DI) with computerized tomography (CT) or magnetic resonance imaging (MRI) has been used to determine response to therapy and timing of delayed resection. We assessed the accuracy of DI in 25 patients (26 total cases) treated over 21 months. Tumor size and location were estimated prior to surgical resection by DI, and the sensitivity and specificity of these studies were determined from operative findings. DI consisted of CT (15), MRI (8), and MRI and CT (3). Discordance between DI and operative findings was found in ten patients (38%). This included three errors of sensitivity (12%), including two false-positives and one falsenegative. Seven errors of specificity were noted; they included a positive scan with no viable tumor identified (3), much more extensive disease (3), or less extensive disease (1). Viable tumor was identified in 18 cases, and in 11 patients, complete resection of macroscopic tumor at the primary site was carried out. Ten of 13 patients operated on within 5 months of beginning chemotherapy were rendered grossly free of neuroblastoma at the primary site after surgery. Eight of 12 patients operated on 6 months or longer after starting chemotherapy were rendered grossly free of tumor at the primary site. Bone marrow transplantation was performed in 21 patients, ten of whom are alive with a median follow up of 20 months. Survival was similar for patients who underwent surgical resection at ≤5 v>6 months after starting chemotherapy. We conclude that the definitive primary tumor status cannot be assessed by DI alone because of errors in sensitivity and specificity that total 38%. Surgical exploration remains the best way to evaluate primary tumor status, and it should be performed even in the face of a negative DI study. There appears to be no benefit in delaying surgical resection for >5 months after beginning chemotherapy.