Sequential Model for Medical Image Vessel and Lesion Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing medical image analysis systems fail to consistently analyze vessel structures and lesions due to independent task execution, neglecting interdependencies, and lack integration of human interventions to adjust analysis results.

Innovation Solution

A sequential model comprising a vessel extraction sub-model and a lesion analysis sub-model, trained independently or jointly, is applied to extract vessel location information and detect lesions, allowing for consistent results and easy incorporation of human edits to refine analysis tasks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple tasks are conducted independently in existing systems, then each task can be processed separately, but the analysis results become inconsistent with each other

Engineering Contradiction:

Improvetask processing efficiencyVSAvoidresult consistency

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent merges multiple independent tasks (vessel segmentation, lesion detection, stenosis analysis) into a unified sequential model where tasks are executed in a coordinated sequence. The vessel segmentation model outputs are directly fed into the lesion detection model, which then feeds into the stenosis analysis model, ensuring that all tasks share consistent vessel location information and produce coherent results.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The vessel segmentation task is performed as a preliminary action before lesion detection and stenosis analysis. By first establishing accurate vessel segmentation results, the subsequent tasks can rely on this pre-computed vessel information, ensuring consistency across all analysis stages without requiring redundant processing.

Inventive Principle:

Principle #10Preliminary action

2Device complexity

If lesion detection is performed without considering vessel information, then the detection process is simpler, but many false positives occur in non-vessel regions

Engineering Contradiction:

Improvedetection process complexityVSAvoidlesion detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The vessel segmentation model serves as an intermediary between the input medical image and the lesion detection model. It extracts and provides accurate vessel location information that acts as a spatial constraint for lesion detection, enabling the system to distinguish between actual lesions and false positives in non-vessel regions while maintaining manageable complexity through modular architecture.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If vessel segmentation is conducted without following vessel centerline or considering existing lesions, then the segmentation process is faster, but broken vessel segmentation masks are produced due to severe lesion stenosis

Engineering Contradiction:

Improvesegmentation processing speedVSAvoidvessel segmentation accuracy

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The system implements feedback by using the outputs of upstream tasks (vessel segmentation and lesion detection) to inform and refine downstream tasks (stenosis analysis). The lesion detection results provide feedback about problematic regions that may indicate severe stenosis, allowing the vessel segmentation to be adjusted or interpreted in context, preventing broken masks while maintaining processing efficiency through the sequential framework.

Inventive Principle:

Principle #23Feedback

4Extent of automation

If existing systems do not include human intervention capability, then the system operation is more automated, but the system cannot adjust analysis results based on expert feedback

Engineering Contradiction:

Improvesystem automation levelVSAvoidhuman intervention flexibility

Core Design Contradiction:

Extent of automationVSAdaptability or versatility

Solution Approach 1:

The system incorporates dynamic adaptability by allowing human experts to intervene at any stage of the sequential analysis process. Users can modify vessel segmentation masks, adjust lesion detection parameters, or correct stenosis measurements, and the system dynamically reprocesses subsequent tasks based on these modifications, combining automation with flexible expert control.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS12086981B2Methods and systems for computer-assisted medical image analysis using sequential model

Publication Date: 2024.09.10 SHENZHEN KEYA MEDICAL TECH CORP

AI summary

Embodiments of the disclosure provide systems and methods for analyzing a medical image containing a vessel structure using a sequential model. An exemplary system includes a communication interface configured to receive the medical image and the sequential model. The sequential model includes a vessel extraction sub-model and a lesion analysis sub-model. The vessel extraction sub-model and the lesion analysis sub-model are independently or jointly trained. The exemplary system also includes at least one processor configured to apply the vessel extraction sub-model on the received medical image to extract location information of the vessel structure. The at least one processor also applies the lesion analysis sub-model on the received medical image and the location information extracted by the vessel extraction sub-model to obtain a lesion analysis result of the vessel structure. The at least one processor further outputs the lesion analysis result of the vessel structure.