To the Editor:

The development of effective COVID-19 vaccines has been essential in slowing the spread of SARS-CoV-2. However, unvaccinated populations as well as those who do not respond to vaccination still remain at risk. Very few cancer patients were included in the COVID-19 mRNA vaccine trials and any individuals receiving chemotherapy or immunotherapy within 6 months were excluded.¹ Consequently, we have an inadequate knowledge of how well these vaccines work in the cancer patient population. However, by extrapolation from other vaccines, we hypothesized that patients with hematologic malignancies, especially those on immunosuppressive therapy, would produce poor serological responses to a COVID-19 vaccine.²

In this single-center, observational cohort study we assessed antibody responses in lymphoma patients receiving a COVID-19 mRNA vaccine (BNT162b2, BioNTech/Pfizer, Germany/New York, NY; or mRNA-1273, Moderna, Cambridge, MA). All patients provided written informed consent to participate in observational research, and this study was approved by the Weill Cornell Medicine institutional review board (IRB 21-02023288). Serum samples were obtained before (when possible) and after vaccination. Post-vaccination samples were collected within 11–70 days of the second dose (median 24.5 days). In the healthcare worker (HCW) control group, the post-vaccination samples were obtained within 10–68 days of the second dose (median 40 days) (Figure S1). We also include data from a healthy control group of 35 HCWs enrolled in the NYP-WELCOME (WEilL COrnell Medicine Employees) observational trial (IRB 20-04021831). The use of this cohort in an mRNA vaccine study as well as the assay to quantify immunoglobulin G (IgG) antibodies to the SARS-CoV-2 S-protein has been described previously.³ Additionally, we determined whether any patients had serum antibodies to the SARS-CoV-2 nucleocapsid (N) protein, a marker for prior infection.

The anti-S protein response to mRNA vaccination was assessed by enzyme-linked immunosorbent assay using sera from 67 patients with lymphoma and 35 healthy HCW controls. The majority of patients in this study were white (74.6%, Table S1). The median age of the study group was 71 (24–90). The most common comorbidities were hypertension (37.3%) and hyperlipidemia (50.7%). All patients were vaccinated with an mRNA vaccine (31 BNT162b2 and 36 mRNA-1273). The patients were categorized as having Hodgkin lymphoma (NHL; n = 4), chronic lymphocytic leukemia (CLL; n = 21), or other non-Hodgkin lymphomas (n = 42). Patients with other non-Hodgkin lymphomas included follicular lymphoma (7), marginal zone lymphoma (10), mantle cell lymphoma (8), diffuse large B-cell lymphoma (8), Waldenstrom macroglobulinemia (7), and other, unclassified lymphomas (2). No SARS-CoV-2 infections were identified during this study (February to April 2021).

The vaccine-induced IgG antibody responses to the SARS-CoV-2 S-protein are shown in Figure 1(A). The median and mean endpoint titers in the HCW control group were higher than in the lymphoma patients, although the difference was not significant. There were also no significant differences in mean titers when patients with different lymphomas were compared. However, while all 35 healthy control group members responded to the vaccine, a substantial proportion of the lymphoma patients did not. Thus, the anti-S endpoint titers in nine of the 21 CLL patients and 17 of the 42 other NHL patients were <10 000 (a cut-off level marked on Figure 1), and were often undetectable. By contrast, the four Hodgkin lymphoma patients all responded to the vaccines. When the data were grouped according to whether the participants received the BNT162b2 or mRNA-1273 vaccine, no differences were apparent.

In total, eight lymphoma patients were anti-N-positive while all members of the HCW control group were anti-N-negative. For four of the eight anti-N-positive lymphoma patients, there was evidence of COVID-19 prior to the start of this study. Thus, three patients had prior documented positive SARS-CoV-2 polymerase chain reaction (PCR) tests, while the fourth was not PCR-tested but later had a positive commercial antibody test. Seven of these eight anti-N-positive patients responded to vaccination. Taken together, anti-N-positive lymphoma patients had significantly higher mean anti-S protein titers than their anti-N-negative counterparts (p < 0.0001) and the HCW group (p = 0.02). However, when anti-N-positive lymphoma patients were separated by treatment status (i.e., naïve, active therapy) the sample sizes were too small for comparisons to the anti-N-negative group.

