Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 19, 2026

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
07:13

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

Which Method Best Predicts Postoperative Complications: Deep Learning, Machine Learning, or Conventional Logistic

Ryosuke Fukuyo1, Masanori Tokunaga1,2, Hiroyuki Yamamoto3

  • 1Department of Gastrointestinal Surgery Institute of Science Tokyo Tokyo Japan.

Annals of Gastroenterological Surgery
|May 18, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Risks and Benefits of Feeding Enterostomy Creation During Minimally Invasive Esophagectomy: A Propensity-Weighted Analysis Using the Japanese National Clinical Database.

Annals of gastroenterological surgery·2026
Same author

Real-World Data of Retroperitoneal Tumor Surgeries Performed by Gastroenterological Surgeons in Japan: Analysis Based on the National Clinical Database.

Annals of gastroenterological surgery·2026
Same author

Annual Report of the 2022 National Clinical Database: Decade-Long Trends and Current Status of Gastroenterological Surgery in Japan.

Annals of gastroenterological surgery·2026
Same author

A simple morphometric algorithm based on nuclear atypia features of invasive breast carcinoma: relationship with nuclear grade, Ki-67 level, and hormone receptor status.

Breast cancer (Tokyo, Japan)·2026
Same author

Preoperative carbohydrate antigen 19-9 as a prognostic biomarker in colorectal cancer with synchronous peritoneal metastasis: multicentre cohort study.

BJS open·2026
Same author

Clinicopathological features and cancer-specific survival in young patients with ulcerative colitis-associated colorectal neoplasia: a nationwide cohort study in Japan.

Inflammatory bowel diseases·2026
This summary is machine-generated.

Deep learning shows promise for predicting surgical complications, especially with complex data like images. However, it sometimes underperforms on standard clinical data and requires further development for better accuracy and interpretability.

Area of Science:

  • Medical Informatics
  • Surgical Oncology
  • Artificial Intelligence in Medicine

Background:

  • Accurate prediction of postoperative complications is crucial for surgical care.
  • Deep learning (DL) is increasingly explored for predictive modeling in medicine.
  • Gastroenterological surgery is a key area for applying predictive analytics.

Purpose of the Study:

  • To compare the performance of logistic regression, machine learning (ML), and deep learning (DL) models in predicting postoperative complications in gastroenterological surgery.
  • To evaluate the impact of different data types (tabular, image, time-series) on predictive model accuracy.
  • To discuss the limitations and potential of DL in clinical decision-making.

Main Methods:

  • Review of existing studies comparing logistic regression, ML (e.g., random forests, gradient boosting), and DL models.
Keywords:
deep learningpostoperative complicationsprediction modelssurgical complications

Related Experiment Videos

Last Updated: May 19, 2026

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
07:13

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

  • Analysis of studies utilizing various data types, including tabular clinical data (e.g., National Clinical Database) and complex data (images, time-series).
  • Discussion of model performance metrics and interpretability challenges.

    • Main Results:

    • Deep learning models show potential, outperforming traditional methods in some studies, particularly with image and time-series data.
    • Machine learning methods like random forests and gradient boosting sometimes outperform deep learning on tabular clinical datasets.
    • Deep learning can underperform on tabular data and faces "black-box" limitations, impacting interpretability.

    Conclusions:

    • Deep learning demonstrates potential for improving postoperative complication prediction, especially when integrating multimodal data (images, time-series).
    • Further research is needed to address deep learning's limitations with tabular data and its "black-box" nature.
    • Integrating complex data types and improving interpretability are key to enhancing deep learning's clinical utility and decision support.