Study: AI May Help Predict Responses to Systemic Therapies in Non-Small Cell Lung Cancer

A study published in Clinical Cancer Research has shown that using standard-of-care computed tomography (CT) scans in patients with advanced non-small cell lung cancer (NSCLC) can predict tumor sensitivity to 3 systematic cancer therapies by using artificial intelligence (AI).

Currently, radiologists can determine whether patients with NSCLC are responding to systemic therapy by quantifying changes in tumor size and the appearance of new tumor lesions. However, this type of evaluation can be limited, especially in patients treated with immunotherapy who can display atypical patterns of response and progression, according to the study authors.

Using data from multiple phase 2 and phase 3 clinical trials that evaluated systemic treatment in patients with NSCLC, the patients were treated with 1 of 3 agents: nivolumab (Opdivo), docetaxel (Taxotere), or gefitinib (Iressa). The researchers analyzed standard-of-care CT images from 92 patients receiving nivolumab in 2 trials: 50 patients receiving docetaxel in 1 trial and 46 patients receiving gefitinib in the other trial.

In order to develop the model, the researchers used the CT images taken at baseline and on first-treatment assessment, which meant 3 weeks for patients treated with gefitinib or 8 weeks for patients treated with either nivolumab or docetaxel. Tumors were then classified as treatment-sensitive or treatment-insensitive based on the reference standard of each trial. Among all 3 cohorts, patients were randomized into training or validation groups.

Machine learning led the researchers to develop a multivariable model to predict treatment sensitivity in the training cohort. Each model could predict a score ranging from 0 (highest treatment sensitivity) to 1 (highest treatment insensitivity) based on the change of the largest measurable lung lesion identified at baseline.

Due to gefitinib’s limited number of patients, the research team built and validated a model using a cohort of patients with metastatic colorectal cancer treated with anti-EGFR therapies. The radiologic features that were used to predict treatment sensitivity identified in the colorectal cancer cohort were then used to build a model in the training cohort of patients with NSCLC treated with gefitinib.

A total of 8 radiologic features were used to build the 3 prediction models; these features included change in tumor volume, heterogeneity, shape, and margin. The nivolumab and gefitinib models used 4 radiologic features and the docetaxel model used 1.

Each signature’s performance was evaluated by calculating the area under the curve (AUC), in which a score of 1 corresponds to perfect prediction. The nivolumab, docetaxel, and gefitinib prediction models achieved an AUC of 0.77, 0.67, and 0.82 in the validation cohorts.

“We observed that similar radiomics features predicted three different drug responses in patients with NSCLC,” associate research scientist Laurent Dercle, MD, PhD of the Department of Radiology at the Columbia University Irving Medical Center, said in a press release. “Further, we found that the same four features that identified EGFR treatment sensitivity for patients with metastatic colorectal cancer could be utilized to predict treatment sensitivity for patients with metastatic NSCLC.”

Although the researchers said that radiomic signatures offer the potential to enhance clinical decision-making, there were limitations to the study, including small sample size.

REFERENCE

Artificial Intelligence May Help Predict Responses to Systemic Therapies in Patients With Non-small Cell Lung Cancer [news release]. Philadelphia, PA; American Association for Cancer Research: March 20, 2020. https://www.aacr.org/about-the-aacr/newsroom/news-releases/artificial-intelligence-may-help-predict-responses-to-systemic-therapies-in-patients-with-non-small-cell-lung-cancer/. Accessed March 24, 2020.