Radiopharmaceutical Therapy FIH: Activity, Dosimetry, and Dose Optimization
放射性藥物治療首次人體試驗：活度、劑量測定與劑量最佳化
English
Radiopharmaceutical therapy (RPT) is the one drug modality in oncology where the word “dose” is systematically ambiguous. When a cardiologist talks about a dose of metoprolol, there is one number: milligrams per day. When a radiation oncologist talks about an external beam dose, there is a number in gray (Gy) delivered to a specific volume. When a nuclear medicine physician or oncologist discusses RPT dosing, they must navigate simultaneously among administered activity (how many becquerels or megabecquerels you inject), mass dose (how many micrograms of the targeting molecule you give), absorbed dose (how many gray a specific tissue actually receives), and the schedule across which multiple cycles are delivered. These are related but not equivalent quantities. A FIH trial that reports only administered activity without dosimetry data is telling you one number when the biology requires four.
The FDA recognized this complexity formally in a 2025 draft guidance titled “Oncology Therapeutic Radiopharmaceuticals: Dosage Optimization During Clinical Development.” This document defines RPT dosage optimization as the optimization of administered activity and schedule — not just activity level alone — and explicitly notes that radiopharmaceuticals were excluded from the 2024 oncology dosage optimization guidance for conventional drugs. The reason for exclusion is not that RPT is exempt from dose optimization principles, but that its dose optimization language and metrics are fundamentally different. The 2025 draft guidance also warns against uncritical borrowing of organ dose limits from external beam radiotherapy (EBRT): EBRT and RPT have different physical mechanisms, dose delivery profiles, and temporal dose-rate characteristics, so EBRT-derived limits may not translate accurately to RPT.
Understanding why MTD logic fails for RPT requires understanding the physical properties of the radionuclides involved. For beta-emitters like lutetium-177 (used in lutetium PSMA therapies and DOTATATE), radiation travels millimeters in tissue, allowing some selectivity between tumor tissue with high target expression and adjacent normal tissue. For alpha-emitters like actinium-225 or bismuth-213, radiation travels only microns but deposits energy approximately 20-fold more densely per unit path length, making them lethal to cells within nanometer proximity but relatively sparing of distant tissue. The DLT concept from conventional oncology assumes that the dose-limiting event happens at a predictable systemic threshold. For RPT, the limiting event may be cumulative absorbed dose to a specific organ — kidney, bone marrow, salivary glands — that accumulates over multiple cycles and may not manifest clinically until weeks or months after the final cycle. A 28-day DLT window catches acute hematological toxicity but cannot capture this cumulative radiation nephropathy pattern.
The BANTAM-01 trial provides the most instructive contemporary teaching case for alpha-emitter FIH design. BAY 3547926 (225Ac-GPC3) is a glypican-3-targeting actinium-225 antibody-chelator conjugate designed for advanced hepatocellular carcinoma. The trial design, published in the Journal of Nuclear Medicine in 2026, is a multicenter, open-label phase 1 dose escalation and expansion study planned for approximately 148 patients. The design includes: prospective confirmation of tumor GPC3 membrane expression as an eligibility criterion; monotherapy dose escalation; monotherapy dose expansion; and a safety run-in plus expansion cohort in combination with standard of care. Dosing cycles are 6 weeks with a maximum of 4 cycles.
What makes BANTAM-01 instructive is that the drug brings together several layers of complexity that conventional drug trials do not face simultaneously. GPC3 target expression must be confirmed prospectively — because alpha-emitter delivery is so spatially localized that heterogeneous target expression could create unpredictable absorbed dose distributions within the tumor. The actinium-225 decay chain produces daughter radionuclides including bismuth-213, which means the radioactive “payload” is not just the parent alpha emitter but also includes decay products with their own tissue distribution and retention patterns. The FIH dose rationale must account for: target expression, antibody mass dose, radioactivity-to-mass ratio, predicted organ biodistribution from animal data, estimated absorbed dose to critical organs (liver, kidney, bone marrow — all relevant for a hepatocellular carcinoma drug), fractionation across a maximum of 4 cycles, and imaging feasibility for dosimetric monitoring.
The SNMMI-FDA 2024 workshop on RPT dose optimization, summarized in the Journal of Nuclear Medicine 2026, articulates a principle that should inform every RPT FIH discussion: for patients with advanced treatment-refractory solid tumors, the greatest near-term health risk is usually the cancer itself, not the potential for late-onset radiation-related effects from therapy. This perspective supports appropriate therapeutic ambition in RPT dose selection — the goal is not to minimize radiation exposure at the cost of therapeutic efficacy, but to optimize the ratio of tumor absorbed dose to organ-at-risk absorbed dose while maintaining acceptable safety margins. Phase 1 exploration of both dose and schedule is critical precisely because it establishes the foundation for all subsequent trials and potential regulatory approval.
