Longitudinal PET-CT Tumor Segmentation for Treatment Response Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical imaging technologies struggle to accurately assess disease progression and treatment response using positron emission tomography (PET) and computed tomography (CT) scans, particularly in identifying lesions and determining metabolic activity changes over time.
Innovation Solution
A machine learning-enabled system applies a longitudinal segmentation model to update tumor masks from PET and CT scans across different timepoints, refining lesion identification and determining treatment responses based on metabolic activity and tumor volume changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional medical imaging analysis methods are used to assess disease progression and treatment response, then the process is simpler and faster, but the accuracy and reliability of lesion identification and metabolic activity assessment deteriorate
Solution Approach 1:
The patent introduces a longitudinal segmentation model as an intermediary between the input medical images and the final assessment results. This model acts as a mediator that processes multiple images across different timepoints and generates refined tumor masks, thereby improving measurement precision while managing system complexity through modular architecture
Solution Approach 2:
The patent replaces traditional manual or rule-based image analysis methods with a machine learning-based longitudinal segmentation model. This substitution transitions from mechanical/manual assessment to an automated intelligent system that can accurately track lesions and assess metabolic activity changes across multiple timepoints
2Reliability
If manual assessment methods are used for analyzing PET and CT scans, then the system is easier to operate, but the reliability and consistency of treatment response assessment deteriorate
Solution Approach 1:
The longitudinal segmentation model performs self-service by automatically processing multiple PET and CT scans across different timepoints, generating refined tumor masks and assessing treatment response without requiring manual intervention. This automation ensures consistent and reliable assessment while reducing operational complexity
Solution Approach 2:
The system incorporates feedback mechanisms by utilizing multiple images from different timepoints to refine tumor masks iteratively. The longitudinal segmentation model learns from historical data and adjusts its assessments based on temporal patterns, thereby improving reliability of treatment response evaluation
3Productivity
If single timepoint analysis is used for PET and CT scans, then the analysis is faster and requires less computational resources, but the ability to assess disease progression and treatment response deteriorates
Solution Approach 1:
The system performs preliminary action by collecting and storing multiple PET and CT scans across different timepoints before conducting the final analysis. This preparatory step ensures that all necessary temporal information is available, enabling comprehensive assessment of disease progression and treatment response without losing critical longitudinal data
Solution Approach 2:
The patent transitions from single-timepoint (2D) analysis to multi-timepoint (4D) analysis by incorporating the time dimension. The longitudinal segmentation model processes images across multiple timepoints, adding temporal depth to the analysis and enabling assessment of disease progression and treatment response over time
Data Source
AI summary
A method may include determining, based on a first positron emission tomography (PET) scan and a first computed tomography (CT) scan from a first timepoint, a first tumor mask corresponding to a first lesion present in the first PET scan and the first CT scan. A second tumor mask corresponding to a second lesion present in the second PET scan and the second CT scan may be determined based on the second PET scan and the second CT scan from a second timepoint. A longitudinal segmentation model may be applied to update, based on the first PET scan, the first CT scan, the second PET scan, and the second CT scan, each of the first tumor mask and the second tumor mask. A response to a treatment for a disease may be determined based on at least one of the first updated tumor mask and the second updated tumor mask. Related systems and computer program products are also provided.