Glasgow Prognostic Score Predicts Prognosis in Patients with Angioimmunoblastic T-Cell Lymphoma

Introduction Glasgow prognosis score (GPS) is an inflammation-based prognostic scoring system that combines clinical tumor markers including C-reactive protein and albumin. However, its role in predicting the prognosis of angioimmunoblastic T-cell lymphoma (AITL) is unclear. The prognosis index for peripheral T-cell lymphoma (PIT) has been widely used to evaluate the prognosis of various types of T-cell non-Hodgkin's lymphoma. In this study, we conducted a retrospective analysis involving 106 patients newly diagnosed with AITL to determine the prognostic value of GPS and compare with PIT model in patients with AITL.

Methods A total of 106 patients newly diagnosed with AITL received standardized chemotherapy at Sun Yat-Sen University Cancer Center and the First Affiliated Hospital of Hainan University between July 2009 and October 2019 were enrolled in our study. The baseline patient characteristics were collected, including Eastern Cooperative Oncology Group performance status, B symptoms, lactate dehydrogenase level, albumin level (ALB), CRP level, extra nodal invasion, bone marrow involvement, and Ann Arbor stage . GPS was calculated according to serum CRP and ALB values. Patients with elevated CRP (>10 mg/L) and hypoalbuminemia (<35 g/L) were rated with a score of 2; patients with CRP >10 mg/L or hypoalbuminemia (<35 g/L) were rated with a score of 1; and patients with neither of these abnormalities were rated with a score of 0. Four variables were used to build up PIT score: age (≤60 versus >60), performance status (ECOG ≤1 versus >2), LDH level (low versus high) and BM involvement (negative versus positive). Depending on the number of adverse prognostic factors (0, 1, 2 or ≥3), patients were classified into low-(0,1) or high-risk (2, ≥3) groups, respectively. Progression-free survival (PFS) and overall survival (OS) rates were the primary endpoints of this study.

Results Patients were divided into three groups by GPS: GPS=0, GPS=1 and GPS=2. Among them, 74 (69.8%), 22 (20.8%), and 10 (9.40%) received cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP or similar CHOP regimens); etoposide, doxorubicin, vincristine, cyclo-phosphamide and prednisone (EPOCH); gemcitabine and oxaliplatin (GEMOX) regimens, respectively. The results showed that the survival outcomes among the three groups were significantly different. The 5-year PFS rate (61.0% vs. 38.7% vs.0.00%; P<0.001) and 5-year OS rate (76.0% vs. 43.4% vs. 0.00%; P<0.001) were significantly better in the patients with GPS=0 than in those with GPS=1 and GPS=2. The OS and PFS of the patients with GPS=1 were significantly lower than those of patients with GPS=0 (P=0.003, 0.007). In addition, the OS of patients with GPS=1 was significantly better than that of patients with GPS=2 (P=0.021). However, no significant difference was found in their PFS (P=0.144).

Furthermore, the proportion of patients with GPS=2 receiving GEMOX and EPOCH was significantly higher than that of patients with GPS=0 or GPS=1. (P=0.001). GEMOX and EPOCH regimens may produce better outcome than CHOP regimen in aggressive peripheral T-cell lymphoma. However, the long-term survival of patients with GPS=2 was significantly worse than those of patients with GPS=0 and GPS=1(P<0.001, =0.021), further confirming the beneficial effect of the GPS model in distinguishing the prognosis of patients.

The GPS prognostic model also could effectively identify patients with poor prognosis in the low-risk group by PIT score. According to the PIT prognostic model, 54.7% of the patients were assigned to the low risk group. In those patients, approximately 28.3%, 44.3%, and 27.4% of patients exhibited GPS=0, 1, 2, respectively. In the low-risk group by PIT score, we get similar results of the GPS model mentioned above to evaluate the prognosis.

Univariate and multivariate analysis were conducted to determine the prognostic value of GPS in AITL. The results showed that GPS≥1 (P<0.001) were an independent adverse prognostic factor for predicting PFS (95% CI: 1.668–6.626) and OS (95% CI: 2.214–14.686).

Conclusion Our study reveals that GPS is an effective prognostic model for patients with AITL. It tends to balance the distribution of patients in the three risk groups and has better prognostic significance than PIT in low-risk groups. Further studies are needed to explain the underlying mechanism of the relationship between high GPS and low survival outcomes in AITL.