Barrett's Oesophagus Quantification via Depth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current endoscopic surveillance for Barrett's oesophagus is costly, time-consuming, and poorly adhered to due to its reliance on operator-dependent measurements, which often underestimate the extent of Barrett's epithelium, particularly ignoring islands of columnar-lined epithelium that can harbor dysplasia or cancer, and lack automated, quantitative assessment tools for risk stratification and treatment monitoring.
Innovation Solution
A method using machine learning techniques, such as feature pyramid networks and encoder-decoder frameworks, for quantifying the area of Barrett's oesophagus from video endoscopy images, performing depth estimation and segmentation to calculate geometrical measures like circumferential and maximal lengths, and providing 3D reconstructions for accurate risk assessment and therapy monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If operator-dependent measurements are used for Barrett's oesophagus surveillance, then the procedure can be performed with standard endoscopic equipment, but the measurement precision deteriorates due to operator variability and underestimation of Barrett's extent
Solution Approach 1:
The patent replaces manual operator-dependent visual estimation and measurement with an automated computer-based image analysis system. The system uses video endoscopy images processed by algorithms to automatically detect the squamo-columnar junction, calculate Prague C&M lengths, and quantify Barrett's extent, eliminating operator variability while maintaining clinical workflow.
Solution Approach 2:
The patent introduces an intermediary automated analysis system that acts as a bridge between the endoscopic imaging and clinical decision-making. This intermediary system processes video images, applies detection algorithms, and provides standardized measurements to guide surveillance intervals and treatment decisions, reducing direct operator dependency.
2Measurement precision
If islands of columnar-lined epithelium are excluded from measurement, then the Prague classification can be applied using standard methods, but the area of Barrett's epithelium is underestimated leading to inadequate risk stratification
Solution Approach 1:
The patent segments the Barrett's epithelium into distinct regions including both continuous segments and discrete islands of columnar-lined epithelium. The automated system identifies and measures each segment separately, then aggregates them to calculate the total Barrett's area, ensuring complete coverage without manual intervention for each island.
Solution Approach 2:
The patent creates a universal measurement system that handles all types of Barrett's presentations (continuous segments, discrete islands, circumferential involvement) through a single automated algorithm. This multi-functional approach eliminates the need for separate manual measurement protocols for different Barrett's morphologies.
3Productivity
If manual endoscopic surveillance is performed, then the procedure can be conducted with existing equipment and training, but the productivity deteriorates due to time-consuming measurements and poor adherence to surveillance protocols
Solution Approach 1:
The patent implements a self-service automated measurement system that performs detection, measurement, and quantification tasks automatically during endoscopy without requiring additional manual steps from the operator. The system processes video images in real-time or post-processing, generating standardized reports that reduce documentation time and improve protocol adherence.
Solution Approach 2:
The patent performs preliminary automated analysis of video endoscopy images to pre-calculate Barrett's measurements and risk stratification before clinical decision-making. This preliminary action provides ready-to-use quantitative data that accelerates surveillance planning and treatment decisions, reducing overall time consumption.
Data Source
AI summary
An area of Barrett's oesophagus in a subject's oesophagus is quantified from a video image signal representing a video image of the subject's oesophagus captured using a camera of an endoscope. Depth estimation on the frames to derive depth maps in respect of frames of the video image. Regions of the frames corresponding to an area of Barrett's oesophagus in the subject's oesophagus are segmented. A value of a geometrical measure of the area of Barrett's oesophagus is calculated using the depth map and segmented region in respect of at least one of the frames.