Hematopoietic stem cell transplantation (HSCT) offers the highest possible curative potential for patients with hematological malignancies. However, the management of post-transplant relapse remains a challenging task. In general, the prognosis of patients with post-transplant relapse is extremely poor, since many of them cannot tolerate or are refractory to commonly used approaches. In view of this, the risks and benefits of salvage treatment must be weighed up, and novel, less toxic, more efficient treatment options are urgently needed.

Venetoclax (VEN), an oral selective inhibitor of anti-apoptotic protein B-cell leukemia/lymphoma-2 (BCL-2), has been approved for the treatment of a variety of hematologic malignancies.^{1, 2} In relapsed/refractory (R/R) myeloid malignancies, the combination of VEN and hypomethylating agent (HMA) has exhibited an encouraging treatment effect.³ Nevertheless, research about VEN-HMA administration for post-transplant relapse is still in a preliminary stage.

Herein, we conducted a multicenter retrospective study, with the aim to evaluate the efficacy and side effects of VEN-HMA for post-transplant relapse and determine which patients may benefit from this combination therapy. Between July 2018 and June 2021, 44 consecutive patients with post-transplant relapse received VEN-HMA inpatient at 5 centers of Zhejiang province. The salvage treatment consisted of VEN for 28 consecutive days (100 mg of VEN for the first day and 200 mg for the second day, then increased to the final dose of 400 mg daily or equivalent to azole co-administration). Either azacytidine (AZA, 75 mg/m², d1-7) or decitabine (DEC, 20 mg/m², d1-5) was used as a VEN partner. During VEN-HMA treatment, hydration and alkalization were performed for the prophylaxis of tumor lysis syndrome (TLS). The response to VEN-HMA was determined according to the 2017 European Leukemia Net (ELN) response criteria. Adverse events were accessed by the Common Terminology Criteria for Adverse Events (CTCAE5.0)

Table S1 summarizes the patient baseline characteristics. Acute myeloid leukemia (AML) (n = 34) and myelodysplastic syndrome (MDS, n = 7) were the most common disease types. Each patient was analyzed for chromosomal and genetic abnormalities either at diagnosis and at post-transplant relapse. A complex/monosomal karyotype was seen in 10 (22.7%) patients. Ten patients (22.7%) had TP53 mutation or deletion, 6 (13.6%) exhibited IDH1/2 mutation (Figure 1A). Based on the 2017 ELN risk stratification, a total of 23 (52.3%) patients had adverse risk profiles. All patients experienced intramedullary relapse except two that relapsed at the chest wall and spine, respectively. Ten (22.7%) patients relapsed within 6 months, 18 (40.9%) relapsed between 6 months to 1 year, and 16 (36.4%) relapsed >1-year post-transplantation. Eighteen (40.9%) patients developed acute graft versus host disease (aGVHD) and 16 (36.4%) developed chronic graft versus host disease (cGVHD) before relapse.

Thirty-nine (88.6%) patients received VEN-HMA at first relapse after transplantation, while 5 (11.4%) received VEN-HMA at second relapse after transplantation. Twenty-six (59.1%) patients were treated with VEN-HMA directly after relapse (first-line therapy). For the remaining 18 (40.9%) patients, VEN-HMA was administrated as second-line therapy after failure of chemotherapy or DLI. Twenty-three (52.3%) patients had a higher tumor burden (bone marrow blasts >20%) at the initiation of VEN-HMA.

