Association of image-defined risk factors, tumor resectability, and prognosis in children with localized neuroblastoma

Background Although localized neuroblastoma has a good prognosis, some cases have undergone treatment failure or recurrence. Apart from biologic features such as MYCN status, we wondered whether some characteristics of growing tumors are prognostic, such as a well-encapsulated mass without infiltration of vital organs. We analyzed the diagnostic utility of image-defined risk factors (IDRFs) to predict successful treatment and prognosis. The overall goal was to achieve maximum cure rates for patients with localized neuroblastoma through a better understanding of clinical characteristics. Methods We retrospectively reviewed the images of patients with localized neuroblastoma who were enrolled between June 1998 and December 2012 at a single institution in Shanghai, China. Unequivocal categorization regarding IDRFs was available in 67 patients. IDRF was assessed at diagnosis and after four cycles of neoadjuvant chemotherapy, on average. The median follow-up period was 84 months (range: 48-132 months) after diagnosis. Results MRI and CT indicated a total of 177 IDRFs in these 67 patients. Logistic regression analysis revealed a highly significant negative correlation between the numbers of IDRFs and the possibility of complete removal of neuroblastoma. Intraspinal extension of the tumor, compression of the trachea, and encasement of the main artery in localized neuroblastoma were predictors for incomplete tumor resection. According to univariate analysis, >= 4 IDRFs and intraspinal extension of the tumor were significant indicators of poor prognosis. Conclusions The number of IDRFs was useful in predicting surgical outcome and event-free survival. The number of IDRFs should be considered in protocol planning, instead of IDRF presence or absence.