9. Sensitivity analyses for image‐based visual inspection or dermoscopy for the detection of invasive melanoma or atypical intraepidermal melanocytic variants.
| Test | Studies | Lesions (cases) |
DOR (95% CI) |
Specificity at 80% sensitivity (95% CI) % |
Sensitivity at 80% specificity (95% CI) % |
RDOR (95% CI) |
| All image‐based evaluations | ||||||
| Clinical (macro) images | 11 | 1740 (305) | 3.2 (1.9 to 5.4) |
42% (28 to 58) |
47% (34 to 59) |
5.6 (3.7 to 8.5) |
| Dermoscopic images | 60 | 13 to 475 (2851) | 17.8 (12.3 to 25.7) |
82% (75 to 87) |
81% (76 to 86) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | +40% (+27 to 57) |
+34% (+24 to +46) |
‐ |
| Image‐based evaluations: direct studies | ||||||
| Clinical (macro) images | 11 | 1740 (305) | 3.6 (1.7 to 7.6) |
48% (25 to 73) |
47% (30 to 64) |
5.3 (3.5 to 8.0) |
| Dermoscopic images | 11 | 1735 (306) | 19.2 (8.7 to 42.0) |
83% (70 to 91) |
83% (68 to 92) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | +34% (+15 to +53) |
+36% (+20 to +52) |
‐ |
| Image‐based evaluations: with histology and follow‐up for those not having surgery | ||||||
| Clinical (macro) images | 0 | ‐ | ‐ | ‐ | ‐ | ‐ |
| Dermoscopic images | 7 | 2612 (523) |
7.4 (4.5 to 12.0) |
67% (58 to 75) |
57% (39 to 74) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ |
| Image‐based evaluations with low risk of bias for the index test | ||||||
| Clinical (macro) images | 3 | 1113 (117) |
1.9 (0.91 to 4.0) |
+32% (17 to 52) |
32% (17 to 52) |
10.4 (5.7 to 19.0) |
| Dermoscopic images | 40 | 11,194 (2318) |
19.8 (12.4 to 31.7) |
83% (76 to 89) |
83% (76 to 89) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | +51% (+35 to +68) |
+51% (+34 to +67) |
‐ |
| Image‐based evaluations with low risk of bias for the reference test | ||||||
| Clinical (macro) images | 9 | 1650 (276) |
3.2 (1.9 to 5.4) |
42% (28 to 58) |
47% (34 to 59) |
5.6 (3.7 to 8.5) |
| Dermoscopic images | 51 | 10,894 (2359) |
17.8 (12.3 to 25.8) |
82% (75 to 87) |
81% (76 to 86) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | +40% (+26 to +53) |
+34% (+24 to +46) |
‐ |
| Image‐based evaluations with low risk of bias for flow and timing | ||||||
| Clinical (macro) images | 1 | 53 (10) |
15.9 (1.6 to 161) |
79% (21 to 98) |
80% (34 to 97) |
0.54 (0.05 to 5.5) |
| Dermoscopic images | 11 | 1391 (410) |
8.6 (4.4 to 16.7) |
65% (42 to 83) |
69% (56 to 80) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | ‐14% (‐67 to +39) |
‐10% (‐48 to +28) |
‐ |
| Image‐based evaluations excluding case‐control studies | ||||||
| Clinical (macro) images | 7 | 964 (183) |
7.2 (3.5 to 14.8) |
62% (40 to 80) |
66% (50 to 78) |
3.4 (1.8 to 6.4) |
| Dermoscopic images | 37 | 10,270 (1923) |
24.3 (15.2 to 39.0) |
86% (79 to 91) |
85% (79 to 90) |
|
| Change replacing visual inspection with dermoscopy (95% CI) | ‐ | ‐ | ‐ | +24% (+4 to +44) |
+20% (+7 to +32) |
‐ |
| CI: confidence interval; DOR: diagnostic odds ratio; RDOR: relative diagnostic odds ratio | ||||||