AAI Internal Wiki

T-cell Engager: Clinical Workflow and Case Studies

閱讀時間約 17 分鐘

T-cell Engager: Clinical Workflow and Case Studies

T 細胞接合器：臨床工作流程與案例研究

English

Understanding the pharmacology of T-cell engager CRS is a necessary prerequisite, but practicing oncologists need something more: a mental model for how to run the clinical workflow when one of these drugs is administered. The workflow is not intuitive, because it bundles activities that are normally separated — drug ordering, inpatient logistics, real-time toxicity assessment, rechallenge decisions, and longer-term infection monitoring — into a single integrated event. Each of the case studies below adds a specific dimension to this workflow picture.

JNJ-80038114: when early termination is the correct dose-optimization decision

JNJ-80038114 is a PSMAxCD3 bispecific antibody for metastatic castration-resistant prostate cancer (mCRPC). The FIH phase 1 study enrolled 39 patients with subcutaneous administration every 3 weeks. The starting dose was set at 0.1 mg based on MABEL, escalation used BOIN design with a target DLT rate not exceeding 30%, and the protocol included step-up dosing, premedication, and at least 36 hours of inpatient monitoring per ASTCT CRS/ICANS grading criteria. In other words, the trial was well-designed by current standards.

The results expose why “well-designed trial” and “successful drug” are different things. CRS occurred in 51.3% of patients, mostly grade 1-2, which was controlled. But DLTs occurred in 7.7%, clinically important neuropathy emerged, anti-drug antibodies developed in 56.8% of patients — dramatically reducing drug exposure over time — and antitumor activity was very limited. The trial was terminated early due to insufficient preliminary efficacy, neuropathy, CRS, and immunogenicity-driven PK disruption.

This case teaches a difficult but important lesson: the correct outcome of a phase 1 trial is not always to find a dose and proceed. The BOIN design worked — it escalated safely and gathered appropriate data. The monitoring protocol worked — it caught adverse events in real time. The conclusion that this drug should not proceed to phase 2 is itself a form of dose optimization: it protected future patients from enrollment in a trial with inadequate therapeutic index. In teaching this case, the useful exercise is to ask students to categorize the reasons for termination into four bins: safety, PK/immunogenicity, target biology adequacy, and efficacy signal adequacy. This structure prevents the oversimplified narrative of “the drug didn’t work.”

Epcoritamab follicular lymphoma: when dose optimization means changing the step-up, not the target dose

The FDA approval of epcoritamab-bysp for relapsed or refractory follicular lymphoma in 2024 offers a different teaching moment: dose optimization as step-up redesign. In EPCORE NHL-1, the primary safety and efficacy population used a 2-step regimen to reach the target dose. A dose optimization cohort of 86 additional patients evaluated a recommended 3-step regimen to reduce CRS. Under the 3-step schedule, CRS occurred in 49% of patients and was exclusively grade 1-2; efficacy appeared comparable to the 2-step population. The FDA approved the 3-step schedule.

The lesson here is that for T-cell engagers, dose optimization can mean adding a step to the ramp-up sequence rather than changing the milligram number. The additional step adds time and, in hospitalization-required protocols, adds cost and logistical burden. But if it reduces severe CRS events without compromising tumor control, it meets the Project Optimus standard: a better benefit-risk profile with adequate supporting evidence. For clinicians, this means that the label schedule for a T-cell engager is not necessarily fixed forever; it can evolve as real-world experience accumulates.

Alnuctamab: route of administration as part of dose optimization

Alnuctamab (ALNUC, CC-93269) is a BCMA × CD3 2+1 IgG1-based bispecific antibody for relapsed or refractory multiple myeloma. The FIH study CC-93269-MM-001, published in Leukemia in 2026, initially evaluated intravenous administration, then transitioned to subcutaneous based on the safety signal and suboptimal response profile seen with IV. The subcutaneous cohort enrolled 95 patients at target doses of 10, 15, 30, and 60 mg, using step-up doses of 3 mg and 6 mg. The dosing frequency de-intensified from weekly to every 2 weeks, then to every 4 weeks.

