Treatment is indicated in Waldenström macroglobulinemia (WM) when clinical manifestations arise due to the IgM paraprotein or lymphoplasmacytic infiltrate.¹ The main classes of drugs used for WM treatment include monoclonal antibodies, chemotherapeutic agents, proteasome inhibitors, and Bruton Tyrosine Kinase inhibitors (BTKi). The most frequently used chemotherapeutic agent is bendamustine, which is largely administered for its efficacy and relatively favorable toxicity profile with low rates of non-hematological adverse events.² Bendamustine plus rituximab (BR) is considered to be especially useful in fit patients in need of rapid disease control.³

Rituximab has been evaluated in combination not only with bendamustine (BR), but also with dexamethasone and cyclophosphamide (DRC).⁴ A retrospective comparison between BR and DRC was performed at Mayo Clinic, reporting a significantly greater two-year progression free survival (PFS) for BR (88% vs. 61%, respectively) but without a significant difference in overall response rate (ORR) at 18 months (93% vs. 96%, respectively).⁴ In contrast, both a superior ORR (98% vs. 85%, respectively) and major response rate (96% vs. 60%, respectively) were reported for BR in the study of Abeykoon et al.⁵

To further evaluate the efficacy and safety of the different chemoimmunotherapy regimens used for the treatment of WM patients, we conducted a retrospective multicenter study involving 14 different Italian centers of the Fondazione Italiana Linfomi (FIL). The study group included unselected, consecutive WM patients who received frontline treatment with a chemoimmunotherapy regimen between January 2008 and December 2022. Primary outcome measures included ORR, PFS, and OS, as defined by the International Workshop on Waldenström's macroglobulinemia.⁶ Primary outcomes were also assessed based on the total received bendamustine dose, which was categorized using the percentage of the relative dose intensity (RDI). The RDI was calculated for each patient as the percentage of drug actually administered compared to that estimated at the beginning of treatment. The starting dose was 90 mg/m²/day, administered on days 1 and 2 of each of the six planned cycles.

The statistical analyses were conducted on four different treatment groups which included BR, DRC, and two other groups that were named “other R-chemo” and “chemo alone.” The “other R-chemo” group included patients treated with a Rituximab-containing regimen other than BR, such as Chl-R (chlorambucil-rituximab), FCR (fludarabine-cyclophosphamide-rituximab), or R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone), whereas the “chemo alone” group included patients treated with chemotherapeutic agents such as chlorambucil or cyclophosphamide as single agents. Categorical variables were analyzed with the Chi-square or Fisher's exact tests, while numerical variables underwent assessment with the Wilcoxon–Mann–Whitney test or Kruskal–Wallis ANOVA test. Survival analysis was conducted with the Kaplan–Meier method. The statistical analyses were performed using the NCSS 2020 Statistical Software by NCSS, LLC, Kaysville, Utah, USA (ncss.com/software/ncss).

We enrolled 547 WM patients, among which 245 received BR, 116 DRC, 86 other R-chemo, and 52 chemo alone; 48 patients treated with Rituximab monotherapy only for IgM-associated symptoms were excluded from the analysis. Baseline characteristics of the four analyzed groups of patients are shown in Table S1. The BR and DRC populations differed with respect to age at treatment initiation (68 vs. 72 years old, respectively, p = .018), comorbidities (higher CIRS and more frequent presence of respiratory disease in DRC, p < .001 and p = .025, respectively), and IPSSWM score and beta2microglobulin levels (higher values in BR, p < .001).

ORR was 93.3% for BR, 79.2% for DRC, 75% for other R-chemo, and 44% for chemo alone (Table S2). The difference in terms of ORR was statistically significant when comparing BR to DRC (HR 3.71 (1.88–7.31), p < .001) and BR to other R-chemo (OR 4.75 (2.34–9.64), p < .001). When analyzing PFS curves, we noted a 4-year PFS of 80% for BR, 68% for other R-chemo, 60% for DRC, and 25% for chemo alone (Figure 1). The difference was significant between BR and DRC (p < .001) but not between BR and other R-chemo (p = .143) or between DRC and other R-chemo (p = .362). OS curves did not differ between the first three main treatment groups (4-year OS was 86% for BR, 89% for DRC, and 93% for other R-chemo), whereas a reduced 4-year OS was observed for patients treated with chemo alone (58%; Figure 1B).

