The role of primary tumor resection in neuroblastoma: When and how much?

Neuroblastoma management is a central problem in pediatric oncology, as the disease remains both challenging and enigmatic. Although great advances have been made in understanding and treating this tumor, the role of surgical resection remains controversial and the debate is hampered by a lack of prospective trials with surgical variables as the specific aim. Controversy centers around two questions: (1) the use of neoadjuvant chemotherapy for localized tumors and (2) the required extent of resection for local and metastatic disease. Currently, patients with metastatic tumors or tumors with MYCN amplification are treated with neoadjuvant chemotherapy. Therefore, the issue of surgical timing is most relevant to patients with localized tumors. The importance of complete versus partial removal of the primary tumor and involved regional lymph nodes is germane to both localized and metastatic neuroblastoma. Cechetto et al., in a 2005 report from the European International Society of Pediatric Oncology Neuroblastoma Group's (SIOPEN's) LNESG1 study, defined a system of preoperatively determined surgical risk factors,1 now known as image-defined risk factors (IDRFs), that could be used to predict surgical efficacy and the frequency of surgical complications.2 In this issue of Pediatric Blood and Cancer, Monclair et al.3 present a more recent analysis of patients from the LNESG1 study, investigating the effect of IDRFs in those who presented with localized neuroblastoma. It is important to note that this study was specifically designed to examine surgical efficacy in patients with localized tumors. The assignment of patients to the L1 (IDRFs absent) or L2 (IDRFs present) category was performed retrospectively and was not a part of the initial study question. Of 905 patients originally enrolled in LNESG1, the authors identified 534 patients comprising three groups: (1) L1 patients treated with surgery alone (n = 291; EFS = 92%, OS = 98%), (2) L2 patients treated with surgery alone (n = 118; EFS = 86%, OS = 95%), and (3) L2 patients who received initial chemotherapy (n = 125; EFS = 73%, OS = 83%). A higher incidence of operative complications occurred in the L2 group compared to L1 patients, but no data were available on surgical morbidity after chemotherapy. Only 18% of INSS stage 3 patients in this study were treated with surgery alone, suggesting a correlation between the number and complexity of IDRF and use of neoadjuvant chemotherapy. The authors conclude that primary surgery for L1 non-MYCN-amplified neuroblastoma is sufficient. The real issue is the most appropriate management of L2 patients and, here, the number and type of IDRFs are important. Low numbers of IDRFs and those not encasing larger vessels may also be amenable to resection. Multivessel encasement or a more complex IDRF with nerve or airway infiltration may benefit from initial chemotherapy. This concept was applied to patients receiving neoadjuvant chemotherapy by Irtan et al., whose report is also published in this issue.4 By assessing the change in the number of IDRFs with systemic therapy, they showed that the preoperative number correlated with extent of resection and operative complications and demonstrated that the median number of IDRFs per tumor fell from four to two. The second and perhaps harder question is determining the necessary extent of tumor removal. For localized disease, there is evidence that some tumors can safely be observed,5, 6 and this management strategy is the subject of present protocol development. Subtotal resection of localized neuroblastoma may also be adequate and a protocol addressing this issue has been completed but not yet published. IDRF data might be useful in guiding such resections. For metastatic or MYCN-amplified tumors, the literature is less clear.7, 8 In this issue, Englum et al.9 present their study of 87 children from 14 institutions, in which they found an overall survival advantage for a >90% resection but no improvement after gross total resection. Curiously, patients undergoing gross total resection (GTR) actually had a worse overall survival compared to those with >90% resection. The extent of resection was surgically determined without imaging corroboration and the authors do not define >90% resection other than as a procedure in which no gross tumor was left. Also, 65% of GTR patients compared to 42% of those with less than a GTR had MYCN amplification. This illustrates some of the problems with studies assessing surgical efficacy in high-risk patients, namely: primary study aims other than resection, and a lack of central imaging review. Recent prospective data from both the Children's Oncology Group 10 and SIOPEN,11 presented at the 2014 Advances in Neuroblastoma Research meeting in Cologne, support more complete resection in high-risk patients. Incorporation of post-resection imaging and categorization analogous to preoperative IDRFs in prospective studies would assess surgical efficacy more accurately.