โก Quick Summary
This study introduces a novel multi-instance learning (MIL) model for classifying subtypes of non-small cell lung cancer (NSCLC), achieving an impressive 97.0% accuracy and an AUC of 0.978. The model demonstrates robust performance across diverse datasets, highlighting its potential as a reliable diagnostic tool. ๐
๐ Key Details
- ๐ Dataset: Cancer Genome Atlas (TCGA) dataset
- ๐งฉ Features used: Pathological images with histological features
- โ๏ธ Technology: Multi-instance learning with an additive attention mechanism
- ๐ Performance: ACC 97.0%, AUC 0.978 on TCGA dataset
๐ Key Takeaways
- ๐ฌ Novel approach: The study proposes a new MIL model for NSCLC subtype classification.
- ๐ High accuracy: Achieved a classification accuracy of 97.0% on the TCGA dataset.
- ๐ Robustness: Generalization experiments showed ACCs of 91.2% and 93.0% on external datasets.
- ๐ AUC performance: The model reached an AUC of 0.978, indicating excellent discrimination ability.
- ๐งช Ablation studies: Validated the contribution of each model component to performance improvement.
- ๐ฅ Clinical relevance: Potential to enhance diagnostic accuracy in diverse clinical settings.
- ๐ Published: In Sci Rep, 2025; 15:43210.
๐ Background
Non-small cell lung cancer (NSCLC) is a leading cause of cancer-related mortality worldwide, primarily comprising two subtypes: lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). Accurate classification of these subtypes is crucial for effective treatment planning, yet it poses significant diagnostic challenges. Recent advancements in machine learning offer promising avenues for improving diagnostic accuracy through enhanced image analysis techniques.
๐๏ธ Study
The researchers developed a multi-instance learning (MIL) model that incorporates an additive attention mechanism and a new category classifier to improve subtype discrimination. The model also employs a cross-scale focal region detection strategy to enhance sensitivity to critical histological features. The study utilized the TCGA dataset for training and evaluated the model’s performance against state-of-the-art methods.
๐ Results
The proposed model achieved a remarkable classification accuracy of 97.0% and an AUC of 0.978 on the TCGA dataset, outperforming existing methods such as ABMIL, CLAM, and others. Generalization tests on external datasets from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) and Yantai Yuhuangding Hospital yielded ACCs of 91.2% and 93.0%, with AUCs of 0.967 and 0.968, respectively, demonstrating the model’s robustness.
๐ Impact and Implications
The findings from this study have significant implications for the field of oncology. By providing a reliable tool for accurate NSCLC subtype classification, this model could enhance treatment planning and patient outcomes. The integration of advanced machine learning techniques into clinical practice may pave the way for more personalized and effective cancer therapies, ultimately improving survival rates and quality of life for patients.
๐ฎ Conclusion
This study highlights the transformative potential of machine learning in the classification of lung cancer subtypes. With a high accuracy and robust performance across various datasets, the proposed MIL model stands as a promising advancement in the diagnostic landscape of NSCLC. Continued research and development in this area could lead to significant improvements in cancer care and patient management. ๐
๐ฌ Your comments
What are your thoughts on the application of machine learning in cancer diagnostics? We would love to hear your insights! ๐ฌ Leave your comments below or connect with us on social media:
Non-small cell lung cancer subtype classification based on cross-scale multi-instance learning.
Abstract
Lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC), the two major subtypes of non-small cell lung cancer (NSCLC), present significant diagnostic challenges with direct implications for treatment planning. In this study, we propose a novel multi-instance learning (MIL) pathological image classification model that incorporates an additive attention mechanism and a new category classifier to enhance subtype discrimination. The model further integrates a cross-scale focal region detection strategy to improve sensitivity to key histological features. Trained on the Cancer Genome Atlas (TCGA) dataset, our model achieved a subtype classification accuracy (ACC) of 97.0% and an area under the ROC curve (AUC) of 0.978, outperforming state-of-the-art methods including ABMIL, CLAM, DS-MIL, DTFD-MIL, FR-MIL, and WIKG-MIL across multiple evaluation metrics. Ablation studies validate the contribution of each module to overall performance improvement. Generalization experiments conducted on the Clinical Proteomic Tumor Analysis Consortium (CPTAC) Cancer Imaging Archive (TCIA) Lung dataset and an external dataset from Yantai Yuhuangding Hospital demonstrate the robustness of our model, achieving ACCs of 91.2% and 93.0%, and AUCs of 0.967 and 0.968, respectively. These results underscore the model’s strong generalization ability and its potential as a reliable tool for accurate NSCLC subtype classification across diverse clinical scenarios.
Author: [‘Jiang P’, ‘Chen W’, ‘Zheng G’, ‘Li X’, ‘Song X’]
Journal: Sci Rep
Citation: Jiang P, et al. Non-small cell lung cancer subtype classification based on cross-scale multi-instance learning. Non-small cell lung cancer subtype classification based on cross-scale multi-instance learning. 2025; 15:43210. doi: 10.1038/s41598-025-27337-7