We studied the CLL and other NHL patients in more detail to understand the implications of their treatment (Figure 1(B)). Every treatment-naïve and remote-therapy (no treatment in over 24 months) CLL patient responded to vaccination, whereas only 40% (6/15) of those currently being treated had anti-S protein titers above the designated cut-off value. A similar pattern was seen for the other NHL patients, although one individual in each of the treatment-naïve and remote-therapy groups failed to respond to the vaccine. Active therapy in this subgroup was again associated with a poor vaccine response, with only 21.4% (3/14) developing anti-S protein titers above the cut-off. The off-therapy subgroup, who had received treatment within 2 years but not at the time of vaccination, also had a lower vaccine response rate of 55.5% (5/9). The four non-responders in this group had all received an anti-CD20 mAb within the previous 2 years; two within 6 months, one within 1 year, and one within 18 months. None of the patients currently on anti-CD20 mAb therapy seroconverted after vaccination.

We next studied the relationship between when anti-CD20 mAb therapy ceased and the vaccine response (Figure 1(C)). None of the 11 CLL and other NHL patients receiving this treatment within 6 months of vaccination had anti-S protein titers above the cut-off, but longer intervals were associated with higher titers. Thus, CLL and other NHL patients who were last treated >24 months before vaccination had response rates of 66.7% (6/9) and 71.4% (10/14), respectively. It is notable that 3/3 CLL and 3/4 other NHL non-responders in this subgroup were receiving a different type of active therapy at the time of vaccination (Table S6). We suggest that even when anti-CD20 mAb therapy ceased >24 months before vaccination, other forms of ongoing active therapy can compromise the vaccine response.

Thus we demonstrated that commonly used lymphoma therapies can adversely influence the performance of COVID-19 vaccines, with anti-CD20 mAbs having the greatest impact. With regard to anti-CD20 mAbs, our results are consistent with a growing number of reports that patients on active, or with recent anti-CD20 mAb treatment do not respond to vaccination.^4-6

Compared with other studies, we report a higher rate of seroconversion in patients on active BTKi monotherapy.^{4, 5} Here, we found that 66.7% (4/6) of CLL patients and 50% (2/4) of other NHL patients did develop high-titer IgG antibodies after mRNA vaccination. In a study by Herishanu et al.⁴ only 16% (8/50) of CLL patients treated with a BTKi responded to vaccination with BNT162b2. In our study, CLL responders on BTKi monotherapy were on treatment for a median length of 53.5 (23–74) months prior to the first vaccine dose. In comparison, CLL non-responders on BTKi monotherapy were on treatment for a median of 2 (1–3) months. The CLL responders were described as having a good response to BTKi monotherapy, with two patients in complete remission and two patients with no progression of disease. All CLL responders were compliant with treatment and only one patient had a recent interruption in therapy. This patient was hospitalized for COVID-19 in April 2020 and treatment was held for approximately 3 weeks after which therapy was restarted. In this study, he was found to be anti-N-positive, consistent with pre-existing serologic immunity from prior infection.

Finally, we studied the avidity of IgG antibodies to the Receptor Binding Domain in the lymphoma and healthy control patients (Figure S1). The avidity was significantly higher (p < 0.0001) for anti-N-positive lymphoma patients than for anti-N-negative lymphoma patients as well as healthy controls, all of whom were anti-N-negative (Figure S1(A)). These findings suggest COVID-19 convalescent patients (i.e., anti-N positive) have had longer to affinity mature their anti-S antibodies, which are boosted by the mRNA vaccines. We noted that patients currently receiving venetoclax or a BTKi had lower avidity S-protein antibodies than the other groups, although the group sizes were too small for statistical significance (Figure S1(B)).

In conclusion, we found that most lymphoma patients respond to vaccination with an mRNA-based COVID-19 vaccine, but a substantial fraction (>40%) do not and therefore may remain at risk of infection and disease. There were no significant differences in the S-protein IgG antibody response rates or titers between the different lymphoma histologic subtypes. Treatment status was, however, a relevant variable. Treatment-naïve lymphoma patients responded to vaccination in a similar manner to the HCW group, as did patients who had not received therapy for at least 2 years. However, this controlled study presents compelling evidence that patients on active therapy for lymphoma may not respond to vaccination. Our results are particularly concerning for patients on anti-CD20 mAb therapy, given that no patients who had received treatment within 6 months responded well to mRNA vaccination. Thus these patients probably remain at risk of infection with SARS-CoV-2. In this patient population, we suggest exploring alternative strategies for protection such as passive immunization with anti-S monoclonal antibody therapy or, if possible, delaying therapy until after vaccination.