The median number of VEN-HMA cycles was 1 (range, 0.5–4). Treatment response was evaluated by bone marrow aspiration or radiological examination after each cycle of VEN-HMA therapy. Most patients were tolerant to the first cycle of VEN-HMA, except 4 (9.1%) of them developed ≥grade 3 non-hematological toxicity and discontinued the treatment on day 14. Fifteen (34.1%) patients fulfilled the criteria of CR/CRi and all of them achieved the best response in the first cycle of VEN-HMA (Table S2). Among 9 patients (including 3 CR patients) who received more than 1 cycle of VEN-HMA, delayed treatment with shortened treatment duration (14 days of VEN) was only applied in 1 CR patient who experienced prolonged myelosuppression in the first cycle. Among CR/CRi patients, 10 patients then received VEN maintenance, 4 patients underwent a second HSCT, and 1 patient received DLI from the original transplant donor after VEN-HMA. One CR patient relapsed at 7.5 months post-VEN-HMA treatment and died of disease progression. Another CR patient died of infection at 9.3 months post-VEN-HMA treatment (Figure 1B). Of those 29 patients who did not achieve CR/CRi, only 5 of them were alive at the writing of this article, 23 patients died of disease progression (16 of them had active infections at the time of death), and 1 patient died of intracranial hemorrhage. Compared with patients who did not achieve CR/CRi, CR/CRi patients had a dramatically improved OS (8.1 vs. 2.8 months, p < .0001).

Considering the adverse events, neutropenia (79.5%), thrombocytopenia (68.2%), and anemia (63.6%) were the three most common regimen-related toxicities. Twenty-one (47.7%) patients developed grade III–IV infection (18 pulmonary infections and 3 bloodstream infections). Eight (18.2%) patients had bleeding episodes. No patients developed TLS during VEN-HMA.

We also investigated the association between treatment effect and patient characteristics (Figure 1C). We found that male patients tended to have a lower CR/CRi rate than female patients (20.8% vs. 50.0%, p = .042). Patients that relapsed within 1-year post-transplantation had a poorer treatment effect than other patients (21.4% vs. 56.3%, p = .019). In addition, patients with ELN adverse risk also had a lower CR/CRi rate (21.7% vs. 47.6%, p = .070). While, other factors such as diagnosis, patient age, status at transplant, tumor burden at the onset of VEN-HMA did not affect the CR/CRi rate. Among patients with a complex/monosomal karyotype, only 1 of them achieved CR/CRi (10.0% vs. 41.2%, p = .127). Patients with TP53 abnormities (10.0% vs. 41.2%, p = .127) and ASXL1 mutation (0.0% vs. 38.5%, p = .149) seemed to have a lower chance to benefit from VEN-HMA therapy. In contrast, DNMT3A mutation (66.1% vs. 31.7%, p = .264), NPM1 mutation (60.0% vs. 30.8%, p = .319), and IDH1/2 mutation (50.0% vs. 31.6%, p = .394) were associated with a favorable CR/CRi rate. In multivariate analysis, TP53 mutation (HR = 17.339, p = .033) and relapse within 1-year post-transplantation (HR = 6.261, p = .026) were identified as independent risk factors that influenced the CR/CRi rate (Table S3).

Altogether, the VEN-HMA treatment yielded a favorable response rate. Interestingly, the CR/CRi rate (56.3%) in patients who relapsed >1-year post-transplantation was very impressive. Meanwhile, the treatment outcome was dismal for patients with a short duration of post-transplant remission. We hypothesize that several reasons such as aggressive tumor biology, incomplete immune reconstitution, and unfitness for cytotoxic therapies may account for the poor prognosis. Consistently with our results, Byrne et al. also reported that among patients exhibiting no response to VEN-HMA, 7 of 8 relapsed within 1-year post-transplantation.⁴ However, a recent study revealed that the combination of VEN and DLI yields a promising response rate in early relapsed AML post-transplantation: half of their patients achieved CR/CRi/MLFS, and low WBC count at relapse and GVHD were associated with high chances of remission.⁵ In addition, VEN-based combination therapy (VEN, low-dose Cytarabine, Actinomycin D, and DLI maintenance) treatment exhibited a 70% CR/CRp rate in relapsed AML patients after HSCT (80% of these patients relapsed within 1-year post-transplantation).⁶ These results suggest that combining VEN with other therapies may have a synergistic treatment effect for early post-transplant relapse.

