Initial radiographic findings of early gastric cancer detected in health check programs and human "dry dock" (multiphasic screening) collective health checks and treated by endoscopic mucosal resection.

Background: The purpose of this study was to evaluate the capacity of detection (detectability), by radiographic examination, of minute gastric cancer that was indicated for endoscopic mucosal resection (EMR).

Methods: Fifty-five lesions in 55 patients with minute gastric cancer were endoscopically resected between 1992 and 2000 at the Foundation for Detection of Early Gastric Carcinoma. Of these 55 patients, 33 patients underwent indirect (with image intensifier and 100-mm roll film) and/or direct radiographic examination of the upper gastrointestinal tract as the initial screening examination, and subsequent upper gastrointestinal endoscopy. The remaining 22 patients underwent upper gastrointestinal endoscopy as the initial screening examination. As the first step, radiographic (indirect and direct) images were reviewed, and the abnormalities which had eventually led to the detection of the minute early gastric cancer (EGC) were reconfirmed. Then, they were analyzed in terms of the reproducibility of abnormal findings which enabled the detection of such a minute EGC, and in terms of whether they could be treated by EMR. The size, morphology, and location of the radiographically detected lesions were analyzed.

Results: Seventeen lesions of EGC were detected in the group in which indirect radiography was initially performed. Of these 17 lesions, some abnormality corresponding to the site of the lesion was pointed out in 7 lesions (2 lesions of type IIa and 5 lesions of type IIc). On the other hand, the remaining 10 lesions (type IIa, 3; type IIc, 5; type IIa + IIc, 1; and type IIb, 1) were detected by endoscopy, which was carried out to confirm the other abnormalities in the indirect radiographic images. Sixteen lesions of EGC were detected in the group in which direct radiography was initially performed. Of these 16 lesions, some abnormality corresponding to the site of the lesion was pointed out in 8 lesions (4 lesions of type IIa and 4 lesions of type IIc). The remaining 8 lesions (3 lesions of type IIa and 5 lesions of type IIc) were not detected by direct radiography.

Conclusion: Indirect and direct radiographic examination enabled the detection of approximately 50% of EGCs which could be treated by endoscopic resection. The significance of radiographic examination should not be underestimated, even in the detection of EGCs which can be endoscopically resected.