Clinical Features Of Primary Gastrointestinal Lymphoma In A Single Center And Prognostic Correlation Of Diffuse Large B-cell Lymphoma Subtype

Background:Lymphoma is the most common blood cancer in my country.According to the primary site,it can be divided into intra-nodal lymphoma and extra-nodal lymphoma.Primary gastrointestinal lymphoma(PGIL)is the most common extranodal lymphoma,accounting for approximately 30-40% of extranodal lymphomas.The most common primary site is the stomach,and the most common pathological type is diffuse large B-cell lymphoma(DLBCL),followed by mucosa-associated lymphoid tissue(MALT)lymphoma.At present,the clinical characteristics and prognosis of PGIL patients reported at home and abroad are quite different,and there is a lack of relevant research on PGIL in the central and western regions.The median age of primary gastrointestinal DLBCL(PGI-DLBCL)is 60-70 years old,and the proportion of male patients is slightly higher than that of female patients.The clinical manifestations are mainly non-specific symptoms of the digestive system,which are easy to be ignored and delay diagnosis and treatment.The 5-year overall survival(OS)rate is 70-80% in foreign countries and 55-85% in China.Therefore,how to optimize the risk classification of PGI-DLBCL Stratification and individualized treatment are still very important issues at present.In recent years,the classification of DLBCL has become more and more refined,from GEP,Hans,NEJM four classifications,Nat Med five classifications to Cancer Cell seven classifications,but this classification is only suitable for about half of the patients with nodal DLBCL.It is worth noting that some studies have shown that there are differences in the pathogenesis and gene mutations between PGI-DLBCL and intranodal DLBCL.The commonly used Ann Arbor staging system,Musshoff staging system,Lugano staging system,and TNM staging system(Paris staging),but which most accurate and simple method is still controversial,and it is difficult to meet its clinical needs.The current prognostic scoring system applied to PGI-DLBCL includes IPI score,R-IPI score,aa IPI score and NCCN-IPI score,etc.,but it is mainly used to evaluate the risk and prognosis of intranodal DLBCL.Therefore,it is necessary to find new simple and effective predictors for PGI-DLBCL patients.Studies in recent years have revealed that pro-inflammatory cytokines(or chemokines)and inflammatory cells in the tumor microenvironment can promote tumor development,DNA damage,corresponding angiogenesis,immunosuppression,and the systemic inflammatory response caused by it,which is closely related to the occurrence and development of various tumors.Inflammation-related indicators can be used to predict the prognosis of tumors,among which NLR,PLR,LMR,SII,SIRI,d NLR,βLR,LLR,AAPR and other indicators have important prognostic value in solid tumors such as colorectal cancer,lung cancer,cervical cancer,and liver cancer.However,the prognostic and clinical significance of these inflammatory indicators in PGI-DLBCL is still unclear.Malnutrition is a common problem in cancer patients,because the rapid growth of tumors will increase the consumption of albumin and cholesterol,and the inflammatory response will also reduce the synthesis of albumin in the liver.Poor nutritional status will also increase the toxicity caused by chemotherapy.Affect therapeutic dose and affect chemotherapy response.At present,nutritional indicators such as PNI and CONUT score are valuable prognostic indicators in esophageal cancer,colorectal cancer,cholangiocarcinoma and other tumors,but there is no relevant research on them in PGIDLBCL.In summary,it is necessary to construct a prognostic prediction model that includes these indicators based on comprehensive analysis of inflammatory and nutritional indicators to accurately and individually predict the prognosis of PGI-DLBCL patients.Objectives:1.Explore the most suitable staging method on the basis of retrospective analysis of the prognostic factors of PGI-DLBCL;2.Explore the impact of inflammatory indexes and nutritional indexes on the prognosis of PGI-DLBCL patients;3.Construct a prognosis prediction model for PGI-DLBCL,to Individualized prediction of patient survival.Methods:1.Used the inpatient medical record system,combined with the inclusion and exclusion criteria,the clinical information of PGIL patients who were treated in our hospital from 2010 to 2022 were collected,followed up by telephone,and the pathological and clinical characteristics of PGIL patients were retrospectively analyzed.2.Select the PGI-DLBCL patients from PGIL patients,conduct univariate and multivariate COX regression survival analysis,and study the prognostic factors and independent risk factors.3.Through multivariate COX regression analysis and construct the clinical prediction model of OS in PGI-DLBCL patients.And visualize the model through the nomogram,evaluate the