Table 2.
ChatGPT’s diagnostic accuracy categorized by tumor and nontumor groups
Correct answer (accuracy rate [%]) | ||||
---|---|---|---|---|
GPT-4-based ChatGPT | GPT-4V-based ChatGPT | |||
Final diagnosis | Differential diagnosis | Final diagnosis | Differential diagnosis | |
Total (n = 106) | 46/106 (43%) | 62/106 (58%) | 9/106 (8%) | 15/106 (14%) |
Tumor group (n = 45) | 14/45 (31%) | 22/45 (49%) | 4/45 (9%) | 5/45 (11%) |
Nontumor group (n = 61) | 32/61 (52%) | 40/61 (66%) | 5/61 (8%) | 10/61 (16%) |
Tumor group (n = 45)a | 14/45 (31%) | 22/45 (49%) | 4/45 (9%) | 5/45 (11%) |
Bone tumor (n = 24) | 8/24 (33%) | 14/24 (58%) | 2/24 (8%) | 3/24 (13%) |
Soft tissue tumor (n = 22) | 6/22 (27%) | 9/22 (41%) | 2/22 (9%) | 2/22 (10%) |
ChatGPT Chat Generative Pre-trained Transformer, GPT-4 Generative Pre-trained Transformer-4, GPT-4V Generative Pre-trained Transformer-4 with vision
aOne case presents both a bone tumor and a soft tissue tumor