The accuracy of staging head and neck MRI versus histopathology in patients with laryngeal cancer
PO-1228
Abstract
The accuracy of staging head and neck MRI versus histopathology in patients with laryngeal cancer
Authors: Issa Mohamad1, Sami Koro1, Ayat Taqash2, Abdelatif Almousa1, Ebrahim Mayta3, Fatenah Ajlouni4, Akram Al-Ibraheem5, Hussam Haddad6, Fawzi Abuhijla1,1, Ramiz Abuhijlih1, Tariq Hussein1, Wisam Al-Gargaz3, Ali Hosni7
1King Hussein Cancer Center, Radiation Oncology Department, Amman, Jordan; 2King Hussein Cancer Center, Biostatistics Departement, Amman, Jordan; 3King Hussein Cancer Center, Surgical Oncology Department, Amman, Jordan; 4King Hussein Cancer Center, Radiology, Amman, Jordan; 5King Hussein Cancer Center, Nuclear Medicine Department, Amman, Jordan; 6King Hussein Cancer Center, Pathology Departemnt, Amman, Jordan; 7 Princess Margaret Cancer Centre, University Health Network, University of Toronto, Radiation Medicine Program, Toronto, Canada
Show Affiliations
Hide Affiliations
Purpose or Objective
Purpose/Objective: To evaluated the diagnostic value of head and neck MRI (HN-MRI) in accurate pre-operative estimation of tumor invasion to laryngeal subsites in patients with laryngeal cancer.
Material and Methods
Methods: Between 2008 and 2021, all patients with primary or recurrent laryngeal cancers who underwent staging HN-MRI and treated with total laryngectomy were evaluated. The preoperative HN-MRI and histopathological findings were independently reviewed by experienced HN-MRI radiologists and dedicated HN pathologists to assess the accuracy of tumor extensions of laryngeal subsites. Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of diagnostic HN-MRI were estimated for all laryngeal subsites. The overall diagnostic accuracy was classified into: excellent (>90%), very good (>80-90%), good (>70-80%), sufficient (>60-70%), bad (>50-60%), not useful (≤50%) (Simundic et al, EJIFCC 2009).
Results
Results: A total of 137 patients were identified. 98(71.5%) had overlapping laryngeal tumors, 32(23.4%) had glottic tumors, and 7(5.1%) had supraglottis tumors. For clinical vs pathological T-category; T4a disease was diagnosed in 116 (84.7%) vs 109(79.6%) patients, T3 in 17 (12.4%) vs 23(16.8%), T2 in 2 (1.5%) vs 3(2.2%), T1 in 2 (1.5%) vs 0(0%), and T0 in 0(0%) vs 1(1.5%) patients. The MRI overall diagnostic accuracy of prediction of tenor extensions was excellent for base of tongue (91%) and arytenoid (92%), very good for vocal cord (82%) and posterior commissure (87%), good for preeiglottis space (72%) and cricoid cartilage (76%), and sufficient for inner thyroid cortex (65%), and subglottis (68%) subsites extension. For regional lymph node involvement MRI showed high rates of false negative findings with overall accuracy of 62% (sensitivity 57.1%, specificity 69.8%).
Conclusion
Conclusion: HN-MRI provides acceptable accuracy results for predicting histopathogical findings of tumor extension, and it is an appealing imaging modality for primary radiation therapy of laryngeal cancer, particularly in the era of MR-linac adaptive radiotherapy.