医学
彩色内窥镜
诊断准确性
预测值
放大倍数
核医学
病理
放射科
胃肠病学
内科学
结直肠癌
结肠镜检查
癌症
计算机科学
计算机视觉
作者
Toyoki Kudo,Shin‐ei Kudo,Kunihiko Wakamura,Yuichi Mori,Masashi Misawa,Takemasa Hayashi,Makoto Kutsukawa,Katsuro Ichimasa,Hideyuki Miyachi,Fumio Ishida,Haruhiro Inoue
摘要
Background and Aim Endocytoscopy (EC) is a next‐generation endoscopic technique that enables diagnostic imaging at 450× magnification. In the present study, we retrospectively evaluated the diagnostic performance of EC and magnifying chromoendoscopy (MCE) for diagnosing the invasion depth of colorectal tumors. Methods We investigated 330 lesions with a ≥10‐mm tumor diameter that could be diagnosed by both MCE and EC. The lesions were classified according to morphological type as follows: laterally spreading type–granular (LST‐G), laterally spreading type–non‐granular (LST‐NG), protruding, or depressed. After all lesions had been classified by both pit pattern and EC, qualitative and quantitative (invasion depth) diagnoses were made. The diagnostic accuracy was then compared between pit pattern classification and EC classification. Results Diagnostic accuracy of EC classification was significantly higher for LST‐NG lesions (90.5%) than for protruding lesions (80.6%) ( P < 0.05). Diagnostic accuracy for LST‐NG lesions was significantly higher with EC classification (90.5%) than with pit pattern classification (79.3%) ( P < 0.001). Comparison of the diagnostic performance of EC3a findings using EC classification between LST‐NG and protruding lesions revealed a sensitivity of 92.9% versus 11.3% ( P < 0.001), positive predictive value of 78.0% versus 27.3% ( P < 0.001), negative predictive value of 95.5% versus 56.1% ( P < 0.001), and diagnostic accuracy of 87.9% versus 51.2% ( P < 0.001), respectively. Conclusion EC is a very useful method for evaluating the invasion depth of LST‐NG lesions.
科研通智能强力驱动
Strongly Powered by AbleSci AI