Download PDF
CC BY 4.0 · Journal of Coloproctology 2024; 44(S 01): S1-S138
DOI: 10.1055/s-0045-1808877
Doença Inflamatória Intestinal
ID – 141843
Open Topics (oral presentation)

ARTIFICIAL INTELLIGENCE AND MINIMALLY INVASIVE ASSESSMENT OF INFLAMMATORY ACTIVITY: A STUDY ON THE DEVELOPMENT AND CORRELATION OF INDICES WITH A CONVOLUTIONAL NEURAL NETWORK MODEL IN PANENDOSCOPY

Maria João Almeida
1   Hospitalar Universitário São João, Porto, Portugal
2   Centro Treino WGO, Porto, Portugal
,
Miguel Mascarenhas
1   Hospitalar Universitário São João, Porto, Portugal
2   Centro Treino WGO, Porto, Portugal
3   Universidade do Porto, Porto, Portugal
,
Pedro Cardoso
1   Hospitalar Universitário São João, Porto, Portugal
2   Centro Treino WGO, Porto, Portugal
,
João Afonso
1   Hospitalar Universitário São João, Porto, Portugal
2   Centro Treino WGO, Porto, Portugal
,
Tiago Ribeiro
1   Hospitalar Universitário São João, Porto, Portugal
2   Centro Treino WGO, Porto, Portugal
,
João Ferreira
3   Universidade do Porto, Porto, Portugal
,
Guilherme Macedo
1   Hospitalar Universitário São João, Porto, Portugal
2   Centro Treino WGO, Porto, Portugal
3   Universidade do Porto, Porto, Portugal
› Author Affiliations
› Further Information
Also available at
 
 PDF Download Permissions and Reprints

    Introduction In patients with Crohn's Disease (CD), capsule endoscopy (CE) is an important method for assessing mucosal inflammatory activity. Traditionally, inflammation is evaluated through the combination of analytical values and validated scores, such as the Lewis score (LS), the Capsule Endoscopy Crohn’s Disease Activity Index (CECDAI), and ELIAKIM. Recent developments in artificial intelligence (AI) have enabled the automatic selection of the most relevant images in CE.

    Objective This study aimed to develop an automated scoring system to objectively classify inflammation in CE images.

    Methods A retrospective study was conducted on panenteric CE videos (PillCam Crohn’s) performed between 09/2020 and 01/2023, and the LS, CECDAI, and ELIAKIM scores were calculated. An automated score based on a convolutional neural network was developed, encompassing the percentage of positive frames selected by the algorithm for both the small intestine and the colon separately. The AI-generated score (AIS) was then correlated with clinical data and the validated scores.

    Results The study included 61 patients. The median scores were LS 225 (0-6006), CECDAI 6 (0-33), ELIAKIM 4 (0-38), and SB_AIS 0.5659 (0-29.45). A strong correlation was found between SB_AIS and the validated scores: LS, CECDAI, and ELIAKIM (Spearman's r = 0.751, r = 0.707, r = 0.655, p = 0.001). Strong correlations were also found between LS and ELIAKIM (r = 0.768, p = 0.001), CECDAI and LS (r = 0.854, p = 0.001), and CECDAI and ELIAKIM (r = 0.827, p = 0.001).

    Conclusion The results of our study demonstrated a strong correlation between the AI-generated score and the clinically validated scores, suggesting that it could be an objective and efficient way to assess inflammatory activity in patients with CD. The initial results of our study provide a solid foundation for the future development of a score that could correlate with prognostic factors and assist in the management of these patients.


    #

    No conflict of interest has been declared by the author(s).

    Publication History

    Article published online:
    25 April 2025

    © 2025. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution 4.0 International License, permitting copying and reproduction so long as the original work is given appropriate credit (https://creativecommons.org/licenses/by/4.0/)

    Thieme Revinter Publicações Ltda.
    Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil

      Permissions and Reprints