Complete surgical resection improves outcome in INRG high-risk patients with localized neuroblastoma older than 18 months

Neuroblastoma is the most common solid extracranial malignancy of childhood and the most common malignant tumor in infants [1]. Despite the advances in multidrug therapy, surgery still plays a major role in treatment of neuroblastoma (NB) [2]. Children presenting with localized disease have an overall better prognosis, mainly depending on the degree of tumor resection [3]. Indeed, completely resected tumors (INSS stage 1) rarely relapse and do not require postoperative chemotherapy [4, 5]. There is, however, still a controversy regarding the value of radical surgery for extended local disease [2]. Biological and clinical prognostic markers may help to stratify risks and guide therapy in these patients, but prospective randomized trials with surgical endpoints are still missing. In 2008, the INRG group implemented a pretreatment risk stratification considering age, MYCN amplification, histological tumor grade, Image Defined Risk Factors and 11q deletion to classify NB patients into very low-, low-, intermediate- and high-risk patients [6]. In this study, we analyzed the impact of the extent of tumor resection on outcome of patients older than 18 months with localized non-metastatic NB aged, who were treated according to the German prospective clinical trial NB97.

All patients registered by the German prospective clinical trial NB97 between November 1, 1996, and September 30, 2004, were included in this analysis when they met the following criteria: stage 1–3 neuroblastoma diagnosed according to the International Neuroblastoma Staging System criteria [7]; age at diagnosis >18 months but less than 21 years. Written informed consent was obtained from patients or their guardians for participation in the study design, data collection and treatment (Registration number: NCT00017225, ClinicalTrials.​gov).

The NB97 trial was a randomized trial comparing ASCT (autologous stem cell transplantation) and oral maintenance chemotherapy in high-risk patients. According to the NB97 protocol, the patients with localized NB were prospectively stratified in NB97 high-risk, NB97 standard-risk or NB97 low-risk subgroups according to INSS stage, age, MYCN-Status and threatening symptoms. NB97 standard-risk patients received 4 cycles of chemotherapy after first surgery and a second resection when necessary whereas NB97 low-risk patients were observed for up to 12 months with examinations and staging every 6 weeks. NB97 high-risk patients were randomized and received ASCT or oral maintenance therapy. The trial protocol had been evaluated by the Institutional Ethical Boards of the University of Cologne and participating hospitals. All patients participated in the trial after informed consent and on voluntary basis. Trial protocol and results of the primary trial end point have been published before [8, 9]. The NB97 protocol also provided clear recommendations on timing and extent of tumor resection. For example, initial or delayed complete resection was advised when no vascular structures or adjacent organs were involved. Incomplete resection was acceptable to reduce the risk of acute complications and long-term organ impairments. Nephrectomy or insertion of vascular prostheses was discouraged. Therefore, complete initial resection was reserved only for patients with well-encapsulated primary tumors. All other patients were recommended to undergo second look operations after four to six cycles of induction chemotherapy. Moreover, radiation therapy of 40 Gy was advised for patients with unresectable residuals of the primary tumor present after induction chemotherapy. Finally, radiation doses less than 40 Gy were applied to protect adjacent structures with low radiation tolerance [10]. Data on extent and complications of resection were collected prospectively using case report forms.

For this analysis, imaging reports, operation notes and pathology reports were retrospectively reviewed by two independent experienced physicians, and discrepant results were clarified after repeated joint review of the patients’ files. In this study, we distinguished between two types of operations, as described before [11]. Briefly, first operation was the tumor operation performed before or within the first six cycles of induction chemotherapy, and best operation was the most extensive removal of primary tumor tissue done at any time during first-line therapy. For outcome analysis, the extent of resection was classified as follows: no operation or biopsy removing less than 50% of tumor tissue; incomplete resection of 50% to less than 90% of tumor volume present at the time of surgery; gross total resection removing more than 90% of the tumor; or complete resection without macroscopic postoperative tumor residuals. It was not possible to include the category of microscopic complete resection because neuroblastomas were rarely removed in toto, and therefore, the pathologist often received several tumor fragments, making confirmation of microscopic complete resection impossible. IDRF were not established in 1997, and therefore, they had to be assessed retrospectively by review of initial imaging and operation notes based on the definitions published by the European International Society of Pediatric Oncology Neuroblastoma Group [12] and the International Neuroblastoma Risk Group [13, 14]. During induction chemotherapy, re-staging, including magnetic resonance imaging of the primary tumor site, bone marrow assessment, and metaiodobenzylguanidine (MIBG) scintigraphy, was scheduled after four to six chemotherapy cycles. If progression or relapse was suspected, complete staging was necessary. Disease status and response to treatment were categorized as complete remission, very good partial remission, partial remission, stable disease, or progression according to the published International Neuroblastoma Response Criteria [7]. Pretreatment Risk Stratification was performed retrospectively according to INRG staging system [6] considering age, MYCN amplification, histological grade, IDRF and 11q deletion stratifying all patients into INRG very low-, INRG low-, INRG intermediate- and INRG high-risk. Histology was defined by INPC guidelines as described before [15].

