Title : Guiding selective neck dissection: Pattern of nodal spread and imaging accuracy in head and neck SCC
Abstract:
Background: Selective neck dissection aims to reduce surgical complications and morbidity while maintaining oncological control of nodal disease in head and neck squamous cell carcinoma (HNSCC). However, its success depends on accurate prediction of level-specific nodal disease, typically guided by primary subsite and pre-operative imaging. This study evaluates the pattern of nodal spread and accuracy of pre-operative imaging within a contemporary Welsh cohort.
Methods: A retrospective single-centre cohort study included patients undergoing neck dissection between 2018–2023 in a district general hospital. Of 110 neck dissections identified, 30 were excluded due to absence of malignancy or non-SCC histology. Primary tumour sites were larynx (n=39), oropharynx (n=18), parotid (n=7), unknown primary (n=7), skin (n=3), and hypopharynx (n=2). Patterns of nodal spread were analysed by tumour subsite. Pre-operative imaging and dissected neck levels were compared with histopathological findings to assess diagnostic accuracy and surgical selectivity.
Results: Eighty neck dissections were analysed. Nodal metastases most frequently involved Levels II and III, particularly in oropharyngeal and laryngeal primaries, with lower rates of involvement at Levels I and V. Pre-operative imaging demonstrated limited performance for detecting level-specific nodal disease (sensitivity 50%, specificity 48%, accuracy 49%), with no significant difference between false positive and false negative rates (p=0.35). Levels I, IV and V were often dissected despite low pathological yield (4%, 11%, 5%), paired analysis confirmed a significant excess of dissected levels relative to histopathological positivity(p<0.001).
Conclusion: Local pre-operative imaging demonstrates limited reliability for level-specific nodal disease in HNSCC and should not guide selective neck dissection in isolation. Surgical planning should remain primarily subsite driven, with imaging serving as an adjunct to identify high-risk or atypical disease. The observed discrepancy between imaging and pathological findings may contribute to a tendency towards over-dissection in current practice, highlighting an opportunity in the future to refine surgical selectivity with advancements in imaging and pathological risk stratification.
