Multiple myeloma (MM) disproportionately affects Blacks compared to individuals belonging to other racial groups.¹ Since the introduction of novel anti-myeloma agents, multiple studies have found disparate survival outcomes among Black patients compared to White MM patients.^{2, 3} Limited access to novel therapies and autologous hematopoietic cell transplantation (auto-HCT) has been considered partly responsible for the worse survival outcomes in Black patients but other disease-related features could also play a contributing role.⁴ Patients receiving auto-HCT probably have equitable access to healthcare, which may nullify the impact of disparate health access on outcomes.⁵ Therefore, to investigate the true impact of race biology on outcome in MM, we compared survival outcomes of Blacks versus Whites who underwent upfront auto-HCT at our center via propensity-matched analysis.

A total of 705 Black and White MM patients, who underwent auto-HCT at our institution from 2007 to 2015 and had complete cytogenetics and fluorescence in-situ hybridization (FISH) information at the time of diagnosis, were included for analysis. Race was self-reported by study participants according to race categories (Black and White). In our study population, we included only patients who self-identified as Black and White. The term “Black” was defined as any individual that descended from any Black racial or ethnic group of Africa. Patients who had disease progression after initial induction treatment or had missing data (n = 37) were excluded from the study. The baseline patient characteristics and survival outcomes of the overall population (n = 668) are presented in Tables S1 and S2 and Figures S1 and S2. In order to obtain bias-corrected comparisons of clinical outcomes among Black and White patients, we conducted 1:1 propensity score matching without replacement. We used the linear logit propensity score as the distance measure and set the caliper of 0.2 standard deviations of the propensity score.⁶ Patients in both groups were matched for age at transplantation, International Staging System stage, serum creatinine, response to induction therapy, induction, consolidation, and preparative regimens, and maintenance therapy. The response was based on the International Myeloma Working Group response criteria and was evaluated before and after auto-HCT.⁷ The primary outcome variables were progression-free survival (PFS) and overall survival (OS). PFS was defined as the elapsed time from the date of transplant to the date of disease progression or death, whichever occurred earlier. OS was defined as the time from the date of transplantation to the date of death or the date of the last follow-up. Patients who were alive and did not experience progressive disease at the time of the last follow-up were administratively censored.

Out of 126 Black patients, a total of 125 (99.2%) were matched to whites via 1:1 propensity-score matching without replacement. Baseline patient characteristics for the matched cohort are summarized in Table S3. Overall, 121 (96.8%) and 119 (95.2%) patients received induction therapy with immunomodulatory drugs (IMiD) and/or proteasome inhibitors (PI) in the Black and the White group, respectively. The use of IMiD-PI-dexamethasone combination as induction was similar in both groups at 60/125 (48.0%) and 61/125 (48.8%) in the Black and White cohort, respectively. Overall response rate (ORR) before auto-HCT was 91.2% (114/125) and 92.0% (115/125) in the Black and White group, respectively. The numbers of patients with complete response (CR), very good partial response, and partial response were 10 (8.0%), 47 (37.6%), and 54 (43.2%), respectively, in the Black group, and 11 (8.8%), 45 (36.0%), and 55 (44.0%), respectively, in the White group. Melphalan alone was used as a preparative regimen in 114 (91.2%) and 113 (90.4%) patients in the Black and White groups, respectively. The combination of busulfan and melphalan was used as a preparative regimen in 11 (8.8%) and 12 (9.6%) patients in the Black and White groups, respectively. Overall, 97 (77.6%) and 99 (79.2%) patients received maintenance therapy in the Black and White groups, respectively. In the Black group, maintenance therapy with either IMiDs or PI was used in 75 (60.0%) and 8 (6.4%) patients, respectively. The corresponding number of patients for the White group was similar at 74 (59.2%) and 8 (6.4%), respectively.

