Prognostic significance of serum aspartic transaminase in diffuse large B-cell lymphoma

We reviewed medical records of 207 patients who were diagnosed, according to the 2016 World Health Organization classification, with de novo DLBCL at the Affiliated Hospital of Jiangnan University and Affiliated Hospital of Nantong University between 2006 and 2016. Blood samples of all patients were collected at the time of diagnosis or before the beginning of treatment. No patients with relapse after treatment were included in this study. Likewise, patients with primary central nervous system lymphoma, primary mediastinal B-cell lymphoma, post-transplant lymphoproliferative disorders, transformed NHL, and HIV-positive DLBCL were excluded from the study. Further cases also excluded from the study, including patients with a history of hepatitis (9/207, 4.3%), chronic alcohol ingestion (5/207, 2.4%), cirrhosis, recent medication for hepatotoxicity (2/207, 1.0%), and fatty liver (2/207, 1.0%), patients with pretreatment liver function injury due to viral infection (10/207, 4.8%), and patients with concomitant or past liver cancer (none). Consequently, 179 patients with DLBCL qualified for the study. All enrolled patients were treated with R-CHOP (rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone) or R-CHOP-like chemotherapy. Because this is a retrospective study, all recruited patients were informed and verbal consent was obtained in accordance with the requirements of the Declaration of Helsinki, and the research project was approved by the review boards of Jiangnan University, Affiliated Hospital of Jiangnan and Nantong Universities.

X-tile 3.6.1 software (Yale University, New Haven, CT, USA) was used to determine the optimal cut-off values for ALT and AST. However, the optimal cutoff value for ALT was not found. Thus, we used 50 U/L for ALT according to the reference value in a liver function test. The optimal cutoff value was 33.3 U/L for AST according to the X-tile software recommendation. Results provided by the X-tile software showed that a dichotomy in the AST level had a better prognostic value than a trichotomy (Additional file 1: Figure S1 and Additional file 2: Figure S2).

The DLBCL microarray and corresponding clinical data in the present study were downloaded from the public GEO database (http://​www.​ncbi.​nlm.​nih.​gov/​geo/​) with the accession number GSE27255. Raw data from the GEO database were extracted using the R Bioconductor limma and impute packages (http://​bioconductor.​org/​biocLite.​R). The data were presented as Volcano and Heatmap plots. The selected mRNAs were examined in cell lines of germinal center B-cell (GCB) and activated B-cell (ABC) subtype to explore if any significant differences in expression exist. Using Mann-Whitney U test, results of fold change > 1 and P < 0.05 between GCB and ABC cell lines were considered significant.

Immunohistochemistry was performed on 4-μm formalin-fixed paraffin-embedded sections. The antibodies used in the study were for CD20 (clone L26, Abcam, cutoff: 30%), CD10 (clone 56C6, Dako, cutoff: 30%), Bcl6 (clone LN22, Dako, cutoff: 30%), MUM1 (clone MUM1p, Dako, cutoff: 30%), and Myc (clone Y69; Abcam, cutoff: 40%). The cutoff values for each antibody have been previously described [7].

Univariable tests were used to compare categorical (Chi-square and Fisher’s exact) and continuous (Student’s t test or Kruskal-Wallis test when appropriated) variables. Survival curves were plotted using the Kaplan–Meier method and were compared using log-rank test. According to Cheson 2014, overall survival (OS) was defined as the time from diagnosis to death; patients who remained alive were censored at the last date of follow-up [8]. Statistical analysis was performed using SPSS software, Version 20.0. For all the tests, P < 0.05 (2-sided) was considered statistically significant.

The median age of the cohort was 57 years (range 18–88 years old). The median follow up time was 28 months (3–112 months). The median treatment cycle was 6 cycles (4–8 cycles). The frequencies with elevated ALT and AST levels were 12.3% (22/179) and 24.6% (44/179), respectively. The baseline clinical parameters of the patients are presented in Table 1.

DLBCL patients with different cells of origin (COO) have distinct outcomes; therefore, we analyzed the differently expressed genes in the GCB and ABC subtypes. We speculated that higher ALT or AST mRNA levels were more frequent in the ABC subgroup. Thus, the Human DLBCL of the Affymetrix Human Gene 1.0 ST Array microarray data were downloaded from the public GEO database. Differently expressed genes in the DLBCL cell lines of the GCB and ABC subtypes (Table 2) were analyzed. A total of 25 genes were differentially expressed as shown in the Volcano and Heatmap plots (Figs. 1 and 2). However, the expression of ALT and AST mRNA was similar between the GCB and ABC cell lines. These results suggest that the mRNA levels of ALT or AST may not be reliant on the COO.

In the present cohort, ALT-positive and ALT-negative patients had similar clinical features. However, AST positivity was significantly associated with advanced Ann Arbor stage (stage III or IV) (P = 0.001), poor performance status (PS) (P = 0.014), elevated lactate dehydrogenase (LDH) level (P < 0.0001), more extra nodal involvement (ENI, ENI ≥ 2) (P = 0.027), high-risk IPI (IPI 3–5) (P = 0.002), presence of B symptoms (P = 0.001), hepatitis B virus surface antigen (HBsAg) positivity (P = 0.045) (Fig. 3a), Myc protein positivity (P < 0.0001) (Fig. 3b), and non-GCB subtypes (P = 0.038) (Table 3).

To extend our findings, survival analysis was performed in our cohort. ALT-positive patients showed similar OS to the negative cases (Fig. 4a). However, AST-positive cases, showed significantly decreased OS compared with the negative ones (2-year OS, 49.2% vs. 81.1%, P < 0.0001) (Fig. 4b). Other clinical and pathological factors were also associated with poor survival, including B symptoms, non-GCB subtype, advanced Ann Arbor stage, poor PS, elevated LDH level, ENI ≥ 2, and high-risk IPI (Fig. 4c-h). In subgroup analysis, patients with higher AST levels showed poorer outcomes in patients without B symptoms ((Fig. 5a) and LDH positive groups (Fig. 6c). Higher AST levels showed poorer survival in patients of different ENI (both ENI < 2 and ENI ≥ 2 groups) (Fig. 5c-d), COO (both GCB and non-GCB groups) (Fig. 5e-f) and IPI risk showed (both IPI 0–2 and IPI 3–5 groups) (Fig. 6a, b). Patients of different AST level had similar OS with either Myc protein positive or negative group (Fig. 6e-f).