The Evaluation And Prediction Of Curative Effect Of Quantitative CT Imaging In Patients With Advanced Lung Squamous Cell Carcinoma

Objective: Imaging tomography is an emerging field,which can be used to noninvasively quantify tumor phenotype by applying advanced imaging features algorithms.This study is aimed to analyze CT features,clinical factors,and therapeutic efficacy of advanced lung squamous cell carcinoma patients before treatment and prognosis,in order to build a prognostic model combining with clinical features of images,and then quantitatively analyze the disease progression time.Eventually providing a method to individually predict the prognosis of patients with advanced lung squamous cell carcinoma.Methods: A retrospective analysis of 96 patients with advanced squamous cell carcinoma of the lung at the First Affiliated Hospital of China Medical University from January 2013 to April 2017 was performed.Patients' CT images before treatment were extracted and the tumor lesions were drawn by physicians.In order to construct the prognostic index model(PIM)using the radiological labeling and clinical risk factors as the parameters of multivariate Cox regression analysis,we use imaging tomography methods to extract features,construct model and analyze.The resulting PIM was used to assess the prognostic difference in overall survival(OS).All statistical algorithms are implemented on the R Development Core Team.Results: 1.No significant TTP difference was found in the PR and non-PR patient groups(p = 0.082).Median TTP for PR and Non-PR were 6.67 and 3.2 months.2.Based on the pre-chemotherapy CT images,eight phenotypic characteristics were extracted,including co-occurrence matrix,run-length matrix and Gabor features.According to the risk stratification,they were divided into Good and Poor groups.P <0.0001).The median TTP for the Good and Poor groups was 6.67 and 3.25 months,respectively.3.Prognosis model PIM establishment: univariate Cox regression analysis showed no significant difference except for AST,ALT and CEA(p <0.05).According to multivariate Cox regression analysis,PIM consisted of three factors : The imaging prognostic label,AST and ALT,according to the established PIM will be divided into low-risk,intermediate risk,high-risk three groups,low risk subgroup of patients(n = 24)benefit from the twocycles of chemotherapy the most,The long median TTP was 7.2 months.The moderate-risk subgroup(median 3.5 months for TTP)was lower than the low-risk subgroup(p <0.0001,HR = 2.31)but higher than the subgroup(HR = 0.42,p <0.0001).Patients in high-risk subgroups often benefited the least from chemotherapy compared with the other two subgroups(median TTP: 1.8 months,p <0.0001).4.The Predictive Value of PIM in Total Survival:OS was significantly lower in high-risk subgroup than in moderate-risk and low-risk subgroups(median OS: 12.0 and 14.0 months respectively)according to PIM(HR : 1.92 and 2.00,p respectively 0.015 and 0.008).There was no significant difference in OS between moderate-risk and low-risk subgroups(p = 0.813).Conclusion: The proposed CT-based prognostic strategy can achieve individualized prediction of TTP in LSC,which holds promise of improving the pre-therapy personalized management of chemotherapy.