Clinical Features,Prognosis Analysis,and Diagnostic Model Establishment Of Lymphoma With The Fever Of Unknown Origin As The Main Manifestation

Background and objectives:Fever of unknown origin(FUO)is a group of heterogeneous and diagnostically challenging diseases,with lymphoma(LYM)being the leading cause of tumor-related etiology.Patients of lymphoma with FUO(FUO LYM)lack specific disease-directed manifestations,also,there are no specific blood tests available to diagnose these conditions,and therefore,doctors often have to perform multiple biopsies to confirm the diagnosis.FUO LYM patients are often confused with other diseases,leading to delayed diagnosis and treatment,and can experience rapid deterioration,making promoting recognition and treatment a great challenge for clinicians.This study retrospectively analyzed the clinical characteristics and prognostic risk factors of patients with lymphoma presenting primarily as fever of unknown origin(FUO LYM).A diagnostic model of multi-indicator was established and its diagnostic efficacy was evaluated to assist clinicians in making a faster diagnosis.As a malignant tumor,the emergence and progression of FUO LYM are closely related to the immune system,and lymphocyte subpopulations are an important component for evaluating immune function.This study attempted to conduct a prospective trial to explore the application value of flow cytometry to detect PD-1(programmed death-1),Ki-67,and other subgroups of lymphocytes in the prognosis assessment and early diagnosis of FUO LYM.Methods:Part 1:Clinical Characteristics and Prognostic Factors of FUO LYM and Establishment of a Diagnostic Model:a Long-term Follow-up Study of a Singlecenter CohortWe conducted a retrospective analysis of 40 patients diagnosed with lymphoma in the FUO cohort of Peking Union Medical College Hospital(designed as the "FUO LYM group")between 2014 and 2018.We matched this group with 40 patients diagnosed with tuberculosis(the "FUO TB group")from the FUO cohort and 40 patients without fever manifestations of lymphoma(the "NF LYM group"),based on the closest admission time.We compared the clinical information of the two lymphoma groups and evaluated the prognostic factors of FUO LYM.We established a multi-index combined diagnostic model by comparing two groups of patients with FUO and drew a nomograph,and then evaluated the diagnostic performance of the model using ROC and calibration curves.Part 2:Exploration of the Value of Peripheral Blood Lymphocyte Subsets Mainly Based on PD-1 and Ki-67 Expression Levels Detection in Prognostic Evaluation and Diagnosis of FUO LYM PatientsProspective exploratory study:A total of 50 hospitalized patients with FUO were recruited at PUMCH between 2018 to 2022,of which 25 were finally diagnosed with lymphoma(FUO LYM group),and 25 were non-lymphoma patients designed as the FUO group.In addition to collecting clinical information,all patients underwent lymphocyte subset detection,measuring the expression levels of indicators including HLA-DR(Human Leukocyte Antigen-DR),CD38(Cluster of differentiation 38),Ki-67,PD-1,etc.in various lymphocyte subtypes.We also recruited 26 healthy individuals(healthy control group)for flow cytometry testing and analyzed the lymphocyte subset characteristics of FUO LYM patients through inter-group comparisons.We evaluated the effects of various lymphocyte subset biomarkers on the prognosis of FUO LYM through multiple factor regression analysis and compared the variables between the FUO LYM and FUO groups,screened for predictive model indicators used for early identification of FUO LYM.We established a multi-index combined diagnostic model based on T lymphocyte Ki-67 and PD-1 expression levels and evaluated the diagnostic performance of the model using ROC and calibration curves.Results:Part 1:Clinical Characteristics and Prognostic Factors of FUO LYM and Establishment of a Diagnostic Model:a Long-term Follow-up Study of a Singlecenter Cohort1 The average time from onset to diagnosis was 361 days,and the proportion of hospitalizations,empirical antimicrobial therapy,diagnostic anti-tuberculosis therapy,and steroid treatment before the diagnosis was higher in FUO LYM compared to NF LYM.FUO LYM patients required more invasive procedures or surgical pathology for definitive diagnosis.Comparison between FUO LYM and FUO TB showed no statistical differences in terms of time from onset to diagnosis,empirical antimicrobial therapy,and use of advanced antibiotics,suggesting similar pre-diagnosis clinical conditions between the two groups.2 The total mortality rates of the three groups at the last follow-up were:FUO LYM 52.5%,NF LYM 27.5%,and FUO TB 10%.Survival analysis showed significant statistical differences between FUO LYM and NF LYM patients,with a shorter survival time and poorer prognosis observed in the FUO LYM group.COX regression analysis showed the following prognostic factors for FUO LYM patients:smoking,IPI score of 35,central nervous system(CNS)involvement,and combined infections for overall survival(OS);and serum natrium(Na),blood urea nitrogen(BUN),IPI score of 3-5,and combined gastrointestinal bleeding for progress free survival(PFS).3 Compared with FUO TB,we established a multi-index combined diagnostic model based on C-reactive protein(CRP),serum ferritin(SF),lactate dehydrogenase(LDH),T8 cell count,T.spot-TB,splenomegaly,and standard uptake value(SVUmax)values for FUO LYM,and draw a nomogram.The ROC curves and calibration curves demonstrated that the predictive model exhibited sound diagnostic performance.Part 2:Exploration of the Value of Peripheral Blood Lymphocyte Subsets Mainly Based on PD-1 and Ki-67 Expression Levels Detection in Prognostic Evaluation and Diagnosis of FUO LYM Patients1 The count of lymphocyte subsets in FUO LYM patients was significantly lower than that in healthy controls,and most subset counts were lower than those in the FUO group.The percentages of PD-1 and Ki-67 expression levels in T cells,B cells,and NK cells were generally higher in FUO LYM patients than in healthy controls and the FUO group,while the proportion of T cell developmental stage cells was significantly lower,and the proportion of sophisticated subsets differed significantly from the majority of the other two groups.2 The COX regression analysis showed that factors related to potential influences on OS in FUO LYM patients were CD 19+Ki-67+/CD 19+and factors related to potential influences on PFS were CD4+Ki-67+/CD4+.3 A predictive model based on the joint detection of CD4+CD28+/CD4+and CD4+PD1+/CD4+showed an area under the ROC curve of 0.9424,a sensitivity of 0.96,a specificity of 0.92,and a calibration curve distribution in the diagonal direction,consistent with the ideal curve,indicating a high diagnostic efficiency for the discriminative diagnostic model.Conclusion:Part 1:Clinical Characteristics and Prognostic Factors of FUO LYM and Establishment of a Diagnostic Model:a Long-term Follow-up Study of a Singlecenter CohortFUO LYM is highly heterogeneous and difficult to differentiate from FUO TB,and these patients have poor prognoses.This study established a multi-indicator joint prediction model to provide a reliable method for the early identification of FUO LYM.Part 2:Exploration of the Value of Peripheral Blood Lymphocyte Subsets Mainly Based on PD-1 and Ki-67 Expression Levels Detection in Prognostic Evaluation and Diagnosis of FUO LYM PatientsPatients with FUO LYM exhibited unique lymphocyte subset changes,characterized by immunosuppression-dominant immune dysregulation,which can also guide the assessment of patient prognosis.A multi-indicator prediction model based on lymphocyte subset analysis can assist in the differentiation of FUO with lymphoma.