Primary study · included by 1 review
A novel AI model for detecting periapical lesion on CBCT: CBCT-SAM.
Chau 2025 · Chau Ka-Kei, Zhu Meilu, AlHadidi Abeer, et al. · Journal of dentistry · 2025
Included by
| Outcome this study informs | Pooled effect | Evidence | Certainty |
|---|---|---|---|
Risk difference for precision in segmentation tasks AI segmentation algorithms (e.g. U-Net, YOLO, SAM) vs Expert annotation (reference standard) | RD 0.98 [0.97, 0.98] | 2 studies · 4039 participants | GRADE not reported |
Pooled specificity AI algorithms for dental detection tasks vs Reference standard (expert clinicians) | Specificity (proportion) 0.94 [0.86, 0.97] | 10 studies · 21516 participants | Moderate |
Precision in segmentation tasks (risk difference) AI algorithms for dental segmentation tasks vs Expert annotation (reference standard) | RD 0.86 [-0.73, 2.46] | 2 studies | GRADE not reported |
Pooled Dice similarity coefficient (DSC/F1) AI segmentation algorithms (e.g. U-Net, YOLO, SAM) vs Expert annotation (reference standard) | Dice similarity coefficient 0.89 [0.13, 1.00] | 3 studies · 4539 participants | Low |
Pooled sensitivity AI algorithms for dental detection tasks vs Reference standard (expert clinicians) | Sensitivity (proportion) 0.85 [0.76, 0.91] | 8 studies · 11452 participants | Moderate |
1 including review in total.