Based on its efficacy in clinical trials, daratumumab has become an integral part of the treatment of multiple myeloma (MM) since its first approval in 2016. Until recently, only a few studies have reported on the real-world use of daratumumab, and these reports were based on small patient cohorts.^1-4 Monoclonal Antibodies in Multiple Myeloma: Outcomes after Therapy Failure (MAMMOTH) was the first and to date only multicenter study that investigated the outcomes of daratumumab-refractory patients.⁵ This work included patients with active MM with evidence of progressive disease during treatment with CD38 antibody-containing regimen and described 275 MM patients from 14 academic centers in the United States. The median overall survival (OS) in the entire cohort after refractoriness to a CD38 antibody, occurring at a median of 50.1 months after diagnosis, was 8.6 months. We have previously described a complete, Danish, nationwide cohort of 635 patients with MM who initiated treatment with daratumumab prior to January 1, 2019.⁶ In the present work, we focus on the clinical course and life expectancy of patients who discontinue daratumumab.

The methods of this study regarding patient identification and data collection have been previously described.⁶ In short, we conducted a retrospective review of the clinical course of all patients treated with a daratumumab-containing regimen prior to January 1, 2019. Treatment data were updated until January 1, 2019. The index regimen (IR) was defined as the first daratumumab-containing line of therapy. T₀ was defined as the date of discontinuation of the IR. Based on the IR, patients were classified into four groups: daratumumab–bortezomib–dexamethasone (DVD), daratumumab–lenalidomide dexamethasone (DRD), daratumumab monotherapy (D-mono), and daratumumab in other combinations (D-other). Previous drug exposure at T₀ was assessed based on four classes of drugs: alkylating agents, proteasome inhibitors (bortezomib, carfilzomib, and ixazomib), immunomodulatory agents (thalidomide, lenalidomide, and pomalidomide), and daratumumab (no other CD38 antibody was used in the study population). Patients were classified into four groups at T₀: double class exposed (exposed to daratumumab and another class of drugs), triple class exposed (exposed to daratumumab and two other classes of drugs), quadruple class exposed (exposed to daratumumab and three other classes of drugs) and alkylator–bortezomib–carfilzomib–daratumumab–lenalidomide–pomalidomide-exposed. Post-T₀ lines of therapy 1–6 were termed P1–P6, respectively. The methods of statistical analysis are presented in Supplementary Methods in Appendix S1.

We identified 474 patients who discontinued the IR. The characteristics of these patients are shown in Table S1. The median age at diagnosis was 66 years and 57% of patients were male. Patients received a median of three lines of therapy prior to the IR and 50% of patients had been treated with high-dose melphalan and autologous stem cell transplantation (HDM-ASCT). The IR was DRD in 44%, D-mono in 32%, DVD in 13%, and D-other in 11% of patients. The median duration of the IR was 5 months. The reasons for discontinuation of the IR are shown in Figure S1. The most frequent reasons for discontinuation were progressive disease (42%), toxicity (11%), and insufficient response (8%). The median time from diagnosis to T₀ was 4 years. At T₀, 12 (3%) patients were double class exposed, 53 (11%) were triple class exposed, 350 (74%) were quadruple class exposed, and 59 (12%) were alkylator–bortezomib–carfilzomib–daratumumab–lenalidomide–pomalidomide-exposed.

The median follow-up after T₀ was 9.2 (interquartile range [IQR]: 1.8–17.6) months. The median OS (mOS) in the entire cohort was 12.2 months (confidence interval [CI]: 9.9–14.7; Figure 1A). According to cytogenetic status, mOS was 8.6 months (CI: 5.8–13.3) in high-risk patients and 14.7 months (CI 11.6–22.3) in standard-risk patients (p = .006; Figure 1B). According to the previous IR, mOS was 11.4 months (CI 8.9–16.1) after DRD, 13.3 months (CI: 10.4–20.7) after D-mono, 17.8 months (CI: 12.7-not reached [NR]) after DVD and 5.8 months (CI 1.5–9.2) after D-other (p = .0003; Figure 1C). According to drug class exposure at T₀, mOS was 15.3 months (CI: 8.9-NR) in double class exposed, 22.5 months (CI: 11.3-NR) in triple class exposed, 12.6 months (CI: 10.0–15.8) in quadruple class exposed, and 8.3 months (CI: 4.1–10.6) in alkylator–bortezomib–carfilzomib–daratumumab–lenalidomide–pomalidomide-exposed patients (p = .002; Figure 1D). According to the cause of discontinuation of the IR, mOS was 10 months (CI: 8.5–12.9) in patients who discontinued due to progressive disease and 15.8 months (CI: 11.4–25.7) in patients who discontinued due to other reasons (p = .009; Figure 1E).

