OCT Tissue Margin Analysis With Two-Tier Artifact Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing surgical oncology practices face challenges in determining whether cancerous lesions have been entirely removed during a surgical procedure due to labor-intensive and time-consuming microscopic pathologic evaluation of excised tissue specimens, which delays the availability of margin results, potentially necessitating additional surgeries and treatments.
Innovation Solution
A two-tiered prediction model system is employed for intraoperative tissue specimen analysis using OCT imaging, where a first model reduces the data set to minimize false negatives and a second model enhances accuracy by identifying false positives, allowing real-time, high-precision margin assessment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single high-accuracy prediction model is used to analyze all tissue images, then prediction accuracy is improved, but computational time and processing speed deteriorate
Solution Approach 1:
The patent divides the tissue image analysis into two distinct stages using two different prediction models. The first model performs initial classification of image patches into positive and negative categories, while the second model provides refined analysis only for ambiguous cases. This segmentation of the analysis process allows the system to achieve high overall accuracy while maintaining fast processing speeds by avoiding the computational intensity of a single comprehensive high-accuracy model applied to all images.
2Measurement precision
If all image patches are analyzed with high computational intensity, then prediction accuracy is improved, but analysis time deteriorates
Solution Approach 1:
The patent applies partial action by using a computationally intensive second prediction model only for a subset of image patches that the first model identified as ambiguous or potentially positive. Rather than applying high computational intensity to all patches, the system selectively applies it only where needed, thereby reducing overall analysis time while maintaining high accuracy for critical cases.
Solution Approach 2:
The first prediction model performs preliminary classification of all image patches into positive and negative categories before the second model analyzes ambiguous cases in detail. This preliminary action filters out clearly negative patches from intensive analysis, reducing the total computational load and analysis time while ensuring that potentially positive cases receive thorough examination.
3Reliability
If false positive predictions are minimized, then prediction reliability is improved, but the number of false negative predictions may increase
Solution Approach 1:
The patent implements a feedback mechanism where the first prediction model's output is fed into the second prediction model for refined analysis of ambiguous cases. The second model's results then feed back into the final diagnosis, allowing the system to correct both false positives and false negatives through iterative refinement. This feedback loop improves overall reliability without sacrificing the ability to detect true positive cases.
Data Source
AI summary
Systems and methods for tissue specimen analysis. Methods for tissue specimen analysis may include: retrieving a primary image data set including a plurality of images representing a tissue specimen margin; generating a reduced data set representing images having suspected artifacts based on a first detection model and the primary image set, the first detection model trained based on pathology-confirmed images and for prioritizing reducing false negative identification of artifacts while minimizing training penalization for false positive identification of artifacts; generating a prediction data set representing a subset of the reduced data set based on a second detection model and the reduced data set; and generating a signal representing the prediction data set for displaying one or more images predicting a true positive identification of a suspected artifact.