Endoscopy Video Feature Enhancement Platform for Adenoma Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional endoscopy systems are limited by variability in physician visualization capabilities, inefficiencies in documentation, and lack of objective real-time feedback, leading to disparities in adenoma detection and increased costs due to reliance on manual biopsies for confirmation.
Innovation Solution
An endoscopy video feature enhancement platform that applies a previously trained detection model to the endoscopy video stream to identify abnormalities, overlays visual indicators, and uses speech-to-text conversion for meta-data generation, providing objective feedback and efficient documentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional endoscopy systems rely on physician visualization capabilities for abnormality detection, then the system is simple and cost-effective, but detection accuracy and consistency vary significantly across different physicians
Solution Approach 1:
The patent introduces an artificial intelligence detection model as an intermediary between the endoscopy video stream and the physician. This AI model processes the video feed in real-time, automatically detecting and highlighting abnormal regions, thereby eliminating the variability in human visualization capabilities while maintaining system simplicity through software-based enhancement rather than complex hardware modifications.
2Productivity
If physicians manually document endoscopy procedure findings, then the system requires minimal additional equipment, but documentation efficiency is low and time-consuming
Solution Approach 1:
The system automatically generates documentation by extracting detection results and procedural data from the AI model's analysis of the endoscopy video stream. This self-service capability eliminates the need for manual documentation by physicians, as the system autonomously captures detection accuracy metrics, abnormality locations, and procedure outcomes, thereby significantly improving productivity and reducing time loss.
3Reliability
If traditional biopsies are used to confirm abnormality detection, then diagnostic reliability is high, but procedural costs increase significantly
Solution Approach 1:
The AI detection model provides real-time feedback to the physician during the endoscopy procedure by highlighting suspected abnormal regions with confidence scores. This immediate feedback allows physicians to make more informed decisions about biopsy necessity, reducing the number of unnecessary biopsies while maintaining diagnostic reliability through the AI's continuous monitoring and detection capabilities throughout the procedure.
4Measurement precision
If endoscopy procedures are performed without real-time feedback, then the procedure is simpler and faster, but adenoma detection rates vary and cancers may be missed
Solution Approach 1:
The AI detection model performs preliminary analysis of the endoscopy video stream in real-time, pre-identifying and marking abnormal regions before the physician completes the examination. This preliminary action ensures that no adenomas or potential cancers are missed, even in complex or time-consuming procedures, by providing continuous automated surveillance that enhances detection precision without requiring complex additional hardware.
Data Source
AI summary
An endoscopy video feature enhancement platform (EVFEP) is connected to the output of any type endoscope system, and inputs and captures the output video. The video is visually augmented live with indicators of possible polyp detection and localization, polyp attributes, and procedure metrics, based on the collective learning of the output results of many different types of endoscopy systems on a large scale. An artificial intelligence model is trained on confirmed polyp detection previously determined by this and other EVFEP devices used with many different types of endoscope systems on a large scale. Augmented video, images and automatically generated short video clips of key procedure segments are passed to a reporting system, and supplemented with meta data.