You might have a look at the following paper:

McIntosh, C., Welch, M., McNiven, A., Jaffray, D. A., & Purdie, T. G. (2017). Fully automated treatment planning for head and neck radiotherapy using a voxel-based dose prediction and dose mimicking method. Physics in Medicine & Biology, 62(15), 5926.

They use a probabilistic method but cite ML-based ones, too. Knowledge-based automatic approaches are, depending on a therapy, more popular, check them as well.

Deep Neural Decision Forests.

Bayesian networks, they can help to show clearly to the patient the potential causes and consequences of the problem. You could also predict the treatment results.

Check this article:

Using a Bayesian Network to Predict L5/S1 Spinal Compression Force from Posture, Hand Load, Anthropometry, and Disc Injury Status by Richard E. Hughes

Dear Mohammad

take a look at Generative Adversarial Network, GANs have shown a great success in training and generating realistic-looking data. have a look at the following paper : Preprint Automated Treatment Planning in Radiation Therapy using Gene...