Patients in the Black group had a similar ORR post-auto-HCT of 97.6% (122/125), compared with 99.2% (124/125) in the White group, and with the same observed CR rate of 27.2% (34/125) (Table S4). The median follow-up time for the matched cohort was 71.5 (interquartile range: 51.6–90.3) months. The median PFS time was 44.6 (95% confidence interval [CI], 35.5–54.7) months and 51.0 (95% CI: 38.3–63.9) months for Black and White patients, respectively (p = .763; Figure 1A). The 4-year PFS rates were 48.0% (95% CI, 39.9%–57.8%) for the African Americans (AA) and 51.2% (95% CI, 43.0%–61.0%) for the White group. The median OS time was 101.1 (95% CI: 94.3–not reached) and 109.8 (95% CI: 86.2–not reached) for Black and White patients, respectively (p = .579; Figure 1B). The 4-year OS rates were 78.5% (95% CI: 71.5%–86.2%) and 80.9% (95% CI: 74.1%–88.2%) for AA and White patients, respectively.

In addition, due to the separation of the PFS curve lines of two groups at longer time-points, we looked at the survival of the two groups at different time intervals. We looked at survival at 12-month time intervals starting at 36 months from transplant up to 96 months (Table S5). We utilized a stratified log-rank test to compare the difference between the two curves. The p-value of .763 indicated that the two curves (Black cohort and White cohort) are not statistically significantly different (Figure S3).

Prior disparity studies have drawn disparate conclusions on treatment outcomes between Blacks and Whites. In a large retrospective study, including more than 15 000 MM patients from the Veteran Affairs hospitals across the United States, the AA patients had superior survival compared to Whites.⁸ This study categorized their patient population as AA and White patients. Although the study had its limitations, with the lack of clinically annotated cytogenetics data and 97.8% of the study population being male, there was a similar utilization rate of novel therapies and auto-HCT in both AA and Whites that contributed to similar survival of both racial groups. In another retrospective study based on Multiple Myeloma Research Foundation CoMMpass registry, the authors concluded that AA have poor survival compared to Whites.⁹ However, in this study, AA were less likely to receive triplet therapy and auto-HCT compared to Whites. Furthermore, when the analysis was limited to patients receiving novel therapies and auto-HCT, the difference in OS between AA and Whites was attenuated.

A previous retrospective study demonstrated that in a comparative study between Black and White MM patients, t(11;14) was associated with a higher OS of Black patients compared to their White counterparts.¹⁰ From our cohort, we identified 10 Black patients and 13 White patients with t(11;14). The standardized mean difference of −0.091, indicated that t(11;14) was balanced between the two matched race subgroups (Table S6). In multivariable analysis adjusting race and t(11;14), neither race nor t(11;14) status was statistically significantly associated with survival outcomes of OS or PFS (Tables S7 and S8; Figures S4 and S5).

While our study showed that Black and White patients have similar outcomes, there are some limitations in our study. The study has a limitation of being a single-center retrospective study; however, our population represents the current real-world population with the majority of patients being treated with PI, IMiD, and auto-HCT. Another limitation of our study is that while we conducted a 1:1 propensity-matched analysis, this may artificially correct biological differences between our two patient populations. We believe social determinants of health are the largest contributing factor for the differences in mortality observed between Black and White patients. While propensity matching creates a balanced data set, allowing a simple and direct comparison between the populations, the social determinants of health between the population may have been artificially corrected for. Another limitation of our study is that our center serves as a referral center from many outside hospitals and therefore, our population who is receiving auto-HCT may not be representative of the general MM population in the community settings. Our study has several strengths, including being the first study to conduct propensity-matched analysis to analyze the racial impact on outcomes in AA and Whites to minimize bias. Our study also included patients who had complete information about cytogenetics and FISH. In conclusion, our study establishes that Black patients, if given optimal MM therapies and auto-HCT, have similar survival as White patients.

