Human-AI collaborations best for skin cancer diagnosis: Study

Sydney, June 29 (IANS) Researchers have found that artificial intelligence (AI) improves skin cancer diagnostic accuracy when used in collaboration with human clinical checks.

The research team tested for the first time if a ‘real world’ collaborative approach involving clinicians assisted by AI improved the accuracy of skin cancer clinical decision making.

“This is important because AI decision support has slowly started to infiltrate healthcare settings, and yet few studies have tested its performance in the real world settings or how clinicians interact with it,” said study researcher Monika Janda from the University of Queensland in Australia.

For the findings, published in the journal ‘Nature Medicine’, the researchers trained and tested an artificial convolutional neural network to analyse pigmented skin lesions, and compared the findings to human evaluations on three types of AI-based decision support.

The study found that the highest diagnostic accuracy was achieved when crowd wisdom and AI predictions were combined, suggesting human-AI and crowd-AI collaborations were preferable to individual experts or AI alone.

Inexperienced evaluators gained the highest benefit from AI decision support and expert evaluators confident in skin cancer diagnosis achieved modest or no benefit, the researchers said.

These findings indicated a combined AI-human approach to skin cancer diagnosis may be the most relevant for clinicians in the future.

Although AI diagnostic software has demonstrated expert-level accuracy in several image-based medical studies, researchers have remained unclear on whether its use improved clinical practice.

“Our study found that good quality AI support was useful to clinicians but needed to be simple, and in accordance with a given task,” Janda said.

“For clinicians of the future, this means that AI-based screening and diagnosis might soon be available to support them on a daily basis,” Janda added.

Implementation of any AI software needs extensive testing to understand the impact it has on clinical decision making, the study noted.

–IANS

bu/arm