Multivariate analysis identified age >75 years (RR 1.83), anemia with Hb <10 g/dL (RR 1.63), and choice of treatment (DRC rather than BR; RR 1.86) as significant variables impacting PFS (Table S3).

Regarding tolerability, a significantly higher rate of hematological toxicities was observed in BR (55%) compared to DRC (38%) and other R-chemo (31%) treated patients, resulting in significantly more frequent dose reductions in the BR (14.3%) than in the DRC (6.0%) or other R-chemo group (4.9%) [HR 0.38 (0.16–0.89), p = .026] (Table S4).

BR patients were also analyzed according to the different dose of bendamustine administered. A first comparison was performed between 129 patients treated with the recommended starting dose of 90 mg/m² and 59 patients treated with a starting dose of 70 mg/m²; both groups of patients received no dose reduction during treatment. Patients receiving the lower dose were older (median age 70 vs. 64 years, p = .005), whereas the other baseline characteristics were similar between the two subgroups (Table S5). No significant difference was observed with respect to 4-year PFS, which was 84% for the 90 mg/m² and 85% for the 70 mg/m² subgroup (Figure 1C). A second comparison was performed between 206 patients treated with an RDI of ≥70% and 34 patients treated with an RDI of <70%. In this case there was a significant difference, with a 4-year PFS of 83% in the first subgroup and 64% in the second subgroup (p = .035, Figure 1D). Finally, we compared the outcome of patients treated with a >30% or ≤30% bendamustine RDI reduction with the outcome of the DRC treated patients. BR-treated patients with a RDI reduction >30% showed the same outcome as DRC-treated patients in terms of PFS (Figure 1E).

Univariate analysis identified the following variables associated with bendamustine RDI reduction of >30%: age (both >65 years and >75 years), presence of a grade 2 neuropathy according to CTCAE, creatinine clearance (CrCl) <70 mmol/L, and ECOG>1. In multivariate analysis, when using age >65 or ≤65 years as a variable, only neuropathy and CrCl were found to be significant, whereas when using age >75 or ≤75 years as a variable, neuropathy and age were significant (Table S6).

To our knowledge, this is one of the largest retrospective real-life studies on treatment-naive WM patients. The comparison between 245 BR patients and 116 DRC patients, even if retrospective, is unique because of the size of the subgroups, allowing us to obtain robust results on efficacy and tolerability for the main chemoimmunotherapy regimens. The obtained data show a significantly higher 4-year PFS for BR compared to DRC, confirming the results of Paludo et al.⁴ over a longer median follow-up of 54 months.

We also investigated the optimal starting bendamustine dose and show that a reduction from the recommended 90 mg/m² to 70 mg/m² is equivalent in terms of PFS. It is worth noting, however, that patients receiving the lower dose were generally older, which could have resulted in better compliance and fewer treatment discontinuation because of toxicities and side effects. In contrast, we show that RDI reduction of >30% resulted in reduced PFS that was comparable to that of DRC. A significant impact of bendamustine dose on PFS was also reported by Arulogun et al.³ although in that study a total bendamustine dose of < or ≥1000 mg/m² was predictive of PFS.

In conclusion, we report that the BR regimen remains a highly effective treatment option for untreated WM patients even in the era of BTKis. This regimen showed a high ORR and exhibited excellent PFS even when the initial bendamustine dose was reduced to 70 mg/m² and dose intensity was decreased up to 30% from the initial dose. Only in case of a relative dose intensity reduction of >30%, the PFS of BR-treated patients was comparable to those treated with DRC. Age over 75 years and CrCl lower than 70 mmol/L were the main risk factors for this significant dose reduction. Patients with these characteristics are likely to benefit less from the BR regimen and should be considered for alternative treatments.

