Diagnosis from Nucleus Pattern Image On Colorectal Tumors with Magnified Chromo-Endoscopy

Diagnosis from Nucleus Pattern Image On Colorectal Tumors with Magnified Chromo-Endoscopy Masafumi Tabuchi, Joji Kitayama, Yoh Kato Introduction: For the diagnosis of colorectal tumors, the pit pattern diagnosis method is now widely used in Japan, which is based on the structure observation of the mucosal glands and so sophisticated way of diagnosis with high accuracy. But it lacks the nucleus observation compared with histology. To get the nucleus image in vivo, the endo-cytoscopy and confocal laser endoscopy were already invented and their studies were published. But they need high cost. The first aim of this study is to check the feasibility of normal type magnifying endoscope to observe the nucleus. The second aim is to evaluate our nucleus pattern classification. Methods: To stain nucleus in less than 3 min, 0.5% methylene blue solution is scattered on colorectal lesions. The scope used was EC490ZH and EC590ZW5 with 7 micro meter maximum resolution. We classified the nucleus pattern images into 3 categories. Type 1 is small size, dot shape. Type 2 is slightly large, elongated shape. Type 3 is large size, oval shape and scattered. The consecutive 146 Colorectal Lesions were surveyed and compared with histology. Results: 146 lesions were consisted of 39 type 1, 96 type 2, and 11 type 3. 39 type 1 lesions were 10 normal or inflammatory mucosa( Z np), 17 hyperplastic nodules( Z hp), 9 metaplastic polyp( Z mp) and 3 adenomas. 96 type 2 lesions were 4 unknown, 9 hn, 10 mp, and 73 adenomas. 11 type 3 lesions were 1 unknown, 7 adenomas and 3 cancers. If we take the diagnosis criteria that nucleus type 1 is not tumor, type 2 is adenoma and type 3 is cancer, the sensitivity of type 1 was 65%, the specificity of type 1 was 92%, the sensitivity of type 2 was 88%, the specificity of type 2 was 79%, the sensitivity of type 3 was 100% , the specificity of type 3 was 30%, and the over all accuracy rate was 79%. Conclusions: Normal magnified chromo-endoscopy was feasible to observe nucleus in daily practise. Our classification was thought to be useful to sophisticate the diagnosis on colorectal tumors. When this method is combined with pit pattern method, endoscopic findings may reach the level of histology and get the more accuracy. Contrary to our predictions, the 70% lesions of type 3 was adenoma. They were small lesions and showed type 3 in a tiny part. The precise section might not be cut out in these lesions. Further technology is needed to get a pin-point section of interest to evaluate this classfication. Reference: In vivo observation of living cancer cells. H.Inoue et al. Gastrointest Endosc Clin N Am. 2004 Jul;14(3):589-94, x-xi. Confocal Laser Endoscopy for Diagnosing Intraepithlial Neoplasma and Colorectal Cancer in vivo. Ralf Kiesslich et al. Gastroenterology 2004;127:706-713.