Utility of MRI with morphologic and diffusion weighted imaging in the detection of post-treatment nodal disease in head and neck squamous cell carcinoma

Purpose: To determine the diagnostic performance of morphologic MRI with diffusion weighted imaging (DWIMRI) for the detection of post-treatment lymph node (LN) recurrence of head and neck squamous cell carcinoma (HNSCC). Methods: This retrospective study is based on 33 HNSCC patients who underwent DWIMRI with apparent diffusion coefficient (ADC) measurements for suspected post-treatment loco-regional failure. Two radiologists, blinded to clinical/histopathological data, analyzed MR images according to established morphologic criteria and measured ADC values by drawing regions of interest on each normal/abnormal looking lymph node (LN). Histopathological findings in 40 neck dissections, 133 LN-levels and 755 LNs served as gold standard. Results: Malignant LNs had lower ADCmean values than benign LNs (1.15 +/- 0.35x10(-3) mm(2)/s versus 1.28 +/- 0.28x10(-3) mm(2)/s, p = .028). The optimal ADCmean threshold to differentiate malignant from benign LNs was 1.1695x10(-3) mm(2)/s. Sensitivity, specificity, positive (PPV) and negative (NPV) predictive values (95% CI in parentheses) of DWIMRI with morphologic criteria and ADCmean<1.1695x10(-3)mm(2)/s were: (a) 100%(86.2; 100), 44.4%(15.3; 77.3), 86.1%(69.7; 94.7), and 100%(39.5; 100) per neck dissection; (b) 83.6%(69.7; 92.2), 91.6%(83.0; 96.2), 85.4%(71.6; 93.4), and 90.5%(81.7; 95.5) per LN-level; (c) 53.1%(43.5; 62.4), 95.5%(93.5; 96.9), 67.4%(56.6; 76.7), and 92.0%(89.6; 93.9) per LN, respectively. Conclusion: The high NPV of DWIMRI irrespective of analysis type (per neck dissection/per neck level/per lymph node) make it a useful follow-up tool after treatment.