When Jack of All Trades Is a Master of None: Comparing the Performance of GPT-4 Omni against Specialised Neural Networks in Identifying Malignant Dermatological Lesions from Smartphone Images and Structured Clinical Data
Although GPT-4o consistently triaged skin lesions with high sensitivity, its very low specificity limits clinical utility. Thus, general-purpose LLMs like GPT-4o are currently unsuitable for clinical dermatological diagnostics without significant field-specific developments and validation.
Authors
- Jiawen Deng (ORCID: https://orcid.org/0000-0002-8274-6468)
- Eddie Guo (ORCID: https://orcid.org/0000-0002-7223-0505)
- Joo Ha Hwang (ORCID: https://orcid.org/0000-0002-7534-230X)
- Heather Jianbo Zhao (ORCID: https://orcid.org/0000-0002-4116-9211)
- Myron Moskalyk (ORCID: https://orcid.org/0009-0003-4136-6981)
- Kiyan Heybati (ORCID: https://orcid.org/0000-0003-4465-2249)
- Aya Alsefaou
Institutions
- Mayo Clinic (US)
- University of British Columbia (CA)
- University of Calgary (CA)
- University of Toronto (CA)
- Queen's University (CA)
- Public Health Ontario (CA)
Publication Details
- Journal
- Dermatology
- Published
- 2025-12-18
- DOI
- https://doi.org/10.1159/000550153
- Citations
- 1
- Primary Topic
- Cutaneous Melanoma Detection and Management
- Type
- article
- Field-Weighted Citation Impact
- 1.05