Waldenström 巨球蛋白血症 （WM） 的治疗指征是由于 IgM 副蛋白或淋巴浆细胞浸润而出现的临床表现。¹ 用于 WM 治疗的主要药物类别包括单克隆抗体、化疗剂、蛋白酶体抑制剂和布鲁顿酪氨酸激酶抑制剂 （BTKi）。最常用的化疗药物是苯达莫司汀，其主要因其疗效和相对有利的毒性特征和非血液学不良事件发生率低而被使用。² 苯达莫司汀加利妥昔单抗 （BR） 被认为对需要快速疾病控制的健康患者特别有用。³

利妥昔单抗不仅与苯达莫司汀 （BR） 联合使用，还与地塞米松和环磷酰胺 （DRC） 联合使用进行了评估。⁴ 妙佑医疗国际对 BR 和 DRC 进行了回顾性比较，结果 BR 的两年无进展生存期 （PFS） 显著更高（分别为 88% 和 61%），但 18 个月时的总缓解率 （ORR） 没有显著差异（分别为 93% 和 96%）。⁴ 相比之下，在 Abeykoon 等人的研究中，BR 的 ORR（分别为 98% 和 85%）和主要缓解率（分别为 96% 和 60%）均较高5。

为了进一步评估用于治疗 WM 患者的不同化学免疫治疗方案的有效性和安全性，我们进行了一项回顾性多中心研究，涉及意大利 Fondazione Italiana Linfomi （FIL） 的 14 个不同的意大利中心。研究组包括 2008 年 1 月至 2022 年 12 月期间接受化学免疫治疗方案一线治疗的未经选择的连续 WM 患者。主要结局指标包括 ORR、PFS 和 OS，由 Waldenström 巨球蛋白血症国际研讨会定义。⁶ 主要结局也根据接受的苯达莫司汀总剂量进行评估，该剂量使用相对剂量强度 （RDI） 的百分比进行分类。每位患者的 RDI 计算为实际给药的药物百分比与治疗开始时估计的百分比。起始剂量为 90 mg/m²/天，在 6 个计划周期中每个周期的第 1 天和第 2 天给药。

对四个不同的治疗组进行了统计分析，其中包括 BR、DRC 和另外两个被命名为 “其他 R-化疗” 和 “单独化疗” 的组。“其他 R-化疗”组包括接受除 BR 以外的含利妥昔单抗方案治疗的患者，例如 Chl-R（苯丁酸氮芥）、FCR（氟达拉滨-环磷酰胺-利妥昔单抗）或 R-CHOP（利妥昔单抗、环磷酰胺、多柔比星、长春新碱和泼尼松），而“单独化疗”组包括接受化疗药物（如苯丁酸氮芥或环磷酰胺）作为单药治疗的患者。分类变量使用卡方或 Fisher 精确检验进行分析，而数值变量使用 Wilcoxon-Mann-Whitney 检验或 Kruskal-Wallis 方差分析检验进行评估。使用 Kaplan-Meier 方法进行生存分析。使用美国犹他州凯斯维尔 （ncss.com/software/ncss） 的 NCSS， LLC 的 NCSS 2020 统计软件进行统计分析。

我们招募了 547 例 WM 患者，其中 245 例接受 BR，116 例接受 DRC，86 例其他 R 化疗，52 例单独化疗;48 例仅接受利妥昔单抗单药治疗 IgM 相关症状的患者被排除在分析之外。表 S1 显示了分析的四组患者的基线特征。BR 和 DRC 人群在治疗开始时的年龄 （分别为 68 岁和 72 岁，p = .018）、合并症 （DRC 中较高的 CIRS 和更频繁的呼吸系统疾病存在，分别为 p < .001 和 p = .025） 以及 IPSSWM 评分和 β2 微球蛋白水平 （BR 中的较高值，p < .001）。

BR 的 ORR 为 93.3%，DRC 为 79.2%，其他 R 化疗为 75%，单独化疗为 44%（表 S2）。将 BR 与 DRC （HR 3.71 （1.88–7.31），p < .001） 和 BR 与其他 R 化疗 （OR 4.75 （2.34–9.64），p < .001 进行比较时，ORR 方面的差异具有统计学意义。在分析 PFS 曲线时，我们注意到 BR 的 4 年 PFS 为 80%，其他 R 化疗为 68%，DRC 为 60%，单独化疗为 25%（图 1）。BR 和 DRC 之间的差异显著 （p < .001），但 BR 与其他 R 化疗之间 （p = .143） 或 DRC 与其他 R-化疗之间 （p = .362） 之间差异不显著。前三个主要治疗组之间的 OS 曲线没有差异 （BR 的 4 年 OS 为 86%，DRC 为 89%，其他 R 化疗为 93%），而观察到单独化疗的患者 4 年 OS 降低 （58%;图 1B）。

