OBJECTIVE: The purpose of our investigation was to determine the usefulness of digital radiography (DR) for diagnosing the depth of invasion of esophageal carcinoma. METHODS: We evaluated 59 patients with esophageal carcinomas who underwent DR. During continuous DR in tangential views, the most distended image of the esophagus was chosen. Percent esophageal stenosis (PES) was based on the diameter across the lesion of maximal narrowing and the average of the normal oral and anal side diameters. The maximal thickness of the tumor was measured on sequentially prepared specimens. We evaluated whether the percent of esophageal stenosis correlated with the maximal thickness of the tumor on histologic findings. Receiver-operating characteristic (ROC) curves were constructed to establish the cut-off level for PES in diagnosing the depth of tumor invasion. Accuracies for the depth of the invasion were calculated based on PES using DR. For the accuracy rate, DR was compared with endoscopy and endoscopic ultrasonography (EUS). RESULTS: There was a close correlation between PES and pathological thickness of the tumor. PES values (mean+/-S.D.) were 2.45+/-0.75% in Tis and T1a tumors, 13.3+/-10.9% in T1b tumors, 35.2+/-11.1% in T2 tumors, 55.2+/-18.1% in T3 tumors, and 86.1+/-7.5% in T4 tumors. Using the ROC analysis, 12.5, 37.5, and 44.4% were the highest cut-off values of PES for differentiating < or =T1a, < or =T1b, and < or =T2 tumors. Regarding T staging, 45 (76%) of 59 lesions were staged correctly with EUS, whereas 47 (80%) were staged correctly with DR. CONCLUSION: DR is useful for diagnosing the depth of the invasion because esophageal stenosis calculated using DR is an objective index of tumor infiltration. The accuracy rate of the depth of invasion with DR was as good as that of EUS.
OBJECTIVE: The purpose of our investigation was to determine the usefulness of digital radiography (DR) for diagnosing the depth of invasion of esophageal carcinoma. METHODS: We evaluated 59 patients with esophageal carcinomas who underwent DR. During continuous DR in tangential views, the most distended image of the esophagus was chosen. Percent esophageal stenosis (PES) was based on the diameter across the lesion of maximal narrowing and the average of the normal oral and anal side diameters. The maximal thickness of the tumor was measured on sequentially prepared specimens. We evaluated whether the percent of esophageal stenosis correlated with the maximal thickness of the tumor on histologic findings. Receiver-operating characteristic (ROC) curves were constructed to establish the cut-off level for PES in diagnosing the depth of tumor invasion. Accuracies for the depth of the invasion were calculated based on PES using DR. For the accuracy rate, DR was compared with endoscopy and endoscopic ultrasonography (EUS). RESULTS: There was a close correlation between PES and pathological thickness of the tumor. PES values (mean+/-S.D.) were 2.45+/-0.75% in Tis and T1a tumors, 13.3+/-10.9% in T1b tumors, 35.2+/-11.1% in T2 tumors, 55.2+/-18.1% in T3 tumors, and 86.1+/-7.5% in T4 tumors. Using the ROC analysis, 12.5, 37.5, and 44.4% were the highest cut-off values of PES for differentiating < or =T1a, < or =T1b, and < or =T2 tumors. Regarding T staging, 45 (76%) of 59 lesions were staged correctly with EUS, whereas 47 (80%) were staged correctly with DR. CONCLUSION: DR is useful for diagnosing the depth of the invasion because esophageal stenosis calculated using DR is an objective index of tumor infiltration. The accuracy rate of the depth of invasion with DR was as good as that of EUS.