Although artificial intelligence (AI)-assisted radiotherapy may help physicians make more informed decisions when treating patients with cancer, there may be challenges in how physicians work with the technology, according to a recent study published by Niraula et al in Nature Communications.
Study Methods and Results
Knowledge-based response-adaptive radiotherapy is a method that uses AI to optimize treatment outcomes by suggesting treatment adjustments based on how well the patient responds to the therapy.
In the study, researchers developed an AI system utilizing patient data such as medical imaging and test results to recommend changes in radiation doses. The researchers asked physicians to make treatment decisions for their patients with cancer without any technological assistance and then to make those decisions with the help of AI. While some physicians found the suggestions beneficial, others preferred to rely on their own judgment.
The researchers found that when physicians used AI to help decide the most effective treatment plan, they made more consistent choices, thereby reducing differences between physicians’ decisions. However, the technology didn't always change physicians’ minds. In some cases, the physicians disagreed with the AI and made treatment decisions based on their experience and patient needs.
The researchers noted that AI can be a helpful tool, but physicians may need to place more trust in the technology for it to work more effectively. Physicians were more likely to follow AI suggestions when they felt confident in its recommendations.
Conclusions
“While AI offers insights based on complex data, the human touch remains crucial in cancer care,” emphasized senior study author Issam El Naqa, PhD, Chair of the Machine Learning Department at the Moffitt Cancer Center. “Every patient is unique, and [physicians] must make decisions based on both AI recommendations and their own clinical judgment,” he indicated.
The researchers hope their findings can lead to better integration of AI tools and collaborative relationships that physicians can use to make more personalized treatment decisions for patients with cancer. They plan to further investigate how AI can support physicians in other medical fields.
“Our research shows that AI can be a powerful tool for [physicians], but it’s important to recognize that AI works best when it’s used as a support, not a replacement, for human expertise,” underscored lead study author Dipesh Niraula, PhD, an applied research scientist in the Machine Learning Department at the Moffitt Cancer Center. “[Physicians] bring their expertise and experience to the table, while AI provides data-driven insights. Together, they can make better treatment plans, but it requires trust and clear communication,” he concluded.
Disclosure: The research in this study was supported by the National Institutes of Health. For full disclosures of the study authors, visit nature.com.
