What question did this study set out to answer?

To enhance diagnostic accuracy for rare cardiac tumours by developing and validating machine learning models that integrate echocardiographic and pathological features from limited datasets.

January 1, 2024Informatics in Medicine UnlockedOpen Access

Precision diagnostics in cardiac tumours: Integrating echocardiography and pathology with advanced machine learning on limited data

View Full Paper

Ask AI

Bookmark

Why the study?

Due to the rarity and complexity of cardiac tumours, more precise, non-invasive, and efficient diagnostic solutions are needed.

Does integrating echocardiography and pathology data with advanced machine learning improve the diagnostic accuracy of cardiac tumours?

Population

399 patients at the Heart Hospital

Comparison

Machine learning models integrating echocardiography and pathology vs traditional diagnostic metrics

Authors

SSSeyed‐Ali Sadegh‐ZadehUniversity of Staffordshire NKNaser Khezerlouy‐AghdamTabriz University of Medical Sciences HSHanieh SakhaTabriz University of Medical Sciences

Discussion

Member takes

Implication

Supports ML-assisted cardiac tumor diagnosis in research settings; leaves open prospective validation before clinical adoption.

Key Points

To enhance diagnostic accuracy for rare cardiac tumours by developing and validating machine learning models that integrate echocardiographic and pathological features from limited datasets.
Evaluated Support Vector Machines, Random Forest, and Gradient Boosting Machines using clinical features from 399 patients at the Heart Hospital.
Identified key diagnostic predictors and conducted real-world clinical validation of the top-performing model at the same institution.
The Random Forest model outperformed other approaches, achieving 96.25% accuracy and an ROC AUC score of 0.99 during initial model evaluation.
Real-world clinical validation yielded a diagnostic accuracy of 94% for the Random Forest model, with patient age, echocardiographic malignancy, and tumor position serving as primary predictive features.

Structured PICO

Does integrating echocardiography and pathology data with advanced machine learning improve the diagnostic accuracy of cardiac tumours?

Population

399 patients with cardiac tumours evaluated at the Heart Hospital

Intervention

Machine learning models (Support Vector Machines, Random Forest, and Gradient Boosting Machines) integrating echocardiography and pathology data

Comparator

Traditional diagnostic metrics

Outcome

Diagnostic accuracy and ROC AUC scoresurrogate

A Random Forest machine learning model integrating echocardiography and pathology data achieved high diagnostic accuracy for cardiac tumours, demonstrating the potential of ML to enhance non-invasive diagnostics.

Limitations

limited dataset
lack of external validation

Cite This Study

Sadegh‐Zadeh et al. (2024) studied this question.

synapsesocial.com/papers/69f53ed4e0fbb6efbd2037f3 https://doi.org/10.1016/j.imu.2024.101544

View Full Paper

Ask AI

Bookmark

Precision diagnostics in cardiac tumours: Integrating echocardiography and pathology with advanced machine learning on limited data

Discussion

Member takes

Implication

Key Points

Structured PICO

Limitations

Cite This Study

Precision diagnostics in cardiac tumours: Integrating echocardiography and pathology with advanced machine learning on limited data

Discussion

Member takes

Implication

Key Points

Structured PICO

Limitations

Cite This Study

Also Consider

Also Consider