B2 microglobulin is a novel prognostic marker of Angioimmunoblastic T-cell lymphoma

The aim of the present study was to analyze features and explore parameters that can help to predict prognosis for angioimmunoblastic T-cell lymphoma (AITL). A total of 117 patients with AITL were retrospectively analyzed. Multivariate analysis showed that β2 microglobulin (β2-M) ≥4.0 mg/L (P = 0.020), rash/pruritus (P = 0.004), performance status (PS) ≥2 (P = 0.006), age >60 years (P = 0.006) and extranodal sites (ENSs) >1 (P = 0.029) were independent risk factors for OS. Rash/pruritus (P = 0.007), age >60 years (P = 0.035) and ENSs >1 (P = 0.006) were independent risk factors for PFS. A novel prognostic model consisting of β2-M, rash/pruritus, PS, age and ENSs >1 was constructed. The model classified patients into 3 risk stratifications: low risk (0 or 1 factor), intermediate risk (2 factors), high risk (≥3 factors) and significantly stratified patients with AITL (P < 0.001). In conclusion, except for PS ≥2, age >60 years and ENSs >1 used in IPI, β2-M and rash/pruritus also indicated adverse prognosis. That we constructed model was commendably prognostic for OS and PFS.

Treatment Response Evaluation. Treatment responses including complete remission (CR), partial remission (PR), stable disease (SD), and progressive disease (PD) were evaluated according to the response criteria of Cheson et al. 17 . The evaluation to prognostic included overall survival (OS) and progression-free survival (PFS). OS was defined as the time from diagnosis to last follow-up or death resulting from any cause. PFS was calculated from the date of diagnosis to the date of disease progression, relapse and death from any cause or the last follow-up. Follow-up of patients not experiencing any of these events was censored at the date of last contact.
Statistical Analysis. Statistical analyses were performed using IBM SPSS statistics software, version 21.0 and GraphPad Prism 6. OS and PFS distributions were estimated using the Kaplan-Meier curve analysis, time-to-event distributions were compared using the log-rank test and two-tailed significance-level of 0.05 was considered statistically significant. Comparisons of clinical and prognostic factors were performed using the Pearson's chi-squared test. Multivariate analysis was performed with a Cox hazards regression model using forward/backward stepwise method with the use of threshold values for removal from and addition to the model of P = 0.10 and P = 0.05 respectively. Results are expressed as hazard ratios (HRs) and 95% confidence intervals (CIs).
Total 64 patients received the anthracycline containing arms as the first-line treatment. Among the 64 patients, 58 patients received CHOP regimen, 2 patients received CHOPE regimen, 2 patients received EPOCH regimen, 1 patients received Hyper CVAD regimen, 1 patients received M-BOCAD (Methotrexate, Bleomycin, Cyclophosphamide, Vincristine and Pirarubicin) regimen. Total 30 patients accepted combination chemotherapy without an anthracycline but based on Gemcitabine and Platinum arms which plus prednisone or thalidomide. Three patients received oral chemotherapy containing Cyclophosphamide and Thalidomide and another 3 patients did never get any therapy and were observed only. Primary treatment failure changed to another therapy including chemotherapy and/or autologous or allogeneic treatment.
In the eligible 100 patients followed up among 117 patients, the median survival time was 22 months. 3-year OS rate for the entire group was 43.3%, and 3-year PFS rate was 27.5% (Fig. 1). Of the 97 cases with treatment, the ORR (overall response rate) was 77.3%, with CR and PR rates of 26.8% and 50.5%. In cases treated with anthracycline containing arms, 3-year OS and PFS rate was 45.1% and 26.4% and with that of 47.5% and 36.5% for gemcitabine and platinum based arms. But there was no statistical significance (P = 0.816) in the two groups. For the patients with oral chemotherapy or without treatment, the longest survival time was 3.2 months.

