Performance of the Neck Imaging Reporting and Data System as applied by general neuroradiologists to predict recurrence of head and neck cancers

Abstract

Background: The Neck Imaging Reporting and Data System (NI-RADS) is used to assess imaging after head and neck cancer treatment. We evaluated NI-RADS with general neuroradiologists rather than with head and neck subspecialists.Methods: Computed tomography and magnetic resonance imaging examinations with/without positron emission tomography from May 2018 to September 2020 were retrospectively identified. NI-RADS scores at the primary site and lymph nodes were provided by 21 neuroradiologists. Recurrence status was based on clinical and imaging findings. Area under the curve (AUC) was used to assess accuracy. Results: We assessed 608 scans from 464 patients. For NI-RADS categories 1, 2, and 3, primary site recurrence rates were 5%, 29%, and 65% with AUC of 0.765, while lymph node recurrence rates were 3%, 10%, and 80% with AUC of 0.820.Conclusions: NI-RADS as used by general neuroradiologists is effective in separating head and neck cancers into discrete categories for predicting recurrent disease. K E Y W O R D Shead and neck cancer, neuroradiology, NI-RADS, recurrence, structured reporting | INTRODUCTIONHead and neck cancer continues to be a disease of great consequence, with approximately 900 000 cases and more than 400 000 deaths annually worldwide. 1 Continually evolving treatment options for cancers of the head and neck include chemotherapy, radiation, and surgery alone or in combination, with regimens varying by institution and by individual patient needs. 2,3 For patients with head and neck cancer, radiologists play an important role in identifying and characterizing the extent of residual or recurrent disease.Radiology societies continue to advocate for organized structured reporting, which includes consistent disease categorizations and standardized templates. [4][5][6]