PURPOSE: To compare the abilities of magnetic resonance (MR) imaging and computed tomography (CT) in detection of lymph node metastasis from head and neck squamous cell carcinoma. MATERIALS AND METHODS: MR imaging and CT were performed with standard protocols in patients with known carcinoma of the oral cavity, oropharynx, hypopharynx, or larynx. Histopathologic examination was performed to validate imaging findings. Between 1991 and 1994, 213 patients undergoing 311 neck dissections were accrued at three institutions. RESULTS: For the upper jugular and spinal accessory regions, the areas under the receiver operating characteristic curves for combined information on size and internal abnormality were 0.80 for CT and 0.75 for MR imaging. Sensitivities, specificities, negative predictive values (NPVs), and positive predictive values (PPVs) were calculated for various size criteria with and without internal abnormality information. With use of a 1-cm size or an internal abnormality to indicate a positive node, CT had an NPV of 84% and a PPV of 50%, and MR imaging had an NPV of 79% and a PPV of 52%. CT achieved an NPV of 90%, correlating with a PPV of 44%, with use of 5-mm size as an indicator of a positive node. CONCLUSION: CT performed slightly better than MR imaging for all interpretative criteria. However, a high NPV was achieved only when a low size criterion was used and was therefore associated with a relatively low PPV.
PURPOSE: To compare the abilities of magnetic resonance (MR) imaging and computed tomography (CT) in detection of lymph node metastasis from head and neck squamous cell carcinoma. MATERIALS AND METHODS: MR imaging and CT were performed with standard protocols in patients with known carcinoma of the oral cavity, oropharynx, hypopharynx, or larynx. Histopathologic examination was performed to validate imaging findings. Between 1991 and 1994, 213 patients undergoing 311 neck dissections were accrued at three institutions. RESULTS: For the upper jugular and spinal accessory regions, the areas under the receiver operating characteristic curves for combined information on size and internal abnormality were 0.80 for CT and 0.75 for MR imaging. Sensitivities, specificities, negative predictive values (NPVs), and positive predictive values (PPVs) were calculated for various size criteria with and without internal abnormality information. With use of a 1-cm size or an internal abnormality to indicate a positive node, CT had an NPV of 84% and a PPV of 50%, and MR imaging had an NPV of 79% and a PPV of 52%. CT achieved an NPV of 90%, correlating with a PPV of 44%, with use of 5-mm size as an indicator of a positive node. CONCLUSION: CT performed slightly better than MR imaging for all interpretative criteria. However, a high NPV was achieved only when a low size criterion was used and was therefore associated with a relatively low PPV.
Authors: P T de Souza Figueiredo; A F Leite; F R Barra; R F Dos Anjos; A C Freitas; L A Nascimento; N S Melo; E N S Guerra Journal: Dentomaxillofac Radiol Date: 2012-01-12 Impact factor: 2.419