After T₀, the number of patients receiving a subsequent line of therapy decreased from 375 (79%) in P1 to 28 (6%) in P6 (Table S2). The overall response rate (ORR) decreased from 44% in P1 to 11% in P6. Median time to next treatment (TNT) decreased from 4.7 months (IQR: 2.0; 8.3) in P1 to 1.8 months (IQR: 0.9; 3.2) in P6. The most frequently used regimens in P1 are shown in Table S3. Figures S2 and S3 show the ORR and TNT in P1 according to the previous IR. Figures S4 and S5 show the ORR and TNT in P1 according to drug class exposure at T₀.

Of the patients who received a subsequent line of therapy after T₀, 192 (51%) were retreated with a daratumumab-containing regimen in P1 (Table S4). The number of patients who had progressed on the IR was 89 (46%) among patients retreated with daratumumab and 147 (80%) among patients treated without daratumumab. The most frequently used regimens in patients retreated with daratumumab were daratumumab–pomalidomide–dexamethasone (DPD; n = 57; 30%), DRD (n = 44; 23%), and DVD (n = 22; 12%). The IR prior to these combinations was D-mono in 41, 23, and 15 of cases, respectively. The most frequently used regimens in patients treated without daratumumab were carfilzomib–dexamethasone (n = 36; 20%), pomalidomide–dexamethasone (n = 22; 12%), and carfilzomib–pomalidomide–dexamethasone (n = 16; 7%). ORR was 48% in patients retreated with and 41% in patients treated without daratumumab (Figure S6). Median TNT was 4.6 months in both groups (Figure S7). Median OS was 23.6 months (CI: 17.5-NR) in patients retreated with and 11.3 months (CI: 9.5–15.1) in patients treated without daratumumab (p < .0001; Figure 1F). After multivariate analysis adjusting for age, previous HDM-ASCT, IR, drug class exposure at T₀, treating site, time from diagnosis to T₀, presence of cytogenetic high-risk markers, and discontinuation of the IR due to progressive disease, daratumumab retreatment was associated with longer OS (Cox regression; hazard ratio 0.59; CI: [0.41–0.87], p = .006).

Based on an unselected nationwide cohort of every patient treated with daratumumab until January 1, 2019, this is the largest real-world study to report the outcomes of patients who discontinue their first daratumumab-containing line of therapy. The mOS of patients who discontinue their first daratumumab-containing line of therapy was 12.2 months, while it was 10 months in patients who discontinued due to progressive disease. Approximately half of the patients who received subsequent therapy were retreated with daratumumab. About 46% of daratumumab-retreated patients were daratumumab-refractory. The most frequently used regimens for daratumumab retreatment were DPD, DRD, and DVD used primarily after daratumumab monotherapy. In our cohort, daratumumab retreatment was associated with a superior OS on multivariate analysis.

A strength of our work is that it investigates a complete, population-based cohort without referral bias. Among the limitations of our work are its retrospective nature and observational design, which resulted in a heterogeneous cohort of patients treated with various daratumumab-containing regimens. Instead of drug class refractoriness, our study reports previous drug class exposure. The life expectancy of patients discontinuing daratumumab was slightly more favorable based on our data than reported in the MAMMOTH study.⁵ This may be explained by differences in methodology. The mOS of 8.3 months in alkylator–bortezomib–carfilzomib–daratumumab–lenalidomide–pomalidomide-exposed patients indicates an unmet medical need. Drug-antibody conjugates, bispecific antibodies, and chimeric antigen receptor T cell therapy are among the promising treatment options for this group of patients. Whether daratumumab retreatment is associated with clinical benefit needs to be explored in prospective randomized clinical trials.

This study was approved on behalf of the Danish Data Protection Agency by The Region of Southern Denmark (Journal 19/52220) and the Danish Patient Safety Authority (Journal 3-3013-2047/1 and 3-3013-2047/2).