The role of dosimetry standardization cannot be overstated. A 2025 EJNMMI Physics technical note evaluated lutetium-177 SPECT/CT-based dosimetry workflows and found that lack of standardized protocols — in reconstruction iterations, calibration methods, voxel-based dosimetry approaches, and target segmentation — produced meaningfully different absorbed dose estimates from the same imaging data. This is a clinical pharmacology problem: if two institutions using the same drug produce different dosimetric measurements, then absorbed-dose-guided dose escalation or individual cycle adjustment is unreliable. For RPT FIH trials, the dosimetry workflow is not a departmental technical matter — it is part of the drug’s clinical pharmacology evidence package, and must be validated and standardized as rigorously as the bioanalytical methods for a small molecule’s plasma concentration.
Phase 0 microdose studies offer a specific tool for early human biodistribution data without the therapeutic complexity of a full RPT trial. The AB001 study, published in the Journal of Nuclear Medicine in 2025, enrolled three patients with metastatic castration-resistant prostate cancer to receive approximately 9.4 MBq of a 212Pb-labeled PSMA radioligand. Post-injection imaging with planar gamma camera and SPECT/CT documented biodistribution across metastatic lesions and normal organs (kidney, bladder, liver, blood pool) for 28 days, with associated safety and biomarker monitoring. No therapeutic activity or efficacy claim was made. The sole purpose was to answer: does this drug reach PSMA-expressing lesions in humans? Which normal organs have the highest uptake? Is the imaging quality sufficient to support dosimetric planning for a therapeutic trial? For an alpha-emitter under development, this phase 0 approach reduces the number of therapeutic-dose patients who receive a drug without confirmed biodistribution in humans — a meaningful clinical and ethical optimization.
中文
放射性藥物治療（RPT）是腫瘤科中「劑量」這個詞系統性地模糊的唯一藥物治療方式。當心臟科醫師談論 metoprolol 的劑量時，只有一個數字：每天毫克數。當放射腫瘤科醫師談論外照射劑量時，有一個以 Gy 為單位的數字，遞送到特定體積。當核醫科醫師或腫瘤科醫師討論 RPT 給藥時，他們必須同時處理：給予放射活性（注射多少 MBq）、質量劑量（給予多少微克標靶分子）、吸收劑量（特定組織實際接受多少 Gy），以及多個週期的時程安排。這些是相關但不等同的數量。一個只報告給予放射活性而無劑量測定資料的 FIH 試驗，是在生物學需要四個數字時只告訴你一個。
FDA 在 2025 年草案指引「腫瘤治療性放射性藥物：臨床開發期間的劑量最佳化」中正式承認了這種複雜性。該文件將 RPT 劑量最佳化定義為給予活性和時程的最佳化——不只是活性水平——並明確指出放射性藥物被排除在 2024 年傳統藥物腫瘤劑量最佳化指引之外。排除的原因不是 RPT 免除劑量最佳化原則，而是其劑量最佳化語言和指標從根本上不同。2025 年草案指引也警告不要不加批判地借用外照射放療（EBRT）的器官劑量限制：EBRT 和 RPT 具有不同的物理機轉、劑量遞送特徵和時間劑量率特性，所以 EBRT 衍生的限制可能無法準確轉化為 RPT。
理解為什麼 MTD 邏輯對 RPT 失敗，需要理解涉及放射性核種的物理性質。對於如鑥-177（用於鑥 PSMA 治療和 DOTATATE）等 beta 發射體，輻射在組織中傳播幾毫米，允許在高標靶表達的腫瘤組織和鄰近正常組織之間有一定選擇性。對於如錒-225 或鉍-213 等 alpha 發射體，輻射只傳播幾微米，但每單位路徑長度的能量沉積密度約高 20 倍，在納米距離內對細胞是致命的，但對遠處組織相對無害。傳統腫瘤科的 DLT 概念假設劑量限制事件在可預測的全身閾值發生。對 RPT，限制事件可能是特定器官的累積吸收劑量——腎臟、骨髓、唾液腺——在多個週期中累積，且可能在最後一個週期後數週或數月才在臨床上表現。28 天的 DLT 窗口可以捕捉急性血液毒性，但無法捕捉這種累積放射性腎病模式。
