Development of an Automated Tool for the Estimation of Histological Remission in Ulcerative Colitis Using Single-Wavelength Endoscopy Technology - Summary - MDSpire

Development of an Automated Tool for the Estimation of Histological Remission in Ulcerative Colitis Using Single-Wavelength Endoscopy Technology

  • By

  • Pieter Sinonquel

  • Matthias Lenfant

  • Tom Eelbode

  • Hiroki Watanabe

  • Belian Callaerts

  • Peter Bossuyt

  • Bram Verstockt

  • João Pedro Guedelha Sabino

  • Gert De Hertogh

  • Frederik Maes

  • Séverine Vermeire

  • Raf Bisschops

  • November 27, 2024

  • 0 min

Share

Objective:

To assess the accuracy of a computer-aided diagnosis (CAD) system for estimating histological activity in ulcerative colitis (UC) using white light endoscopy (WLE) and single-wavelength endoscopy (SWE), highlighting its potential clinical significance.

Key Findings:
  • SWE-CAD outperformed WLE-CAD with a mean sensitivity of 88.0% vs 73.9% and diagnostic accuracy of 83.3% vs 67.5%, with p-values indicating statistical significance.
  • Consecutive training of SWE-CAD on the entire dataset achieved 95.2% accuracy, 96.4% sensitivity, and 92.9% specificity, reinforcing the model's reliability.
Interpretation:

SWE provides enhanced capillary visibility, allowing for more accurate detection of histological remission in UC compared to WLE, with significant implications for improving patient outcomes.

Limitations:
  • Study limited to a single center, which may affect generalizability and applicability to broader populations.
  • Potential biases in histological scoring due to subjective interpretation, which could influence results.
Conclusion:

Automated CAD using SWE demonstrates high diagnostic accuracy for assessing histological remission in UC, potentially reducing inter-reader variability and suggesting avenues for future research.

Original Source(s)

Development of an Automated Tool for the Estimation of Histological Remission in Ulcerative Colitis Using Single-Wavelength Endoscopy Technology

  • by Pieter Sinonquel, Matthias Lenfant, Tom Eelbode, Hiroki Watanabe, Belian Callaerts, Peter Bossuyt, Bram Verstockt, João Pedro Guedelha Sabino, Gert De Hertogh, Frederik Maes, Séverine Vermeire, Raf Bisschops

  • November 27, 2024

Related Content

Diagnostic utility of chest wall vessel involvement sign on ultra-high-resolution CT for primary lung cancer infiltrating the chest wall

Who Decides When a Doctor Steps Back?

Cognitive decline, credentialing, and why “I’ll know” often isn’t enough.

Application and prospect of artificial intelligence in diagnostic imaging of prostate cancer