Material and Method: The present study retrospectively analyzed the 64-slice MDCT of twenty-one patient’s diagnosis with esophageal cancer who received surgical treatment in Siriraj Hospital. Patients were enrolled between June 1, 2004 and Dec 31, 2009. Twenty-one CT images of chest were evaluated by two radiologists without knowing each patient’s history as determined by surgical and pathology findings. Image analysis was determined for evaluating tumor location, wall appearance, findings of direct mediastinal extension. Accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were evaluated for tumor staging and nodal staging.

Results: Concordance between the two different readers for the interpretations with a kappa coefficient to assess interobserver variation of 0.2 to 0.9 suggests only slight agreement between the two readers. The overall sensitivity, specificity, PPV, NPV, and accuracy of CT T3 staging were 75%, 78%, 66.7%, 84.6%, and 77.3% respectively. The CT T4 staging had sensitivity 75%, specificity 85.7%, PPV 75%, NPV 85.7%, and accuracy 81.8%. In N staging, N0 staging from CT study had sensitivity 50%, specificity 33%, and accuracy 38%. N1 staging from CT study had sensitivity 33%, specificity 50%, and accuracy 38%.

Conclusion: 64-slice MDCT can be evaluated for mediastinal tumor invasion of esophageal cancer with high sensitivity, specificity, and accuracy. The metastatic node and reactive lymphadenopathy in esophageal cancer were equivocal to discriminating from CT findings, especially using 1 cm in diameter of short axis as cut point.

Keywords: Esophageal cancer, Multidetector tomography, Picus’ angle