AstraZeneca presented initial Phase I PRIMAVERA data for AZD-3470 at ASCO 2026 showing a tolerable safety profile and encouraging responses in heavily pretreated relapsed/refractory classical Hodgkin lymphoma. In a 39-patient dose-escalation cohort (median six prior therapies), TEAEs occurred in 85% but were mostly low grade and no dose-limiting toxicities or TEAE-related deaths were reported. Clinical activity was observed at ≥450 mg (ORR 58%, CR 35%) and at 600 mg (ORR 63%, CR 44%). The study will expand to ~161 patients and longer-term outcomes are pending.
ASCO 2026: AstraZeneca’s AZD-3470 Demonstrates Promising Activity and Acceptable Safety in Relapsed/Refractory Classical Hodgkin Lymphoma

At the American Society of Clinical Oncology (ASCO) 2026 Annual Meeting on 30 May, AstraZeneca presented initial Phase I data for AZD-3470 from the PRIMAVERA study (NCT06137144), evaluating the investigational PRMT5 inhibitor in adults with relapsed or refractory (R/R) classical Hodgkin lymphoma (cHL).
Background
AZD-3470 inhibits protein arginine methyltransferase 5 (PRMT5), an enzyme frequently upregulated when methylthioadenosine phosphorylase (MTAP) is epigenetically silenced — a change reported in more than 80% of cHL cases. PRMT5 inhibitors are an emerging class across oncology, with multiple companies advancing programmes in haematologic and solid tumours.
Study Design
The presented dataset covers the first dose-escalation cohort of 39 heavily pretreated R/R cHL patients (aged ≥18 years) with a median of six prior lines of therapy, including brentuximab vedotin and anti–PD-1 antibodies. AZD-3470 was administered as monotherapy in escalating oral doses from 25 mg up to 600 mg, continuing until unacceptable toxicity or disease progression.
Safety
By the data cutoff, median exposure was 15 weeks. Treatment-emergent adverse events (TEAEs) were reported in 85% of patients and were predominantly low grade. The most common TEAEs were anaemia (28%) and nausea (15%). Grade 3 or higher events occurred in 28% of patients; neutropenia and hypokalemia were each observed in two patients. Importantly, no dose-limiting toxicities (DLTs), treatment discontinuations due to TEAEs, or TEAE-related deaths were reported. At the time of reporting, 51% of patients remained on treatment, supporting a tolerable safety profile up to the maximum tested dose.
Efficacy
Clinical activity emerged at doses ≥450 mg (n=31), with an objective response rate (ORR) of 58% and a complete response (CR) rate of 35%. Among patients receiving the 600 mg dose (n=16), the ORR was 63% and CR rate 44%. Secondary endpoints — including progression-free survival (PFS), durability of response and overall survival (OS) — require longer follow-up and will be reported as the trial matures.
Context And Competitive Landscape
These response rates and the safety profile are notable in a heavily pretreated R/R cHL population and compare broadly with historical single-agent anti–PD-1 studies in similar cohorts (for example, pembrolizumab and nivolumab trials reporting ORRs in the high 60s to low 70s). Multiple companies, including Bristol Myers Squibb, Tango Therapeutics, Johnson & Johnson and Bayer, are developing PRMT5 inhibitors. BMS’s navlimetostat and Tango’s vopimetostat have advanced solid-tumour programmes and hold FDA fast-track designations. Market analysis from GlobalData estimates the PRMT5 inhibitor opportunity could approach roughly $1.2 billion globally by 2031.
Next Steps
The PRIMAVERA trial will continue enrollment with planned expansion to approximately 161 patients across 29 global sites. AstraZeneca has also opened PRIMROSE (NCT06130553), a Phase I/II study of AZD-3470 in solid tumours. Further dose optimisation, cohort expansion and longer follow-up will determine durability and confirmatory efficacy as the programme advances toward registrational testing.
Disclosure: This article summarises data presented at ASCO 2026. The findings are preliminary from a Phase I trial and should be interpreted cautiously until larger, controlled studies provide confirmatory results.
Help us improve.




























