AI Analysis of PET/CT Images Predicts Side Effects of Immunotherapy in Lung Cancer

This condition is characterized by lung scarring (fibrosis) and can be life-threatening due to respiratory failure. Predicting the onset of interstitial lung disease caused by immunotherapy remains a challenge. As a result, there is a need for reliable methods to assess the risk of developing this condition following immunotherapy. Now, a new study has found that PET/CT image analysis combined with artificial intelligence (AI) can predict the likelihood of developing interstitial lung disease.

In this retrospective study, researchers from Niigata University (Niigata, Japan) analyzed data from 165 patients with primary lung cancer who were undergoing immunotherapy. It is generally believed that interstitial lung disease might develop when inflammation triggered by immunotherapy affects healthy lung tissue in addition to cancer cells. Based on this, the researchers hypothesized that patients with significant inflammation in their healthy lungs prior to treatment might be at greater risk of developing interstitial lung disease.

To test this hypothesis, the researchers utilized PET/CT scans, a nuclear imaging technique that can detect inflammation throughout the body. They used AI to assess the level of inflammation in noncancerous regions of the lungs—areas not affected by the cancer. The study, published in Academic Radiology, found that patients with high levels of inflammation in the healthy lung tissue were about 6.5 times more likely to develop interstitial lung disease after immunotherapy compared to those with lower inflammation.

"PET/CT is generally performed to detect cancer metastasis, but it would potentially be useful for estimating the risks of side effects associated with cancer treatment,” said researcher Motohiko Yamazaki. “The results of our study may not only help to predict the occurrence of interstitial lung disease after immunotherapy, but also to elucidate the mechanism of this serious side effect. We should conduct a multicenter prospective study for further investigation."