AI-Enabled Early-PET Acquisition via ML Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current nuclear medicine imaging techniques require lengthy scan times and patient waiting periods due to the statistical nature of isotope decay, which limits image quality and diagnostic efficiency, and are constrained by the limited administrable tracer amount for radiation safety and cost reasons.
Innovation Solution
A data processing system utilizing a pre-trained machine learning module to predict or estimate nuclear medicine images based on early projection data, allowing for reduced waiting and acquisition times, and iterative reconstruction to enhance image quality without extending traditional time periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If scan time is extended to improve image quality, then image quality improves, but patient comfort deteriorates and imaging throughput decreases
Solution Approach 1:
The machine learning module is pre-trained on historical PET data to learn the relationship between early projection data and final images. During acquisition, the model predicts later-stage images from early data, effectively performing the heavy computational work in advance so that shorter scan times can produce acceptable image quality without requiring extended patient exposure time.
Solution Approach 2:
Instead of collecting all projection data for a full scan duration, the system collects early projection data and uses the machine learning model to generate a predicted copy of what the final image would look like after complete acquisition. This predicted image serves as a surrogate for the full-length acquisition, reducing actual scan time while maintaining diagnostic quality.
2Measurement precision
If tracer amount is increased to improve image quality, then image quality improves, but radiation safety and cost constraints are violated
Solution Approach 1:
The machine learning model is trained in advance on historical data to predict image quality outcomes. This pre-computed knowledge allows the system to optimize tracer dosing by predicting which lower doses will still achieve diagnostic quality images, thereby reducing actual radiation exposure while maintaining image quality standards.
Solution Approach 2:
The system changes the parameter of tracer administration dose by using machine learning predictions to identify optimal dosing levels. Instead of using fixed high doses to ensure image quality, the model enables dose reduction by predicting which lower doses will produce sufficient image quality, thus lowering radiation exposure and cost.
3Productivity
If acquisition time is reduced to improve throughput, then imaging throughput improves, but image quality deteriorates
Solution Approach 1:
The machine learning module creates a virtual copy of the final image from incomplete early projection data. This predicted image represents what the complete acquisition would have produced, allowing the system to achieve high imaging throughput by using these predicted images instead of requiring full-duration acquisitions for each patient.
Solution Approach 2:
The system replaces the mechanical process of extended data collection with a computational prediction mechanism. Instead of physically extending the scan duration to gather more data, the machine learning model substitutes the missing data through intelligent prediction, enabling shorter scans to produce images of acceptable quality.
4Measurement precision
If iterative reconstruction is used to enhance image quality, then image quality improves, but computational complexity increases
Solution Approach 1:
Instead of performing computationally intensive iterative reconstruction from scratch, the machine learning model predicts the final image directly from early projection data. This approach replaces complex iterative algorithms with a trained neural network that can generate similar or better images faster, reducing computational complexity while maintaining or improving image quality.
Solution Approach 2:
The system substitutes traditional iterative reconstruction mechanics with machine learning prediction. The neural network model, trained on historical data, performs the image reconstruction task through pattern recognition rather than mathematical iteration, significantly reducing computational time and complexity while producing high-quality images.
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
Significantly shortens nuclear medicine imaging time periods, improving patient comfort and imaging throughput by enabling accelerated protocols that estimate later imagery from early data, reducing the likelihood of local minima trapping in reconstruction algorithms and enhancing image quality.
Implementation Method 1
a previously trained ('pre-trained') machine learning module, configured to predict or estimate, based on the first projection data or on the first image, a second projection data or a second image associable with a second waiting period
Implementation Method 2
Photons emitted (in opposite spatial direction) by proton annihilation with a local electron during the isotope decay are detected by, typically ring-shaped, detector arrangements outside the body
Implementation Method 3
The patient is administered a positron-emitting radio-pharmaceutical... during the isotope decay
Implementation Method 4
an iterative image reconstructor to reconstruct a first image in a given iteration step (i) from projection data acquired by a nuclear medicine imaging apparatus
Data Source
AI summary
Data processing systems (DPS) and related methods for nuclear medicine imaging. At an input interface (IN), first projection data (λ), or a first image (V) reconstructable from the first projection data, is received. The first projection data is associated with a first waiting period (ΔT*). The first waiting period indicates the time period from administration of a tracer agent to a start of acquisition by a nuclear medicine imaging apparatus (IA) of the projection date. A trained machine learning module (MLM) estimates, based on the first projection data (λ) or on the first image (V), a second projection data (λ′) or a second image (V′) associable with a second waiting period (ΔT), longer than the first waiting period (ΔT*). Nuclear imaging can thus be conducted quicker. Similar machine learning based data processing systems and related methods are also envisaged to reduce acquisition time periods or the time it takes to reconstruct imagery.