In our cohort, extended surgery of the primary tumor site improved local control rate and outcome in patients older than 18 months with localized neuroblastoma. After complete resection during treatment the patients had a better local-progression-free survival (LPFS), event-free survival (EFS) and OS than the other groups. The wide range of clinical manifestations of neuroblastomas, their localization and their resectability leads to a high heterogeneity of our cohort. The NB97 protocol clearly recommended less aggressive surgical approach particularly during first operation in an effort to avoid surgical complications. Most surgeons choose a less aggressive approach when the disease was considered to be unresectable during operation. Thus, the question whether patients did not undergo complete resection because of less aggressive surgical approaches, or simply due to unresectable disease cannot be answered retrospectively. Moreover, less aggressive surgeons are more likely to classify patients as unresectable and vice versa. Prospective studies with a statement of the surgeon’s decision in his/hers operation notes are needed to address different approaches.

However, subgroup analyses showed better EFS and OS for patients with complete resection in INSS stage 3 and MYCN amplification. Patients with MYCN amplification belong to the INRG high-risk stratification and are therefore of special interest since the role of surgery in high-risk patients remains controversial. Other studies of high-risk patients especially evaluating stage 4 patients show no advantage of complete resection on patient’s outcome [11, 16], while lower local recurrence rates and better OS are reported after extended surgery by La Quaglia et al. [17, 18]. But analyzing localized neuroblastoma in patients >1 year, radical surgery is recommended which is in line with our findings [2]. A meta-analysis of 2599 patients with stage 3 and 4 neuroblastomas from 33 studies demonstrated that relative risk of mortality was decreased in patients who underwent >90% resection [19]. Our study indicates that complete surgical resection of the primary tumor site should be attempted in INRG high-risk patients.

Risk of complications at first and best operation is comparable to previously published data [2]. Since the presence of IDRF predicts the risk of complications and the extent of surgery, the impact of IDRF should become an integrated part of therapy planning as shown before [20‐22], but cannot be safely used as an independent risk factor for outcome. Therefore, INRG classification considering IDRF, histological category, MYCN amplification, grade of tumor differentiation and chromosomal aberrations, was applied in our study and showed a significant difference in high-risk patients with complete surgical resection of the tumor. Future trials should encourage a complete resection of the primary tumor in INRG high-risk patients with a more aggressive approach to improve outcome with an acceptable risk of complication, but probable life-threatening complications should still be avoided.

In this study postoperative staging to assess the presence and extent of possible residual tumors was carried out by computed tomography, magnetic resonance imaging or ultrasound, but MRI was recommended as gold standard for follow-up. The interval between surgery and imaging varied between one and three months postoperatively, which can result in differences of the extent of resection described by imaging studies and by operation notes. Whenever there was inconsistency (26.7% of all cases), joint review with a radiologist was performed.

Even in patients with localized neuroblastoma older than 18 months, spontaneous regression or differentiation can occur at any time throughout therapy. This should be considered in patients with incomplete resection or gross total resection achieved at first operation. In this study, only few patients with incomplete/gross total resection after first operation underwent second operations to achieve complete resection, whereas most patients did not (9 vs. 18). Due to the small number of patients, a benefit of further surgery in these cases remains unclear. Possibly, molecular characterization of the primary tumors will be able to answer these questions in the future.

In patients with localized neuroblastoma >18 months, especially with INRG high-risk classification, extended surgery improves EFS, LPFS and OS and should therefore be attempted.