Table 4 Meta-regression analysis of deep learning algorithm performance based on patient-based analysis in external validation cohorts for diagnosing microsatellite instability-high (MSI-H) in colorectal cancer using whole slide images
Subgroup | Studies, n | Sensitivity(95%CI) | Meta-regression P-value | Specificity(95%CI) | Meta-regression P-value |
---|---|---|---|---|---|
Center | 0.03 | 0.21 | |||
Single center | 2 | 0.77 (0.51–0.1.00) | 0.88 (0.66–1.00) | ||
Multi-center | 19 | 0.93 (0.90–0.97) | 0.68 (0.55–0.82) | ||
Optimal AI algorithm | 0.23 | 0.06 | |||
CNN | 16 | 0.93 (0.89–0.97) | 0.65 (0.50–0.80) | ||
Non-CNN | 5 | 0.91 (0.83–1.00) | 0.80 (0.69–1.00) | ||
Reference standard | <0.001 | 0.23 | |||
Only PCR | 7 | 0.89 (0.80–0.98) | 0.62 (0.37–0.86) | ||
non-only PCR | 14 | 0.94 (0.91–0.98) | 0.75 (0.61–0.89) | ||
Tile size | 0.87 | <0.001 | |||
224*224 or 256*256 | 15 | 0.94 (0.90–0.98) | 0.58 (0.44–0.72) | ||
512*512 | 4 | 0.88 (0.75–1.00) | 0.91 (0.81–1.00) | ||
Magnification | 0.71 | 0.53 | |||
20* | 15 | 0.95 (0.91–0.98) | 0.66 (0.50–0.82) | ||
40* | 3 | 0.86 (0.70–1.00) | 0.76 (0.47–1.00) |