โก Quick Summary
This study explores the use of dual time point CT scans combined with a foundation model to predict survival in patients with non-small cell lung cancer (NSCLC). The findings indicate that analyzing temporal changes in feature vectors significantly enhances survival predictions compared to traditional methods.
๐ Key Details
- ๐ Dataset: 102 NSCLC patients treated with radiation therapy
- ๐งฉ Features used: Pre-treatment and post-treatment CT scans
- โ๏ธ Technology: Foundation model for feature extraction
- ๐ Statistical methods: Random forest and gradient boosted survival models
๐ Key Takeaways
- ๐ Temporal analysis of CT scans can improve survival predictions.
- ๐ก Foundation models provide high-dimensional feature vectors for analysis.
- ๐ Euclidean distance and element-wise subtraction of feature vectors were key metrics.
- ๐ฅ Traditional clinical data models are less effective in predicting survival.
- ๐ This approach could lead to more personalized treatment plans for NSCLC patients.
- ๐ Study published in Sci Rep, 2025.
- ๐ PMID: 41339367
๐ Background
Lung cancer, particularly non-small cell lung cancer (NSCLC), is a leading cause of cancer-related mortality worldwide. Traditional methods for predicting patient survival often rely on clinical data, which can be limited in scope and accuracy. Recent advancements in deep learning and imaging technologies present new opportunities to enhance predictive capabilities, particularly through the analysis of temporal changes in imaging data.
๐๏ธ Study
The study utilized a dataset of 102 NSCLC patients who underwent radiation therapy, with both pre-treatment and post-treatment CT scans available for analysis. By employing a foundation model, researchers extracted high-dimensional feature vectors from the scans, which were then summarized for further statistical analysis. The aim was to assess whether these temporal changes could yield better survival predictions than traditional single-time point analyses.
๐ Results
The results indicated that the use of temporal changes in feature vectors significantly improved survival predictions. Specifically, the analysis of Euclidean distance and element-wise subtracted feature vectors outperformed traditional clinical data models, demonstrating the potential of this approach in enhancing predictive accuracy for NSCLC patient outcomes.
๐ Impact and Implications
This study’s findings could have profound implications for the management of lung cancer. By integrating advanced imaging techniques with machine learning, healthcare providers may be able to offer more accurate and personalized treatment plans for NSCLC patients. This could ultimately lead to improved survival rates and better quality of life for those affected by this challenging disease.
๐ฎ Conclusion
The research highlights the transformative potential of using foundation models and dual time point CT scans in predicting lung cancer survival. As we continue to explore the intersection of technology and healthcare, such innovations could pave the way for more effective patient management strategies. Continued research in this area is essential to fully realize the benefits of these advanced predictive models.
๐ฌ Your comments
What are your thoughts on the use of advanced imaging techniques for predicting cancer survival? We would love to hear your insights! ๐ฌ Leave your comments below or connect with us on social media:
Foundation model based prediction of lung cancer survival using temporal changes in dual time point CT scans.
Abstract
Lung cancer remains a significant cause of mortality, with non-small cell lung cancer (NSCLC) representing most cases. Currently, clinical data based models fall short in predicting survival while more advanced deep learning based image models require vast amounts of data and are often limited to predictions based on single time points. This study uses dual time point CT scans and features derived from a foundation model to predict survival. A dataset containing 102 NSCLC patients treated with radiation therapy was used, with each patient having both pre-treatment and post-treatment CT scans. A foundation model applied to the scans generated high-dimensional feature vectors and these vectors were then further summarized. Statistical analyses, including random forest and gradient boosted survival models, were then used to predict survival. The results demonstrated that temporal changes in feature vectors, specifically the Euclidean distance and element-wise subtracted feature vectors, can offer improved prediction of survival over single-time point features and clinical data.
Author: [‘Petrochuk J’, ‘Pai S’, ‘He J’, ‘Haugg F’, ‘Xu Y’, ‘Christiani D’, ‘Mak R’, ‘Aerts H’]
Journal: Sci Rep
Citation: Petrochuk J, et al. Foundation model based prediction of lung cancer survival using temporal changes in dual time point CT scans. Foundation model based prediction of lung cancer survival using temporal changes in dual time point CT scans. 2025; 15:43042. doi: 10.1038/s41598-025-26365-7