Pre-treatment Enhanced CT Imaging To Predict The Efficacy Of Immunotherapy In Non-small Cell Lung Cancer

Objective:To explore the value of pre-treatment enhanced CT imaging to predict the efficacy of immunotherapy for advanced non-small cell lung cancer,so as to provide an important decision-making basis for clinical development of individualized treatment plans,so as to achieve accurate diagnosis and treatment of advanced lung cancer.methods:A total of 96 patients who were diagnosed with advanced NSCSC and received immunotherapy in Yunnan Cancer Hospital from January 2016 to December 2020 were retrospectively collected.The clinical and pathological data and pretreatment Computed Tomography(CT)scan images were collected.Patients underwent enhanced CT examination within1 week before treatment,and CT imaging follow-up was performed after the first cycle of treatment.The development of tumor lesions after treatment was evaluated based on the RECIST1.1 for immune based therapeutics(iRECIST).The complete response(CR),partial response(PR)and stable disease(SD)were regarded as the effective group,while the progressive disease(PD)was regarded as the effective group.PD)were treated as the ineffective group.The Radiomics program of Siemens syngo.via was used to delineate the whole lesion on the venous phase images of enhanced CT,and the radiomics features were extracted.least absolute shrinkage and selection operator(LASSO)method was used to reduce the feature dimension and retain the 9 radiomics features with the most predictive value for the response of advanced NSCLC immunotherapy,and the weight coefficient was calculated.The radiomics model was established by combining the radiomics score.Univariate and multivariate logistic regression analysis were used to analyze the correlation between clinical characteristics,laboratory tests,pathological data and the prognosis of patients after immunotherapy,and a clinicopathological prediction model was established.The radiomics model was added to the clinicopathological model to establish a hybrid prediction model.Receiver Operating Characteristic Curve(ROC)and Delong test were used to study the predictive performance between the three models,and the Areas Under the Curve(AUC)were calculated.results: A total of 236 target lesions from 96 patients were analyzed according to the inclusion and exclusion criteria,including primary and metastatic lesions.The response of each lesion to the first immunotherapy was evaluated and predicted.From 854 radiomics features extracted from the enhanced CT images of all lesions,9 radiomics features were selected to establish a radiomics model.The AUC value of the training set model was 0.794(95% CI(confidence interval): 0.715～0.872),the AUC of the validation dataset was 0.812(95%CI: 0.693～0.931).Univariate and multivariate analysis showed that clinical stage,white blood cell count,red blood cell count and platelet count were significantly different between the effective group and the ineffective group(P < 0.05).The above factors were included in the construction of the clinicopathological prediction model,and the AUC value of the training set was 0.614(95%CI:0.524～0.704),the AUC value of the validation set model was 0.572(95%CI: 0.438～0.706).The mixed model was constructed by combining the radiomics signature and clinicopathological features,the AUC of the training set and validation set of the mixed model were 0.643(95%CI:0.554～0.733)and 0.522(95%CI: 0.399～0.655),respectively.The prediction performance of the mixed model was equivalent to that of the clinicopathological model.Subgroup analysis according to the location of primary and metastatic lesions showed that: For patients with advanced lung cancer(left and right lung),the AUC values of the radiomics prediction model in the training set and validation set were 0.766(95%CI:0.620～0.913)and 0.783(95%CI: 0.612～0.953),respectively.For lymph node lesions,the AUC values of the radiomics prediction model in the training set and validation set were 0.760(95%CI: 0.612～0.907)and0.847(95%CI: 0.542-1.000),respectively.For other metastases(brain,liver,adrenal gland,sternothorax,deep abdominal fat layer,buttery muscle),the AUC values of the radiomics prediction model in the training and validation cohorts were 0.868(95%CI: 0.739～0.996)and 0.792(95%CI:0.483～1.000),respectively.Conclusion:1.The radiomics model based on the pre-treatment contrast-enhanced CT can effectively predict the efficacy of immunotherapy in advanced NSCLC,And the efficiency is better than the clinicopathological model and the mixed model,and can be used as an independent factor to predict the efficacy.2.The pre-treatment contrast-enhanced CT radiomics model can predict the response to immunotherapy for advanced NSCLC patients with primary lung lesions and metastases to other sites.