The subcutaneous route produced important differences. CRS occurred in 57.9% of SC patients but was entirely grade 1-2. Grade 3/4 neutropenia was 43.2% and infection (any grade) 64.2%. ORR for the SC cohort was 58.9%, with the 30 mg cohort achieving 71.4%. MRD negativity was achieved in 47 of 95 SC patients — a deep response signal. The selected RP2D was 30 mg, not 60 mg, because the incremental benefit of the higher dose did not justify the additional risk given the 30 mg cohort’s efficacy data.

The subcutaneous-versus-intravenous dimension here is not trivial. SC delivery produces a more gradual concentration-time profile, which reduces the cytokine peak that drives severe CRS during the first dose. It also changes the trial operations: SC injection can be given more quickly in clinic without a long infusion suite booking. But SC delivery does not eliminate CRS, neutropenia, or infection. The clinical workflow for SC TCEs still requires careful step-up scheduling, premedication, post-injection monitoring, and infection prophylaxis (including IVIG monitoring and antifungal/antibacterial coverage). Neutropenia in a patient who receives immunoglobulin-lowering therapy is a compounded infection risk that must be anticipated, not managed reactively.

Pavurutamab and the RP2D without DLT

Pavurutamab (AMG 701) is a BCMA-targeting bispecific T-cell engager for triple-class relapsed/refractory multiple myeloma. The Blood 2026 phase 1/1b study enrolled 172 patients across a dose range of 5 to 18,000 µg given weekly IV, with week-1 step-up dosing. Seventy-three patients received the RP2D of 18,000 µg. Across the entire study, 12 patients had DLTs including CRS and transaminase elevation, but no DLTs were observed at the RP2D. CRS occurred in 74.4% of patients overall, grade 3+ infection in approximately 34.9%. At the RP2D, ORR was 65.8%, median duration of response 36.6 months, and median PFS 16.8 months — exceptional efficacy for a triple-class refractory population. Importantly, the phase 1b also evaluated two different step-up dosing regimens, adding comparative data on CRS mitigation strategies.

The pavurutamab case illustrates that a high CRS rate at the RP2D does not automatically make the RP2D unacceptable. When CRS is predominantly grade 1-2, is predictably concentrated in the step-up period, responds reliably to tocilizumab/steroids, and the efficacy is as deep and durable as pavurutamab’s, the benefit-risk calculus can still favor treatment. What the trial design has to provide — and what this trial does provide — is granular data on which dose period produces the most CRS, what medications were required, whether any CRS events led to treatment discontinuation, and how infection risk evolved over the longer treatment period.

Tobemstomig: checkpoint bispecific FIH requires proof-of-mechanism, not just safety

Tobemstomig (RO7247669) is a PD-1/LAG-3 bispecific antibody — not a CD3 engager but a checkpoint bispecific — published in Clinical Cancer Research 2026. This distinction matters for teaching because checkpoint bispecifics require a different dose optimization logic. A CD3-engaging TCE needs to reach effective T-cell redirection with acceptable CRS. A PD-1/LAG-3 checkpoint bispecific needs to demonstrate dual checkpoint blockade, immune reinvigoration, and clinical activity, with the step-up dosing question largely replaced by a proof-of-mechanism question.

The tobemstomig FIH enrolled 35 dose-escalation patients and 69 expansion patients. MTD was not reached; the recommended dose for expansion (RDE) was 2100 mg every 2 weeks. In the CPI-experienced melanoma expansion cohort, ORR was 15%. More importantly, the investigators documented CD8-positive T-cell increases, stem-like CD8-positive T-cell expansion, and cytotoxic effector function acquisition — all consistent with the drug’s proposed mechanism of releasing dual checkpoint suppression. Treg changes were limited, suggesting tumor-specific rather than systemic immune activation.

For the clinician reading this paper, the key question is whether proof-of-mechanism evidence rises to the level of dose optimization evidence. An ORR of 15% in a checkpoint-experienced population is modestly encouraging. The immune pharmacodynamic data suggests that the drug is doing what it was designed to do. But these are exploratory endpoints in a phase 1 trial. The dose at which receptor occupancy plateaus, the relationship between immune activation and clinical response duration, and whether dose intensification or combination would improve outcomes — none of these are answered. The RDE is the highest tested safe dose that also showed mechanism-consistent pharmacodynamics, not the dose that maximizes clinical benefit. The next stage must answer the latter question.