基于其在临床试验中的疗效，自 2016 年首次获批以来，达雷妥尤单抗已成为治疗多发性骨髓瘤 (MM) 不可或缺的一部分。直到最近，只有少数研究报告了达雷妥尤单抗的实际使用情况，而这些报告基于小型患者队列。^1-4多发性骨髓瘤中的单克隆抗体：治疗失败后的结果 (MAMMOTH) 是第一项也是迄今为止唯一一项调查 daratumumab 难治性患者结果的多中心研究。⁵这项工作包括在使用含 CD38 抗体的方案治疗期间有进展性疾病证据的活动性 MM 患者，并描述了来自美国 14 个学术中心的 275 名 MM 患者。对 CD38 抗体耐药后，整个队列的中位总生存期 (OS) 为 8.6 个月，发生在诊断后 50.1 个月的中位时间。我们之前描述了一个完整的丹麦全国性队列，该队列由 635 名 MM 患者组成，他们在 2019 年 1 月 1 日之前开始接受 daratumumab 治疗。6^在目前的工作中，我们关注停用 daratumumab 患者的临床病程和预期寿命。

本研究关于患者识别和数据收集的方法已在前面描述过。⁶简而言之，我们对 2019 年 1 月 1 日之前接受含达雷妥尤单抗方案治疗的所有患者的临床过程进行了回顾性审查。治疗数据更新至 2019 年 1 月 1 日。指数方案 (IR) 定义为第一个含有达雷妥尤单抗的治疗线。T ₀被定义为 IR 的终止日期。根据 IR，将患者分为四组：达雷妥尤单抗-硼替佐米-地塞米松 (DVD)、达雷妥尤单抗-来那度胺地塞米松 (DRD)、达雷妥尤单抗单药治疗 (D-mono) 和达雷妥尤单抗其他组合 (D-other)。以前在 T _{0时的药物暴露}根据四类药物进行评估：烷化剂、蛋白酶体抑制剂（硼替佐米、卡非佐米和艾沙佐米）、免疫调节剂（沙利度胺、来那度胺和泊马度胺）和达雷木单抗（研究人群中未使用其他 CD38 抗体）。_{患者在 T 0}被分为四组：双类暴露（暴露于 daratumumab 和另一类药物）、三类暴露（暴露于 daratumumab 和其他两类药物）、四类暴露（暴露于 daratumumab 和其他三类）药物）和烷基化剂-硼替佐米-卡非佐米-达雷木单抗-来那度胺-泊马度胺暴露。后 T ₀线治疗 1-6 分别称为 P1-P6。统计分析方法见附录 S1 的补充方法。

我们确定了 474 名停止 IR 的患者。这些患者的特征见表S1。诊断时的中位年龄为 66 岁，57% 的患者为男性。在 IR 之前，患者接受了中位三线治疗，50% 的患者接受了高剂量美法仑和自体干细胞移植 (HDM-ASCT) 治疗。IR 为 DRD 占 44%，D-mono 占 32%，DVD 占 13%，D-other 占 11%。IR 的中位持续时间为 5 个月。停止 IR 的原因如图 S1 所示。最常见的停药原因是疾病进展（42%）、毒性（11%）和反应不足（8%）。从诊断到 T ₀的中位时间为 4 年。在 T ₀, 12 (3%) 名患者接受双类暴露，53 (11%) 名患者接受三类暴露，350 (74%) 名患者接受四类暴露，59 (12%) 名患者接受烷化剂-硼替佐米-卡非佐米-达雷妥尤单抗-来那度胺-泊马度胺-裸露。

_{T 0}后的中位随访时间为 9.2（四分位距 [IQR]：1.8-17.6）个月。整个队列的中位 OS (mOS) 为 12.2 个月（置信区间 [CI]：9.9–14.7；图 1A）。根据细胞遗传学状态，高危患者的 mOS 为 8.6 个月（CI：5.8-13.3），标准风险患者的 mOS 为 14.7 个月（CI 11.6-22.3）（p  = .006；图 1B）。根据之前的 IR，mOS 为 DRD 后 11.4 个月（CI 8.9-16.1），D-mono 后 13.3 个月（CI：10.4-20.7），DVD 后 17.8 个月（CI：12.7-未达到 [NR]）和 5.8 D-other ( p  = .0003；图 1C)后几个月 (CI 1.5–9.2 )。根据 T _{0时的药物类别暴露}, mOS 为 15.3 个月 (CI: 8.9-NR) 两次暴露, 22.5 个月 (CI: 11.3-NR) 在三级暴露, 12.6 个月 (CI: 10.0-15.8) 在四级暴露, 8.3 个月 (CI : 4.1–10.6) 在烷化剂-硼替佐米-卡非佐米-达雷妥尤单抗-来那度胺-泊马度胺暴露的患者中 ( p  = .002；图 1D)。根据停用 IR 的原因，因疾病进展而停用的患者的 mOS 为 10 个月（CI：8.5-12.9），因其他原因停用的患者的 mOS 为 15.8 个月（CI：11.4-25.7）（p  = .009；图 1E)。

