头颈部弥漫大B细胞淋巴瘤患者临床特征及预后模型分析研究

背景与目的：头颈部弥漫大B细胞淋巴瘤（head and neck diffuse large B-cell lymphoma，HN-DLBCL）是该解剖部位常见侵袭性非霍奇金淋巴瘤，随着含有利妥昔单抗免疫化疗方案以及局部放疗和细胞免疫治疗的成功应用，HN-DLBCL患者的缓解率和无病生存率较以往有显著提高，然而仍有一部分患者成为复发/难治性病例。前期临床研究发现，基于预后模型分层治疗可以使高危淋巴瘤患者得到早期充分干预而显著降低其复发或成为难治性病例的概率。系统分析HN-DLBCL患者的临床特点与其预后的相关性，并比较不同预后模型的分层能力。方法：回顾性分析2010年1月—2018年12月在上海交通大学医学院附属第九人民医院诊治的134例HN-DLBCL患者的临床资料，应用Kaplan-Meier法计算生存率并绘制生存曲线，log-rank检验和COX回归模型进行单因素及多因素生存预后分析，进一步采用国际预后指数（International Prognostic Index，IPI）、美国国家综合癌症网络-IPI（National Comprehensive Cancer Network-IPI，NCCN-IPI）和西班牙淋巴瘤/自体骨髓移植工作组-IPI（Spanish Lymphoma/Autologous Bone Marrow Transplant Working Group-IPI，GELTAMO-IPI）进行危险度分层，对比各预后分层系统不同危险程度患者3年生存率的差异。结果：单因素分析结果显示，Ann Arbor临床分期、非生发中心B细胞（non-germinal center B-cell，non-GCB）亚型、血清乳酸脱氢酶（lactate dehydrogenase，LDH）和β2微球蛋白水平是HN-DLBCL患者预后的影响因素（P<0.05），而多因素COX分析发现，LDH水平是影响患者预后的独立危险因素。根据IPI、NCCN-IPI和GELTAMO-IPI评分系统，高危HN-DLBCL患者3年总生存率（overall survival，OS）分别为44.7%、36.8%和32.8%，3年无进展生存率（progression-free survival，PFS）分别是44.3%、20.7%和16.3%，与IPI预后模型相比，NCCN-IPI和GELTAMO-IPI评分系统更能甄别高危HN-DLBCL患者。结论：在免疫化疗治疗时代，血清LDH水平是HN-DLBCL患者预后的独立危险因素，而改良国际预后模型NCCN-IPI和GELTAMO-IPI较IPI评分更能分辨高危患者。

Retrospective analytical study on clinical features and prognostic models in patients with head and neck diffuse large B-cell lymphoma

Background and purpose: Head and neck diffuse large B-cell lymphoma (HN-DLBCL) is one of the most common aggressive non-Hodgkin's lymphoma involving this anatomical site. With the clinical application of rituximab-containing immunochemotherapy and local radiotherapy as well as cellular immunotherapy, treatment outcomes of HN-DLBCL patients, including not only rates of response, but also progression-free survival (PFS), have been dramatically improved. However, a number of patients become relapsed/refractory cases. Previous clinical studies showed that risk-adapted treatment strategies could allow prompt and tailored intervention for high-risk patients thereby significantly reducing probabilities of being relapsed or refractory cases. The present study aimed to systematically analyze clinical features and their prognostic values in HN-DLBCL patients andcompare stratification efficacies of different prognostic models. Methods: A total of 134 patients with HN-DLBCL in Shanghai Ninth People’s Hospital from Jan. 2010 to Dec. 2018 were retrospectively identified and systemically analyzed. Survival curves were estimated by using Kaplan-Meier survival methodology. The prognostic influences of clinical factors on survival were studied by univariate log-rank test and multivariate Cox proportional hazards regression models. Three prognostic models including International Prognostic Index (IPI), National Comprehensive Cancer Network-IPI (NCCN-IPI) and Spanish Lymphoma/Autologous Bone Marrow Transplant Working Group-IPI (GELTAMO-IPI) were applied for risk stratification to compare 3-year survival rates among patients in different risk groups. Results: Univariate analysis identified prognostic factors for HN-DLBCL patients, including Ann Arbor stage, non-germinal center B-cell (non-GCB) subtype, serum lactate dehydrogenase (LDH) and β2 microglobulin value, while multivariate Cox regression analysis revealed that serum LDH level was the independent prognostic indicator. The predicted 3-year overall survival (OS) of high-risk HN-DLBCL patients stratified by IPI, NCCN-IPI and GELTAMO-IPI were 44.7%, 36.8% and 32.8%, while their 3-year PFS were 44.3%, 20.7% and 16.3%, respectively. NCCN-IPI and GELTAMO-IPI demonstrated enhanced discrimination compared with IPI for high-risk HN-DLBCL cases ［3-year PFS (IPI vs NCCN-IPI vs GELTAMO-IPI): 44.3% vs 20.7% vs 16.3%］. Conclusion: In immunochemotherapy era, LDH level was an independent prognostic factor in HN-DLBCL patients. Modified prognostic models including NCCN-IPI and GELTAMO-IPI might be more powerful than IPI in predicting high-risk patients.