These scripts can be used to generate results for the Analytical and clinical validation of AIM-NASH: A Digital Pathology Tool for Artificial Intelligence-based Measurement of Nonalcoholic Steatohepatitis Histology manuscript
- The first set of analysis is used for computing true positive and false positive success rates,per pathologist, for the overlay validation. The run script for this analysis is:
run_overlay_analysis_manuscript.py- NOTE: The package rpy2 is needed to run this script to compute Wilson's score CIs.
- The second set of analysis is used for computing mean agreement rates for testing repeatability and reproducibility of the AIM-NASH algorithm. Additionally, one can also compute mean inter-reader agreement for manual pathologists. The run script for this analysis is:
run_av_analysis_manuscript.py - The final set of analysis is used for compare accuracy of the AIM-NASH and the AI-assisted workflows to that of the manual pathologists. The run script for this analysis is:
run_cv_analysis_manuscript.py