Lung cancer is one of the most common types of cancers and radiation therapy is targeted in nearly one-half of lung cancer cases. This process can take from days to weeks. Researchers have developed and validated a deep learning algorithm that can identify non-small cell lung cancer tumor on computed tomography (CT) scan. This newly developed deep learning algorithm can turn a time-intensive task to process that takes only seconds. This helps in saving valuable time of clinicians treating non-small cell lung cancer patients. The new analysis has been published in Lancet Digital Health and it showed that those who utilized algorithm were able to precisely target tissue for radiation therapy at times up to 65% faster when compared to traditional methods.
Brigham and Women’s Hospital researchers work as a part of the Massachusetts Genera Brigham Medical Artificial Intelligence Program have developed this deep learning algorithm to identify and delineate non-small cell lung cancer. In this study, they used CT images of 787 patients to train their model to distinguish between tumors and other tissues. Also, the team tested performance of algorithm using scans from over 1300 patients from increasingly external datasets. They plan to combine this work with artificial intelligence models to identify organs that are at risk of receiving unwanted radiation during cancer treatment.
Raymond Mack, MD, of the Brigham Division of Radiation Oncology said, “The biggest gap in the translation of AI in medicine is the failure to learn how to use AI to improve clinicians and vice versa.” “We are exploring how to build partnerships and collaborations between humans and artificial intelligence that lead to better outcomes for patients. Benefits of this approach for patients include greater consistency in tumor segmentation and faster treatment times. Benefits for clinicians include a reduction in routine hard work at the computer, which can reduce burnout and increase the time they can spend with patients,” he added.
This study presents a new strategy for evaluating artificial intelligence models to highlight the importance of human-AI collaboration. The approach may help pave the way for clinical deployment.
The information shared in this blog is for educational purposes only. You should always consult your medical practitioner for medical issues.