Student Help – Computational Assistance – Human Oncology (deadline 1/1/22)

The Department of Human Oncology provides comprehensive radiation oncology services with dedication to improve the outcome for cancer patients using state-of-the-art treatment techniques combined with highly compassionate care. Our clinical focus is on each individual cancer patient and their families; however, our mission extends more broadly to include cutting-edge research and advanced training for the next generation of MD and PhD students to ensure cancer therapy improvements for the future.

This position requires some work to be performed in-person, onsite, at a designated campus work location. Some work may be performed remotely, at an offsite, non-campus work location.

This position will be up to 20 hours per week with a flexible schedule. The pay is $15/hour.

Deadline for application: 1/1/22

Qualifications: Computer sciences and/or image processing with interest and experience in AI/machine learning, in particular coding for the project

Knowledge and skills: Coding for AI and Matlab/machine learing. Completed CS AI courses. Programming languages: Python,R.

Position summary/job duties:
Specific Aim: To develop and implement an auto-contouring tool for female breast treatments.
Approach: From expertly contoured clinical cases, identify and compute key features for each volume of interest. The key features will include first order statistics. Workflow to be followed:
1. Compile CT images, contours and treatment plans for past 75-100 breast plans.
2. Run in-house and open source codes to extract features
3. Using leave one-out-cross comparison, generate test contours using  MIM’s auto-contouring
4. Validate auto-generated contours using difference metrics.
5. Use AI toolkit to develop a expert contours and thus generate a JCAC toolkit.
6. Validate the toolkit generated contours safety checklist.

More info and to apply: