Onpode
Cover art for SELLAS's leukemia vaccine cut death risk 41% in Phase 3—but won't file with FDA until 2027

SELLAS's leukemia vaccine cut death risk 41% in Phase 3—but won't file with FDA until 2027

November 20, 2026 · 11 min

Juniper Vale & Hope Sterling

Juniper, hi — how's your week, actually, wait, forget that, I need to tell you about this patient I've been thinking about. An actual patient or a — ? No, like, a hypothetical — someone I invented in my head to understand this news. Late sixties, sitting in a clinic chair, and the doctor is…

On November 17, 2026, SELLAS Life Sciences Group (NASDAQ: SLS) announced positive topline results from the pivotal randomized Phase 3 REGAL trial (NCT04229979) evaluating galinpepimut-S (GPS), an immunotherapeutic cancer vaccine targeting Wilms tumor 1 (WT1), in patients with acute myeloid leukemia (AML).

0:0010:57
Get the next episode on Health →

Follow it free — new episodes land in your feed.

Or make your own — any topic, in minutes

More Onpode episodes on Health →

About this episode

On November 17, 2026, SELLAS Life Sciences Group (NASDAQ: SLS) announced positive topline results from the pivotal randomized Phase 3 REGAL trial (NCT04229979) evaluating galinpepimut-S (GPS), an immunotherapeutic cancer vaccine targeting Wilms tumor 1 (WT1), in patients with acute myeloid leukemia (AML).

Grounded in 12 sources
The current socioeconomic and regulatory landscape of immune effector cell therapies ↗ · doi.org
Management of Acute Myeloid Leukemia: A Review ↗ · pmc.ncbi.nlm.nih.gov
99.1 - SEC.gov ↗ · sec.gov
SELLAS cancer vaccine cuts death risk 41% in Phase III AML trial, targets 2027 BLA filing ↗ · allsci.com
Approach to Patients with Acute Myeloid Leukemia in First Relapse: A Practical Guide for Canadian Hematologists | Canadian Hematology Today ↗ · canadianhematologytoday.com
Galinpepimut-S Improves Overall Survival in Phase 3 AML Maintenance Trial | CancerNetwork ↗ · cancernetwork.com
Episode 2: Clinical Challenges and Unmet Needs in Relapsed/Refractory NPM1-Mutated AML - Cancer Therapy Advisor ↗ · cancertherapyadvisor.com
Galinpepimut-S Reduces Risk of Death by 41% in Acute Myeloid Leukemia Following Second Remission ↗ · hmpgloballearningnetwork.com
Leukemia Vaccine Galinpepimut-S Cut Death Risk 41% for AML Patients in Remission in Unreviewed Phase 3 Results ↗ · medicaldaily.com
Optimizing Care in Relapsed/Refractory AML ↗ · oncologynewscentral.com
AML: 24.0-Mo OS With Galinpepimut-S Vs 14.7-Mo BAT In CR2 (REGAL 2026) | OncologyTube ↗ · oncologytube.com
Galinpepimut-S Cuts Risk of Death by 41% in Phase 3 REGAL Trial | Pharmacy Times ↗ · pharmacytimes.com
Read transcript

Hope Sterling: Juniper, hi — how's your week, actually, wait, forget that, I need to tell you about this patient I've been thinking about.

Juniper Vale: An actual patient or a — ?

Hope Sterling: No, like, a hypothetical — someone I invented in my head to understand this news. Late sixties, sitting in a clinic chair, and the doctor is saying: your leukemia came back, and your body literally cannot survive a bone marrow transplant. And that doctor — on November 17th, 2026 — could have pulled up a press release from SELLAS Life Sciences and said, there is a drug called galinpepimut-S, and it just changed what I can offer you.

Juniper Vale: The REGAL Phase 3 trial results.

Hope Sterling: Yes — and I want to explain what this drug actually does before we get into numbers, because the concept is kind of beautiful? Acute myeloid leukemia cells carry this specific marker on them, it's called WT1, it's basically a flag. And galinpepimut-S is a vaccine that shows your immune system that exact flag — like, here, memorize this face — so your body learns to target those cells itself. You're not poisoning everything, you're programming the bouncer.

Juniper Vale: And the REGAL trial tested exactly that, in patients who'd relapsed at least once and couldn't get a transplant.

Hope Sterling: 24 months versus 14.7 months. That's what they found. The galinpepimut-S arm versus Best Available Treatment — 41% lower hazard of death.

Juniper Vale: Nine point three months. For someone sitting in that chair you described — that's not a statistic, that's a different kind of year entirely.

Hope Sterling: But wait — that chair, that person, is that actually who REGAL enrolled? Because I've been picturing like, the worst-case AML patient, and I'm not sure that's right.

Juniper Vale: That's exactly the thing I want to sit with. REGAL's population was transplant-ineligible, second or later remission — which sounds dire, and it is, but it's actually a relatively selected group. And the number that tells you that? The BAT arm. Best Available Treatment got 14.7 months median survival. That's not nothing.

Hope Sterling: Wait — 14.7 months is actually... decent?

Juniper Vale: For relapsed AML, yeah. It tells you these patients weren't in the most catastrophic situation. And then — I mean, compare that to what happens after azacitidine-venetoclax relapse. Aza-ven is the big frontline regimen right now, venetoclax is a BCL-2 inhibitor, and it's become almost standard. But when patients relapse after it? MD Anderson published data showing median survival of approximately two to three months — even with gilteritinib, even with the new menin inhibitors like revumenib or ziftomenib.

Hope Sterling: Two to three months. That's shorter than most people's maternity leave.

Juniper Vale: It is. And those post-aza-ven patients — that's the highest unmet-need subgroup in this whole disease. REGAL didn't specifically enroll them. The trial wasn't designed around post-venetoclax relapse.

Hope Sterling: So the trial may not even be testing the people who need it most?

Juniper Vale: That's the uncomfortable nuance, yeah. And I don't want to undersell what REGAL actually is — it's the first randomized Phase 3 evidence for a WT1-targeted cancer vaccine in AML. That's a genuine landmark. SELLAS did something nobody has done before in this disease. But landmark in a defined slice is different from breakthrough for the whole spectrum.

Hope Sterling: So galinpepimut-S might be the most exciting thing to happen in AML immunotherapy in years, and it still might not reach the patients dying in two months. That's — I don't even know what to do with that.

Juniper Vale: And that gap — between what the science proved and who actually gets it — that's exactly where the timeline becomes the real story. Because SELLAS announced those REGAL results on November 17th, 2026. And their plan is to initiate FDA BLA filing discussions in 2027. Not file. Initiate discussions about filing.

Hope Sterling: Wait — start talking about filing. That's the milestone.

Juniper Vale: That's the milestone. So once you begin those conversations, you've still got the full BLA review process, potential advisory committee meetings — realistically you're looking at approval decisions in 2028 or later. Maybe well into 2028.

Hope Sterling: Okay but — no, I mean, I need to sit with this — an oncology nurse in a community clinic, early 2027, has a newly relapsed patient in front of her. And she knows galinpepimut-S exists. She's read the data. And she has to say... you just can't have it. The trial worked, we know it worked, but — you just can't.

Juniper Vale: And SELLAS's public disclosures don't fully explain the gap. I want to be fair — CMC manufacturing scale-up, assembling a complete BLA package, that's genuinely hard work. But the specific reasons for this particular lag aren't spelled out. The sourcing here is thin. We don't have visibility into what they're managing internally.

Hope Sterling: So we're not saying bad faith.

Juniper Vale: Not saying bad faith. But I'm saying — the structural fact stands regardless. Relapsed AML. Median survival under six months for the broader relapsed and refractory population. And the regulatory clock doesn't move in emergency time. It just doesn't.

Hope Sterling: It's like — galinpepimut-S is a completely different category of thing than CAR-T, it's not a living cell product, it targets WT1 like a flag, it should be more manufacturable — and yet somehow we're still here, staring down 2028.

Juniper Vale: And honestly — the question of what patients would even be comparing it against by then is something we need to get into. Because revumenib, ziftomenib, those menin inhibitors got approved for relapsed AML during and after REGAL's enrollment, and that changes what 'best available treatment' even means now. Whether GPS's edge holds, shrinks, or looks additive — that's the uncomfortable math waiting for us.

Hope Sterling: Wait, so like — revumenib and ziftomenib got approved while REGAL was still running? That's the part I need to understand, because that means the people in the BAT arm might have been getting something we'd now consider outdated.

Juniper Vale: Yeah, and that's actually the uncomfortable math. Both menin inhibitors — revumenib for NPM1-mutated or KMT2A-rearranged patients, and ziftomenib, which is Komzifti, specifically for NPM1-mutated — both got FDA-approved for relapsed/refractory AML. During and after REGAL's enrollment window. So 'best available treatment' in 2022 is not the same as best available treatment right now.

Hope Sterling: So the 24 versus 14.7 comparison — does that get weaker?

Juniper Vale: Maybe. I mean — honestly, I don't know, and I want to be careful not to oversell either direction. If BAT today includes revumenib or ziftomenib, the control arm might perform better than 14.7 months. Which could mean GPS's 41% edge looks smaller against a stronger comparator. That's the uncomfortable read. But — wait, actually — there's a different way to frame it.

Hope Sterling: Okay, what's the other frame.

Juniper Vale: GPS targets WT1. The menin inhibitors target a completely different pathway — they're blocking the menin-KMT2A interaction. FLT3 inhibitors like gilteritinib are yet another mechanism entirely. These aren't competing for the same biological real estate. So in theory — and this is genuinely unknown, we don't have combination trial data — GPS could be additive on top of a menin inhibitor. A patient gets ziftomenib, achieves remission, then receives GPS as maintenance. That's at least biologically plausible.

Hope Sterling: Oh — okay that reframes it completely. It's not GPS versus Komzifti, it's maybe GPS plus Komzifti.

Juniper Vale: Possibly. And that's where the CAR-T contrast actually matters — CAR-T runs over $400,000 per patient, requires individualized manufacturing, the logistics are brutal. GPS is off-the-shelf. If combination regimens become the standard and you need something scalable that a community oncologist can actually access, that changes the math considerably. But I want to be honest — that scalability advantage is still hypothetical until SELLAS demonstrates manufacturing at volume.

Hope Sterling: So picture a transplant center pharmacist in, like, Tucson — not a major academic center — opening a GPS vial in 2029 instead of coordinating a CAR-T transfer. That's the actual version of 'accessible' we're talking about.

Juniper Vale: And that pharmacist in Tucson — that's the whole access question in one image. Because CAR-T is FDA-approved. It works. It costs over $400,000 per patient, and most community hospitals can't touch it. SELLAS is a small-cap company, SLS on NASDAQ — their manufacturing scale, their pricing strategy, none of that is in any public disclosure yet. So GPS being off-the-shelf doesn't automatically mean it gets to that pharmacist. Approval and access are genuinely not the same event.

Hope Sterling: And the patients REGAL enrolled — transplant-ineligible, second or later remission, older, often with other stuff going on — like, those are exactly the people least likely to be at a major academic center. They're in community hospitals. They're the ones who'd need it to be easy to get, and they're the ones for whom it probably won't be.

Juniper Vale: The trial proved something real. Whether that something reaches the right people — that's a different trial, and nobody's enrolled in it yet.

Hope Sterling: I keep thinking about the person I invented at the start. Late sixties, in the chair, doctor says there's something new. That image is — it's still true. The data is real. I just don't know when that person gets to actually be in the room where it's available.

Juniper Vale: Neither do I. Thank you for thinking through this one with me — it's a hard one to hold.