Tusamitamab ravtansine in 2L NSCLC: when the antigen is real but the ADC is not enough

What was tried

CARMEN-LC03 was a Phase 3, open-label, randomized study comparing the anti-CEACAM5 antibody-drug conjugate tusamitamab ravtansine (SAR408701) against docetaxel in adults with metastatic non-squamous non-small-cell lung cancer (NSCLC) who had progressed on platinum chemotherapy and a checkpoint inhibitor. Enrollment required tumors expressing CEACAM5 by immunohistochemistry at moderate-to-high levels in at least 50% of cells. The trial enrolled 389 patients, 194 to tusamitamab ravtansine and 195 to docetaxel, and had dual primary endpoints of progression-free survival (PFS) and overall survival (OS). The ClinicalTrials.gov record for NCT04154956 now lists the study as terminated with the brief explanation, "Sponsor decision, the decision is not related to any safety concern."

Tusamitamab ravtansine is a humanized IgG1 antibody against CEACAM5 conjugated through a cleavable SPDB linker to DM4, a maytansinoid tubulin inhibitor with a drug-to-antibody ratio of about 3.8 (Gazzah et al., Annals of Oncology 2022). The dosing regimen was 170 mg/m² intravenously every two weeks. Docetaxel was given at the standard 75 mg/m² every three weeks. Treatment continued to progression or intolerable toxicity.

The Phase 3 rationale rested on a Phase 1 dose-expansion cohort of 64 patients with CEACAM5-high non-squamous NSCLC, in which tusamitamab ravtansine produced a confirmed response rate of 20.3% (95% CI 11.3 to 32.2) and a median PFS of 5.4 months. That signal was enough for the sponsor to design a head-to-head registrational trial with docetaxel as comparator, a deliberately high bar in a setting where docetaxel itself is widely viewed as a beatable benchmark.

The biological hypothesis

CEACAM5 (CEA cell adhesion molecule 5) is a glycosylphosphatidylinositol-anchored member of the carcinoembryonic antigen family. It is expressed at low levels in normal columnar epithelium and adult colon, and at substantially higher levels in adenocarcinomas of the colon, pancreas, stomach, breast, and lung. The original tumor-mapping work by Blumenthal and colleagues showed that high CEACAM5 staining is present in roughly 70% of colorectal cancers, 50% of non-squamous lung cancers, and 30% to 40% of gastric cancers, with comparatively limited normal-tissue expression (Blumenthal et al., BMC Cancer 2007). For an antibody-drug conjugate program, that ratio of tumor to normal expression is the engineering target.

The clinical rationale was sharpened by lung-cancer-specific surveys. Lefebvre and colleagues quantified CEACAM5 expression across 1,378 NSCLC samples and found that roughly 20% of non-squamous NSCLC met the threshold of moderate-to-high staining in at least 50% of tumor cells used to enrich the trial (Lefebvre et al., Lung Cancer 2023). Internalization rates in those lines were sufficient to deliver the DM4 payload to a mitotically active cell.

The mechanism is conceptually clean. CEACAM5 on the tumor cell surface binds tusamitamab. The complex internalizes through clathrin-coated pits, traffics to the lysosome, the cleavable disulfide linker is reduced, and DM4 diffuses into the cytosol where it binds the vinca domain of tubulin and arrests cells in mitosis. The Phase 1 program documented a tolerable safety profile dominated by reversible corneal keratopathy and asthenia, with very little neutropenia, a feature distinct from docetaxel (Tabernero et al., Cancer Research Communications 2023).

What actually happened

The dual primary endpoint failed. Median PFS was 5.39 months for tusamitamab ravtansine versus 5.85 months for docetaxel (hazard ratio 1.14, p = 0.82). Median OS was 12.81 months versus 11.53 months (hazard ratio 0.85, p = 0.11), a directionally favorable signal that fell well short of the prespecified significance threshold. Confirmed objective response rates were 21.7% (95% CI 16.0 to 28.3) and 24.1% (95% CI 18.1 to 30.8). The data were presented at WCLC 2024 and the trial closed shortly after Sanofi's program-discontinuation announcement in late 2023 (Sanofi press release, 21 December 2023).

The safety profile was the expected story for a non-cytotoxic backbone. Treatment-emergent serious adverse events occurred in 28.4% of tusamitamab patients versus 38.4% of docetaxel patients. Grade ≥3 events were 40.7% versus 57.6%. Dose reductions were less than half as common in the ADC arm (16.5% versus 36.2%). Tusamitamab was clearly easier to tolerate, but no patient population can be treated with a drug that has matched efficacy at a high price unless they cannot tolerate the comparator.

A prespecified subgroup analysis using a higher CEACAM5 expression cutoff (at least 80% of cells) suggested a more favorable hazard ratio of 0.87 for PFS and 0.71 for OS, but this was not part of the primary statistical plan and the absolute numbers in that subgroup were modest. Sanofi cited that residual signal when pivoting to a different anti-CEACAM5 asset.

Failure mechanism, best guess

Three explanations fit the data. First, the chosen CEACAM5 expression threshold may have been too permissive. The trial allowed any tumor with moderate-to-high staining in 50% of cells, but the post hoc 80% subgroup analysis hints that the antigen density required for meaningful payload delivery is higher than the marketing-friendly 50% cutoff. ADCs are not on-off switches. Receptor density, internalization kinetics, and bystander effect all scale, and the drug failed precisely where the antigen was dilute.

