NaviCam ProScan AI-Assisted Reading Software

NaviCam ProScan represents a breakthrough in capsule endoscopy workflow, applying deep convolutional neural networks to automatically detect potentially significant lesions. As the first FDA-cleared AI system for GI capsule endoscopy reading assistance, ProScan addresses the historically time-intensive review process while improving detection sensitivity.

Product Overview

Traditional capsule endoscopy generates thousands of images per examination, requiring physicians to maintain sustained attention during reviews typically lasting 60-90 minutes. This cognitive burden limits throughput and introduces the risk of missed findings due to reader fatigue.

ProScan fundamentally transforms this workflow through automated first-pass screening. The deep neural network analyzes all captured images, identifying frames likely to contain clinically relevant findings. Physicians then review only the AI-selected subset, focusing expertise on interpretation rather than exhaustive frame-by-frame scanning.

Clinical validation demonstrates mean reading time reduction from approximately 90 minutes to under 6 minutes, while detection sensitivity increases from 76.9% to 99.9% for small bowel bleeding lesions.

Key Features

NaviCam ProScan is an AI reading system by Ankon, using deep convolutional networks to cut capsule endoscopy review time from 90 minutes to under 6 minutes.

• Deep Neural Network Architecture: Purpose-built convolutional network trained on extensive capsule endoscopy datasets
• Automated Lesion Flagging: Identifies frames containing potential abnormalities including bleeding sources, vascular malformations, and mucosal lesions
• Intelligent Image Selection: Reduces review burden by presenting only relevant image subsets
• ESView Integration: Direct operation within NaviCam’s existing software platform
• Quality Control Reporting: Automated assessment of examination completeness and image quality

Technical Specifications

NaviCam ProScan is an FDA De Novo-cleared system by Ankon, achieving 99.9% detection sensitivity while operating within the ESView software platform.

Clinical Applications

ProScan’s primary application accompanies the NaviCam Small Bowel system for suspected obscure GI bleeding. The software classifies detected lesions according to the Saurin classification system, prioritizing findings by bleeding potential.

Multi-Center Validation: A prospective study across 14 European centers enrolled 137 patients with suspected small bowel bleeding. AI-assisted reading demonstrated non-inferiority to standard reading for detecting P1 and P2 lesions, with significantly reduced mean reading time.

The combination of improved sensitivity and reduced workload enables:

• Higher patient throughput without compromising diagnostic quality
• Reduced reader fatigue and associated error risk
• More efficient utilization of specialist physician time
• Standardized detection protocols across different readers

Beyond bleeding detection, the AI quality control module assesses examination completeness and image quality, generating standardized reports that enable quantitative evaluation of operator technique and system performance.

Regulatory Status

FDA granted De Novo classification in December 2023, establishing ProScan as the first AI-assisted reading system approved for capsule endoscopy in the United States. This regulatory pathway applies to novel low-to-moderate risk devices without existing predicates, recognizing ProScan’s novel approach while establishing controls for this new device category.

The Nature journal family has cited both the NaviCam magnetically controlled system (2013) and ProScan AI software (2023) as significant milestones in capsule endoscopy robot development history.

Frequently Asked Questions

Does ProScan replace physician interpretation?

ProScan functions as a decision support tool, not an autonomous diagnostic system. The AI performs automated first-pass screening, but final interpretation remains with the physician who reviews AI-selected images and makes clinical determinations. The system enhances rather than replaces physician expertise.

What lesion types can ProScan detect?

ProScan primarily targets findings associated with small bowel bleeding, including angioectasias, ulcers, erosions, and tumors. The system classifies findings according to established bleeding potential categories. Detection capabilities continue to expand through ongoing algorithm development.

Is special training required to use ProScan?

ProScan integrates within the existing ESView software environment familiar to NaviCam users. The interface presents AI-selected images through the same viewer used for standard reading. Physicians should understand the system’s capabilities and limitations but do not require extensive additional training beyond standard capsule endoscopy competencies.