Regarding treatment effects and chromosomal/genetic profiles, response rates differ greatly among patient groups. The relationship between TP53 mutation and treatment response has also been described in other publications. Byrne et al. reported that none of 4 patients with complex karyotype and TP53 mutation responded to VEN-HMA.⁴ Stahl et al. also suggested that TP53 was associated with a poor CR/CRi rate.³ Nevertheless, results from Joshi et al. did not show a difference in response rate between TP53 mutated and wild type patients.⁷ Additionally, patients with IDH1/2 and NPM1 mutations may be sensitive to VEN-based therapy, while patients with NOTCH1, FLT3, and PTPN11 mutations are more likely to be resistant.¹ In the future, the associations between cytogenetic/mutational alterations and VEN response still need to be validated by large-scale prospective studies.

Remarkably, 3 of 21 (14.3%) responders in our cohort lost their initial response to VEN-HMA. One patient relapsed after achieving CR, while failing to attain a second CR status to VEN-HMA. The other two patients (1 PR and 1 blast reduction) had disease progression in the following cycles of VEN-HMA treatment. Recently, several studies have revealed that potent mechanisms of VEN resistance. For instance, the mutations of BCL-2 protein could reduce the binding affinity of VEN; and the upregulation of other anti-apoptotic signals such as MCL-1 and BCL-XL could block the apoptosis of AML cells.¹ Based on these findings, combination strategies such as VEN plus MCL-1 inhibitors have been applied and exhibited synthetic effects in some pre-clinical studies.⁸ Since there are limited clinical experiences for the treatment of VEN-resistant patients, monitoring emergent genetic mutations and combining VEN with targeting regimens may overcome drug resistance and improve the anti-leukemia activity.

In conclusion, VEN-HMA is a safe and efficacious treatment option for patients with myeloid malignancies relapsing after transplantation. Patients without TP53 mutation or relapsing >1-year post-transplantation may stand a greater chance to achieve CR/CRi.

造血干细胞移植 (HSCT) 为血液系统恶性肿瘤患者提供了尽可能高的治愈潜力。然而，移植后复发的管理仍然是一项具有挑战性的任务。一般而言，移植后复发患者的预后极差，因为他们中的许多人不能耐受或对常用方法无效。有鉴于此，必须权衡抢救性治疗的风险和收益，迫切需要新的、毒性更小的、更有效的治疗方案。

Venetoclax (VEN) 是一种口服选择性抗凋亡蛋白 B 细胞白血病/淋巴瘤-2 (BCL-2) 抑制剂，已被批准用于治疗多种血液系统恶性肿瘤。^{1, 2}在复发/难治性 (R/R) 髓系恶性肿瘤中，VEN 和低甲基化剂 (HMA) 的组合显示出令人鼓舞的治疗效果。³然而，关于移植后复发的 VEN-HMA 给药的研究仍处于初步阶段。

在此，我们进行了一项多中心回顾性研究，旨在评估 VEN-HMA 对移植后复发的疗效和副作用，并确定哪些患者可能受益于这种联合治疗。2018 年 7 月至 2021 年 6 月，连续 44 例移植后复发患者在浙江省 5 个中心接受了 VEN-HMA 住院治疗。补救治疗包括连续 28 天的 VEN（第一天 100 mg VEN，第二天 200 mg，然后增加到每天 400 mg 的最终剂量或相当于唑类共同给药）。氮胞苷 (AZA, 75 mg/m ² , d1-7) 或地西他滨 (DEC, 20 mg/m ², d1-5) 被用作 VEN 伙伴。在 VEN-HMA 治疗期间，进行水化和碱化以预防肿瘤溶解综合征 (TLS)。对 VEN-HMA 的反应是根据 2017 年欧洲白血病网 (ELN) 反应标准确定的。通过不良事件通用术语标准 (CTCAE5.0) 访问不良事件

表 S1 总结了患者的基线特征。急性髓性白血病 (AML) ( n  = 34) 和骨髓增生异常综合征 (MDS, n = 7) 是最常见的疾病类型。在诊断时和移植后复发时分析每位患者的染色体和遗传异常。10 名 (22.7%) 患者出现复杂/单体核型。10 名患者（22.7%）有 TP53 突变或缺失，6 名（13.6%）有 IDH1/2 突变（图 1A）。根据 2017 年 ELN 风险分层，共有 23 名 (52.3%) 患者有不良风险特征。除了分别在胸壁和脊柱复发的两名患者外，所有患者都经历了髓内复发。10 名 (22.7%) 患者在 6 个月内复发，18 名 (40.9%) 患者在 6 个月至 1 年内复发，16 名 (36.4%) 患者在移植后超过 1 年复发。18 名 (40.9%) 患者在复发前发展为急性移植物抗宿主病 (aGVHD)，16 名 (36.4%) 患者发展为慢性移植物抗宿主病 (cGVHD)。

