Table 1. Demographic data for each histological diagnosis and the performance of dermatologists and CADx.
| No. of Correct/Incorrect Diagnoses2 | ||||||
| Pathology | N1 | % | Sex (F/M) | Mean Age (Year) | Dermatologist | CADx |
| Benign | 595 | 77.37 | 358/237 | 40.88 | 511/84 | 521/74 |
| Blue nevus | 22 | 2.86 | 16/6 | 35.64 | 19/3 | 20/2 |
| Compound nevus | 53 | 6.89 | 35/18 | 28.68 | 50/3 | 47/6 |
| Congenital nevus | 9 | 1.17 | 3/6 | 22.89 | 9/0 | 6/3 |
| Dermatofibroma* | 49 | 6.37 | 30/19 | 36.84 | 47/2 | 48/1 |
| Dysplastic nevus | 2 | 0.26 | 1/1 | 39.5 | 1/1 | 1/1 |
| Epidermal nevus* | 3 | 0.39 | 2/1 | 17 | 3/0 | 2/1 |
| Hemangioma* | 38 | 4.94 | 22/16 | 47.21 | 35/3 | 23/15 |
| Intradermal nevus | 240 | 31.21 | 161/79 | 36.98 | 209/31 | 225/15 |
| Junctional nevus | 46 | 5.98 | 29/17 | 35.85 | 39/7 | 42/4 |
| Lentigo simplex | 1 | 0.13 | 0/1 | 72 | 0/1 | 1/0 |
| Nevus lipomatosus superficialis* | 1 | 0.13 | 0/1 | 57 | 1/0 | 0/1 |
| Nevus sebaceous* | 2 | 0.26 | 2/0 | 30.5 | 2/0 | 2/0 |
| Nevus spilus | 2 | 0.26 | 2/0 | 23.5 | 2/0 | 2/0 |
| Seborrheic keratosis* | 127 | 16.51 | 55/72 | 57.69 | 94/33 | 102/25 |
| Malignant | 174 | 22.63 | 81/93 | 68.86 | 145/29 | 149/25 |
| BCC* | 110 | 14.3 | 57/53 | 68.14 | 97/13 | 99/11 |
| Cutaneous melanoma | 8 | 1.04 | 5/3 | 63.25 | 6/2 | 6/2 |
| Kaposi's sarcoma* | 14 | 1.82 | 2/12 | 70.5 | 11/3 | 10/4 |
| Keratoacanthoma* | 22 | 2.86 | 7/15 | 64.82 | 14/8 | 18/4 |
| SCC* | 20 | 2.6 | 10/10 | 78.35 | 17/3 | 16/4 |
| All lesions | 769 | 100 | 439/330 | 47.21 | 656/113 | 670/99 |
1
N: number of images.
2
Indeterminate diagnoses by dermatologists were considered as incorrect. The diagnoses by CADx were made using the final 16-feature model with a cutoff value 0.3972.
*
Non-melanocytic skin lesions.