在 T ₀之后，接受后续治疗的患者人数从 P1 的 375 人（79%）减少到 P6 的 28 人（6%）（表 S2）。总体反应率 (ORR) 从 P1 的 44% 下降到 P6 的 11%。到下一次治疗的中位时间 (TNT) 从 P1 的 4.7 个月 (IQR: 2.0; 8.3) 减少到 P6 的 1.8 个月 (IQR: 0.9; 3.2)。P1 中最常用的方案见表 S3。图 S2 和 S3 显示了根据之前的 IR 在 P1 中的 ORR 和 TNT。_{图 S4 和 S5 根据 T 0}时的药物类别暴露显示了 P1 中的 ORR 和 TNT 。

_{在 T 0}之后接受后续治疗的患者中，192 名（51%）在 P1 中接受了含达雷妥尤单抗的治疗方案（表 S4）。在接受 daratumumab 治疗的患者中，IR 进展的患者数量为 89 人 (46%)，在未接受 daratumumab 治疗的患者中为 147 人 (80%)。达雷妥尤单抗复治患者最常用的方案是达雷妥尤单抗-泊马度胺-地塞米松（DPD；n  = 57；30%）、DRD（n  = 44；23%）和 DVD（n  = 22；12%）。这些组合之前的 IR 在 41、23 和 15 例中分别为 D-mono。在未接受达雷妥尤单抗治疗的患者中，最常用的方案是卡非佐米-地塞米松（n = 36; 20%）、泊马度胺-地塞米松（n  = 22；12%）和卡非佐米-泊马度胺-地塞米松（n  = 16；7%）。接受达雷妥尤单抗治疗的患者的 ORR 为 48%，而未接受达雷妥尤单抗治疗的患者的 ORR 为 41%（图 S6）。两组的中位 TNT 为 4.6 个月（图 S7）。接受 daratumumab 治疗的患者中位 OS 为 23.6 个月（CI：17.5-NR），而未接受 daratumumab 治疗的患者中位 OS 为 11.3 个月（CI：9.5-15.1）（p  < .0001；图 1F）。在调整年龄、既往 HDM-ASCT、IR、T ₀时药物类别暴露、治疗部位、从诊断到 T _{0时间的多变量分析后}、细胞遗传学高风险标志物的存在以及由于疾病进展而中断 IR，daratumumab 再治疗与更长的 OS 相关（Cox 回归；风险比 0.59；CI：[0.41–0.87]，p  = .006）。

基于截至 2019 年 1 月 1 日之前接受 daratumumab 治疗的每位患者的未经选择的全国队列，这是报告停止其第一个含 daratumumab 治疗线的患者结果的最大真实世界研究。停止首次含 daratumumab 治疗的患者的 mOS 为 12.2 个月，而因疾病进展而停止治疗的患者为 10 个月。大约一半接受后续治疗的患者接受了达雷妥尤单抗治疗。约 46% 的达雷妥尤单抗复治患者为达雷妥尤单抗难治性。达雷妥尤单抗复治最常用的方案是主要在达雷妥尤单抗单药治疗后使用的 DPD、DRD 和 DVD。在我们的队列中，达雷妥尤单抗再治疗与多变量分析的优越 OS 相关。

我们工作的一个优势在于它调查了一个完整的、基于人群的队列，没有转诊偏倚。我们工作的局限性之一是其回顾性和观察性设计，这导致了一组接受各种含达雷妥尤单抗治疗的患者的异质性队列。我们的研究报告了以前的药物类别暴露，而不是药物类别的耐火度。根据我们的数据，停用达雷妥尤单抗的患者的预期寿命比 MAMMOTH 研究报告的要好一些。⁵这可以通过方法论的不同来解释。烷基化剂-硼替佐米-卡非佐米-达雷妥尤单抗-来那度胺-泊马度胺暴露患者的 mOS 为 8.3 个月，表明医疗需求未得到满足。药物-抗体偶联物、双特异性抗体和嵌合抗原受体 T 细胞疗法是这类患者有希望的治疗选择。需要在前瞻性随机临床试验中探索达雷妥尤单抗再治疗是否与临床获益相关。

本研究由丹麦南部地区（期刊 19/52220）和丹麦患者安全局（期刊 3-3013-2047/1 和 3-3013-2047/2）代表丹麦数据保护局批准。