EUVIMEDEUVIMED entdecken
Crossref

RADIÔMICA E INTELIGÊNCIA ARTIFICIAL EM TOMOGRAFIA COMPUTADORIZADA PARA PROGNÓSTICO DE APENDICITE AGUDA COMPLICADA: IMPACTO NA INDICAÇÃO CIRÚRGICA E ESTRATÉGIA

Isadora André de Camargo Pacheco, Guilherme Silva Moreira, Matheus de Silvio Cobucci Cirino

LUMEN ET VIRTUS · 10.08.2026 · 17 (63) · e13994

DOI: 10.56238/levv17n63-021 ↗

Kurz gefasst

Fachartikel in der Zeitschrift LUMEN ET VIRTUS, online erschienen am 10.08.2026. Verfasst von Isadora André de Camargo Pacheco und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Differentiating uncomplicated from complicated acute appendicitis before treatment is essential because severity classification influences the feasibility of nonoperative management, urgency of appendectomy, antimicrobial therapy, and operative planning. Conventional computed tomography interpretation may fail to identify a substantial proportion of gangrenous or perforated cases, creating interest in quantitative radiomics and artificial intelligence-assisted classification.

Objective

To evaluate the diagnostic performance, methodological quality, and potential surgical applications of computed tomography-based radiomics and artificial intelligence models for predicting complicated acute appendicitis.

Methods

PubMed, Scopus, Web of Science, the Cochrane Library, Latin American and Caribbean Health Sciences Literature, ClinicalTrials.gov, and the International Clinical Trials Registry Platform were searched for recent studies evaluating computed tomography-based severity classification. Diagnostic performance, model construction, validation, risk of bias, and potential effects on treatment selection were qualitatively synthesized because methodological heterogeneity precluded meta-analysis. Results and Discussion: Four studies met the strict eligibility criteria, including two studies evaluating conventional imaging or physician classification and two radiomics-based investigations. Conventional interpretation demonstrated high recognition of uncomplicated disease but limited sensitivity for complicated appendicitis. A single-center combined radiomics model achieved an area under the receiver operating characteristic curve of 0.817 in its test cohort, whereas a multicenter clinical-radiomics random forest model achieved values of 0.84 in testing and 0.76 in external validation. Radiomics therefore improved discrimination over conventional computed tomography assessment, but retrospective designs, inconsistent definitions, manual segmentation, limited external validation, and uncertain decision thresholds reduced confidence in immediate clinical implementation.

Conclusion

Computed tomography radiomics and artificial intelligence may improve preoperative identification of complicated acute appendicitis and support decisions regarding urgent surgery, antibiotic-first treatment, drainage, and operative preparation. These tools should currently complement rather than replace multidisciplinary clinical and radiological assessment.

Zitieren

APA

Pacheco, I. A. D. C., Moreira, G. S., & Cirino, M. D. S. C. (2026). RADIÔMICA E INTELIGÊNCIA ARTIFICIAL EM TOMOGRAFIA COMPUTADORIZADA PARA PROGNÓSTICO DE APENDICITE AGUDA COMPLICADA: IMPACTO NA INDICAÇÃO CIRÚRGICA E ESTRATÉGIA. LUMEN ET VIRTUS, 17(63), e13994. https://doi.org/10.56238/levv17n63-021

Vancouver

Pacheco IADC, Moreira GS, Cirino MDSC. RADIÔMICA E INTELIGÊNCIA ARTIFICIAL EM TOMOGRAFIA COMPUTADORIZADA PARA PROGNÓSTICO DE APENDICITE AGUDA COMPLICADA: IMPACTO NA INDICAÇÃO CIRÚRGICA E ESTRATÉGIA. LUMEN ET VIRTUS. 2026;17(63):e13994. doi:10.56238/levv17n63-021

BibTeX
@article{Pacheco2026,
  author = {Pacheco, Isadora André de Camargo and Moreira, Guilherme Silva and Cirino, Matheus de Silvio Cobucci},
  title = {RADIÔMICA E INTELIGÊNCIA ARTIFICIAL EM TOMOGRAFIA COMPUTADORIZADA PARA PROGNÓSTICO DE APENDICITE AGUDA COMPLICADA: IMPACTO NA INDICAÇÃO CIRÚRGICA E ESTRATÉGIA},
  journal = {LUMEN ET VIRTUS},
  year = {2026},
  volume = {17},
  number = {63},
  pages = {e13994},
  doi = {10.56238/levv17n63-021}
}
Kostenlos starten →