Risks and benefits for patients with relapsed or refractory diffuse large B-cell lymphoma in early-phase clinical trials: a systematic review and meta-analysis.

We aimed to evaluate antitumour activity and safety outcomes across drug classes in early-phase trials in relapsed or refractory diffuse large B-cell lymphoma since 2000.

我们旨在评估自2000年以来，在复发或难治性弥漫性大B细胞淋巴瘤早期试验中，不同药物类别的抗肿瘤活性和安全性结果。

Methods

We did a systematic reviewand meta-analysisof phase 1-2 trials.

我们进行了1-2期试验的系统评价和荟萃分析。

We searched PubMed, Embase.com, Web of Science and Wiley/Cochrane Library from database inception to May 9, 2025.

我们从数据库创建开始至2025年5月9日，检索了PubMed、Embase.com、Web of Science以及Wiley/Cochrane图书馆。

We included English-language studies published between Jan 1, 2000, and May 9, 2025, enrolling adults aged 18 years or older with relapsed or refractory diffuse large B-cell lymphoma treated with experimental agents alone or combined with CD20-antibodies; trials including other B-cell malignancies were eligible if diffuse large B-cell lymphoma-specific responses could be extracted.

我们纳入了2000年1月1日至2025年5月9日期间发表的英文研究，招募了18岁或以上的成人，这些成人患有复发或难治性弥漫性大B细胞淋巴瘤，并接受实验性药物单独治疗或与CD20抗体联合治疗；如果可以提取出针对弥漫性大B细胞淋巴瘤的特定反应，那么包括其他B细胞恶性肿瘤的试验也是符合条件的。

Trials restricted to highly-selected subgroups, supportive-care, administration-routes, country-specific approvals, and conference abstracts were excluded.

排除了仅限于高度选择性亚组、支持性护理、给药途径、国家特定批准和会议摘要的试验。

Two investigators independently extracted summary data.

两位研究员独立提取了汇总数据。

The primary outcomes were objective response rateand complete response rate, and were pooled using random-effects generalised linear mixed models.

主要结果是客观缓解率和完全缓解率，并使用随机效应广义线性混合模型进行了汇总。

Adverse events were secondary outcomes.

不良事件是次要结果。

Prespecified subgroup analyses evaluated drug class and publication period.

预先指定的亚组分析评估了药物类别和发表时间。

The study was registered with PROSPERO, CRD42023394451.

该研究已在PROSPERO注册，注册号为CRD42023394451。

Results

We identified 2797 citations, of which 1824 unique records remained after removal of duplicates. 132 trials including 7786 patients were eligible for analysis. 3375 (43%) of 7786 patients were female and 4411 (57%) were male.

我们确定了2797个引文，去除重复项后，剩余1824个独特记录。共有132项试验，包括7786名患者，符合分析资格。

Objective response ratewas 30·5% (95% CI 26·0-35·5, I2=84·7%) and complete response rate 14·3% (11·5-17·7, I2=82·2%).

客观缓解率为30.5%（95%置信区间26.0-35.5，I2=84.7%），完全缓解率为14.3%（11.5-17.7，I2=82.2%）。

Response rates varied across drug classes, with the highest objective response rateor complete response rate for cellular therapies (70·0%, 95% CI 61·0-77·0 and 51·0%, 95% CI 43·0-59·0), followed by bispecific antibodies (46·0%, 38·0-53·0 and 30·0%, 24·0-36·0) and antibody-drug conjugates (40·0%, 32·0-47·0 and 18·0%, 13·0-24·0).

不同药物类别之间的反应率有所不同，细胞治疗的客观缓解率或完全缓解率最高（70.0%，95%置信区间61.0-77.0和51.0%，95%置信区间43.0-59.0），其次是双特异性抗体（46.0%，38.0-53.0和30.0%，24.0-36.0）和抗体药物偶联物（40.0%，32.0-47.0和18.0%，13.0-24.0）。

Objective response rateincreased over time, from 16·6% (95% CI 9·0-29·0) in 2000-08 to 36·8% (30·0-45·0) in 2018-25.

客观缓解率随时间增加，从2000-08年的16.6% (95% 置信区间 9.0-29.0) 增加到2018-25年的36.8% (30.0-45.0)。

The overall rate of dose-limiting-toxicities or discontinuations was 6·0% (95% CI 4·7-7·6).

剂量限制性毒性或停药的总体发生率为6.0% (95% 置信区间 4.7-7.6)。

The rate of grade 3-4 adverse events was 61·5% (95% CI 54·2-68·3), treatment-related-mortality was 0·6% (0·4-1·0), and non-relapse-mortality was 3·6% (2·9-4·5).

3-4级不良事件的发生率为61.5%（95%置信区间54.2-68.3），治疗相关死亡率为0.6%（0.4-1.0），非复发死亡率为3.6%（2.9-4.5）。

Treatment-related mortality remained below 1% over time.

随时间推移，治疗相关死亡率保持在1%以下。

interpretation

Since the year 2000, early-phase trials in relapsed or refractory diffuse large B-cell lymphoma have shown more than a doubling of response rates, driven primarily by cellular and bispecific antibody therapies, while maintaining low treatment-related mortality.

自2000年以来，复发或难治性弥漫性大B细胞淋巴瘤的早期试验显示反应率翻了一番多，主要由细胞和双特异性抗体疗法推动，同时保持了低治疗相关死亡率。

These results provide risk-benefit trends in early-phase trials and define contemporary benchmarks for clinicians, investigators and regulators.

这些结果为早期试验中的风险-效益趋势提供了依据，并为临床医生、研究者和监管机构定义了当代基准。

funding

None.

无。