Detection of recurrent nasopharyngeal carcinoma: MR imaging versus CT

To compare the use of magnetic resonance (MR) imaging and computed tomography (CT) in detection of recurrent nasopharyngeal carcinoma.

Forty-five sets of CT and MR images were obtained in 34 patients. The images were placed in three categories: (a) clinically or radiologically abnormal findings in patients who underwent biopsy (n = 16), (b) clinically normal and radiologically borderline findings in patients who were followed up clinically and radiologically (n = 10), and (c) clinically and radiologically normal findings in patients who were monitored only clinically (n = 19). All images were read by two observers independently.

There were nine positive and seven negative biopsy results. All patients in the latter two categories had normal findings at followup. CT had a sensitivity of 45% and 67% and a specificity of 64% and 70% for each of the two observers. MR imaging had a sensitivity of 56% (for both observers) and a specificity of 78% and 83%. The kappa test for interobserver concordance was 0.53 for CT and 0.66 for MR imaging.

Both modalities have relatively low sensitivity and moderate specificity in detection of tumor recurrence and in distinguishing recurrence from post-radiation therapy changes.