Author: Adebayo Abe, Samuel Olaolu Adeneye, Eben Aje, Bidemi I. Akinlade, Inioluwa Damilola Ariyo, Lilian Ekpo, Muhammad Habeebu, Adedayo O. Joseph, Charles S. Mayo, Noah Ndianaobong, Ikechi S Ozoemelam, Margaret Dideolu Taiwo, Godwin Uwagba 👨🔬
Affiliation: University of Michigan, University of Ibadan, University of Lagos, Missouri University of Science and Technology, NSIA-LUTH Cancer Center, University of Lagos, NSIA-LUTH Cancer Centre, NSIA-LUTH Cancer Center 🌍
Purpose:
Head and neck cancers (HNC) present significant challenges in radiotherapy due to complex anatomy and the proximity to critical organs at risk (OARs). These challenges are compounded in nations with more limited resources for planning and greater need to shorten time to treatment. This retrospective study evaluates the efficacy of the open-source application Automatic Virtual Integrative (AVI) Planner, an AI-driven radiotherapy planning system, against expert manual planning for nasopharyngeal and laryngeal cancer cases treated with Intensity Modulated Radiation Therapy (IMRT) and Volumetric Modulated Arc Therapy (VMAT).
Methods:
Fourteen patients treated at NSIA-LUTH Cancer Center, with initial plans created using Varian Eclipse treatment planning system (TPS) prescribed to deliver 70 Gy in 35 fractions. These plans were re-evaluated using the AVI Planner, which utilizes AI for contouring and dose optimization. Key dosimetric parameters, such as the D98% and D2%, as well as OAR sparing metrics were compared between AVI and manual plans.
Results:
The AVI Planner achieved equivalent or superior target coverage compared to manual plans for IMRT and VMAT in head and neck cancer treatments. In nasopharyngeal cancer cases, AVI plans achieved significantly higher mean target coverage with D98% (67.76 Gy vs. 51.23 Gy, p < 0.001) and D2% (73.75 Gy vs. 54.35 Gy, p < 0.001). For laryngeal cancers, VMAT plans created with AVI enhanced OAR sparing, achieving lower spinal cord doses (43.83 Gy vs. 45.47 Gy), although not all observed differences reached statistical significance.
Conclusion:
The AVI Planner reduced plan generation time from 16 hours to 50 minutes, thereby facilitating expedited therapy initiation without compromising plan quality. Overall, the AVI Planner demonstrates substantial potential to optimize radiotherapy workflows and enhance treatment outcomes for complex head and neck cancer cases. Future prospective validation studies are recommended to validate these findings and evaluate the real-world impact on patient outcomes.