BANTAM-01 試驗為 alpha 發射體 FIH 設計提供了最具教學價值的當代案例。BAY 3547926（225Ac-GPC3）是靶向 glypican-3 的錒-225 抗體螯合物偶聯體，設計用於晚期肝細胞癌。2026 年發表於 Journal of Nuclear Medicine 的試驗設計，是計畫納入約 148 位病人的多中心、開放標籤一期劑量升量和擴增研究。設計包括：前瞻性確認腫瘤 GPC3 膜表達作為入組標準；單藥劑量升量；單藥劑量擴增；以及與標準治療聯合的安全性啟動加擴增 cohort。給藥週期為 6 週，最多 4 次。
讓 BANTAM-01 具有教學意義的是，這個藥物帶來了傳統藥物試驗不需要同時面對的幾層複雜性。GPC3 標靶表達必須前瞻性確認——因為 alpha 發射體遞送空間高度局部化，異質性標靶表達可能在腫瘤內部創造不可預測的吸收劑量分布。錒-225 衰變鏈產生包括鉍-213 在內的子核素，這意味著放射性「payload」不只是母體 alpha 發射體，還包括具有自身組織分布和滯留模式的衰變產物。FIH 劑量理由必須考慮：標靶表達、抗體質量劑量、放射活性與質量比、來自動物資料的預測器官生物分布、對關鍵器官（肝臟、腎臟、骨髓——對肝細胞癌藥物均相關）的估計吸收劑量、最多 4 個週期的分次給藥，以及劑量測定監測的影像可行性。
SNMMI-FDA 2024 年 RPT 劑量最佳化研討會（摘要於 2026 年 Journal of Nuclear Medicine 發表）闡明了一個應告知每次 RPT FIH 討論的原則：對於晚期治療後難治的實體腫瘤病人，最大的近期健康風險通常是癌症本身，而非治療的潛在遲發放射相關效應。這個視角支持 RPT 劑量選擇的適當治療雄心——目標不是以損害治療療效為代價最小化輻射暴露，而是在維持可接受安全邊際的同時最佳化腫瘤吸收劑量與危及器官吸收劑量的比率。一期對劑量和時程的探索至關重要，正是因為它建立了所有後續試驗和潛在監管核准的基礎。
劑量測定標準化的作用不可誇大。2025 年 EJNMMI Physics 的技術說明評估了鑥-177 SPECT/CT 劑量測定工作流程，發現缺乏標準化方案——在重建迭代次數、校正方法、體素劑量測定方法和靶區分割方面——從相同影像資料中產生有意義不同的吸收劑量估計。這是臨床藥理問題：如果使用相同藥物的兩家機構產生不同的劑量測定結果，那麼基於吸收劑量指導的劑量升量或個別週期調整是不可靠的。對 RPT FIH 試驗，劑量測定工作流程不是部門技術事務——它是藥物臨床藥理證據包的一部分，必須像小分子血漿濃度的生物分析方法一樣嚴格驗證和標準化。
零期微劑量研究提供了在沒有完整 RPT 試驗的治療複雜性的情況下獲得早期人體生物分布資料的特定工具。AB001 研究（2025 年發表於 Journal of Nuclear Medicine）納入三位轉移性去勢抗性攝護腺癌病人，接受約 9.4 MBq 的 212Pb 標記 PSMA 放射性配體。注射後使用平面伽馬相機和 SPECT/CT 的成像記錄了 28 天內轉移病灶和正常器官（腎臟、膀胱、肝臟、血池）的生物分布，伴隨安全性和生物標記監測。沒有提出治療活性或療效主張。唯一目的是回答：這個藥物是否在人體中到達 PSMA 表達的病灶？哪些正常器官的攝取最高？影像質量是否足以支持治療試驗的劑量測定計劃？對於正在開發的 alpha 發射體，這種零期方法減少了在人體中確認生物分布之前接受藥物的治療劑量病人數量——這是有意義的臨床和倫理最佳化。
Key Concepts | 核心概念
- “Dose” is four quantities in RPT | RPT 中的「劑量」是四個數量: administered activity, mass dose, absorbed dose (organ-specific), and schedule — all must be reported
- MTD logic fails for RPT | MTD 邏輯對 RPT 無效: Cumulative kidney/marrow absorbed dose manifests weeks after last cycle, beyond standard DLT window
- Alpha emitter complexity | Alpha 發射體複雜性: Daughter radionuclide decay chains (e.g., 225Ac → 213Bi) create additional biodistribution and dosimetry variables
- Dosimetry standardization | 劑量測定標準化: Non-standardized SPECT/CT reconstruction protocols produce meaningfully different absorbed dose estimates from identical data
- Phase 0 microdose role | 零期微劑量的作用: Answers biodistribution questions in humans without therapeutic-level radiation exposure
- RPT excluded from 2024 FDA guidance: RPT has its own 2025 draft guidance; direct application of conventional oncology dose optimization criteria is inappropriate
Related Pages | 相關頁面
- three-modern-fih-platforms-compared — How RPT dose optimization differs from ADC and TCE approaches
- adc-fih-case-studies-b7h3-trop2 — BANTAM-01 (225Ac-GPC3) vs. ifinatamab deruxtecan (B7-H3 ADC): different modality, same tumor target
- adc-fih-clinical-pharmacology — Parallel pharmacology framework for ADCs