Machine Learning CT Workflow for Pre-Scan Radiation Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CT imaging systems expose patients to significant radiation, particularly in cases where conditions like pregnancy are undetected until after image acquisition, necessitating retrospective estimation and remedial actions.
Innovation Solution
A trained classification model is used to analyze a pre-scan radiograph to detect conditions such as pregnancy, allowing adjustment of CT scan parameters to avoid unnecessary radiation exposure by modifying or aborting the scan.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a CT image is acquired to detect internal structures and abnormalities, then diagnostic information is obtained, but radiation exposure to the patient increases significantly
Solution Approach 1:
The system performs preliminary detection of pregnancy status using a classification model trained on radiograph images before the CT scan is executed. This preliminary action allows the system to identify vulnerable conditions and adjust imaging parameters or abort the scan beforehand, preventing unnecessary radiation exposure while maintaining diagnostic capability when appropriate
2Object-affected harmful factors
If CT imaging parameters are adjusted to reduce radiation exposure, then patient safety is improved, but diagnostic image quality may deteriorate
Solution Approach 1:
The system applies different imaging strategies to different anatomical regions based on detected patient conditions. When pregnancy is detected, the system specifically adjusts parameters for the abdominal and pelvic regions to minimize fetal radiation exposure while maintaining diagnostic quality for other areas if clinically necessary
Solution Approach 2:
The system dynamically changes CT imaging parameters such as tube current, voltage, and scan range based on the detected patient condition. The classification model output triggers automated adjustment of these parameters to reduce radiation dose while preserving essential diagnostic information
3Object-affected harmful factors
If a machine learning model is integrated into the CT workflow to detect patient conditions, then radiation exposure is reduced, but system complexity increases
Solution Approach 1:
The system introduces a classification model as an intermediary component that processes radiograph images and provides condition detection results to the CT imaging system. This intermediary layer enables intelligent decision-making about scan execution and parameter adjustment without requiring complex integration throughout the entire imaging workflow
Solution Approach 2:
The classification model performs condition detection during the radiograph acquisition phase, which is a preliminary step before the main CT scan. This timing allows the system to make informed decisions about subsequent imaging actions without adding complexity to the critical scan execution phase
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables efficient detection of patient conditions prior to CT scanning, reducing radiation exposure by adapting the scan process accordingly, thereby safeguarding vulnerable areas.
Implementation Method 1
A trained classification model is used to analyze a pre-scan radiograph to detect conditions such as pregnancy
Implementation Method 2
a narrow beam of x-rays is emitted towards a patient and detected on an opposite side of the patient while the emitter and detector are rotated around the patient
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method include acquisition of a two-dimensional radiograph of a patient, input of the radiograph to a trained machine learning model to generate a classification, performance of a computed tomography scan of the patient based on the two-dimensional radiograph if the classification indicates that the patient does not have a first condition, and determination to modify the computed tomography scan if the classification indicates that the patient has the first condition.