AI-based body composition analysis predicts lung cancer treatment outcomes

Tafadzwa Chaunzwa, MD, an investigator in the Artificial Intelligence in Medicine (AIM) Program at Mass General Brigham and a senior physician in the Radiation Oncology Program at Harvard, is the lead author of a paper published in JAMA Oncology. Chaunzwa and senior author Hugo Aerts, PhD, director of the AIM program and associate professor at Harvard University, shared highlights from their work.

How would you summarize your study for a lay audience?

As treatments such as immunotherapy improve cancer survival rates, there is a growing need for clinical decision support tools that predict treatment response and patient outcomes. This is particularly important for lung cancer, which remains the leading cause of cancer death worldwide. Previous studies have linked body mass index (BMI) to lung cancer outcomes and side effects of immunotherapy drugs. However, BMI is a limited measure that does not capture details about different body tissues and their interaction with cancer treatments. We used medical imaging and artificial intelligence (AI) to analyze body composition in a large cohort of patients with lung cancer that has spread to other parts of the body. Our study found that changes in muscle mass and fat quality during treatment are important predictors of outcomes for this population.

What knowledge gaps does your studies help fill?

As we continue to improve the treatment of advanced lung cancer with different systemic agents, including immunotherapy drugs, biomarkers that are both prognostic and predictive of treatment response are increasingly needed to inform clinical decisions. Previous studies have identified an association between BMI and lung cancer outcomes. A correlation between BMI and the incidence of adverse events with immunotherapy has also been elucidated. However, BMI alone is a crude measurement that does not capture the distribution and relative contributions of different body tissues. Medical imaging-based analyzes of body composition are increasingly being explored. However, in the setting of advanced non-small cell lung cancer (NSCLC), studies have been limited by small sample sizes and manual and difficult-to-reproduce methodologies.

How did you do your study?

We set out to perform comprehensive body composition analysis in large cohorts of subjects treated for advanced or metastatic lung cancer with different systemic drugs. We have developed a powerful end-to-end AI-based platform to help you with this task.

What did you find?

We found that while the distribution of different tissue compartments at the start of cancer-directed therapy had some value, the change in these metrics during treatment was more strongly associated with patient outcomes. Specifically, we found that muscle loss was a poor prognostic factor in patients who received chemotherapy, immunotherapy, or chemoimmunotherapy. Interestingly, among patients receiving immunotherapy, either alone or in combination with chemotherapy, changes in the quality of the fat under the skin (subcutaneous adipose tissue), as seen on CT scans, were also associated with the risk of cancer progression or mortality of the lung.

What are the consequences;

This study presents key discoveries that will help advance the prognosis and surveillance of patients receiving immunotherapy for NSCLC. The first breakthrough is the application of an automated AI-based pipeline for comprehensive body composition analysis at scale in a diverse patient population receiving immunotherapy and cytotoxic chemotherapy for advanced NSCLC. This is the largest and most extensive such study, incorporating real-world data and prospective clinical trial cohorts, with longitudinal multimodal data collection and extended follow-up to monitor disease response to treatment. Our results demonstrate the potential of this analysis framework to provide a more nuanced understanding of the relationship between body composition and response to immunotherapy in NSCLC compared with raw BMI measurements. This can have important clinical implications for patient selection, treatment and follow-up. The second contribution is the sharing of this powerful end-to-end deep learning pipeline for automated slice selection and body compartment segmentation in cross-sectional imaging.

What are the next steps?

We offer the software as an open source artificial intelligence tool that seamlessly integrates with platforms for image analysis and also distributes it on the modelhub.ai platform. By making this algorithm publicly available, we hope to accelerate future studies in this area and further facilitate the development of new approaches to analyze complex cancer imaging datasets. This will allow further investigation of significant associations determined using BMI or CT-based manual body composition measures. More broadly, advances in this research area will help guide the management of different cancers and improve our ability for precision oncology.