Table 3.
Endoscopic diagnostic performance for the differentiating between SSLs and SSLDs
| Author | Year | Study design | Diagnostic method | SSL with dysplasia/carcinoma (n) | SSL without dysplasia (n) | Sensitivity | Specificity | Accuracy |
|---|---|---|---|---|---|---|---|---|
| Sano | 2018 | Retrospective |
WLI (only morphology) |
26 (dysplasia) | 300 | 46.2% | 97.3% | 93.3% |
| Murakami | 2017 | Retrospective | WLI |
41 (dysplasia), 7 (carcinoma) |
414 | 91.7% | 85.3% | 85.9% |
| Tate | 2018 | Prospective | WLI + NBI (Non-magnifying) |
36 (dysplasia, size ≥ 8 mm) |
105 (size ≥ 8 mm) |
93.9% | 95.4% | 95.0% |
| Murakami | 2021 | Retrospective |
Magnifying NBI (JNET classification) |
52 (dysplasia), 10 (carcinoma) |
647 | 83.9% | 95.5% | 94.5% |
| Burgess | 2016 | Prospective | Pit pattern* |
66 (dysplasia, size ≥ 20 mm) |
135 (size ≥ 20 mm) |
66.7% | 72.6% | 70.6% |
| Murakami | 2017 | Retrospective |
Magnifying Chromoendoscopy (Pit pattern) |
30 (dysplasia), 6 (carcinoma) |
278 | 94.4% | 100.0% | 99.4% |
| Tanaka | 2017 | Prospective |
Magnifying Chromoendoscopy (Pit pattern) |
33 (dysplasia) | 90 | 93.9% | 87.8% | 89.4% |
*The use of dyes was not always required in the study
SSL sessile serrated lesion, WLI white light imaging, NBI narrow band imaging, JNET Japan NBI expert team