Table 2.
Overview of study characteristics
| First author, year | Patient population | Public availability of dataset | Image type/Image No. | Validation (H = histology, C = clinical diagnosis) | Deep learning system | Model output | |
|---|---|---|---|---|---|---|---|
| ethnicity/ancestry/race/location | dataset information | ||||||
| Piccolo et al. (2002) [23] | Fitzpatrick I–V | Lesion n = 341 Patient n = 289 F 65% (n = 188) Average age 33.6 |
No | Dermoscopy Total 341 Training 0 Testing 341 Validation 0 |
All – H | Network architecture DEM-MIPS (artificial neural network designed to evaluate different colorimetric and geometric parameters) | Binary (melanoma, non-melanoma) |
| Iyatomi et al. (2008) [24] | Italian Austrian Japanese |
n/a | No | Dermoscopy Total 1,258 Training 247 Testing NA Validation 1,258 |
Dataset A and B– H Dataset C – H + C |
Network architecture ANN (back-propagation artificial neural networks) |
Binary (malignant or benign) Malignancy (risk) score (0–100) |
| Chang et al. (2013) [25] | Taiwanese | Lesion n = 769 Patient n = 676 F 56% (n = 380) Average age 47.6 |
No | Clinical Total 1,899 Training NA Testing NA Validation NA |
Benign – C Malignant – H |
Network architecture Computer-aided diagnosis (CADx) system built on 91 conventional features of shape, texture, and colors (developing software – MATLAB) |
Binary (benign or malignant) |
| Chen et al. (2016) [26] | American Indian, Alaska Native, Asian, Pacific Islander, black or African, American, Caucasian | Community dataset Patient n = 1,900 F 52.3% (n = 993) Age >50% under 35 |
Partially DermNet NZ – Yes Community – No |
Clinical Total 12,000 Training 11,780 Testing 337 Validation NA |
Community dataset (benign and malignant) – C DermNet – H |
Network architecture Patented image-search algorithm that builds on proven computer vision methods from the field of CBIR |
Binary (melanoma and non-melanoma) |
| Yang et al. (2017) [27] | Korean | Patient n = 110 F 50% (n = 55) |
No | Dermoscopy Total 297 Training 0 Testing 297 Validation 0 |
All – H | Network architecture 3 stage algorithm, pre-processing, stripe pattern detection and automatic discrimination (MATLAB) |
Binary (LM, nevus) |
| Han et al. (2018) [28] | Korean Caucasian |
n/a | Partially MED-NODE, Atlas, Edinburgh, Dermofit – yes Others – no |
Clinical Total 182,044 Training 178,875 Testing 1,276 Validation 22,728 |
ASAN – C+ H Multiple other dataset (5) used with unclear validation |
Neural network – CNN Network Architecture Microsoft ResNet – 152 Google Inception |
12-class skin tumor types |
| Yu et al. (2018) [29] | Korean | Lesion n = 275 | No | Dermoscopy Total 724 Training 372 Testing 362 Validation 109 |
All – H | Neural network – CNN Network architecture modified VGG – 16 |
Binary (melanoma/non-melanoma) |
| Zhang et al. (2018) [30] | Chinese | Lesion n = 1,067 | No | Dermoscopy Total 1,067 Training 4,867 Testing 1,142 Validation NA |
Benign and malignant – C Three dermatologists disagree – H |
Neural network – CNN Network architecture GoogLeNet Inception v3 |
Four class classifier (BCC, SK, melanocytic nevus, and psoriasis) |
| Fujisawa et al. (2019) [31] | Japanese | Patients n = 2,296 | No | Clinical Total 6,009 Training 4,867 Testing 1,142 Validation NA |
Melanocytic nevus, split nevus, lentigo simplex – C Others – PE |
Neural network – DCNN Network architecture GoogLeNet DCNN model |
1. Two class classifier (benign vs. malignant) 2. Four class classifier (malignant epithelial lesion, malignant melanocytic lesion, benign epithelial lesion, benign melanocytic lesion) 3. 14 class classification 4. 21 class classification |
| Jinnai et al. (2019) [38] | Japanese | Patient n = 3,551 | No | Dermoscopy Total 5,846 Training 4,732 Testing 666 Validation NA |
Malignant – H Benign tumor – C |
Neural network – FRCNN Network architecture – VGG-16 |
Binary (benign/malignant) Six-class classifications (6 skin conditions) |
| Zhao et al. (2019) [32] | Chinese | n/a | No | Clinical Total 4,500 Training 3,375 Testing 1,125 Validation NA |
Benign – C Malignant – PE |
Neural network – CNN Network architecture – Xception |
Risk (low/high/dangerous) |