39 名 (88.6%) 患者在移植后第一次复发时接受了 VEN-HMA，而 5 名 (11.4%) 患者在移植后第二次复发时接受了 VEN-HMA。26 名 (59.1%) 患者在复发后直接接受 VEN-HMA 治疗（一线治疗）。对于其余 18 名 (40.9%) 患者，在化疗或 DLI 失败后给予 VEN-HMA 作为二线治疗。23 名 (52.3%) 患者在 VEN-HMA 开始时肿瘤负荷较高（骨髓原始细胞 >20%）。

VEN-HMA 循环的中位数为 1（范围，0.5-4）。在每个 VEN-HMA 治疗周期后，通过骨髓抽吸或放射学检查评估治疗反应。大多数患者对第一个周期的 VEN-HMA 耐受，但其中 4 名（9.1%）患者出现≥3 级非血液学毒性并在第 14 天停止治疗。15 名（34.1%）患者符合 CR/CRi 标准并且它们都在 VEN-HMA 的第一个周期中获得了最佳响应（表 S2）。在接受超过1个周期的VEN-HMA治疗的9例患者（包括3例CR患者）中，仅在1例在第一个周期出现骨髓抑制延长的CR患者中采用缩短治疗时间（VEN治疗14天）的延迟治疗。在 CR/CRi 患者中，10 名患者随后接受了 VEN 维持，4 名患者接受了第二次 HSCT，1 名患者在 VEN-HMA 后接受了原始移植供体的 DLI。一名 CR 患者在 VEN-HMA 治疗后 7.5 个月复发并死于疾病进展。另一名 CR 患者在 VEN-HMA 治疗后 9.3 个月死于感染（图 1B）。在未达到CR/CRi的29例患者中，撰写本文时仅存活5例，23例死于疾病进展（其中16例在死亡时有活动性感染），1例死于疾病进展颅内出血。与未达到 CR/CRi 的患者相比，CR/CRi 患者的 OS 显着改善（8.1 个月 vs. 2.8 个月，VEN-HMA 治疗后 3 个月（图 1B）。在未达到CR/CRi的29例患者中，撰写本文时仅存活5例，23例死于疾病进展（其中16例在死亡时有活动性感染），1例死于疾病进展颅内出血。与未达到 CR/CRi 的患者相比，CR/CRi 患者的 OS 显着改善（8.1 个月 vs. 2.8 个月，VEN-HMA 治疗后 3 个月（图 1B）。在未达到CR/CRi的29例患者中，撰写本文时仅存活5例，23例死于疾病进展（其中16例在死亡时有活动性感染），1例死于疾病进展颅内出血。与未达到 CR/CRi 的患者相比，CR/CRi 患者的 OS 显着改善（8.1 个月 vs. 2.8 个月，p  < .0001)。

考虑到不良事件，中性粒细胞减少（79.5%）、血小板减少（68.2%）和贫血（63.6%）是三种最常见的与方案相关的毒性。21 名 (47.7%) 患者出现 III-IV 级感染（18 名肺部感染和 3 名血流感染）。8 名 (18.2%) 患者出现出血事件。在 VEN-HMA 期间没有患者出现 TLS。

