Efficiency and Clinical Utility of AI-Assisted Radiotherapy Planning Using RatoGuide for Oropharyngeal Cancer: A Case Report.

This study evaluates the efficiency and dosimetric performance of RatoGuide, an artificial intelligence (AI)-assisted radiotherapy planning tool, by comparing AI-generated and manually created treatment plans for a 50-year-old male with right-sided oropharyngeal cancer (cT2N2bM0, cStage IVA) who underwent concurrent chemoradiotherapy. Treatment plans were created using volumetric-modulated arc therapy (VMAT) following the approach used by the Japanese Clinical Oncology Group (JCOG) protocol. RatoGuide generated two plans: one prioritizing the planning target volume (PTV) and the other focusing on organs at risk (OAR), while an experienced radiation oncologist manually developed a plan using a treatment planning system (TPS). Dosimetric comparisons focused on target coverage, OAR sparing, and dose homogeneity. Results showed that both AI-generated and TPS plans achieved comparable PTV coverage, with nearly identical values for Dmin, Dmean, and Dmax. The TPS plan exhibited slightly better dose homogeneity, whereas the AI-generated plan provided superior OAR sparing, particularly for the spinal cord and parotid glands, reducing the spinal cord's intermediate-dose volume (V30) by approximately 40%. However, the AI plan yielded slightly higher mean doses to both submandibular glands, though still within clinically acceptable thresholds. Additionally, the AI planning workflow was completed in just 30 minutes, significantly reducing the time required for manual planning. RatoGuide demonstrated efficiency in generating high-quality treatment plans, achieving comparable PTV coverage, and improving OAR sparing in certain areas. However, minor refinements are needed to optimize dose homogeneity and further minimize submandibular gland exposure. These findings suggest that AI-assisted planning has the potential to enhance radiotherapy efficiency and consistency.