↑

Study Suggests the Feasibility of Automating Radiation Therapy Planning for Early-Stage Lung Cancer

AiRato Inc. (Headquarters: Sendai, Miyagi Prefecture; CEO: Yuto Kimura), which develops and commercializes AI-based radiation therapy planning support services, announces that a study conducted by the Department of Radiation Oncology, Graduate School of Medicine, Tohoku University, and the Department of Radiation Therapy, University of Yamanashi Hospital has suggested the feasibility of automating treatment planning for high-precision radiation therapy (SBRT-VMAT)*¹ for early-stage lung cancer using RatoGuide, the radiation therapy planning support software developed and provided by the Company.

This study evaluated the clinical validity of an automated, deep-learning-based approach with the aim of reducing the human and time burdens associated with treatment planning while enabling the creation of irradiation plans with consistent quality. AiRato was not involved in the design or analysis of the study and supported the research solely through the provision of RatoGuide.

*1 Stereotactic Body Radiotherapy (SBRT) and Volumetric Modulated Arc Therapy (VMAT):

SBRT (Stereotactic Body Radiotherapy) is a form of stereotactic radiation therapy that delivers a high radiation dose precisely concentrated on the tumor. VMAT (Volumetric Modulated Arc Therapy) is a type of rotational intensity-modulated radiation therapy in which irradiation is delivered while rotating the linear accelerator gantry, optimizing the dose distribution throughout the treatment arc.

The SBRT-VMAT approach used in this study combines these two techniques, representing a form of stereotactic radiation therapy delivered via rotational intensity-modulated radiation therapy and is considered one of the representative modalities of high-precision radiation therapy. This approach enables both a high degree of dose concentration to the target lesion and substantial dose reduction to surrounding normal tissues.

Key Research Findings

  • The automated treatment plans achieved quality equivalent to manual plans, with no significant differences observed in dose–volume histogram (DVH) evaluation metrics.
  • In the clinical evaluation, two radiation oncologists judged all automated plans to be clinically acceptable. In particular, centrally located (hilar/central-type) lung tumors received more favorable evaluations than peripherally located tumors.
  • The average planning time was reduced to approximately 4.6–4.8 minutes, representing a substantial improvement in efficiency compared with conventional automated planning methods, which typically require 30–60 minutes.
  • This study is the world’s first to demonstrate automated SBRT-VMAT treatment planning for early-stage lung cancer that reproduces an AI-predicted ideal dose distribution.

Background and Study Overview

SBRT-VMAT therapy, which combines stereotactic body radiotherapy (SBRT) with volumetric modulated arc therapy (VMAT), is an advanced technique that delivers high radiation doses to tumors in a small number of treatment sessions. However, creating such highly precise treatment plans requires repeated calculations involving numerous parameter adjustments, often taking several hours to days to complete. In addition, plan quality can vary depending on the planner’s level of experience, making it a significant challenge to balance efficiency with consistent quality.

To address these challenges, this study employed artificial intelligence. Using CT images along with tumor and organ contour data as inputs, RatoGuide predicts an ideal dose distribution, which is then converted into ring-shaped dose structures. Based on these structures, the final irradiation plan is automatically generated using the treatment planning system RayStation. This approach enables a substantial acceleration of the planning process while maintaining consistent plan quality without reliance on the planner’s individual experience.

Publication Information

  • Title:“Evaluation of Deep Learning–Based Automated Radiotherapy Planning for Early-Stage Lung Cancer Using SBRT-VMAT: A Comparison with Manual Planning”
  • Journal:Journal of Applied Clinical Medical Physics (2025)
  • Link:https://pubmed.ncbi.nlm.nih.gov/41088571/

Future Outlook

The findings of this study demonstrate the effectiveness of RatoGuide in automating radiation therapy planning and suggest its potential applicability to other treatment sites and irradiation conditions.

Going forward, AiRato will collaborate with medical institutions in Japan and overseas to expand the methodology validated in early-stage lung cancer to other disease sites, such as prostate cancer, further advancing the accuracy and clinical application of AI-driven treatment planning.

Through the development of products that leverage AI-based treatment planning with RatoGuide, AiRato aims to achieve substantial reductions in planning time and optimization of therapeutic outcomes, thereby contributing to both the quality and efficiency of radiation therapy.

News