致编辑：

开发有效的 COVID-19 疫苗对于减缓 SARS-CoV-2 的传播至关重要。然而，未接种疫苗的人群以及对疫苗接种无反应的人群仍然处于危险之中。COVID-19 mRNA 疫苗试验包括极少数癌症患者，任何在 6 个月内接受化疗或免疫治疗的个体都被排除在外。¹因此，我们对这些疫苗在癌症患者群体中的效果知之甚少。然而，通过从其他疫苗推断，我们假设患有血液系统恶性肿瘤的患者，尤其是接受免疫抑制治疗的患者，会对 COVID-19 疫苗产生较差的血清学反应。²

在这项单中心观察性队列研究中，我们评估了接受 COVID-19 mRNA 疫苗（BNT162b2，BioNTech/Pfizer，德国/纽约，纽约；或 mRNA-1273，Moderna，Cambridge，MA）的淋巴瘤患者的抗体反应。所有患者都提供了参与观察性研究的书面知情同意书，该研究得到了威尔康奈尔医学机构审查委员会 (IRB 21-02023288) 的批准。在疫苗接种之前（如果可能）和之后获得血清样本。在第二次接种后的 11-70 天内（中位 24.5 天）收集了疫苗接种后的样本。在医护人员 (HCW) 对照组中，接种后样本是在第二次接种后的 10-68 天内（中位数 40 天）获得的（图 S1）。我们还包括来自 NYP-WELCOME（威尔康奈尔医学员工）观察性试验 (IRB 20-04021831) 的 35 名 HCW 的健康对照组的数据。之前已经描述了在 mRNA 疫苗研究中使用该队列以及量化针对 SARS-CoV-2 S 蛋白的免疫球蛋白 G (IgG) 抗体的测定。³此外，我们确定是否有任何患者对 SARS-CoV-2 核衣壳 (N) 蛋白（先前感染的标志物）有血清抗体。

使用来自 67 名淋巴瘤患者和 35 名健康 HCW 对照的血清通过酶联免疫吸附试验评估抗 S 蛋白对 mRNA 疫苗接种的反应。本研究中的大多数患者是白人（74.6%，表 S1）。研究组的中位年龄为 71 岁（24-90 岁）。最常见的合并症是高血压（37.3%）和高脂血症（50.7%）。所有患者均接种了 mRNA 疫苗（31 个 BNT162b2 和 36 个 mRNA-1273）。患者分为霍奇金淋巴瘤（NHL；n  = 4）、慢性淋巴细胞白血病（CLL；n  = 21）或其他非霍奇金淋巴瘤（n = 42)。其他非霍奇金淋巴瘤患者包括滤泡性淋巴瘤 (7)、边缘区淋巴瘤 (10)、套细胞淋巴瘤 (8)、弥漫性大 B 细胞淋巴瘤 (8)、Waldenstrom 巨球蛋白血症 (7) 和其他未分类的淋巴瘤。 2）。在本研究期间（2021 年 2 月至 2021 年 4 月）未发现 SARS-CoV-2 感染。

疫苗诱导的 IgG 抗体对 SARS-CoV-2 S 蛋白的反应如图 1(A) 所示。HCW 对照组的中位和平均终点滴度高于淋巴瘤患者，尽管差异不显着。当比较不同淋巴瘤的患者时，平均滴度也没有显着差异。然而，虽然所有 35 名健康对照组成员都对疫苗有反应，但很大一部分淋巴瘤患者没有。因此，21 名 CLL 患者中的 9 名和 42 名其他 NHL 患者中的 17 名的抗 S 终点滴度<10 000（图 1 中标记的截止水平），并且通常无法检测到。相比之下，四名霍奇金淋巴瘤患者都对疫苗有反应。

总共有 8 名淋巴瘤患者为抗 N 阳性，而 HCW 对照组的所有成员均为抗 N 阴性。对于八名抗 N 阳性淋巴瘤患者中的四名，在本研究开始之前就有 COVID-19 的证据。因此，三名患者先前记录的 SARS-CoV-2 聚合酶链反应 (PCR) 检测呈阳性，而第四名患者未进行 PCR 检测，但后来商业抗体检测呈阳性。这八名抗 N 阳性患者中有七名对疫苗接种有反应。综上所述，抗 N 阳性淋巴瘤患者的平均抗 S 蛋白滴度显着高于抗 N 阴性患者 ( p  < 0.0001) 和 HCW 组 ( p = 0.02)。然而，当抗 N 阳性淋巴瘤患者按治疗状态（即初始治疗、积极治疗）分开时，样本量太小，无法与抗 N 阴性组进行比较。