Discussion
Lymphoma is a kind of malignant tumor with strong heterogeneity and includes many subtypes with distinct prognosis. Though being a rare disease, AITL is the second most common subtype (18.5%) of peripheral T cell lymphoma or natural killer/T-cell lymphoma and has typical clinical and pathologic features different from other nodal peripheral T-cell lymphomas 3,5 . Studies showed that AITL originated from germinal center T-helper cells characteristically expressing CXCL13 and presented EBV infected cells. Use of the new immunostains for CXCL13 might help to improve diagnostic accuracy of AITL 18,19 . In present study, CXCL 13 (+) was in 74.0% and 71.0% cases contained a variable number of EBER-positive cells. However, it exerted no effect on survival whether or not presence of EBV-infected cells and expressing CXCL13, which was accordant with previous reports 3,4,19 .
The characteristics can be summarized as followers: More than half of the patients encountered high Ann Arbor stages (stage III/IV), extranodal involvement, B symptoms, anemia, lymphocytopenia, monocytoza, elevated LDH, elevated β2-M or hypoalbuminemia respectively, but all of them has no statistical significance to OS in univariate analysis except elevated β2-M. Rash/pruritus were observed in 29.9%.The two symptoms of rash/pruritus and rash/pruritus showed a significant correlation (P = 0.015) in present analysis. Spleen involvement (45.1%) is commonly seen than liver involvement (13.3%) or bone marrow involvement (16.0%), but only liver involvement, rather than the other two, is of statistical significance to OS in univariate analysis. Unlike other subtypes T cell lymphoma, AITL is inclined to suffer serious immunodeficiency. In recent 3 years, studies have certified that the frequent coexistence of somatic mutations in the Rho GTPase RhoA (RhoAG17V) and loss-of-function mutations in the 5-methylcytosine oxidase TET2 which may account for immunoinflammatory responses were associated with AITL 20,21 . On account of immunodeficiency, opportunistic infections increase and intense chemotherapy is difficult to conduct which also result in poor prognosis. In this study, the median survival time was 22 months and 3-year OS rate was 43.3%.
Although there are many known prognostic factors, such as the IPI and PIT, there are still no effective clinical indices that can be used for the prognostic stratification of AITL patients. IPI and PIT are controversial for the application to AITL. In this analysis, for IPI and PIT, it seems to be of statistical significance to prognostic strati- So we aimed to explore prognostic factors and novel model to better evaluate prognosis. In this study, statistical analysis proved that elevated β2-M, rash/pruritus, PS ≥2, age ≥60 years and ENSs >1 were independent prognostic factors for AITL while β2-M and rash/pruritus were not used in IPI or PIT prognostic model. That we constructed novel prognosis model with β2-M (P = 0.020), rash/pruritus (P = 0.004), PS ≥2 (P = 0.006), age ≥60 years (P = 0.006) and ENSs >1(P = 0.029) perfectly stratified patients with AITL (Fig. 2a,b).  Several factors may support the established prognostic model. First of all, rash/pruritus is a relatively common clinical feature of AITL in the scale of lymphoma and was also regarded as one of the prognostic factors associated with shorter survival for AITL by Archimbaud 24,25 . In terms of β2-M, it has been reported to be adverse prognostic index in various subtypes of lymphoma [26][27][28] . Several decades years ago, Swan F, Jr. 29 questioned Ann Arbor staging system and insisted that the serum levels of β2-M might provide a more precise system for defining risk groups in large-cell lymphomas. Khouri IF and Rodriguez J. 30,31 also evaluated that the high β2-M at transplantation was an adverse prognostic factor in patients with diffuse mantle cell lymphoma and peripheral T-cell lymphoma after ASCT. B2-M being a potentially prognostic risk factor for AITL, the possible mechanisms are as follows: on the one hand, β2-M participates in immune recognition and loss of functional β2-M through genomic alterations or other yet uncharacterized mechanisms leads to lack of HLA-I expression and escape from CTL 11 . On the other hand, β2-M mirrors tumor burden in many tumors 29 . Above all, β2-M and rash/pruritus used in our

Characteristic
No.   In this study, that the median survival time was 22 months indicates adverse prognosis. To date, the optimal therapy for patients with AITL remains unestablished. Although anthracycline-based regimens were commonly used as the first-line therapy, these regimens had disappointing results 8,23,32 . Consolidation ASCT at the first CR remains under debate 3,32-35 . A prospective randomized controlled trial about the efficacy and safety of GDPT with standard CHOP regimen for patients with newly diagnosed PTCL was carried out at our center between 2010 and 2016 and Li L demonstrated that GDPT arms were better than CHOP arms 36 . In this study, 64 patients were treated with Anthracycline-based regimens and 30 patients were treated with Gemcitabine and Platinum regimens. There was no difference between the two arms (P = 0.816). The negative result different from Li L may result from disequilibrium of the two treatment groups.  New treatment regimens have also been studied. It has been evaluated the efficacy and safety of Chidamide in relapsed or refractory PTCL and shown a favorable efficacy and an acceptable safety profile [37][38][39] . Because of the EBV-positive B cells in AITL, Rituximab has been reported effective and might provide a potential therapeutic target 40,41 . Yang J. 42 discovered that some monoclonal antibodies (m-Abs) specific to human β2-M could induce apoptosis via recruiting MHC class I molecules to lipid rafts and activating Lyn and PLCg2. Such m-Abs may offer a possibly therapeutic approach to AITL on account of its immunodeficiency and aberrant β2-M protein expression. Shi C. also suggested that β2-M was a promising new therapeutic target for human cancers 43 . In recent 3 years, studies has demonstrated that recurrently hypermethylated genes involved in T-cell receptor signaling and T-cell differentiation likely contribute to lymphomagenesis in AITL and speculated that targeting of TCR-related events may hold promise for the treatment of TFH-derived lymphomas 44,45 .
In conclusion, aggressive AITL had poor prognosis and anthracycline-based regimens failed to improve survival. It is necessary to confirm whether immune modulation, antiangiogenic agents and therapeutic targets are effective to improve survival for AITL. Further studies will be needed to explore the pathogenesis and search new therapeutic targets. B2-M is an adverse prognostic factor in AITL independent of IPI or PIT and m-Abs specific to human β2-M may provide a potential therapeutic target. Our prognostic model significantly stratify patients with AITL, which makes it possible to accurately evaluate prognosis. Further studies are welcome to evaluate our model.