中文

理解 T 細胞接合器 CRS 的藥理學是必要前提，但實踐中的腫瘤科醫師需要更多：一個當這些藥物給藥時如何運行臨床工作流程的心智模型。這個工作流程不是直覺性的，因為它將通常分開的活動——藥物下單、住院後勤、即時毒性評估、再給藥決定和長期感染監測——整合成一個事件。以下每個案例研究為這個工作流程圖景增加了一個特定維度。

JNJ-80038114：早期終止是正確的劑量最佳化決定

JNJ-80038114 是靶向 PSMAxCD3 的雙特異性抗體，用於轉移性去勢抗性攝護腺癌（mCRPC）。FIH 一期研究納入 39 位病人，皮下注射每 3 週一次。起始劑量根據 MABEL 設定為 0.1 mg，升量使用 BOIN 設計（目標 DLT 率不超過 30%），方案包括逐步升量給藥、預處置和根據 ASTCT CRS/ICANS 分級標準至少 36 小時的住院監測。換言之，試驗按目前標準設計良好。

結果揭示了「設計良好的試驗」和「成功的藥物」是不同的事情。CRS 在 51.3% 的病人中發生，多為 1-2 級，得到控制。但 7.7% 出現 DLT，出現臨床重要的神經病變，56.8% 的病人出現抗藥抗體——隨時間顯著降低藥物暴露——且抗腫瘤活性非常有限。試驗因初步療效不足、神經病變、CRS 和免疫原性驅動的 PK 破壞而提前終止。

這個案例教導了一個困難但重要的教訓：一期試驗的正確結果並不總是找到一個劑量並繼續進行。BOIN 設計有效——它安全升量並收集了適當資料。監測方案有效——它即時發現不良事件。決定這個藥物不應進入二期本身就是劑量最佳化的一種形式：它保護了未來病人免受治療指數不足的試驗入組。在教授這個案例時，有用的練習是要求學員將終止原因分類到四個框：安全性、PK/免疫原性、靶點生物學充分性和療效訊號充分性。這種結構防止了「藥物無效」的過度簡化敘事。

Epcoritamab 濾泡性淋巴瘤：劑量最佳化意味著改變逐步升量而非目標劑量

FDA 2024 年核准 epcoritamab-bysp 用於復發或難治性濾泡性淋巴瘤提供了不同的教學時刻：劑量最佳化作為逐步升量重新設計。在 EPCORE NHL-1 中，主要安全性和療效族群使用 2 步方案到達目標劑量。一個 86 位額外病人的劑量最佳化 cohort 評估建議的 3 步方案以降低 CRS。在 3 步時程下，CRS 在 49% 的病人中發生且均為 1-2 級；療效看起來與 2 步族群相當。FDA 核准了 3 步時程。

這裡的教訓是：對 T 細胞接合器，劑量最佳化可能意味著在加速序列中增加一步，而非改變毫克數。額外的步驟增加時間，在需要住院的方案中增加費用和後勤負擔。但如果它在不損害腫瘤控制的情況下減少嚴重 CRS 事件，它符合 Project Optimus 標準：具有足夠支持證據的更好效益—風險比。對臨床醫師而言，這意味著 T 細胞接合器的標籤時程不一定永遠固定；隨著真實世界經驗積累，它可以演進。

Alnuctamab：給藥途徑作為劑量最佳化的一部分

Alnuctamab（ALNUC，CC-93269）是靶向 BCMA × CD3 的 2+1 IgG1 型雙特異性抗體，用於復發或難治性多發性骨髓瘤。FIH 研究 CC-93269-MM-001 於 2026 年發表在 Leukemia，最初評估靜脈注射，然後根據 IV 的安全性訊號和次優反應特徵轉為皮下注射。皮下 cohort 納入 95 位病人，目標劑量為 10、15、30 和 60 mg，使用 3 mg 和 6 mg 的逐步升量劑量。給藥頻率從每週逐步降低到每 2 週，再到每 4 週。

