Background Accurately predicting response to immunotherapy in non-small cell lung cancer (NSCLC) patients is important in establishing a precision treatment strategy. This study aims to investigate chest CT-based radiomic features using artificial intelligence (AI) algorithms to predict tumor response and biomarkers in patients with NSCLC.
Methods Clinicopathologic data from 132 patients with stage III-IV NSCLC treated with immunotherapy were collected, and the tumor response was evaluated based on RECIST 1.1 and immune-related RECIST (irRECIST). Patients were classified into two groups: durable responders (which include complete response [CR], partial response [PR], or stable disease [SD]) and non-responders (progressive disease [PD]). Segmentation was performed by three physicians using LIFEx software (IMIV/CEA, Orsay, France). 3D-radiomic features were collected for the tumor and peritumoral region from contrast enhanced CT scans. The lesion size was measured as a volume-of-Interest (VOI). Linear mixed-effects (LME) regression model was used to evaluate the association between radiomic features and the size of the VOI, with radiation-related and immunotherapy-related pneumonitis status specific variables for slope and intercept. The chemotherapy containing regimen was considered as a random factor. The corresponding p-values were used to assess differences in LME regression slope and/or intercept between all groups.
Results Among 132 patients, 61 (46.2%) were male and 71 (53.8%) were female (median age 66.5 years [range, 29–89]). Association between some CT-based radiomic features and the tumor size demonstrates statistically significant between responders and non-responders per RECIST 1.1 and irRECIST criteria. Significant difference in the intercept (-1.09E8, 95%CI[-1.5E8, -0.7E8], p<0.001) and the slope (31.1E6, 95%CI[20.1E6, 41.4E6], p<0.001) of MORPHOLOGICAL Integrated Intensity computed for primary tumors was found with irRECIST criteria (intercept difference -6.6E7 95%CI[-11.6E7, -1.5E7], p<0.01 and slope difference (15.8E6, 95%CI[36.3E6, 28E6], p<0.01). The same radiomics feature was statistically significant between responders and non-responders per RECIST 1.1 criteria with the intercept (-1.17E8, 95%CI[-1.55E8, -0.8E8], p<0.001) and the slope (32.8E6, 95%CI[23.7E6, 41.9E6], p<0.001) with the intercept difference (-6.1E7 95%CI[-10.9E7, -1.3E7], p<0.013) and slope difference (14.7E6, 95%CI[32.3E6, 26.2E6], p<0.02). For the peritumoral space, Gray Level Size Zone Matrix (GLSZM) Large Zone Low Grey Level Emphasis, demonstrated significant difference in the intercept (-0.73, 95%CI [-1.63, 0.16], p<0.1) and the slope (0.31, 95%CI [0.9, 0.53], p<0.005) and irRECIST associated difference in the intercept was 1.9 (95%CI[0.84, 2.97], p<0.01) and slope difference was -0.5 (95%CI[-0.76, -0.25], p<0.001).
Conclusions CT-based radiomic features may help predict potential responders to immunotherapy in NSCLC patients.
Ethics Approval Northwestern IRB approved: STU00207117
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.