College
Rowan-Virtua School of Osteopathic Medicine
Keywords
melanoma, skin cancer, Artificial Intelligence, early detection, dermoscopy, underserved communities
Date of Presentation
5-1-2025 12:00 AM
Poster Abstract
Background:
Melanoma, while comprising only 1% of skin cancer cases, accounts for a high mortality rate due to its ability to metastasize if not detected early. Traditional face-to-face dermatologic evaluations provide limited access to care, resulting in delayed diagnoses.
Significance:
Artificial Intelligence (AI) offers a novel approach to improve diagnostic accuracy and accessibility by alleviating strain on healthcare resources in medically underserved areas.
Methods:
A literature review was conducted using studies published in the last ten years. Eligible resources included case reports, clinical studies, clinical trials, meta-analyses, reviews, and systematic reviews in English. Studies that were repetitive or inapplicable were excluded.
Results:
AI-assisted diagnostic tools, particularly deep learning algorithms, exhibited high sensitivity and specificity in melanoma detection. These technologies have shown melanoma detection to be comparable to, and in some aspects better than traditional methods.
Discussion:
AI-assisted diagnosis shows the capability to improve melanoma outcomes through standardization. Despite promising findings, limitations persist, including variability in AI training datasets and reduced generalizability due to controlled study environments.
Future Directions:
Future research involves expanding AI training across diverse populations, integrating AI into routine clinical practice, and assessing its impact in primary care and medically underserved settings.
Disciplines
Dermatology | Diagnosis | Medicine and Health Sciences | Neoplasms | Oncology | Pathological Conditions, Signs and Symptoms
Included in
Dermatology Commons, Diagnosis Commons, Neoplasms Commons, Oncology Commons, Pathological Conditions, Signs and Symptoms Commons
The Effective Utilization of Artificial Intelligence (AI) in Early Melanoma Detection
Background:
Melanoma, while comprising only 1% of skin cancer cases, accounts for a high mortality rate due to its ability to metastasize if not detected early. Traditional face-to-face dermatologic evaluations provide limited access to care, resulting in delayed diagnoses.
Significance:
Artificial Intelligence (AI) offers a novel approach to improve diagnostic accuracy and accessibility by alleviating strain on healthcare resources in medically underserved areas.
Methods:
A literature review was conducted using studies published in the last ten years. Eligible resources included case reports, clinical studies, clinical trials, meta-analyses, reviews, and systematic reviews in English. Studies that were repetitive or inapplicable were excluded.
Results:
AI-assisted diagnostic tools, particularly deep learning algorithms, exhibited high sensitivity and specificity in melanoma detection. These technologies have shown melanoma detection to be comparable to, and in some aspects better than traditional methods.
Discussion:
AI-assisted diagnosis shows the capability to improve melanoma outcomes through standardization. Despite promising findings, limitations persist, including variability in AI training datasets and reduced generalizability due to controlled study environments.
Future Directions:
Future research involves expanding AI training across diverse populations, integrating AI into routine clinical practice, and assessing its impact in primary care and medically underserved settings.
