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: Apr 4, 2026

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.7K

Towards Interactive and Interpretable Image Retrieval-Based Diagnosis: Enhancing Brain Tumor Classification with LLM

Pranav Manjunath1, Brian Lerner1, Timothy Dunn1

  • 1Duke University, Durham, NC 27005, USA.

Artificial Intelligence in Medicine. Conference on Artificial Intelligence in Medicine (2005- )
|April 3, 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

A multi-site evaluation of clinical pharmacist time, motion, interventions and interruptions.

International journal of clinical pharmacy·2025
Same author

Preimplantation genetic testing and disorders of placental implantation: a systematic review and meta-analysis.

Journal of assisted reproduction and genetics·2025
Same author

Variation in the relationship between fasting glucose and HbA1c: implications for the diagnosis of diabetes in different age and ethnic groups.

BMJ open diabetes research & care·2024
Same author

Impaired bone morphogenetic protein (BMP) signaling pathways disrupt decidualization in endometriosis.

Research square·2023
Same author

<b>In vitro fertilization outcomes in VACTERL association (vertebral defects, anal atresia, cardiac defects, tracheoesophageal fistula, renal anomalies and limb anomalies): report of 2 cases</b>.

F&S reports·2022
Same author

Construct validity of the Surgical Autonomy Program for the training of neurosurgical residents.

Neurosurgical focus·2022

This study introduces an interactive system for brain tumor classification using deep learning and content-based image retrieval (CBIR). The system offers accurate, interpretable results, enhancing clinician decision-making.

Area of Science:

  • Artificial Intelligence
  • Medical Imaging
  • Oncology

Background:

  • Clinicians rely on experience for diagnoses, but complex cases like brain tumors benefit from broader data access.
  • Traditional content-based image retrieval (CBIR) systems for medical images have seen limited clinical adoption due to performance and interpretability issues.
  • Deep learning improves CBIR performance but often sacrifices interpretability.

Purpose of the Study:

  • To develop an interactive image retrieval system for accurate and interpretable brain tumor classification.
  • To enhance the accessibility and usability of medical CBIR systems for clinicians.
  • To create a clinician-ML system that clinicians can trust and interact with.

Main Methods:

  • Utilized supervised contrastive learning to train image encoders, preserving structure in the retrieval space.
Keywords:
Brain Tumor ClassificationContrastive LearningDeep LearningLarge Language ModelsMedical CBIR

More Related Videos

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
04:09

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

8.9K

Related Experiment Videos

Last Updated: Apr 4, 2026

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.7K
Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
04:09

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

8.9K
  • Integrated off-the-shelf large language models (LLMs) for report summarization and user Q&A.
  • Developed an interactive framework for clinician-ML collaboration in medical image analysis.

    • Main Results:

    • Image encoders demonstrated classification performance comparable to or exceeding conventional black-box classifiers.
    • The system provides accurate and interpretable brain tumor classification.
    • LLM integration enhanced system accessibility through natural language interactions.

    Conclusions:

    • Supervised contrastive learning enables interpretable and high-performing medical image retrieval.
    • The developed system augments clinician performance by mirroring natural thought processes.
    • This framework facilitates faster, more trustworthy interactions with medical CBIR systems.