Reliability of diffusion tensor tractography of facial nerve in cerebello-pontine angle tumours

• Autor(a) principal: Szmuda, Tomasz
• Data de Publicação: 2020
• Outros Autores: Sloniewski, Pawel, Ali, Shan, Gonçalves Pereira, Pedro M., Pacholski, Mateusz, Timemy, Fanar, Sabisz, Agnieszka, Szurowska, Edyta, Kieronska, Sara
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.1/16641
• topic: Cerebello-pontine angle; Facial nerve; Magnetic resonance imaging; Diffusion tensor imaging; Vestibular schwannoma

Aim of the study. This study aimed to verify the accuracy of preoperative visualisation of the facial nerve (FN) by magnetic resonance-based (MR) diffusion tensor imaging-fibre tracking (DTI-FT) with neuronavigation system integration in patients with cerebello-pontine angle (CPA) tumours. Clinical rationale for the study. Complete excision with preservation of the FN remains the critical goal of today's vestibular schwannoma (VS) surgery. DTI-FT of the FN with neuronavigation is yet to be fully evaluated, and could make surgery safer. Materials and methods. This was a prospective cohort study in which 38 consecutive patients with a CPA tumour (32 VSs, five meningiomas and one epidermoid cyst) were operated on via the retrosigmoid route from 2013 to 2019. The course of the FN was simulated before surgery using StealthViz and the images were transferred to the Medtronic S7 neuronavigation system. The FN location reconstructed by DTI-FT was verified during the surgery. Results. MR acquisition was inappropriate in three patients (7.9%). DTI-FT correctly predicted the course of the FN in 31 of the 38 patients; the discordance rate was 18.4%. The accuracy of DTI-FT was 81.6% (95% CI: 65.67-92.26), sensitivity 88.57% (95%CI: 73.26-96.80) and positive predictive value was 91.18% (95% CI: 90.17-92.09). The reliability of the neuronavigation-integrated visualisation of the FN did not depend on the tumour size (p = 0.85), but the method was more accurate when the nerve was compact in shape (p = 0.03, area under curve (AUC) 0.87, 95% CI: 0.60-1.00) and in females (p = 0.04, AUC 0.78, 95% CI: 0.56-1.00). Following surgery, 86.5% ofthe patients presented with useful facial function (House-Brackmann grades I-III). Correct simulation of the FN did not prevent postoperative facial palsy (p = 0.35). Conclusions. The accuracy of DTI-FT of the FN integrated with neuronavigation remains unsatisfactory. This method does not provide any clinical benefit over non-integrated DTI-FT in terms of nerve function preservation. Clinical implications. Due to the low reliability of the predictions, further technical advances in predicting the course of the FN are awaited by clinicians. However, DTI-FT images in the operating theatre would make tumour excision more comfortable for the surgeon.