Artificial intelligence (AI) is revolutionizing early skin cancer detection
Last Updated on October 19, 2023 by Joseph Gut – thasso
October 14, 2023 – Artificial Intelligence (AI) is revolutionizing various areas of medicine, especially in the identification and analysis of medical phenotypes of patients. Thasso has previously published articles on this topic, once regarding genetics and the recognition of common characteristics in portrait photos of affected individuals
using AI, and once in the context of establishing systems in AI that enable researchers to detect a variety of genetic diseases early on affected patients.
In fact, artificial intelligence (AI), especially since ChatGPT was released, is on the rise and is increasingly playing an important role in numerous areas of life, including in medicine and especially in cancer research. And as a new study shows, AI is set to revolutionize cancer research.
This is the conclusion reached by this study recently presented at the 2023 Congress of the European Academy of Dermatology and Venereology (EADV), which examined 22,356 patients with suspected skin cancer over a period of 2.5 years. In addition to a sensitivity of 100% (59/59 cases identified) in detecting melanoma, the most serious form of skin cancer, the current AI software correctly detected 99.5% (189/190) of all skin cancers and 92.5% (541 /585) of precancerous lesions.
This illustrates the enormous progress in AI research and application: just two years ago (2021), the first version of this AI system detected only 85.9% (195/227) of all melanomas in tests, 83.8% (903/ 1078) of all skin cancers and 54.1% (496/1078) of all skin cancers were detected. The current third version of the same AI system represents a significant improvement over the first model and allows early detection of skin cancer in almost 100% of cases. Even precancerous lesions such as precancerous lesions and melanoma cases have not escaped the system.
According to the researchers involved, the study showed how AI is rapidly improving and learning, with the high accuracy directly attributable to improvements in AI training techniques and the quality of the data used to train the AI. In everyday clinical practice, these advances in AI significantly benefit other patients, especially in cases where clinical endpoints are set with the highest precision in large numbers of patients. This is how Dr. According to Andrew, lead author of the study, over 1000 face-to-face consultations will be saved in secondary care between April 2022 and January 2023. This gives treating doctors more time for patients who urgently need attention.
However, the researchers emphasize that the software cannot replace a dermatologist. Nevertheless, AI will increasingly play an important role in medicine. Although the
data is incredibly encouraging, the research team notes that AI should not be used as a stand-alone detection tool without the support of a consultant dermatologist. Of the basal cell carcinoma cases, a single case out of 190 was missed, which was later identified during a second safety net examination by a dermatologist. This further demonstrates the need for appropriate clinical monitoring of AI.
The co-author Dr. Irshad Zaki, consultant dermatologist at University Hospitals Birmingham NHS Foundation Trust, also explains: “We would like to emphasize that AI should not be used as a standalone tool for skin cancer detection and that AI is not a replacement for consultant dermatologists. “The role of AI in dermatology and the most appropriate path are being discussed,” says Dr. Andrew. “Further research with appropriate clinical oversight
could enable the use of AI as a triage tool. AI is not yet an independent tool in dermatology; However, the data from the study presented show the great promise of the role that AI will play in future healthcare, in this case in the early detection of skin cancer.
A very exciting question will be how AI can help to prospectively predict the further course and clinical outcome of correctly diagnosed (skin) cancer in each individual patient.
Disclaimer: Images and/or videos (if any) on this blog may be subject to copyright. All rights remain with the owner of these rights.