Predicting The Efficacy Of Immunotherapy For Non-small Cell Lung Cancer Based On Radiomics And Clinical Features

【Background and Purpose】Lung cancer is the most common malignancy with the highest risk of tumor-related death,and the prognosis is very poor.The emergence of immunotherapy provides a new treatment option for patients with non-small cell lung cancer.But studies have shown that not all patients benefit from immunotherapy.Currently,biomarkers commonly used to predict the efficacy of immunotherapy include programmed death receptor ligand 1(PD-L1)and high tumor mutation load(TMB).However,the detection of the above indexes needs to obtain the tumor tissue of patients and the prediction efficiency is not high.Therefore,it becomes more and more important to find a convenient and reliable new indicator in clinical practice.There are many models based on imaging omics or clinical features to predict immunotherapy efficacy.The purpose of this study was to comprehensively analyze the imaging and clinical features related to immunotherapy in patients,find out the features related to immunotherapy efficacy and evaluate its predictive efficacy,so as to provide new markers for immunotherapy in clinical patients with non-small cell lung cancer.【Method】Patients with non-small cell lung cancer who were diagnosed and received immunotherapy in the Affiliated Cancer Hospital of Guangzhou Medical University from August 2018 to August 2021 were collected in this study.The clinical data and imaging data of 53 patients met the inclusion criteria.In this study,the manual segmentation method based on the image J platform was used to delineate the ROI region of the tumor and extract the image omics features.A total of 144 imaging features were extracted,and 32 imaging features were selected for analysis,including5 shape features,7 first-order statistical features and 20 GLCM texture features.Patients were divided into response group(PR/CR)and non-response group(PD)to evaluate the efficacy of immunotherapy according to RECIST criteria.Normal test was performed on the statistical images and clinical data respectively.Independent sample t test was used for data conforming to normal distribution or approximately normal distribution.Mann-Whitney U test was used for quantitative data of skewed distribution,and Chi-square test,corrected Chi-square test and Fisher’s exact test were used for categorical variables.Finally,binary Logistic regression analysis was carried out to find out the independent influence indicators of statistically significant indicators.All statistical analyses were performed using SPSS 27.0 version.P < 0.05 was considered statistically significant.【Result】Three of the clinical features analyzed were associated with immunotherapy efficacy: absolute neutrophil count,absolute monocyte count,and ratio of derived neutrophil lymphocytes.The AUC of neutrophils was 0.574(95% confidence interval:0.395-0.755).The AUC for monocytes was 0.533(95% confidence interval:0.331-0.735).The AUC corresponding to the rate of derived neutral lymphocytes was0.577(95% confidence interval: 0.414-0.739).Other clinical features such as sex,age,smoking history,pathological type,clinical stage,ECOG score,BMI,lung related tumor markers,PD-L1 expression,leukocyte number,absolute lymphocyte count,neutral lymphocyte ratio,lymphomonocyte ratio,lactate dehydrogenase,albumin and immunotherapy efficacy were not statistically significant.In the analysis of image-related features,the Area,IDM(0 degrees,180 degrees),ASM(0 degrees,180degrees)of the lung window interface on CT plain scan had statistical significance with immunotherapy efficacy.The AUC corresponding to Area was 0.742(95%confidence interval: 0.579 to 0.905).The AUC corresponding to IDM(0 degrees)was0.702(95% confidence interval: 0.527-0.877).The AUC corresponding to IDM(180degrees)is 0.708(95% confidence interval: 0.533-0.883).The AUC value of ASM(0degrees)was 0.700(95%CI 0.517-0.883).The AUC of ASM(180 degrees)is 0.702(95%CI 0.520-0.884).Other image omics features such as kurtosis,entropy,contrast,and correlation were not found to be statistically significant.Further analysis showed that ASM(180 degree)was an independent factor for evaluating the efficacy of immunotherapy.【Conclusion】Preimmunotherapy clinical features and imaging omics can be used to predict the efficacy of immunotherapy in patients with non-small cell lung cancer.Neutrophil number,monocyte number,d NLR,Area,ASM and IDM were correlated with immunotherapy efficacy.The imaging features and clinical features of patients have certain guiding significance for clinical selection of immunotherapy patients.