Enhancing Radiotherapy Workflow Efficiency with Artificial Intelligence - EFAI RT Suite CT HCAP-Segmentation System

Radiation therapy is widely used in modern cancer treatment. Physicians must delineate tumors and critical organs based on individual patient imaging data to proceed with treatment planning. Manual contouring is time-consuming and affects treatment timeliness. To improve speed and plan accuracy, developing an automated contouring system is critical.

This system utilizes AI to assist radiation therapy by enhancing image quality, fusing multiple imaging modalities, and offering critical organ delineation before treatment aiming toward a complete tumor targeting solution.

After data development and validation by the hospital, the technology was transferred to Ever Fortune. AI Co. , Ltd, obtaining medical device licensing and enabling continuous updates and applications.

• The latest version of the auto-contouring system integrates numerous critical anatomical structures commonly encountered in clinical practice, covering multiple regions including the head, male pelvis, thorax, and abdomen, and supports automated contouring tasks for both CT and MR imaging. including Key structures such as the optic nerve, hippocampus, vertebral bodies (VB), gastrointestinal tract, adrenal glands, ribs, and iliopsoas muscles have been incorporated, significantly enhancing the completeness of multi-regional organ delineation.
• The latest version added structures were selected based on clinical needs, with direct involvement of radiation oncology specialists in both selection and annotation of training data. The models were strengthened through the application of a hybrid 2D/3D deep learning architecture combined with visual transformer models, improving segmentation accuracy, particularly in scenarios characterized by high anatomical variability or unclear anatomical boundaries.
• Planning and risk assessment for high-risk tumor-adjacent areas
  e.g. PRV creation
• Organ differentiation and segmentation under special lesion configurations or multimodal imaging assistance
• Clinical support for complex treatment plans
  e.g. head and neck or pelvis combined with abdominal radiotherapy
• The latest Update, the auto-contouring system promote and balance between speed, precision, and clinical adaptability, helping clinical teams improve overall treatment quality and efficiency.

Journal Article

• Developing an AI-assisted planning pipeline for hippocampal avoidance whole brain radiotherapy.
  https://pubmed.ncbi.nlm.nih.gov/36773828/
• Image Segmentation in 3D Brachytherapy Using Convolutional LSTM.
  https://link.springer.com/article/10.1007/s40846-021-00624-0
• Association of radiation dose to cardiac substructures with major ischaemic events following breast cancer radiotherapy.
  https://pubmed.ncbi.nlm.nih.gov/37585426/
• Automated Hypofractionated IMRT treatment planning for early-stage breast cancer.
  https://pubmed.ncbi.nlm.nih.gov/32178694/

Awards

2023

• National Healthcare Quality Award: Good quality mark
• Symbol of National Quality: Bronze Award

2022

• 19th NATIONAL INNOVATION AWARD: Smart Health Technology Category

Patent

• TFDA Medical Device License (MOHW Medical Device Manufacturing No. 007449)
• TFDA Medical Device License (MOHW Medical Device Manufacturing No. 007989)
• U.S. FDA Medical Device License(K220264)
• U.S. FDA Medical Device License(K231928)

Contact Window

Department of Radiation Oncology
+886-4-2205-2121 ext. 17459
094872@tool.caaumed.org.tw