与属于其他种族群体的个体相比，多发性骨髓瘤 (MM) 对黑人的影响更大。¹自从引入新型抗骨髓瘤药物以来，多项研究发现黑人患者与白人 MM 患者的生存结果不同。^{2, 3}获得新疗法和自体造血细胞移植 (auto-HCT) 的机会有限被认为是黑人患者生存结果较差的部分原因，但其他与疾病相关的特征也可能起作用。⁴接受 auto-HCT 的患者可能可以公平地获得医疗保健，这可能会抵消不同的健康获得对结果的影响。⁵因此，为了研究种族生物学对 MM 结果的真正影响，我们通过倾向匹配分析比较了在我们中心接受预先自动 HCT 的黑人与白人的生存结果。

共有 705 名黑人和白人 MM 患者被纳入分析，他们于 2007 年至 2015 年在我们机构接受了自体 HCT，并且在诊断时具有完整的细胞遗传学和荧光原位杂交 (FISH) 信息。种族由研究参与者根据种族类别（黑人和白人）自我报告。在我们的研究人群中，我们只包括自我认定为黑人和白人的患者。“黑人”一词被定义为来自非洲任何黑人种族或族裔群体的任何个人。初始诱导治疗后出现疾病进展或数据缺失的患者 ( n  = 37) 被排除在研究之外。总体人群的基线患者特征和生存结果 ( n= 668) 列在表 S1 和 S2 以及图 S1 和 S2 中。为了获得黑人和白人患者临床结果的偏差校正比较，我们进行了 1:1 倾向评分匹配，无需替换。我们使用线性 logit 倾向得分作为距离度量，并将倾向得分的标准差设为 0.2。⁶两组患者在移植年龄、国际分期系统分期、血清肌酐、诱导治疗反应、诱导、巩固和准备方案以及维持治疗方面进行匹配。响应基于国际骨髓瘤工作组响应标准，并在 auto-HCT 之前和之后进行评估。⁷主要结果变量是无进展生存期（PFS）和总生存期（OS）。PFS 定义为从移植日期到疾病进展或死亡日期（以较早者为准）的经过时间。OS定义为移植日期至死亡日期或末次随访日期的时间。在最后一次随访时还活着并且没有经历疾病进展的患者被行政审查。

