Radiologist-Assisted Machine Learning Diagnostic System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CAD systems in medical imaging lack optimization for clinical impact and quantification of performance efficiency, particularly in diagnostic radiology, where AI and machine learning are used for detecting conditions like breast cancer.
Innovation Solution
A method and apparatus that continuously updates a training dataset using a medical image diagnostic computer with machine-learning capability, employing a three-dimensional cursor to select sub-volumes for human-generated and machine-generated analyses, resolving disagreements, and incorporating patient-specific data to improve diagnostic accuracy and reporting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CAD systems are used to detect conditions in medical imaging, then diagnostic support is provided, but the systems lack optimization for clinical impact and quantification of performance efficiency
Solution Approach 1:
The system implements continuous feedback loops where machine learning models generate predictions, radiologists provide annotations and corrections, and these corrections are fed back to retrain and improve the models. This closed-loop feedback mechanism optimizes clinical impact by ensuring the system learns from real-world diagnostic outcomes and adapts to improve performance over time.
Solution Approach 2:
The system performs self-improvement through automated retraining processes where the machine learning model automatically updates its parameters and structure based on accumulated training data and radiologist corrections. This self-service capability allows the system to continuously optimize its diagnostic accuracy without requiring manual reconfiguration, thereby improving both reliability and adaptability.
2Reliability
If machine learning models are trained on static datasets, then initial diagnostic capability is achieved, but the systems lack continuous improvement through radiologist feedback
Solution Approach 1:
The system maintains continuous improvement by implementing ongoing training processes where the machine learning model is repeatedly trained on updated datasets that incorporate radiologist annotations. This continuous training action ensures the model remains current and improving, transforming the static dataset limitation into a dynamic, continuously evolving knowledge base that enhances diagnostic accuracy over time.
Data Source
AI summary
A computerized medical diagnostic system uses a training dataset that is updated based on reports generated by a radiologist. AI and/or CAD is used to make an initial determination of no finding, finding, or diagnosis based on the training dataset. Normal results with a high confidence of no finding are not reviewed by the radiologist. Low confidence results, findings, and diagnosis are reviewed by the radiologist. The radiologist generates a report that associates terminology and weighting with marked 3D image volumes. The report is used to update the training dataset.