predictive ability of the model through the time-dependent ROC curve and survival analysis,evaluate the accuracy of the model through the calibration chart,and evaluate the clinical usefulness of the model through the decision curve analysis(DCA)curve.Results:1.The first part,clinical features of PGIL patients:A total of 215 PGIL patients were included,with a median age of 56 years and a male to female ratio of 1.56:1.The most common pathological type was firstly DLBCL,accounting for 58.1% of the total,followed by MALT lymphoma,accounting for 19.1% of the total.Compared with primary gastric lymphoma(PGL),primary intestinal lymphoma(PIL)is more likely to penetrate the intestinal wall,has B symptoms,more number of extranodal lesions,higher ECOG-PS score,and easier perforation,anemia and hypoalbuminemia are more likely to be combined,and the stage is higher.The 5-year OS rate of PIL patients was 52.4%,while the 5-year OS rate of PGL patients was 76.2%,which was higher than that of PIL patients.Compared with B lymphocyte PGIL,T lymphocyte PGIL has more extranodal lesions,higher ECOG-PS score,more perforation,hypoalbuminemia and abnormal β2 microglobulin(β2-MG)are more likely.2.The second part,pathology,clinical characteristics and survival analysis of PGIDLBCL patients:2.1 125 PGI-DLBCL patients were included,whose median age was 57 years and male to female ratio was 1.27:1.Among the affected sites,the stomach was the most common,accounting for 64.8%.PFS rate: 75.7% in 1 year,67.6% in 3 years,60.5% in 5 years,57.4%in 10 years;OS rate: 83.3% in 1 year,74.9% in 3 years,70.1% in 5 years,66.5% in 10 years.Among them,the PFS rate and OS rate of patients before 2016 were lower.2.2 The results of univariate survival analysis showed that the common prognostic factors of PFS and OS in patients with PGI-DLBCL were: number of extranodal lesions,TNM-N stage,PS score,IPI score,NCCN-IPI score,LDH,d NLR,SIRI,PNI;the independent prognostic influencing factor of PFS is: β2-MG;the independent prognostic influencing factors of OS are: peripheral lymph node invasion,surrounding tissue invasion,Musshoff stage,LMR,LLR,and treatment methods.Univariate survival analysis suggested that patients with higher d NLR had poorer prognosis.Patients with higher SIRI had poorer prognosis.Patients with a higher PNI had a better prognosis.Patients with higher LMR had longer OS.Patients with higher LLR had shorter OS.In the staging system,the stages in the Musshoff stage are more balanced,and there are more differences in the prognosis of the staging groups,which are prognostic factors for OS.Relatively speaking,the Musshoff stage is better.Among the prognostic risk stratification scores,the NCCN-IPI prognostic score has more differences among risk stratifications,and NCCN-IPI is better.2.3 Multivariate survival analysis showed that the independent prognostic factors of PFS in PGI-DLBCL patients were: ECOG-PS,LDH,d NLR;the independent prognostic factors of OS were: NCCN-IPI,SIRI,d NLR.3.The third part,the construction and verification of the prognosis prediction model of PGI-DLBCL:A clinical prediction model of OS in PGI-DLBCL patients based on NCCN-IPI,SIRI and d NLR was constructed and a nomogram was drawn.It can be seen from the calibration figure that the accuracy of the model is good,and it can predict the survival of patients more accurately.By comparing the time ROC curve,the predictive ability of the model(NCCNIPI+SIRI+d NLR)is significantly improved for NCCN-IPI,and it is also better than the(NCCN-IPI+SIRI)model.The DCA decision curve shows that the clinical benefit of the research model(NCCN-IPI+SIRI+d NLR)is better,and it is better than the(NCCNIPI+SIRI)model and NCCN-IPI.Conclusion:1.215 PGIL patients were followed up in our center for 12 years.The 5-year and 10-year PFS rates were 53.1% and 51.3%,and the OS rates were 67.1% and 61.4%.The most common site of disease is the stomach,followed by the intestinal tract,and the most common pathological type is DLBCL.Most of them are B-NHL,with less T-NHL and NKTNHL.T and NKT-NHL have more adverse clinical features,higher stage and worse prognosis.PIL has more adverse clinical features,higher stage and worse prognosis.2.In the stage of PGI-DLBCL,Musshoff stage is better than other stage,but the value of its staging is still not significant.The stratification of the NCCN-IPI score in the prognostic score of PGI-DLBCL is relatively good,but there are still problems such as unbalanced stratification and no difference between some stratifications.3.This study found that d NLR,PNI,and LLR were the prognostic factors of PGIDLBCL,and d NLR was an independent risk factor for PGI-DLBCL.SIRI and LMR are also prognostic factors.4.A clinical prognosis prediction model composed of NCCN-IPI,SIRI and d NLR was constructed with good accuracy.Compared with(NCCN-IPI+SIRI)model and NCCN-IPI,the predictive ability and clinical usefulness of this model are better.