多变量分析确定了年龄 >75 岁 （RR 1.83）、贫血伴 Hb <10 g/dL （RR 1.63） 和治疗选择（DRC 而不是 BR;RR 1.86）作为影响 PFS 的重要变量（表 S3）。

在耐受性方面，与 DRC （55%） 和其他 R 化疗组 （38%） 相比，BR （31%） 的血液学毒性发生率显着更高，导致 BR （14.3%） 的剂量减少频率显著高于 DRC （6.0%） 或其他 R 化疗组 （4.9%） [HR 0.38 （0.16-0.89），p = .026]（表 S4）。

还根据苯达莫司汀的不同剂量分析 BR 患者。对 129 例接受推荐起始剂量 90 mg/m² 治疗的患者和 59 例接受 70 mg/m² 起始剂量治疗的患者进行了首次比较;两组患者在治疗期间均未接受剂量减少。接受较低剂量的患者年龄较大 （中位年龄 70 岁 vs. 64 岁，p = .005），而两个亚组之间的其他基线特征相似 （表 S5）。在 4 年 PFS 方面未观察到显著差异，90 mg/m² 亚组为 84%，70 mg/m² 亚组为 85%（图 1C）。在 206 例 RDI 为 ≥70% 的患者和 34 例 RDI 为 <70% 的患者之间进行了第二次比较。在这种情况下存在显着差异，第一个亚组的 4 年 PFS 为 83%，第二个亚组为 64%（p = .035，图 1D）。最后，我们将 >30% 或 ≤30% 苯达莫司汀 RDI 降低治疗患者的结局与 DRC 治疗患者的结局进行了比较。RDI 降低 >30% 的 BR 治疗患者在 PFS 方面显示出与 DRC 治疗患者相同的结果（图 1E）。

单变量分析确定了以下与苯达莫司汀 RDI 降低 >30% 相关的变量：年龄 （>65 岁和 >75 岁）、根据 CTCAE 存在 2 级神经病变、肌酐清除率 （CrCl） <70 mmol/L 和 ECOG>1。在多变量分析中，当使用年龄 >65 或 ≤65 作为变量时，仅发现神经病变和 CrCl 是显着的，而当使用年龄 >75 或 ≤75 作为变量时，神经病变和年龄是显着的（表 S6）。

据我们所知，这是针对初治 WM 患者的最大回顾性真实研究之一。245 名 BR 患者和 116 名 DRC 患者之间的比较，即使是回顾性的，也是独一无二的，因为亚组的规模很大，使我们能够获得关于主要化学免疫治疗方案的疗效和耐受性的可靠结果。获得的数据显示，与 DRC 相比，BR 的 4 年 PFS 显著更高，证实了 Paludo 等人 4 在较长的中位随访 54 个月中的结果。

我们还研究了最佳起始苯达莫司汀剂量，并表明从推荐的 90 mg/m² 减少到 70 mg/m² 在 PFS 方面是等效的。然而，值得注意的是，接受较低剂量的患者通常年龄较大，这可能会导致更好的依从性并减少因毒性和副作用而停止治疗的次数。相比之下，我们表明 >30% 的 RDI 降低导致 PFS 降低，与 DRC 相当。Arulogun 等人也报告了苯达莫司汀剂量对 PFS 的显着影响³，尽管在该研究中，苯达莫司汀总剂量为 < 或 ≥1000 mg/m² 可预测 PFS。

总之，我们报告说，即使在 BTKis 时代，BR 方案仍然是未经治疗的 WM 患者的高效治疗选择。该方案显示出高 ORR，并且即使初始苯达莫司汀剂量减少至 70 mg/m² 并且剂量强度从初始剂量降低高达 30% 时，也表现出出色的 PFS。仅在相对剂量强度降低 >30% 的情况下，BR 治疗患者的 PFS 与 DRC 治疗的患者相当。年龄超过 75 岁和 CrCl 低于 70 mmol/L 是剂量显着减少的主要危险因素。具有这些特征的患者可能从 BR 方案中获益较少，应考虑替代治疗。