Robotic Catheter Navigation With AI Confidence Risk Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic catheter navigation systems face risks due to out-of-domain input data, model limitations, and uncertainties in training data annotation, leading to potential hemorrhage or hematoma from catheter/guidewire pressure or punctures.
Innovation Solution
A colored dynamic roadmap with uncertainty quantification is generated using a trained segmentation network, and an AI agent navigates the catheter while evaluating confidence levels and optimizing trajectories to mitigate risks, with OOD detection and catheter status evaluation to prevent high-risk movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If robotic catheter navigation systems are used to assist surgeons, then surgeon training difficulty is reduced and radiation exposure is reduced, but risk and uncertainty increase leading to limited adoption
Solution Approach 1:
The system implements continuous feedback loops where the AI agent evaluates confidence levels of navigation actions in real-time. When confidence drops below thresholds, the system provides feedback to switch to manual control or re-plan the trajectory, creating a closed-loop control system that adapts to uncertainty dynamically
Solution Approach 2:
The navigation system dynamically adjusts control modes between autonomous and manual based on real-time confidence assessments. The AI agent can transition control authority dynamically during the procedure, allowing the system to be highly autonomous in confident situations while seamlessly handing off to human operators when uncertainty increases
2Productivity
If AI agent navigates with high autonomy, then navigation efficiency is improved, but risk from out-of-domain input data and model limitations increases
Solution Approach 1:
The system performs preliminary actions by pre-training the AI agent on extensive synthetic datasets that cover diverse anatomical variations and navigation scenarios. This pre-training prepares the AI to handle out-of-domain situations by exposing it to a wide range of potential inputs before actual deployment
Solution Approach 2:
The system implements multiple layers of safety checks and confidence threshold mechanisms that act as cushioning against AI errors. When the AI encounters uncertain situations or low-confidence predictions, these pre-established safety mechanisms trigger manual review or alternative navigation strategies, preventing potentially harmful autonomous actions
3Loss of time
If training data annotation uncertainty is high, then system development time is reduced, but navigation accuracy and reliability deteriorate
Solution Approach 1:
The system employs self-service mechanisms where the AI agent automatically identifies and flags regions of high annotation uncertainty during navigation. The system then adapts its behavior in these regions, either by requesting additional imaging, switching to alternative navigation strategies, or increasing safety margins, allowing it to compensate for annotation uncertainties without requiring manual review of every annotation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods for navigating a catheter in a patient using a robotic navigation system with risk management are provided. An input medical image of a patient is received. A trajectory for navigating a catheter from a current position to a target position in the patient is determined based on the input medical image using a trained segmentation network. One or more actions of a robotic navigation system for navigating the catheter from the current position towards the target position and a confidence level associated with the one or more actions are determined by a trained AI (artificial intelligence) agent and based on the generated trajectory and a current view of the catheter. In response to the confidence level satisfying a threshold, the one or more actions are evaluated based on a view of the catheter when navigated according to the one or more actions. The catheter is navigated from the current position towards the target position using the robotic navigation system according to the one or more actions based on the evaluation.