| Lung cancer is the most common and fatal malignant tumor in the world.Nonsmall cell lung cancer(NSCLC)is the most common histologic type of lung cancer,accounting for approximately 87%of lung cancer cases.Due to the high invasiveness of NSCLC,the 5-year survival rate is only about 24%.In recent years,with the optimization of chemotherapy regimen and the rise of targeted therapy,the treatment regimen for advanced NSCLC has gradually diversified.However,the treatment effect of lung cancer patients has been unsatisfactory due to tumor heterogeneity,drug resistance and other factors.Therefore,predicting the efficacy of NSCLC patients in advance and accurately differentiating the histological subtypes of NSCLC patients have important clinical significance for guiding clinical treatment.First,tumor cells can become resistant to platinum drugs and targeted drugs in the course of treatment,so it becomes very important to predict drug resistance in advance.Tumor response is used to measure the effectiveness of chemotherapy and targeted therapy.Predicting tumor response at the time of diagnosis can determine whether to continue,escalate,discontinue,or change a patient's therapeutic schedule.Currently,tumor response is clinically evaluated based on Response Evaluation Criteria in Solid Tumors(RECIST 1.1),which are ignorant of a substantial amount of information within the radiographic image and can cause delays in creating appropriate treatment schedules for patients.Radiomics can be used for estimating the tumor and its microenvironment,and longitudinal estimating of tumor evolution.Some studies have developed the radiomics signature for the prediction of efficacy in advance,and found that radiomics signature could predict the tumor response of chemotherapy as well as targeted therapy.However,the real world was far more complicated than former researches,with many patients undergoing both chemotherapy and targeted therapy(sequential or concurrent).To our knowledge,no researchers have combined these three treatments(chemotherapy,targeted therapy,and a combination of both)for a real-world radiomics analysis.Second,as two major histologic subtypes of NSCLC,adenocarcinoma(ADC)and squamous cell carcinoma(SCC)have very different treatment regimens.Accurate differentiation of ADC and SCC is the basis for precision therapy.Pathological diagnosis is commonly regarded as the gold standard for distinguishing ADC from SCC.Bronchoscopy and percutaneous biopsy are the main methods to obtain the pathological diagnosis of NSCLC.However,both are invasive tests and may cause severe complications.For lesions near the airway or large vessels,the risk of puncture is high.Also,tumors are often heterogeneous,and biopsy may not capture the full tumor form and phenotype.These limitations of pathological diagnosis prompt us to develop non-invasive and accurate alternative methods.The prediction model based on radiomics analysis has the potential to quantify tumor phenotypic characteristics non-invasively.Several studies have focused on the identification of histologic subtype of NSCLC based on radiomics.However,these radiomics studies generally had small size datasets,and were pure imaging studies,which did not take clinical factors into account,limiting the generalization performance and stability of the constructed models.In order to predict the response of NSCLC patients after first-line chemotherapy and targeted therapy in advance,and predict the histologic subtypes of NSCLC and then personalize the treatment of NSCLC patients,and ultimately reduce the mortality and prolong the survival,this study includes the following two contents:Part I:CT-based radiomics signatures can predict the tumor response of nonsmall cell lung cancer patients treated with first-line chemotherapy and targeted therapyObjective:To evaluate the effectiveness of radiomics signatures on pre-treatment computed tomography(CT)images of lungs to predict the tumor responses of nonsmall cell lung cancer(NSCLC)patients treated with first-line chemotherapy,targeted therapy,or a combination of both.Materials and Methods:This retrospective study included 322 NSCLC patients who were treated with first-line chemotherapy,targeted therapy,or a combination of both.Of these patients,224 were randomly assigned to a cohort to help develop the radiomics signature.A total of 1,946 radiomics features were obtained from each patient's CT scan.The top-ranked features were selected by the Minimum Redundancy Maximum Relevance(MRMR)feature-ranking method and used to build a lightweight radiomics signature with the Random Forest(RF)classifier.The independent predictive(IP)features(AUC>0.6,p-value<0.05)were further identified from the top-ranked features and used to build a refined radiomics signature by RF classifier.Its prediction performance was tested on the validation cohort,which consisted of the remaining 98 patients.Results:The initial lightweight radiomics signature constructed from 15 top-ranked features had an AUC of 0.721(95%CI,0.619-0.823).After six IP features were further identified and a refined radiomics signature was built,it had an AUC of 0.746(95%CI,0.646-0.846).Conclusions:Radiomics signatures based on pre-treatment CT scans can predict tumor response in NSCLC patients after first-line chemotherapy or targeted therapy treatments.Radiomics features could be used as promising prognostic imaging biomarkers in the future.Part ?:CT-based radiomics signatures can predict the histologic subtypes of non-small cell lung cancerObjective:To evaluate the effectiveness of radiomics signatures on pre-treatment computed tomography(CT)images of lungs to predict the histological subtypes of non-small cell lung cancer(NSCLC),and to explore the value of a combined model based on radiomics features and clinical factors to predict the histologic subtypes of NSCLC.Materials and Methods:This retrospective study included 301 NSCLC patients,who consisted of 215 adenocarcinoma(ADC)patients and 86 squamous cell carcinoma(SCC)patients.Of these patients,210 were randomly assigned to a cohort to help develop the radiomics signature.A total of 1,046 radiomics features were obtained from each patient's CT scan.Least Absolute Shrinkage and Selection Operator(Lasso)were used to select the radiomics features,and three classifiers,Logistics Regression(LR),Support Vector Machines(SVM)and Random Forest(RF),were used to build the radiomics signatures.LR classifier was used to build a combined model of radiomics features and serum markers,as well as a combined model of radiomics features and clinical factors(including age,sex,and serum markers).Their prediction performances were tested on the validation cohort,which consisted of the remaining 91 patients.Results:The radiomics signatures of LR,SVM and RF constructed from 18 most discriminating features had an AUC of 0.725(95%CI,0.612-0.838),0.732(95%CI,0.625-0.839)and 0.722(95%CI,0.613-0.832),respectively.The combined signature of radiomics features and serum markers had an AUC of 0.825(95%CI,0.732-0.917).The combined signature of radiomics features and clinical factors had an AUC of 0.914(95%CI,0.857-0.971).Conclusions:Radiomics signatures based on pre-treatment CT scans can predict the histologic subtypes of non-small cell lung cancer.The combined signature based on the radiomics features and clinical factors had a better performance on the classification of ADC and SCC than the radiomics signature. |
PDF Full Text Request