Second, DM4 may be the wrong payload for an NSCLC line of therapy that has already seen platinum and frequently taxanes. DM4 is a tubulin inhibitor. Most of these patients had been exposed to a different microtubule-targeting agent within the prior year. Cross-resistance through P-glycoprotein upregulation, beta-tubulin isoform shifts, and altered microtubule dynamics is a known liability. The contrast with topoisomerase I payload ADCs in the same disease, where trastuzumab deruxtecan and patritumab deruxtecan have moved the needle, is striking.

Third, recent single-cell work shows that surface-antigen heterogeneity within a treated tumor can rise sharply under selection pressure, with phenotype switching driven by neuroendocrine and EMT-like programs (Zaidi et al., PNAS 2024). The same biology has been described in lung adenocarcinoma. A monovalent antibody targeting a single surface antigen at a fraction of the cell population will not produce durable disease control if the surviving cells are antigen-low.

What this means for similar programs

CEACAM5 is not dead as a target. Sanofi has redirected to a topo-payload anti-CEACAM5 ADC, and at least three independent programs targeting the same antigen with different linker-payload chemistry are in clinical development, including labetuzumab govitecan, MEDI-565 bispecific T-cell engager, and several preclinical T-cell engagers and CAR-T constructs. Any of these benefits from CARMEN-LC03 in two ways: the registrational enrichment threshold needs to be tighter than 50% staining, and the comparator math has to anticipate that docetaxel after platinum and immunotherapy is no longer the soft target it appeared to be a decade ago.

Open questions

• Does the CEACAM5-high subgroup hazard ratio of 0.71 for OS survive a properly powered confirmatory trial, or is it noise?
• Does the same antibody backbone perform differently when paired with a topoisomerase I inhibitor payload rather than DM4?
• What is the rate of antigen loss and lineage switching in CEACAM5-positive NSCLC under prior taxane exposure?
• Why did interstitial lung disease emerge as a class signal across DM1 and DM4 maytansinoid ADCs but not at concerning rates in this trial (Maurier et al., Revue des Maladies Respiratoires 2025)?
• Should second-line NSCLC trials still use docetaxel as control given current real-world outcomes?

Sources

1. Barlesi F, et al. CARMEN-LC03: a Phase 3 trial of tusamitamab ravtansine in second-line non-squamous NSCLC. World Conference on Lung Cancer 2024. Trial record at ClinicalTrials.gov NCT04154956.

2. Sanofi. Press release, 21 December 2023. End of program evaluating tusamitamab ravtansine after a 2L NSCLC Phase 3 trial did not meet a primary endpoint. sanofi.com.

3. Gazzah A, Bedard PL, Hierro C, Kang YK, Abdul Razak A, Ryu MH, et al.. Safety, pharmacokinetics, and antitumor activity of the anti-CEACAM5-DM4 antibody-drug conjugate tusamitamab ravtansine (SAR408701) in patients with advanced solid tumors: first-in-human dose-escalation study. Ann Oncol. 2022;33(4);416-425. PMID: 35026412.

4. Tabernero J, Bedard PL, Bang YJ, Vieito M, Ryu MH, Fagniez N, et al.. Tusamitamab Ravtansine in Patients with Advanced Solid Tumors: Phase I Study of Safety, Pharmacokinetics, and Antitumor Activity Using Alternative Dosing Regimens. Cancer Res Commun. 2023;3(8);1662-1671. PMID: 37645622.

5. Lefebvre AM, Adam J, Nicolazzi C, Larois C, Attenot F, Falda-Buscaiot F, et al.. The search for therapeutic targets in lung cancer: Preclinical and human studies of carcinoembryonic antigen-related cell adhesion molecule 5 expression and its associated molecular landscape. Lung Cancer. 2023;184;107356. PMID: 37660479.

6. Blumenthal RD, Leon E, Hansen HJ, Goldenberg DM. Expression patterns of CEACAM5 and CEACAM6 in primary and metastatic cancers. BMC Cancer. 2007;7;2. PMID: 17201906.

7. Zaidi S, Park J, Chan JM, Roudier MP, Zhao JL, Gopalan A, et al.. Single-cell analysis of treatment-resistant prostate cancer: Implications of cell state changes for cell surface antigen-targeted therapies. Proc Natl Acad Sci U S A. 2024;121(28);e2322203121. PMID: 38968122.

8. Maurier L, Chéné AL, Hulo P, Chen J, Sagan C, Pons-Tostivint E. [Diffuse interstitial lung disease induced by antibody-drug conjugates]. Rev Mal Respir. 2025;42(5);274-285. PMID: 40263022.

9. Pouzin C, Tod M, Chadjaa M, Fagniez N, Nguyen L. Covariate analysis of tusamitamab ravtansine, a DM4 anti-CEACAM5 antibody-drug conjugate, based on first-in-human study. CPT Pharmacometrics Syst Pharmacol. 2022;11(3);384-394. PMID: 35191618.

10. CARMEN-LC03. Randomized, Open-label, Phase 3 Study of SAR408701 Versus Docetaxel in Previously Treated, Metastatic Nonsquamous, Non-small-cell Lung Cancer Patients With CEACAM5-positive Tumors. ClinicalTrials.gov. NCT04154956.