Background
Phase III open-label RCT (OPTiM), N=436 unresectable stage IIIB/IIIC/IV melanoma with injectable cutaneous, subcutaneous, or nodal lesions; prior systemic therapy allowed. Stratified by disease stage, prior therapy, ECOG PS. 64 sites, enrolled 2009–2011. T-VEC (talimogene laherparepvec) is a modified HSV-1 expressing GM-CSF — the first oncolytic virus therapy approved in oncology.
Interventions and follow up
Arm A: T-VEC intralesional injection: initial 10⁶ PFU/mL, then 10⁸ PFU/mL q2wk up to 26 wk (injected into accessible lesions)
Arm B: GM-CSF 125 mcg/m² SC days 1–14 of each 28-day cycle
Primary endpoint: Durable response rate (DRR; CR/PR ≥6 mo, beginning within 12 mo)
mFollow up: 49 mo
Arm B: GM-CSF 125 mcg/m² SC days 1–14 of each 28-day cycle
Primary endpoint: Durable response rate (DRR; CR/PR ≥6 mo, beginning within 12 mo)
mFollow up: 49 mo
Results
DRR: 16.3% vs 2.1% (T-VEC vs GM-CSF), P<.0001
ORR: 26.4% vs 5.7%
mOS: 23.3 vs 18.9 mo, HR 0.79 (95% CI 0.62–1.00), P=.051 — borderline
Stage III (IIIB/IIIC) DRR: 33% (T-VEC) vs 0% (GM-CSF) — substantially higher in locoregional disease
Non-injected lesion responses: observed in a subset, suggesting immune-mediated abscopal effect
ORR: 26.4% vs 5.7%
mOS: 23.3 vs 18.9 mo, HR 0.79 (95% CI 0.62–1.00), P=.051 — borderline
Stage III (IIIB/IIIC) DRR: 33% (T-VEC) vs 0% (GM-CSF) — substantially higher in locoregional disease
Non-injected lesion responses: observed in a subset, suggesting immune-mediated abscopal effect
Adverse events
Grade ≥3 AEs: 11.0% (T-VEC) vs 4.8% (GM-CSF)
Most common any-grade (T-VEC): fatigue 50.3%, chills 48.6%, pyrexia 42.8%, nausea 35.6%, influenza-like illness 30.5% — injection-related/flu-like
Grade ≥3 cellulitis: 2.1% (injection site)
Other: herpetic reactions possible from viral shedding (precautions for immunocompromised contacts); no treatment-related deaths
Most common any-grade (T-VEC): fatigue 50.3%, chills 48.6%, pyrexia 42.8%, nausea 35.6%, influenza-like illness 30.5% — injection-related/flu-like
Grade ≥3 cellulitis: 2.1% (injection site)
Other: herpetic reactions possible from viral shedding (precautions for immunocompromised contacts); no treatment-related deaths
Conclusions
T-VEC significantly improved durable response rate (16.3% vs 2.1%, P<.0001) and ORR (26.4% vs 5.7%) vs GM-CSF in unresectable melanoma with injectable lesions. OS improvement was clinically meaningful but did not reach formal significance (HR 0.79, P=.051). Benefit was most pronounced in stage IIIB/IIIC disease, supporting T-VEC as a regional therapy option.
Key Limitations
Control arm (GM-CSF SC) is not a meaningful active comparator — no established activity in melanoma. OS did not reach significance (P=.051), limiting OS benefit claims. Responses primarily in injected/regional lesions; modest distant response (limited abscopal effect in unselected patients). Requires directly injectable lesions — visceral/uninjectable disease excluded. Benefit strongly stage-specific (minimal in M1b/M1c). PD-1 inhibitors now show far superior OS (ORR 30–45% even in visceral disease). MASTERKEY-265 (T-VEC + pembrolizumab) was negative vs pembrolizumab alone.
Clinical Context
FDA approved T-VEC (Imlygic) for unresectable melanoma limited to skin and lymph nodes (stage IIIB/IIIC/IVM1a) in October 2015 — the first oncolytic virus therapy approved in the US. Reserved for patients with accessible injectable lesions and stage IIIB–IVM1a disease, or as palliative locoregional treatment; utility substantially limited by PD-1 inhibitor adoption. ESMO guidance positions T-VEC as a later-line option for injectable, unresectable regionally advanced melanoma. Rarely used as monotherapy in the modern IO era.