MRG003 Added to Ongoing Treatment After Progression in EGFR-Positive Advanced Thyroid Cancer

This prospective, open-label, single-center, nonrandomized phase II study will investigate MRG003 as an add-on treatment after RECIST 1.1 disease progression on a previously beneficial treatment regimen. Eligible participants must have EGFR-positive tumors by immunohistochemistry and documented prior clinical benefit consisting of complete response, partial response, or sustained stable disease, followed by radiographic progression. Continued use of the prior regimen must remain clinically appropriate and tolerable.

Cohort A includes anaplastic or poorly differentiated thyroid carcinoma after prior benefit from lenvatinib plus an anti-PD-1 inhibitor and subsequent progression. Lenvatinib and the same anti-PD-1 inhibitor will be continued, and MRG003 will be added. Cohort B includes radioactive iodine-refractory differentiated thyroid carcinoma after prior benefit from lenvatinib and subsequent progression. Lenvatinib will be continued, and MRG003 will be added.

MRG003 is the only newly introduced anticancer agent in both cohorts and will be administered intravenously at 2.0 mg/kg every 3 weeks. Lenvatinib will be continued at the tolerated dose used at progression; in Cohort A, the same anti-PD-1 inhibitor will continue according to the prior regimen. Dose interruptions, modifications, and discontinuation will follow protocol-defined safety rules. Add-on treatment will continue until subsequent disease progression, unacceptable toxicity, withdrawal of consent, or another protocol-defined discontinuation criterion.

This is an exploratory phase II study with planned enrollment of 39 participants: 24 in Cohort A and 15 in Cohort B. The cohort allocation reflects anticipated recruitment feasibility. Each cohort includes its initial safety lead-in participants. Early confirmed objective response rate will be summarized separately for each cohort using response counts, full-analysis-set denominators, percentages, and two-sided exact 95% Clopper-Pearson confidence intervals. The full analysis set includes enrolled participants who receive at least one dose of MRG003. Participants without a qualifying confirmed response are counted as nonresponders. The study is not designed to provide separately powered efficacy hypothesis tests for both cohorts.

Each cohort begins with 3 participants in a fixed-starting-dose safety lead-in. The primary dose-limiting toxicity (DLT) assessment window is Days 1-42 after the first MRG003 dose, covering two 21-day cycles. The first participant in each cohort must complete the first 21-day cycle and undergo safety review before the second participant starts; the third starts at least 7 days after the second. The first participants in the two cohorts are also staggered by at least 7 days.

If 0 of 3 DLT-evaluable participants experience a DLT and there are no other unresolved safety concerns, cohort expansion may proceed after documented safety review. One suspected DLT triggers immediate suspension of cohort enrollment and study dosing pending review; if the event is manageable and no major safety signal is present, the lead-in may expand to 6 DLT-evaluable participants within the planned cohort enrollment. Expansion requires no more than 1 of 6 evaluable participants to experience a DLT and a favorable safety review. DLTs in 2 or more participants during the cohort safety lead-in stop expansion of the current cohort regimen and trigger suspension of relevant study dosing pending review. Multiple DLT events in one participant count as one participant for these boundaries. A potentially treatment-related or unexplained death, life-threatening event, or another prespecified major safety signal triggers immediate study-wide suspension of enrollment and study dosing pending urgent review. Necessary medical care and individual risk assessment continue during a suspension. Safety monitoring continues beyond the DLT window.

Participants who receive MRG003 remain in the full analysis and safety sets regardless of toxicity, early discontinuation, or incomplete DLT assessment. Any participants enrolled to complete the DLT-evaluable lead-in are counted within the planned cohort and total enrollment. DLT assessment is based on new or worsened CTCAE v5.0 toxicities with a reasonable possibility of relationship to any component of the study regimen or their interaction.

A new tumor imaging baseline will be obtained within 28 days before the first MRG003 dose. Tumor assessments will be performed at Week 6 and every 6 weeks thereafter within the 12-month efficacy observation period, until progression or an applicable censoring condition. Necessary imaging follow-up continues after treatment discontinuation without progression. Complete and partial responses must be confirmed by a subsequent assessment at least 4 weeks later. Unscheduled imaging may be performed for suspected progression.

The primary outcome is early confirmed objective response rate according to RECIST 1.1, evaluated separately in each cohort: CR or PR must first be documented within 12 weeks after the first MRG003 dose and confirmed at least 4 weeks later, no later than Week 18. Both the initial response and confirmation must precede progression or new anticancer therapy. Responses first documented after Week 12 do not contribute to the primary outcome. Scheduled assessments at Weeks 6, 12, and 18 support this endpoint. Participants without a response meeting these requirements are counted as nonresponders in the full analysis set.

Disease control rate is the proportion of each cohort with CR, PR, or stable disease at the Week 12 assessment. It describes status at that time point, rather than best overall disease control over a period. Subsequent confirmation of CR or PR is not required for this measure. Progression, death, or new anticancer therapy before the Week 12 assessment, or absence of an adequate Week 12 assessment, is counted as no disease control.

Time to response and duration of response are evaluated among participants whose initial response and confirmation both occur within 12 months after the first MRG003 dose, before progression or new anticancer therapy. Eligible late responders beyond Week 12 may contribute to these secondary analyses. Time to response ends at the initial response date; duration of response starts on that date and ends at subsequent progression or death. Duration-of-response follow-up ends 12 months after the first MRG003 dose, not 12 months after the response.

Progression-free survival is measured from the first MRG003 dose to subsequent RECIST 1.1 progression or death, whichever occurs first, with follow-up capped at 12 months after the first MRG003 dose. Progression establishing eligibility is not an event for this outcome. For progression-free survival and duration of response, participants without an event are censored at the last adequate tumor assessment within the observation window; if new anticancer therapy begins first, censoring is at the last adequate assessment before that therapy.

Overall survival is measured from the first MRG003 dose to death from any cause, through 12 months after that dose, with censoring at the last date known alive within this window. Survival follow-up continues after progression or new anticancer therapy, every 12 weeks after treatment discontinuation, with a planned 12-month status assessment. Analyses are performed separately by cohort; 12-month survival and Kaplan-Meier estimates are reported, and median overall survival is reported only if estimable. Completion of the early primary endpoint does not end the remaining efficacy or safety follow-up. Safety is assessed during each treatment cycle and follows the separate collection rules below.

All adverse events (AEs), including serious adverse events (SAEs), are collected from informed consent, throughout study treatment, and through 90 days after the last dose of any protocol-defined study treatment, irrespective of attribution. The reference date is the latest administration date across all components administered within the study treatment regimen. Pre-existing conditions are recorded at baseline; treatment-emergent AEs are new or worsened events after the first MRG003 dose. Events are coded using MedDRA and graded using NCI CTCAE v5.0, and AEs and SAEs are summarized by cohort. The date of discontinuation of study treatment and any transition to routine care are documented; subsequent routine-care treatment does not reset the 90-day window.

Unresolved AEs and SAEs are followed until resolution, return to baseline, or medical stabilization with the subsequent care plan documented. Suspected treatment-related SAEs learned of after the 90-day window continue to be evaluated, recorded, and reported under the applicable safety reporting procedures. Initiation of other anticancer therapy does not automatically end safety follow-up; subsequent treatments and event attribution are documented. This safety follow-up is separate from the 42-day DLT assessment and the efficacy outcome assessment windows.