| Cho et al. (2020) [33] | Korean | Patient n = 404 | No | Clinical Total 2,254 Training 1,629 Testing 625 Validation NA |
Benign – C Malignant – H |
Neural network – DCNN Network architecture – Inception-ResNet-V2 |
Binary classification (benign or malignant) |
| Han et al. (2020) [36] | Korean Caucasian |
ASNA, Normal, SNU Patient n = 28,222 F 55% (n = 15,522) Average age 40 MED-NODE, Web, Edinburgh NA |
Partially Edinburgh – Yes SNU – upon request |
Clinical Total 224,181 Training 220,680 Testing 2,441 Validation 3,501 |
ASAN – C+ H Edinburgh – H Med-Node – H SNU – C + H Web – image finding |
Neural network – CNN Network architecture SENet Se-ResNet-50 Visual geometry group (VGG-19) |
Binary (malignant, non-malignant) Binary (steroids, antibiotics, antivirals, antifungals) Multiple class classification (134 skin disorders) |
| Han et al. (2020) [35] | Korean | Patients n = 673 Lesions n = 673 F 54% (n = 363) Average age 58 |
No | Clinical Total 185,192 Training 182,348 Testing NA Validation 2,844 |
All – H | Neural network – CNN Network architecture SENet SE-ResNeXt-50 SE-ResNet-50) |
Risk output (risk of malignancy) |
| Han et al. (2020) [34] | Korean | Patient n = 9,556 Lesion n = 10,426 F 55% (5,255) Average age 52 |
No | Clinical Total 40,331 Training 1106,886a Testing NA Validation 40,331 |
All – H | Neural network – RCNN Network architecture SENet Se-ResNeXt-50 |
Binary (malignant, non-malignant) 32 class classification |
| Huang et al. (2020) [37] | Chinese | Lesion n = 1,225 | No | Clinical Total 3,299 Training 2,474 Testing 825 Validation NA |
All – PE | Neural network - CNN Network architecture Inception V3 Inception-ResNet V2 DenseNet 121 ResNet 50 |
Binary (SK/BCC) |
| Li (2020) [44] | Chinese | Patient n = 106 | No | Dermoscopy and clinical Total 212 Training 200,000a Testing 212 Validation NA |
All – H | Network architecture Youzhi AI software (Shanghai Maise Information Technology Co., Ltd., Shanghai, China) |
Binary (benign or malignant) 14 class classification |
| Liu et al. (2020) [39] | Fitzpatrick type I–VI | Patient n = 15,640 Lesion n = 20,676 |
No | Clinical Total 79,720 Training 64,837 Testing NA Validation 14,483 |
Benign – C Malignant – H |
Neural network – DLS Network architecture Inception – v4 |
26 class classification (primary output) 419 class classification (secondary output) |
| Wang et al. (2020) [40] | Chinese, with Fitzpatrick type IV | n/a | No | Dermoscopy Total 10,307 Training 8,246 Testing 1,031 Validation 1,031 |
BCC – C+H Others – C |
Neural network – CNN Network architecture GoogLeNet Inception v3 |
Binary classification (psoriasis and others) Multi-class classification |
| Huang et al. (2021) [37] | Taiwanese Caucasian |
KCGMH Patient no. 1,222 F 52.4% (n = 640) Average age 62 HAM10000 n/a |
Partially KCGMH – no HAM10000 – yes |
Clinical Total 1,287 Training 1,031 Testing 128 Validation 128 |
All – H | Neural network – CNN Network architecture DenseNet 121 – binary classification EfficientNet E4 – five class classification |
Binary (benign/malignant) 5 class classification (BCC, BK, MM, NV, SCC) 7 class (AK, BCC, BKL, SK, DF, MM, NV) |
| Minagawa et al. (2021) [42] | Caucasian Japanese |
Patient n = 50 | Partially ISIC–yes Shinshu – no |
Dermoscopy Total 12,948 Training 12,848 Testing 100 Validation NA |
Benign – C Malignant – H |
Neural network - DNN Neural architecture – Inception-ResNet-V2 |
4 class classification (MM/BCC/MN/BK) |
| Yang et al. (2021) [43] | Chinese | n/a | No | Clinical Total 12,816 Training 10,414 Testing 300 Validation 2,102 |
All – C | Neural network – DCNN Neural architecture DenseNet-96 ResNet-152 ResNet-99 Converged network (DenseNet – ResNet fusion) |
6 class classification (Nevi, Melasma, cafe-au-lait, SK, and acquired nevi) |
BCC, basal cell carcinoma; BK, benign keratosis; CNN OR DCNN, convolutional neural network; DF, dermatofibroma; SCC, squamous cell carcinoma; SK, seborrheic keratosis; MM, melanoma; MN, melanocytic; PE, pathological examination (insufficient information provided whether histopathology and/or clinical evaluation was used); n/a, not available; n, number; CBIR, content-based image retrieval.
aAlgorithm previously trained on an different dataset; therefore, dataset numbers are not included.