Zachary M. Soler, MD, is a consultant for Ora (Andover, MA). Timothy L. Smith, MD, is funded by grant support from the NIH/NIDCD for research unrelated to this effort. Timothy L. Smith is also a consultant for Intersect ENT (Palo Alto, CA) and ENTrigue Surgical (San Antonio, TX), neither of which provided financial or material support for this article. The authors have no other funding, financial relationships, or conflicts of interest to disclose.
Esthesioneuroblastoma is a malignant neoplasm originating from the specialized neuroepithelium of the olfactory cleft. Surgical resection through an open craniofacial approach has been the accepted treatment modality for decades, together with adjuvant radiotherapy and sometimes chemotherapy. Over the last 10 to 15 years, surgical techniques for esthesioneuroblastomas have evolved to include endoscopically assisted craniofacial resections, and more recently, purely endoscopic skull base resections (Fig. 1). However, the introduction of these newer techniques has not been without significant debate. This controversy stems largely from important limitations in the currently available published literature.
A meta-analysis by Dulguerov et al. summarized the outcomes of 76 studies published between 1990 and 2000.1 Although these data predate endoscopic techniques, they establish tumor stage (Kadish or Dulguerov), histopathologic grade (Hyam's classification), and treatment modality (surgery ± adjuvant therapy) as important determinants of disease-specific survival. Additionally, the authors stress that esthesioneuroblastoma can recur up to 10 years following treatment, with many authors reporting an average time to recurrence of 6 years. Therefore, promising short-term results must be confirmed over many years of follow-up.
As endoscopic techniques were applied to malignant sinonasal disease, concerns arose regarding piecemeal resection of tumors as opposed to formal en bloc resections. The only study that addresses this question was published by Wellman et al. in 1999.2 The authors present 30 malignant neoplasms originating in the paranasal sinuses resected through an open, midline anterior craniofacial approach, of which 16 (53.3%) had an en bloc resection and 14 required a piecemeal resection. The piecemeal resection group had fewer complications (21% vs. 31%) and improved survival at 3 years (70% vs. 56%) compared to the en bloc group, although formal statistical testing was not done. The authors concluded that survival is dependent on margins, but independent of the approach used to obtain clear margins.
An updated systematic review with meta-analysis was published by Devaiah and Andreoli in 2009, incorporating studies published between 1992 and 2008 that described five or more patients with esthesioneuroblastoma.3 The goal of that analysis was to compare survival outcomes after endoscopic resection to those following traditional open craniofacial resection. A total of 23 studies met inclusion criteria, and individual data were pooled from 379 subjects, with an average follow-up ranging from 27.0 to 65.6 months. Summary Kaplan-Meier analysis showed improved survival with endoscopic resection compared to open craniofacial resection (hazard ratio, 3.56; 95% confidence interval, 1.61-7.91), with similar follow-up in the endoscopic and open groups (54.5 vs. 51.0 months). However, this significant finding is not without important caveats. First, all of the included studies are retrospective case series, usually from single institutions. The source data for this meta-analysis were thus susceptible to all of the limitations and biases typical of retrospective reviews. Another important limitation is the high likelihood of publication bias, namely that those institutions with positive or flattering results are more likely to publish their findings, especially with regard to new techniques or technologies. Perhaps most importantly, the analysis was not controlled for tumor stage, a known predictor of outcome. As might be expected, those patients undergoing open craniofacial resection were more likely to have a higher stage tumor than those treated with a purely endoscopic approach (P = .0012), potentially confounding the overall results.
Because of that meta-analysis, additional single-institution case series have been published reporting outcomes after open and endoscopic resections. De Gabory et al. reported long-term outcomes in a French case series spanning 35 years of open craniofacial resections. In these 28 patients, the 5-year disease-specific survival was 88.6%, dropping to 66.4% at 20 years.4 Most recently, Gallia et al. published their series of nine patients with a mix of Kadish A-D tumors undergoing a purely endoscopic resection followed by radiotherapy.5 In each case negative margins were obtained, and all patients were disease free at the last clinic visit, although follow-up only averaged 27 months.
The optimal treatment of esthesioneuroblastoma involves complete surgical resection followed in most instances by radiotherapy. Definitive prospective studies comparing open craniofacial resection with purely endoscopic techniques will probably never be available, owing to the relative rarity of esthesioneuroblastomas and their proclivity for late recurrence. The currently available evidence suggests that equivalent short-term outcomes can be achieved with both techniques, with purely endoscopic techniques more likely to be performed in lower-stage tumors and open techniques in higher-stage cases. The completeness of the surgical resection is probably more important than the specific surgical technique used to achieve negative margins. The best practice for cases of esthesioneuroblastoma would thus be for the individual surgeon to use the technique that is most likely to result in complete tumor removal, whether it is a traditional open procedure, endoscopically assisted, or purely endoscopic. With this paradigm, the specific surgical technique would be expected to vary based on tumor stage and the individual surgeon's training/preference.
LEVEL OF EVIDENCE
The available evidence regarding surgical techniques for esthesioneuroblastoma includes single-institution, retrospective case series (level 4) and two systematic reviews with meta-analyses of individual case series (level 3a-).