皮下途徑產生了重要差異。SC 病人中 CRS 發生率 57.9%，但全為 1-2 級。3/4 級嗜中性球低下為 43.2%，感染（任何級別）64.2%。SC cohort 的 ORR 為 58.9%，30 mg cohort 達到 71.4%。95 位 SC 病人中 47 位達到 MRD 陰性——深度反應訊號。選擇的 RP2D 為 30 mg 而非 60 mg，因為更高劑量的邊際獲益在 30 mg cohort 療效資料的情況下不能證明額外風險合理。

皮下與靜脈注射的維度不是瑣碎的。SC 給藥產生更漸進的濃度—時間曲線，降低在第一劑期間驅動嚴重 CRS 的細胞激素峰值。它也改變試驗操作：SC 注射可以在門診更快給予，無需長時間的輸注室預訂。但 SC 給藥不能消除 CRS、嗜中性球低下或感染。SC TCE 的臨床工作流程仍需要仔細的逐步升量時程安排、預處置、注射後監測和感染預防（包括 IVIG 監測和抗黴菌/抗菌覆蓋）。接受免疫球蛋白降低療法的病人中的嗜中性球低下是複合感染風險，必須預先計畫而非被動管理。

Tobemstomig：免疫檢查點雙特異性 FIH 需要機轉證明，不只是安全性

Tobemstomig（RO7247669）是 PD-1/LAG-3 雙特異性抗體——不是 CD3 接合器而是免疫檢查點雙特異性——2026 年發表在 Clinical Cancer Research。這個區別對教學很重要，因為免疫檢查點雙特異性需要不同的劑量最佳化邏輯。CD3 接合型 TCE 需要在可接受 CRS 的情況下達到有效的 T 細胞重定向。PD-1/LAG-3 免疫檢查點雙特異性需要證明雙重檢查點阻斷、免疫重新活化和臨床活性，逐步升量問題在很大程度上被機轉證明問題所取代。

Tobemstomig FIH 納入 35 位劑量升量病人和 69 位擴增病人。MTD 未達；建議擴增劑量（RDE）為每 2 週 2100 mg。在 CPI 經歷的黑色素瘤擴增 cohort 中，ORR 為 15%。更重要的是，研究者記錄了 CD8 陽性 T 細胞增加、幹樣 CD8 陽性 T 細胞擴增和細胞毒性效應功能獲得——均與藥物提出的雙重免疫檢查點抑制釋放機轉一致。Treg 變化有限，提示腫瘤特異性而非全身性免疫活化。

對閱讀這篇論文的臨床醫師，關鍵問題是機轉證明證據是否上升到劑量最佳化證據的水平。在免疫檢查點經歷族群中 15% 的 ORR 適度令人鼓舞。免疫藥效學資料表明藥物正在做它設計的事情。但這些是一期試驗中的探索性終點。受體佔有率穩定的劑量、免疫活化與臨床反應持續時間之間的關係，以及劑量增強或聯合治療是否能改善結果——這些都未被回答。RDE 是最高測試安全劑量，同時顯示與機轉一致的藥效學，而不是最大化臨床獲益的劑量。下一階段必須回答後者。

Key Concepts | 核心概念

  • Early termination as dose optimization | 早期終止作為劑量最佳化: JNJ-80038114 terminated for immunogenicity + neuropathy + insufficient efficacy — protecting future patients IS the optimization
  • Step-up redesign | 逐步升量重設計: Epcoritamab 3-step vs 2-step shows dose optimization can mean adding a step, not changing the mg number
  • Route of administration | 給藥途徑: SC alnuctamab reduces CRS Cmax but does not eliminate infection or myelosuppression risk
  • Checkpoint bispecific vs. CD3 TCE dosing logic: Checkpoint bispecifics require proof-of-mechanism (receptor occupancy + immune PD), not step-up CRS management
  • Infection risk in BCMA × CD3 therapies: Hypogammaglobulinemia, neutropenia, and tumor burden → compound infection probability requiring prophylactic protocol

Related Pages | 相關頁面

關係圖譜

反向連結