Accuracy of magnetic resonance imaging in diagnosing thyroid cartilage and thyroid gland invasion by squamous cell carcinoma in laryngectomy patients

Objectives: We examined the accuracy of magnetic resonance imaging in assessing thyroid cartilage and thyroid gland invasion in patients undergoing total laryngectomy for squamous cell carcinoma, by comparing histopathology results with imaging findings. Study design: A retrospective study reviewed histology and magnetic resonance scan results for all total laryngectomies performed between 1998–2008 at University Hospital Aintree, Liverpool. Methods: Pre-operative magnetic resonance images were reviewed independently by two consultant head and neck radiologists masked to the histology; their opinions were then compared with histology findings. Results: Eighty-one magnetic resonance scans were reviewed. There were 22 laryngectomy patients with histologically verified thyroid cartilage invasion and one patient with thyroid gland invasion. There were 31 patients with apparent radiological thyroid cartilage invasion pre-operatively (with 17 false positives), giving sensitivity, specificity, and positive and negative predictive values of 64, 71, 45 and 84 per cent, respectively. On assessing thyroid gland invasion, there were nine false positive scans and no false negative scans, giving sensitivity, specificity, and positive and negative predictive values of 100, 89, 10 and 100 per cent, respectively. Conclusion: Magnetic resonance scanning over-predicts thyroid cartilage and gland invasion in patients undergoing total laryngectomy. Magnetic resonance scans have limited effectiveness in predicting thyroid cartilage invasion by squamous cell carcinoma in laryngectomy patients.