New study highlights AI’s potential to transform lung cancer screening

3 minutes, 49 seconds Read


A recent study published cancer conducted a meta-analysis to evaluate the potential of artificial intelligence (AI) in early lung cancer detection.

Study: AI-powered models to predict diagnosis and outcome in lung cancer: a systematic review and meta-analysis.  Image credit: metamorworks/Shutterstock.comStudy: AI-powered models for diagnosing and predicting lung cancer outcomes: a systematic review and meta-analysis. Image credit: metamorworks/Shutterstock.com

Background

Lung cancer is a global health problem with high mortality due to late diagnosis. Current techniques for early detection include computed tomography (CT) scans; However, benign lesions and radiologists experience impact performance. Innovative strategies are needed to improve prognosis and survival rates.

AI models can enhance these methods by increasing accuracy and efficiency, reducing false positive and negative cases, and providing complementary strategies to existing ones.

AI-assisted diagnostic systems in healthcare, especially for lung cancer, can increase diagnostic accuracy, stability and work efficiency.

About the study

In the current meta-analysis, researchers evaluated the effectiveness of AI models in the early detection of pulmonary cancer, highlighted their potential for improved diagnostic accuracy, and analyzed their strengths, limitations, and comparative advantages over traditional methods.

The team searched the PubMed, Science Direct, Embase and Google Scholar databases to retrieve relevant records published in English up to October 2023.

Two researchers independently screened records using predefined criteria to select high-quality studies and resolved discrepancies by consensus or consultation with a third researcher.

The team included original research articles on studies evaluating AI effectiveness in detecting early-stage lung cancer and reporting results as performance metrics such as specificity, sensitivity, and accuracy.

They excluded studies with inadequate information about the performance of AI models, commentaries, conference abstracts lacking primary data, and reviews.

The team extracted data on study setting, design, AI models used, data sources, performance metrics, validation methods, and outcomes.

They analyzed using the Diagnostic Accuracy Studies (QUADAS-2) tool and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for assessing study quality.

Bias risk domains assessed in the study related to patient selection, index tests, reference standards, flow, and timing were reported as low, clear, or high risk.

Researchers assessed heterogeneity across studies using the I2 statistic and the chi-square test. They performed random-effects modeling for meta-analysis and forest plot analysis of pooled diagnostic metrics for AI models.

result

Initially, the team identified 1,024 records, assessed 116 for eligibility, and excluded 326 duplicates and 28 studies published in non-English languages.

Only 39 records met the eligibility criteria, demonstrating various AI model applications for lung cancer detection and highlighting various strengths among studies.

The study reveals the potential of artificial intelligence to diagnose lung cancer at an early stage, with a pooled sensitivity of 0.87 and a specificity of 0.87, indicating high accuracy in detecting true positives and negatives.

However, the team noted heterogeneity across studies due to differences in study populations, data sources, and model specifications. The team found low risk of bias in patient selection, index tests and reference standards but high risk of bias in flow and timing.

Studies have shown that AI models, especially recurrent neural networks (RNN) and convolutional neural networks (CNN), can improve lung cancer prediction accuracy, reduce false positives, and reduce the impact of missing data.

Other AI models used include DL, DBN, machine learning (ML), logistic regressions (LL), random forest classifiers (RF), naive Bayesian systems (NBS), Bayesian networks (BN), and decision trees.

A study in China applied three-dimensional deep learning models to computed tomography scans, achieving sensitivity, specificity, and overall diagnostic accuracy of 75%, 82%, and 89%, respectively.

Another study in China used Support Vector Machine-List Absolute Linkage and Selection Operator (SVM-LASSO) on Lung Image Database Consortium (LIDC)-Image Database Resource Initiative (IDRI) data and achieved 85% accuracy, 12% higher than lung. Reporting and Data Systems (RADS).

Another study conducted in China used a three-dimensional customized mixed link network (CMixNet) for data derived from the LIDC-IDRI and lung nodule analysis (LUNA-16) datasets, achieving a sensitivity and specificity of 94% and 91%, respectively. Better results than existing techniques used to detect lung cancer.

The results highlighted the trade-offs between specificity and sensitivity and the power of AI-based techniques.

While some studies have highlighted the potential of AI to overcome particular challenges, others have emphasized the reliability and efficiency of artificial intelligence in pulmonary cancer screening, benefiting healthcare professionals and patients.

Conclusion

Research results have shown that AI models effectively detect lung cancer at an early stage, detect positives and negatives, and improve prognosis.

However, heterogeneity in studies emphasizes the need for standardized protocols. Future research should focus on refining AI models, consider challenges, and collaborate with researchers, clinicians, and policymakers to establish guidelines and standards for AI systems in pulmonary cancer screening.

Addressing these challenges will drive AI technology forward, ultimately facilitating early lung cancer diagnosis and rapid management.



Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *