DESCRIPTION (provided by applicant): Radiation therapy has been shown to be a cost-effective therapy for curative cancer treatments in low- and middle-income countries (LMICs). For many malignancies, including locoregionally advanced carcinomas of the uterine cervix and head/neck, radiation therapy is the only effective treatment modality; there is an extremely poor prognosis for these patients, with no reasonable curative alternative if radiation therapy is not available It is estimated that by 2020 there will be a deficit in radiation therapy staff in LMIICsof more than 50,000 full-time equivalent staff, including almost 10,000 medical physicists. Our central hypothesis is that by fully automating the treatment planning process, we can reduce the shortfall of needed medical physicists by close to 50% (5000 people - equivalent to more than 20,000 training-years) while providing robust, high-quality radiation treatment plans in LMICs. The long-term goal of our project is to develop and clinically implement a fully automated treatment planning system ("Radiation Planning Assistant") that meets the following criteria: a. A person educated to the level of a high-school diploma can be fully trained (using video and online tutorials) to use the system in one half day. This would address staff retention and re- training issues. b. Once trained, treatment plans can be created by the treatment planner from scratch, including transfer to the local record-and-verify software, in less than 30 minutes. This would address patient throughput. c. All steps in the process include automated QA tools and purpose-designed plan documentation. This would address the need for effective QA to ensure safe patient treatments (especially given staff inexperience). We will focus on curative radiation treatment of breast, head/neck, and uterine cervical cancers, as these demonstrate improved survival with radiation therapy and account for the majority of malignancies in LMICs. The Radiation Planning Assistant will include automatic QA of all planning tasks by, in most cases, repeating the tasks using a parallel, independent algorithm. For example, auto contouring of normal structures will be carried out twice, the results compared, and any discrepancies flagged to the user. During the initial testing and deployment, we will partner with institutions in South Africa and The Philippines. After that we propose to deploy to up to 40 sites in Southeast Asia and Sub-Saharan Africa. Metrics for success of the overall project will include reductions in wait time and an increase in the number of patients who receive treatment plans.