在 126 名黑人患者中，共有 125 名 (99.2%) 通过 1:1 倾向评分匹配与白人匹配，无需替换。匹配队列的基线患者特征总结在表 S3 中。总体而言，黑人和白人组分别有 121 名 (96.8%) 和 119 名 (95.2%) 患者接受了免疫调节药物 (IMiD) 和/或蛋白酶体抑制剂 (PI) 的诱导治疗。使用 IMiD-PI-地塞米松组合作为诱导在两组中相似，在黑人和白人队列中分别为 60/125 (48.0%) 和 61/125 (48.8%)。自动 HCT 前的总体反应率 (ORR) 在黑白组中分别为 91.2% (114/125) 和 92.0% (115/125)。完全缓解（CR）、极好部分缓解和部分缓解的患者数分别为 10 例（8.0%）、47 例（37.6%）和 54 例（43.2%），在黑色组中，分别为 11 (8.8%)、45 (36.0%) 和 55 (44.0%) 个在白色组中。分别在黑人和白人组的 114 名 (91.2%) 和 113 名 (90.4%) 患者中单独使用美法仑作为准备方案。白消安和美法仑的组合分别用作黑人和白人组 11 名 (8.8%) 和 12 名 (9.6%) 患者的准备方案。总体而言，黑人组和白人组分别有 97 名 (77.6%) 和 99 名 (79.2%) 患者接受了维持治疗。在黑色组中，分别有 75 名 (60.0%) 和 8 名 (6.4%) 患者使用 IMiD 或 PI 维持治疗。白组的相应患者数量相似，分别为 74 (59.2%) 和 8 (6.4%)。分别在黑人和白人组的 114 名 (91.2%) 和 113 名 (90.4%) 患者中单独使用美法仑作为准备方案。白消安和美法仑的组合分别用作黑人和白人组 11 名 (8.8%) 和 12 名 (9.6%) 患者的准备方案。总体而言，黑人组和白人组分别有 97 名 (77.6%) 和 99 名 (79.2%) 患者接受了维持治疗。在黑色组中，分别有 75 名 (60.0%) 和 8 名 (6.4%) 患者使用 IMiD 或 PI 维持治疗。白组的相应患者数量相似，分别为 74 (59.2%) 和 8 (6.4%)。分别在黑人和白人组的 114 名 (91.2%) 和 113 名 (90.4%) 患者中单独使用美法仑作为准备方案。白消安和美法仑的组合分别用作黑人和白人组 11 名 (8.8%) 和 12 名 (9.6%) 患者的准备方案。总体而言，黑人组和白人组分别有 97 名 (77.6%) 和 99 名 (79.2%) 患者接受了维持治疗。在黑色组中，分别有 75 名 (60.0%) 和 8 名 (6.4%) 患者使用 IMiD 或 PI 维持治疗。白组的相应患者数量相似，分别为 74 (59.2%) 和 8 (6.4%)。6%) 黑人和白人组的患者，分别。总体而言，黑人组和白人组分别有 97 名 (77.6%) 和 99 名 (79.2%) 患者接受了维持治疗。在黑色组中，分别有 75 名 (60.0%) 和 8 名 (6.4%) 患者使用 IMiD 或 PI 维持治疗。白组的相应患者数量相似，分别为 74 (59.2%) 和 8 (6.4%)。6%) 黑人和白人组的患者，分别。总体而言，黑人组和白人组分别有 97 名 (77.6%) 和 99 名 (79.2%) 患者接受了维持治疗。在黑色组中，分别有 75 名 (60.0%) 和 8 名 (6.4%) 患者使用 IMiD 或 PI 维持治疗。白组的相应患者数量相似，分别为 74 (59.2%) 和 8 (6.4%)。

黑色组患者在自动 HCT 后的 ORR 为 97.6% (122/125)，而白色组为 99.2% (124/125)，观察到的 CR 率为 27.2% (34/125)。 125）（表S4）。匹配队列的中位随访时间为 71.5（四分位距：51.6-90.3）个月。黑人和白人患者的中位 PFS 时间分别为 44.6（95% 置信区间 [CI]，35.5-54.7）个月和 51.0（95% CI：38.3-63.9）个月（p  = .763；图 1A）。非裔美国人 (AA) 的 4 年 PFS 率为 48.0%（95% CI，39.9%–57.8%），白人组为 51.2%（95% CI，43.0%–61.0%）。黑人和白人患者的中位 OS 时间分别为 101.1（95% CI：94.3-未达到）和 109.8（95% CI：86.2-未达到）（p = .579；图 1B)。AA 和白人患者的 4 年 OS 率分别为 78.5%（95% CI：71.5%–86.2%）和 80.9%（95% CI：74.1%–88.2%）。

此外，由于两组在较长时间点的 PFS 曲线线的分离，我们观察了两组在不同时间间隔的生存情况。我们观察了从移植后 36 个月到 96 个月的 12 个月时间间隔的存活率（表 S5）。我们使用分层对数秩检验来比较两条曲线之间的差异。的p的0.763 -值表示，两条曲线（黑色队列和白队列）都没有统计学显著不同（图S3）。

先前的差异研究对黑人和白人之间的治疗结果得出了不同的结论。在一项大型回顾性研究中，包括来自美国各地退伍军人事务部医院的 15000 多名 MM 患者，AA 患者与白人相比具有更高的生存率。⁸该研究将他们的患者群体分类为 AA 和白人患者。尽管该研究有其局限性，缺乏临床注释的细胞遗传学数据且 97.8% 的研究人群为男性，但 AA 和白人对新疗法和 auto-HCT 的利用率相似，这对两者的存活率相似种族群体。在另一项基于多发性骨髓瘤研究基金会 CoMMpass 注册的回顾性研究中，作者得出结论，与白人相比，AA 的生存率较差。⁹然而，在这项研究中，与白人相比，AA 不太可能接受三联疗法和自体 HCT。此外，当分析仅限于接受新疗法和 auto-HCT 的患者时，AA 和 Whites 之间的 OS 差异会减弱。

