A groundbreaking artificial intelligence system named iSeg is poised to transform radiation oncology by automating the precise mapping of lung tumors, a task traditionally performed manually by clinicians. Developed by researchers at Northwestern Medicine, iSeg not only matches the accuracy of expert physicians in delineating tumors on CT scans but also identifies regions that may be overlooked, potentially improving patient outcomes.
Tumor segmentation—the process of outlining the exact boundaries of a tumor—is critical in radiation therapy to ensure that high-dose radiation targets cancerous tissues while sparing healthy ones. Historically, this task has been manual, time-consuming, and subject to variability among clinicians, leading to potential inconsistencies in treatment planning. iSeg addresses these challenges by automating the segmentation process, providing consistent and accurate tumor outlines.
What sets iSeg apart is its ability to account for tumor movement due to respiration. Unlike previous AI tools that analyzed static images, iSeg employs three-dimensional deep learning techniques to segment tumors dynamically as they shift with each breath. This motion-resolved approach offers a more physiologically accurate representation, crucial for effective radiation therapy planning.
In a study published in npj Precision Oncology, iSeg was trained on CT scans and physician-drawn tumor outlines from hundreds of lung cancer patients across nine clinics within the Northwestern Medicine and Cleveland Clinic health systems. The AI system was then tested on new patient scans, and its tumor outlines were compared to those drawn by physicians. The results demonstrated that iSeg consistently matched expert outlines and identified additional tumor regions that some doctors missed. These missed areas, when left untreated, were associated with worse outcomes, underscoring the clinical significance of iSeg’s enhanced detection capabilities.
The integration of iSeg into clinical practice could lead to more precise and standardized radiation therapy, reducing delays and ensuring fairness across hospitals. By automating tumor contouring, iSeg has the potential to minimize human error and improve patient care.
The research team is currently testing iSeg in real-time clinical settings and plans to expand its capabilities to segment tumors in other organs, such as the liver, brain, and prostate. Additionally, efforts are underway to adapt the tool for use with other imaging modalities, including MRI and PET scans. If these developments proceed as anticipated, iSeg could be integrated into clinical workflows within the next few years, serving as a valuable decision-support tool for radiation oncologists.