我们更详细地研究了 CLL 和其他 NHL 患者，以了解他们治疗的意义（图 1(B)）。每个初治和远程治疗（超过 24 个月未治疗）CLL 患者都对疫苗接种有反应，而目前正在接受治疗的患者中只有 40% (6/15) 的抗 S 蛋白滴度高于指定的临界值。其他 NHL 患者也出现了类似的情况，尽管在初治组和远程治疗组中各有一人对疫苗没有反应。该亚组的积极治疗再次与较差的疫苗反应相关，只有 21.4% (3/14) 的人产生高于临界值的抗 S 蛋白滴度。在 2 年内接受治疗但未在接种疫苗时接受治疗的非治疗亚组的疫苗应答率也较低，为 55.5% (5/9)。该组中的四名无反应者在过去 2 年内均接受过抗 CD20 mAb；6个月内2个，1年内1个，18个月内1个。目前接受抗 CD20 mAb 治疗的患者在接种疫苗后均未发生血清转化。

我们接下来研究了何时停止抗 CD20 mAb 治疗与疫苗反应之间的关系（图 1（C））。在接种疫苗后 6 个月内接受这种治疗的 11 名 CLL 和其他 NHL 患者中，没有一个人的抗 S 蛋白滴度高于临界值，但较长的间隔与较高的滴度相关。因此，在接种疫苗前最后一次治疗超过 24 个月的 CLL 和其他 NHL 患者的反应率分别为 66.7% (6/9) 和 71.4% (10/14)。值得注意的是，该亚组中 3/3 的 CLL 和 3/4 的其他 NHL 无应答者在接种疫苗时正在接受不同类型的积极治疗（表 S6）。我们建议，即使抗 CD20 mAb 治疗在疫苗接种前 24 个月停止，其他形式的正在进行的积极治疗也会损害疫苗反应。

因此，我们证明了常用的淋巴瘤疗法会对 COVID-19 疫苗的性能产生不利影响，其中抗 CD20 mAb 的影响最大。关于抗 CD20 mAb，我们的结果与越来越多的报告一致，即正在积极或最近接受抗 CD20 mAb 治疗的患者对疫苗接种没有反应。^4-6

与其他研究相比，我们报告了接受活性 BTKi 单药治疗的患者的血清转化率更高。^{4, 5}在这里，我们发现 66.7% (4/6) 的 CLL 患者和 50% (2/4) 的其他 NHL 患者在接种 mRNA 后确实产生了高滴度 IgG 抗体。在 Herishanu 等人的一项研究中。⁴接受 BTKi 治疗的 CLL 患者中只有 16% (8/50) 对 BNT162b2 疫苗接种有反应。在我们的研究中，BTKi 单药治疗的 CLL 应答者在第一次疫苗接种前接受治疗的中位时间为 53.5（23-74）个月。相比之下，BTKi 单药治疗的 CLL 无反应者接受治疗的中位数为 2 (1-3) 个月。CLL 反应者被描述为对 BTKi 单一疗法有良好的反应，两名患者完全缓解，两名患者没有疾病进展。所有 CLL 反应者都对治疗依从，只有一名患者最近中断了治疗。该患者于 2020 年 4 月因 COVID-19 住院，治疗持续约 3 周，之后重新开始治疗。在这项研究中，他被发现是抗 N 阳性，

最后，我们研究了淋巴瘤和健康对照患者中 IgG 抗体对受体结合域的亲和力（图 S1）。抗 N 阳性淋巴瘤患者的亲和力显着高于 抗 N 阴性淋巴瘤患者和健康对照（图 S1（A））。这些发现表明，COVID-19 恢复期患者（即抗 N 阳性）的抗 S 抗体亲和力成熟的时间更长，而这些抗体由 mRNA 疫苗增强。我们注意到，目前接受 venetoclax 或 BTKi 的患者的 S 蛋白抗体亲和力低于其他组，尽管组规模太小而无统计学意义（图 S1(B)）。

总之，我们发现大多数淋巴瘤患者对接种基于 mRNA 的 COVID-19 疫苗有反应，但很大一部分（>40%）没有反应，因此可能仍有感染和疾病的风险。不同淋巴瘤组织学亚型之间的 S 蛋白 IgG 抗体反应率或滴度没有显着差异。然而，治疗状态是一个相关变量。初治淋巴瘤患者对疫苗接种的反应与 HCW 组相似，至少 2 年未接受治疗的患者也是如此。然而，这项对照研究提供了令人信服的证据，表明接受淋巴瘤积极治疗的患者可能对疫苗接种没有反应。我们的结果特别关注抗 CD20 mAb 治疗的患者，鉴于在 6 个月内接受治疗的患者没有对 mRNA 疫苗反应良好。因此，这些患者可能仍有感染 SARS-CoV-2 的风险。在这个患者群体中，我们建议探索替代策略来保护，例如用抗 S 单克隆抗体治疗进行被动免疫，或者如果可能的话，将治疗推迟到疫苗接种后。