EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Automatic textual description of colorectal polyp features: explainable artificial intelligence

Ayla Thijssen iD, Ramon-Michel R Schreuder iD, Roger Fonollà, Quirine van der Zander iD, Thom Scheeve iD, Bjorn Winkens, Sharmila Subramaniam iD, Pradeep Bhandari, Peter de With, Ad Masclee, Fons van der Sommen iD, Erik Schoon

Endoscopy International Open · 01.05.2023 · 11 (05) · E513-E518

DOI: 10.1055/a-2071-6652 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Endoscopy International Open, erschienen am 01.05.2023. Verfasst von Ayla Thijssen und 11 weiteren Autor:innen. Beteiligte Einrichtung: Maastricht University Medical Center, Division of Gastroenterology and Hepatology, Maastricht, Netherlands. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 10 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: KWF Kankerbestrijding.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Computer-aided diagnosis systems (CADx) can improve colorectal polyp (CRP) optical diagnosis. For integration into clinical practice, better understanding of artificial intelligence (AI) by endoscopists is needed. We aimed to develop an explainable AI CADx capable of automatically generating textual descriptions of CRPs. For training and testing of this CADx, textual descriptions of CRP size and features according to the Blue Light Imaging (BLI) Adenoma Serrated International Classification (BASIC) were used, describing CRP surface, pit pattern, and vessels. CADx was tested using BLI images of 55 CRPs. Reference descriptions with agreement by at least five out of six expert endoscopists were used as gold standard. CADx performance was analyzed by calculating agreement between the CADx generated descriptions and reference descriptions. CADx development for automatic textual description of CRP features succeeded. Gwet’s AC1 values comparing the reference and generated descriptions per CRP feature were: size 0.496, surface-mucus 0.930, surface-regularity 0.926, surface-depression 0.940, pits-features 0.921, pits-type 0.957, pits-distribution 0.167, and vessels 0.778. CADx performance differed per CRP feature and was particularly high for surface descriptors while size and pits-distribution description need improvement. Explainable AI can help comprehend reasoning behind CADx diagnoses and therefore facilitate integration into clinical practice and increase trust in AI.

Autor:innen und Institutionen

  • Maastricht University Medical Center, Division of Gastroenterology and Hepatology, Maastricht, Netherlands.
  • Maastricht University, GROW School for Oncology and Reproduction, Maastricht, Netherlands.
  • Catharina Hospital, Division of Gastroenterology and Hepatology, Eindhoven, Netherlands.
  • Eindhoven University of Technology, Department of Electrical Engineering, Eindhoven, Netherlands.
  • Maastricht University, Department of Methodology and Statistics, Maastricht, Netherlands.
  • Maastricht University, CAPHRI, Care and Public Health Research Institute.

ORCID-Profile: Ayla Thijssen, Ramon-Michel R Schreuder, Quirine van der Zander, Thom Scheeve, Sharmila Subramaniam, Fons van der Sommen

Förderung

  • KWF Kankerbestrijding

Zitieren

APA

Thijssen, A., Schreuder, R. R., Fonollà, R., van der Zander, Q., Scheeve, T., Winkens, B., Subramaniam, S., Bhandari, P., de With, P., Masclee, A., van der Sommen, F., & Schoon, E. (2023). Automatic textual description of colorectal polyp features: explainable artificial intelligence. Endoscopy International Open, 11(05), E513-E518. https://doi.org/10.1055/a-2071-6652

Vancouver

Thijssen A, Schreuder RR, Fonollà R, van der Zander Q, Scheeve T, Winkens B, et al. Automatic textual description of colorectal polyp features: explainable artificial intelligence. Endoscopy International Open. 2023;11(05):E513-E518. doi:10.1055/a-2071-6652

BibTeX
@article{Thijssen2023,
  author = {Thijssen, Ayla and Schreuder, Ramon-Michel R and Fonollà, Roger and van der Zander, Quirine and Scheeve, Thom and Winkens, Bjorn and Subramaniam, Sharmila and Bhandari, Pradeep and de With, Peter and Masclee, Ad and van der Sommen, Fons and Schoon, Erik},
  title = {Automatic textual description of colorectal polyp features: explainable artificial intelligence},
  journal = {Endoscopy International Open},
  year = {2023},
  volume = {11},
  number = {05},
  pages = {E513--E518},
  doi = {10.1055/a-2071-6652}
}
Kostenlos starten →