Establishment And Validation Of Prognostic Nomogram For Advanced Non-small Cell Lung Cancer Patients Treated With PD-1 Inhibitors

Objective Immune checkpoint inhibitors reactivated the tumor-killing function of T cells,had greatly changed the clinical practice patterns of therapy regimens for advanced NSCLC patients.Nonetheless,given the low efficiency of immunotherapy and the complexity interaction of tumor cells and immune system that a single biomarker is unable to perfectly predict immunotherapy responses.In this study,a survival prediction model for advanced NSCLC patients receiving PD-1 inhibitors was established by combining clinical pathological features and peripheral blood biomarkers,and its application value was evaluated to assist in therapeutic decisions making process.Methods 1.Data acquisition Clinical information of 126 advanced NSCLC patients who received PD-1 inhibitors injection in our hospital from July 2017 to September 2019 was collected.A total of 110 patients fulfilled the inclusion criteria and were enrolled in this study.Data of clinicopathologic features,peripheral blood indicators and treatment records were extracted from the electronic inpatient record system or acquired by telephone follow-up,including gender,age,ECOG PS score,smoking status,lesion maximum diameter,site of metastasis,EGFR/ALK/ROS1 mutation status,tumor staging,PD-L1 expression level and other clinical pathological features;baseline complete blood cell count and its ratio,LDH,CRP,albumin,CEA,CA-199,CYFRA21-1,SCC and other peripheral blood indicators;immunotherapy regimen of patients and receiving PD-1 inhibitors’ time and cycles for patients.2.Evaluation criteria for therapeutic efficacy and side effects The optimal response pattern and disease progression were evaluated according to RESIST1.1 standard and recording the progressive site.Toxicity was graded using Common Terminology Criteria for Adverse Events(Version 4.0)and documenting the dealing methods.factors The mean standard deviation or median(min-Max)represents the continuity variable,and the percentage represents the classification variable.Chi-square or t-test was used to describe differences in demographic characteristics and clinical variables;COX univariate and multivariate survival analysis was applied to analyze the relationship between peripheral blood biomarkers and PFS,OS.4.Establishment and evaluation of prognostic model In COX univariate survival analysis,OS-related variables(P<0.05)were included in Stepwise AIC analysis to screen out factors for the establishment of OS survival probability prediction model nomograms of immunotherapy patients with advanced NSCLC;the receiver operating characteristic(ROC)curve was used to evaluate the discrimination performance of the model;calibration of the model was exhibited by calibration curve;the clinical utility of the model was assessed by clinical decision analysis curve;the nomogram was subjected to bootstrapping validation(500 Bootstrap resamples)to validate this model;Total Point Score(TPS)of each patient was calculated based on the nomogram,and X-tile software was applied to determine cut-off points of TPS,patients with low,medium and high scores was compared by Kaplan–Meier curves.3.Statistical analysis of clinical characters and prognosticResults 1.Demographic characteristics of immunotherapy patients with advanced NSCLC This study of patients with an average age of 63.8±9.7 years and 80.0% were males,former or current smokers constitute 63.6% of the subjects.Half of patients were lung squamous carcinoma,49 cases with adenocarcinoma,6 cases with large cell carcinoma or mixed adeno-squamous cancer.There were 39(35.5%),24(21.8%)and 47(42.7%)patients received immunotherapy as first-line,second-line,or subsequent to second-line(≥three-line)treatment,respectively.Sixty-nine patients received PD-1 inhibitor combining with radiotherapy,antiangiogenic drugs or chemotherapy,and mono-immunotherapy was applied in 41 patients.2.Efficacy and immune related adverse events of immunotherapy At the end of follow-up,the optimal response mode of progression disease(PD),stable disease(SD),partial response(PD)was assessed in thirty-seven,forty-three,thirty patients,respectively.The objective response rate(ORR),disease