Overview of Model for Assessing the value of AI (MAS-AI)

* The above figure is from the upcoming article “Updating the Model for ASsessing the value of AI (MAS-AI): Expanding beyond imaging to broader clinical AI use”. The article is currently in peer review in International Journal of Technology Assessment in Health Care.

Legend version for Fig : The updated MAS-AI retains three parts: two domain-based steps (Step 1 with four domains; Step 2 with six domains) and four process factors. Step 1 functions as a feasibility or pre-screening (i.e. a “go / no go”) and must be completed before Step 2. The tool is iterative, and the process factors apply throughout; completing Step 1 yields an “early MAS-AI,” while completing both steps constitutes a full MAS-AI assessment. Within the two steps, the order of domains has no particular significance. 

MAS-AI:

Through a series of workshops and extensive collaborative work, we have updated the first part of EUcanScreen’s MAS-AI so that it is now tailored to screening. This has been an exciting process, resulting in an updated version of MAS-AI entitled “Updating the Model for ASsessing the value of AI (MAS-AI): Expanding beyond imaging to broader clinical AI use.” The article is currently under peer review in the International Journal of Technology Assessment in Health Care.

The next step is to test the updated MAS-AI in practice through eight confirmed pilot cases. These cases cover several clinical areas, including skin cancer detection, lung cancer, breast cancer, and prostate cancer.

Author

Iben Fasterholdt