A Study on the Clinical Application of AI-based Auto-Segmentation for Target Volumes in Patients with Breast Cancer after Breast-Conserving Surgery
Source abstract
Abstract
We evaluated artificial intelligence (AI)-based auto-segmentation models in patients with breast cancer undergoing surgery and radiation therapy. Radiation oncologists manually defined clinical target volume (CTV) and planning target volume (PTV) in 100 cases to train Acculearning 2.2.3.182 and OncoStudio 2.0.4, with models automatically contouring CTV and PTV, showing acceptable agreement with manual contours using Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), and mean surface distance (MSD). In four cases, manual contouring took 548 ± 205 s, whereas AI-assisted contouring with manual revision took 187 ± 40 s. The paired t-test revealed significant accuracy improvements with the 3D approach for RNI cases (p < 0.05) and laterality-dependent differences in WBI cases (p < 0.05). These findings highlight the need to consider treatment extent and anatomical laterality in AI auto-segmentation. Deep learning segmentation speeds up contouring and enhances workflow efficiency in radiation therapy planning.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Advanced Radiotherapy Techniques · AI in cancer detection · Breast Cancer Treatment Studies
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References
Reference metadata was not supplied for this record.