The NBI International Colorectal Endoscopic (NICE) classification
Performance characteristics of the NICE classification for high-confidence prediction
NICE 1 vs. 2 | NICE 2 vs. 3 | |
---|---|---|
Prediction | Adenoma | Deep submucosal invasive cancer |
Rater, n | Experts, 2 | Medical students, 5 |
Polyps (n) | 236 | 400 |
Evaluation | Real time during colonoscopy | Still images |
Performance, % (95% CI) | ||
Accuracy | 89 (83–93) | 90 (85–93) |
Sensitivity | 98 (94–100) | 92 (87–95) |
Specificity | 69 (55–80) | 88 (84–92) |
PPV | 87 (80–92) | 88 (83–91) |
NPV | 95 (84–99) | 92 (87–95) |
HC rate, % | 75 | 50 |
The NICE classification has been established by the international experts and disseminated across the world as the classification can easily be applied without optical magnification. However, there remain some issues regarding the NICE classification. First, sessile serrated polyp (SSP) was not included in the NICE classification, because the standard histopathological diagnostic criteria for SSP were not fully established. Second, NICE 2 category includes two types of polyp: benign low-grade adenomas and submucosal invasive cancer with malignant behavior. These two types should be diagnosed individually in terms of selecting an appropriate treatment strategy. The former should be treated by polypectomy or piecemeal resection; however the latter should be treated by en bloc deep resection such as EMR or ESD to examine accurate depth of cancer invasion which is necessary for judgment of curative or non-curative resection. Therefore, the Japanese endoscopists decided to separate the type 2 category into type 2A (low-grade adenoma) and type 2B (high-grade adenoma and submucosal cancer) by using the advantage of magnifying endoscopy and developed the JNET classification as an advanced version of the NICE classification.