Analysis Of Risk Factors And Simple Prognostic Model To Predict Survival In Early Relapse/progression With Newly Diagnosed Follicular Lymphoma

Background and objective:Follicular lymphoma(FL)represents the most common indolent non-Hodgkin’s lymphoma in Europe and America,about 22～35%of all new diagnostic NHL.The incidence has risen to 5/100 000 in recent years.FL is a highly heterogeneous disease,and the prognosis for patients with follicular lymphoma(FL)has improved significantly since the introduction of immunochemotherapy,with an OS approaching 80%10 years after FL treatment,but there is still a subset of patients with a worse prognosis.In recent years,foreign studies have found that approximately 17%～28%of patients with FL experience progression of disease(POD)within 2 years of first-line therapy with poor prognosis.clinical indices of risk have been created to identify patients with more aggressive disease and shorter responses to treatment.These mainly include the International Prognostic Index for Follicular Lymphoma(FLIPI),the International Prognostic Index for Follicular Lymphoma 2(FLIPI-2),and PRIMA-PI,help to predict overall survival of FL,but do not accurately predict the risk of POD24.To our knowledge,the domestic research on POD24prediction factors is relatively few and stratification of risk in the setting of disease progression and its influence on subsequent patient survival has not been previously validated in FL.Clinical POD24 prediction tools are not simple and effective.Therefore,there is an urgent need to establish a clinical prediction model for POD24,which can help identify the high-risk population of POD24 and adopt different treatment strategies for intervention.We retrospectively analyzed 144 cases of NDFLpatients data from our center to determine whether POD24 is predictive of inferior OS in this disease.And the study analyzed the prognostic value and risk factors of POD24,aiming to construct a predictive model of POD24 to establish a prognostic score system suitable for the newly diagnosed follicular lymphoma population through our cohort study.Methods:Retrospective analysis was performed on 144 NDFL patients admitted to the First Hospital of Bethune,Jilin University from January 2010 to September 2019.All pat ients met the latest World Health Organization(WHO)diagnostic and classification standards(2016edition)and staged according to the Ann Arbor staging system.Chi-square test or Fisher’s exact probability method were used to analyze the correlation with clinical characteristics.The prognostic factors OS survival analysis and Cox regression analysis were carried out by Kaplan-Meier curve.Logistic regression was used to analyze the influencing factors of POD24,and a prediction model was constructed.Results:1.POD24 incidence and the influence factors of OS in NDFL patients:For the 144patients with newly diagnosed Follicular Lymphoma,median age was 53years(range,28 to88 years),and 52.8%of patients were male.Variables predictive for OS were elevated lactate dehydrogenase level(LDH),elevatedβ2-microglobulin(β2-MG),absolute lymphocyte count(ALC)≤1×10~9/L,Grade 3 FL histologic finding,Stage III/IV,longest diameter of the largest involved node(Lo DLIN)>6cm,high-risk FLIPI scores,high-risk FLIPI-2 scores,high-risk PRIMA-PI,immunofree chemotherapy,POD24.In multivariate analysis,POD24were independent predictors of survival with a hazard ratio(HR)of 409.269(95%CI:5.496～30034.2,P=0.006).Of the 85 patients enrolled in the Pod24 study,36 developed POD24(42.4%),and the median OS was only 44 months in Pod24 patients,while the median OS was not reached in non-POD24 patients(HR:12.12,95%CI:4.397～33.41,P<0.0001).2.Risk factor analysis for POD24:Univariate Logistic regression analysis showed that elevated LDH,high-risk PRIMA-PI,Ki-67≥30%,and negative expression of CD10 were risk factors for POD24.Multivariate analysis showed that high-risk PRIMA-PI(OR=3.697,95%CI:1.211～11.293,P=0.022),Ki-67≥30%(OR=6.056,95%CI:1.412～25.968,P=0.012),negative expression of CD10(OR=4.410,95%CI:1.096-15.637,P=0.036)was an independent risk factor for POD24.3.Establishment of a predictive model for POD24:high-risk PRIMA-PI,Ki-67≥30%,and negative CD10 expression were independent prognostic factors for POD24.After further inclusion in the prediction model,the Logistic regression coefficientβvalue was 1.308,1.801 and 1.421,respectively,and Patients were assigned a score of 2 for Ki-67≥30%,and 1for the other two variables.The ROC curve was drawn according to the scores and survival status,and the area under the curve was 0.763(95%CI:0.650～0.876).The optimal cut-off value of 3 points was taken as the critical point.According to the scores,the patients were divided into low-risk group(0～2 scores)and high-risk group(3～4 scores).4.Verification of POD24 prediction model:In 73 NDFL patients with complete clinical data,46 patients with NDFL score<3,of which 35 patients(76.1%)did not develop POD24;POD24 occurred in 20 of 27 patients(74.1%)with NDFL score≥3.Therefore,the positive prediction rate(PPV)was 74.1%,and the negative prediction rate(NPV)was 76.1%.The positive likelihood ratio(PLR)was 5.65,suggesting that patients with NDFL score≥3 were5.65 times more likely to develop POD24 than those with score 0 to 2.In addition,the POD24 predictive model can also predict patients’OS.The median OS of NDFL patients in the low-risk group was not reached;The median OS for high-risk NDFL patients was 35months,and the difference was statistically significant(P<0.0001).Conclusions:1.In multivariate analysis,POD24 Was independent predictors of survival which was associated with markedly reduced OS compared with the reference group.2.High-risk PRIMA-PI,Ki-67≥30%,and negative expression of CD10 were independent prognostic factors for POD24.The POD24 prediction model was established The patients were divided into low-risk group(0～2 scores)and high-risk group(3～4 scores).3.Validation of prognostic models:The risk of POD24 in the high-risk group was 5.65times higher than that in the low-risk group.The median OS of the high-risk NDFL patients was only 35 months,and the median OS of the low-risk NDFL patients had not yet reached.4.Based on high-risk group of PRIMA-PI combined with high expression of Ki-67≥30%and negative expression of CD10,a POD24 prediction model was established to help identify the risk of POD24 and predict OS,providing a theoretical basis for timely individualized treatment strategies.