我们还调查了治疗效果与患者特征之间的关联（图 1C）。我们发现男性患者的 CR/CRi 率往往低于女性患者（20.8% 对 50.0%，p  = .042）。移植后 1 年内复发的患者的治疗效果比其他患者差（21.4% 对 56.3%，p  = .019）。此外，具有 ELN 不良风险的患者的 CR/CRi 率也较低（21.7% vs. 47.6%，p  = .070）。而其他因素，如诊断、患者年龄、移植状态、VEN-HMA 发病时的肿瘤负荷并不影响 CR/CRi 率。在具有复杂/单体核型的患者中，只有 1 人达到 CR/CRi（10.0% vs. 41.2%，p = .127)。TP53 异常（10.0% 对 41.2%，p  = .127）和 ASXL1 突变（0.0% 对 38.5%，p  = .149）的患者从 VEN-HMA 治疗中获益的机会似乎较低。相比之下，DNMT3A 突变（66.1% 对 31.7%，p  = .264）、NPM1 突变（60.0% 对 30.8%，p  = .319）和 IDH1/2 突变（50.0% 对 31.6%，p  = .394) 与有利的 CR/CRi 率相关。在多变量分析中，TP53 突变（HR = 17.339，p  = .033）和移植后 1 年内复发（HR = 6.261，p  = .026）被确定为影响 CR/CRi 率的独立危险因素（表S3)。

总之，VEN-HMA 治疗产生了良好的反应率。有趣的是，移植后 > 1 年复发的患者的 CR/CRi 率 (56.3%) 令人印象深刻。同时，移植后缓解期短的患者的治疗结果令人沮丧。我们假设几个原因，例如侵袭性肿瘤生物学、不完全的免疫重建和不适合细胞毒疗法，可能是预后不良的原因。与我们的结果一致，Byrne 等人。还报告说，在对 VEN-HMA 没有反应的患者中，8 人中有 7 人在移植后 1 年内复发。⁴然而，最近的一项研究表明，VEN 和 DLI 的组合在移植后早期复发 AML 中产生了有希望的反应率：他们一半的患者达到 CR/CRi/MLFS，复发时 WBC 计数低和 GVHD 与高机会相关的缓解。⁵此外，基于 VEN 的联合治疗（VEN、低剂量阿糖胞苷、放线菌素 D 和 DLI 维持）治疗在 HSCT 后复发的 AML 患者中表现出 70% 的 CR/CRp 率（这些患者中有 80% 在 1 年内复发移植后）。⁶这些结果表明，将 VEN 与其他疗法相结合可能对移植后早期复发具有协同治疗效果。

关于治疗效果和染色体/遗传谱，患者组之间的反应率差异很大。TP53 突变与治疗反应之间的关系也在其他出版物中有所描述。伯恩等人。据报道，4 名具有复杂核型和 TP53 突变的患者均未对 VEN-HMA 作出反应。⁴斯塔尔等人。还表明 TP53 与较差的 CR/CRi 率有关。³然而，Joshi 等人的结果。没有显示 TP53 突变和野生型患者之间的反应率差异。⁷此外，具有 IDH1/2 和 NPM1 突变的患者可能对基于 VEN 的治疗敏感，而具有 NOTCH1、FLT3 和 PTPN11 突变的患者更有可能产生耐药性。¹未来，细胞遗传学/突变改变与 VEN 反应之间的关联仍需要通过大规模前瞻性研究来验证。

值得注意的是，我们队列中的 21 名 (14.3%) 响应者中有 3 名失去了对 VEN-HMA 的初始响应。一名患者在达到 CR 后复发，但未能达到 VEN-HMA 的第二个 CR 状态。其他两名患者（1 名 PR 和 1 名爆炸减少）在随后的 VEN-HMA 治疗周期中出现疾病进展。最近，几项研究揭示了 VEN 抗性的有效机制。例如，BCL-2 蛋白的突变会降低 VEN 的结合亲和力；其他抗凋亡信号如 MCL-1 和 BCL-XL 的上调可以阻断 AML 细胞的凋亡。¹基于这些发现，VEN 加 MCL-1 抑制剂等联合策略已被应用，并在一些临床前研究中表现出合成效果。⁸由于治疗 VEN 耐药患者的临床经验有限，监测紧急基因突变并将 VEN 与靶向方案相结合可能会克服耐药性并提高抗白血病活性。

总之，对于移植后复发的髓系恶性肿瘤患者，VEN-HMA 是一种安全有效的治疗选择。没有 TP53 突变或移植后复发 > 1 年的患者可能有更大的机会实现 CR/CRi。