Pilot applications of GPT-4 in radiation oncology: Summarizing patient symptom intake and targeted chatbot applications

"For patients receiving cancer radiotherapy, the weekly on-treatment visit is important for discussing any occurring radiation toxicity with their radiation oncologist and for asking questions about treatment. Manual reporting involves asking the patient specific questions to screen for urinary and digestive toxicities that could be related to radiation therapy, within the time frame of a typical on-treatment visit (10 to 20 min). It has been shown that physicians underreport patient toxicity during these visits, to the possible detriment of patients. In addition, the process of symptom gathering and documentation is burdensome on the physician and may detract from time spent discussing other important patient concerns.

Newly released large language models (LLMs) have demonstrated human-level performance on various professional and academic benchmarks, such as the USMLE STEP 1,2, and 3 exams. , The acceleration of LLM performance holds great promise for the development of new, cost-effective tools that can be easily deployed in the healthcare sector. Past examples of such tools include chatbots and symptom checkers, which provide assistance in both diagnosis and triage with varying degrees of accuracy, The capabilities of these tools have been steadily increasing with advances in artificial intelligence (AI)..." Lire la suite