之前的一项回顾性研究表明，在黑人和白人 MM 患者之间的比较研究中，与白人相比，t(11;14) 与黑人患者更高的 OS 相关。¹⁰从我们的队列中，我们确定了 10 名黑人患者和 13 名白人患者，其中 t(11;14)。-0.091 的标准化平均差异表明 t(11;14) 在两个匹配的种族亚组之间是平衡的（表 S6）。在调整种族和 t(11;14) 的多变量分析中，种族和 t(11;14) 状态与 OS 或 PFS 的生存结果均无统计学显着相关性（表 S7 和 S8；图 S4 和 S5）。

虽然我们的研究表明黑人和白人患者具有相似的结果，但我们的研究存在一些局限性。该研究具有作为单中心回顾性研究的局限性；然而，我们的人群代表了当前的真实世界人群，其中大多数患者正在接受 PI、IMiD 和 auto-HCT 治疗。我们研究的另一个限制是，虽然我们进行了 1:1 倾向匹配分析，但这可能会人为地纠正我们两个患者群体之间的生物学差异。我们认为，健康的社会决定因素是造成黑人和白人患者死亡率差异的最大因素。虽然倾向匹配创建了一个平衡的数据集，允许在人群之间进行简单和直接的比较，人口之间健康的社会决定因素可能已被人为校正。我们研究的另一个限制是我们的中心是许多外部医院的转诊中心，因此，我们接受 auto-HCT 的人群可能不能代表社区环境中的一般 MM 人群。我们的研究有几个优势，包括是第一个进行倾向匹配分析的研究，以分析种族对 AA 和白人结果的影响，以尽量减少偏见。我们的研究还包括具有完整的细胞遗传学和 FISH 信息的患者。总之，我们的研究表明，如果接受最佳 MM 治疗和 auto-HCT，黑人患者的生存率与白人患者相似。我们研究的另一个限制是我们的中心是许多外部医院的转诊中心，因此，我们接受 auto-HCT 的人群可能不能代表社区环境中的一般 MM 人群。我们的研究有几个优势，包括是第一个进行倾向匹配分析的研究，以分析种族对 AA 和白人结果的影响，以尽量减少偏见。我们的研究还包括具有完整的细胞遗传学和 FISH 信息的患者。总之，我们的研究表明，如果接受最佳 MM 治疗和 auto-HCT，黑人患者的生存率与白人患者相似。我们研究的另一个限制是我们的中心是许多外部医院的转诊中心，因此，我们接受 auto-HCT 的人群可能不能代表社区环境中的一般 MM 人群。我们的研究有几个优势，包括是第一个进行倾向匹配分析的研究，以分析种族对 AA 和白人结果的影响，以尽量减少偏见。我们的研究还包括具有完整的细胞遗传学和 FISH 信息的患者。总之，我们的研究表明，如果接受最佳 MM 治疗和 auto-HCT，黑人患者的生存率与白人患者相似。我们的研究有几个优势，包括是第一个进行倾向匹配分析的研究，以分析种族对 AA 和白人结果的影响，以尽量减少偏见。我们的研究还包括具有完整的细胞遗传学和 FISH 信息的患者。总之，我们的研究表明，如果接受最佳 MM 治疗和 auto-HCT，黑人患者的生存率与白人患者相似。我们的研究有几个优势，包括是第一个进行倾向匹配分析的研究，以分析种族对 AA 和白人结果的影响，以尽量减少偏见。我们的研究还包括具有完整的细胞遗传学和 FISH 信息的患者。总之，我们的研究表明，如果接受最佳 MM 治疗和 auto-HCT，黑人患者的生存率与白人患者相似。