Gogol Bhattacharya

Background:  Artificial intelligence (AI) has seen remarkable growth in recent years, fueled by significant advances in computing power and data availability. AI systems can now tackle tasks traditionally requiring human intelligence across multiple disciplines, including healthcare. ^[1-4] The diagnosis and treatment of diseases has been a focus area for AI since the 1970s, with AI heralded as a technology to improve healthcare access and outcomes.^[1,2]

Methods:  A literature search was conducted in TAMU library’s database searching: “AI in Healthcare”, “ML in Healthcare”, “’AI’ AND ‘Physician Burnout’”, “’AI AND Physician’”. Results limited to papers published in the last 10 years.

Results:  In dermatology, convolutional neural networks (CNNs) have achieved near human-level accuracy in classifying skin lesions as benign, malignant or non-neoplastic from dermoscopic images. The Inception v3 model outperformed board-certified dermatologists on a validation set. ^[5,6] Nugroho et al consistently highlights CNNs providing highly accurate diagnostic classification, though human expertise remains crucial for complex cases.^[5,6]  Commercially available deep learning algorithms (DLAs) for detecting abnormalities on chest radiographs have outperformed radiologists. One study found a DLA achieved an AUC of 0.974, higher than radiologists at 0.949, with greater sensitivity for malignancies. The DLA showed consistent performance across patient subgroups, though further validation in varied clinical settings is needed.^[7] Beyond diagnostics, AI can assist non-expert physicians interpret medical images. When provided with explainable AI advice on abnormalities in x-rays, diagnostic accuracy of non-radiologists significantly improved versus controls or AI findings alone.^[8] Furthermore, the group that received explainable AI advice maintained a higher accuracy, despite incorrect AI findings,^[8] demonstrating AI’s role as a decision support tool. AI shows significant potential for reducing physician burnout from administrative burdens. In a large primary care study, doctors saved 9 minutes per visit by accepting AI assistant recommendations 84% of the time. They reported decreased burnout, chart review burden, and increased job satisfaction.^[9]

Conclusions: In summary, AI is transforming healthcare – improving disease detection, clinical decision support, workflow efficiency and physician well-being. While challenges remain, the evidence highlights AI’s capacity to positively impact both patient care and the provider experience. In the words of Jesse Ehrenfeld, MD, President of the AMA, “[AI] will never replace physicians — but physicians who use AI will replace those who don’t.”

Works Cited:

1. Alowais SA, Alghamdi SS, Alsuhebany N, et Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Medical Education. 2023;23(1).
2. Davenport T, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthcare Journal. 2019;6(2):94–98.
3. Younis HA, Eisa TAE, Nasser M, et al. A Systematic Review and Meta-Analysis of Artificial Intelligence Tools in Medicine and Healthcare: Applications, Considerations, Limitations, Motivation and Challenges. 2024;14(1):109.
4. Bohr A, Memarzadeh The rise of artificial intelligence in healthcare applications:25–
5. Elsevier 2020.
6. Nugroho ES, Ardiyanto I, Nugroho HA. Systematic literature review of dermoscopic pigmented skin lesions classification using convolutional neural network (CNN). Inter- national Journal of Advances in Intelligent Informatics. 2023;9(3):363.
7. Esteva A, Kuprel B, Novoa RA, et al. Dermatologist-level classification of skin cancer with deep neural networks. 2017;542(7639):115–118.
8. Kim EY, Kim YJ, Choi WJ, et Performance of a deep-learning algorithm for referable thoracic abnormalities on chest radiographs: A multicenter study of a health screening cohort. PLOS ONE. 2021;16(2):e0246472.
9. Gaube S, Suresh H, Raue M, et al. Non-task expert physicians benefit from correct explainable AI advice when reviewing X-rays. Scientific Reports. 2023;13(1).
10. Waldren SE, Billings E. Artificial Intelligence Assistant for Clinical Review and Value- based American Academy of Family Physicians Innovation Labs Report. 2024.