control rate(DCR),median progression-free survival(PFS),median overall survival(OS)were 27.3%(all patients from PR),67.4%(SD+PR patients),3.5(95%CI: 0.2-19.9)months and 5.5(95%CI: 0.2-31.8)months,respectively.The differences in baseline PD-L1,NLR,PLR,Alb levels and ECOG PS scores between different optimal response modes(PR,SD,PD)were significantly different.The median PFS(5.8 VS 8.7 months,P = 0.396)and median OS(2.6 VS 20.4 months,P = 0.137)were longer in immune-related adverse events(ir AEs)group than without ir AEs group,but statistical difference wasn’t observed.Twenty-five patients had adverse events of any grade,with immune-related side,and three(2.7%)patients with grade 3-4 grades adverse events.The frequency of various adverse events was followed by 14 cases(12.7%)of immune-related dermatitis,3 cases of fatigue,3 cases of immune-related pneumonia,2 cases of hypothyroidism,1 case of immune-related myocarditis,and 1 case of immune-related colitis.3.Predictors related to PFS and OS in immunotherapy COX univariate analysis showed elevated baseline ANC,NLR,PLR,CRP,CA199,CEA,CYFRA21-1 levels and second-line or third-line treatment,ECOG PS score two was divided into risk factors for PFS and OS in immunotherapy with(P <0.05),but smoking was a protective factor.Multivariate analysis exhibited that smoking was an independent protective factor for PFS and OS(OS HR = 0.1,95% CI: 0.0-0.3,P = 0.001;PFS HR=0.1,95% CI: 0.0-0.4,P <0.001);Elevated baseline CRP levels,lung metastases,and second-line or third-line treatment were significantly associated with shortened PFS and OS.Elevated baseline CYFRA21-1 level was an independent risk factors for dismal prognosis of OS;baseline CA-199,brain metastases,and pleural metastases are significantly associated with shortened PFS.4.OS prognostic nomogram for advanced NSCLC patients treated with PD-1 Inhibitors In the COX univariate analysis,variables related to OS(P <0.05)were screened through stepwise AIC regression.Smoking,liver metastasis,number of treatment lines,ECOG PS score,baseline CYFRA21-1,and CRP levels were selected to establish the OS survival probability prediction model in PD-1 inhibitors treated advanced NSCLC,and plotting 3 months,6 months,and 12 months OS survival probability nomogram.The AUC for ROC curve was 0.81,C-index for the current OS model was 0.81(95% CI: 0.72-0.90),indicated the current OS model had a distinguished discrimination.Decision curve analysis for 6-months survival revealed that patients could benefit from the application of current model to guide clinical treatment decisions.5.Model validation and comparison Results of computer resampling(Bootstrap)500 times implied that the AUC for ROC curve was 0.82(95% CI 0.72-0.89)in resampling model,which indicated that current model had a outstanding prediction performance in validation.Calculate the total score(TPS)of each patient according to Nomogram,divide patients into three subgroups according to the X-tile software determined cut-off points: low risk group(TPS≤118),medium risk group(118<TPS≤189),and high risk group(TPS(>189).Kaplan-Meier survival curve analysis manifested statistically significant differences in OS between groups.The C-index for the current OS model was 0.81(95% CI: 0.72-0.90),which is higher than the previous OS prediction model(Botticelli model,C-index 0.76,95%CI: 0.68-0.81)for immunotherapy NSCLC patients,indicated the current OS model had an distinguished discrimination,and the prediction accuracy wasn’t inferior to Botticelli model.Decision curve analysis for 6-months survival revealed that the current nomogram had a higher net benefit rate than Botticelli nomogram implied that patients could benefit from using current model to guide clinical treatment decisions.Conclusions This study developed a Nomogram scoring tool based on easily available and inexpensive clinical parameters for predicting the OS survival probability of immunotherapy patients with advanced NSCLC.By evaluating the discrimination,calibration and clinical significance of the model,it was shown that this prediction model scoring tool could be valuable to clinicians in individual risk assessment of advanced NSCLC patients before receiving PD-1 inhibitors,and in more accuracy decision of treatment strategies.