Stem-cell transplant beat standard consolidation in primary CNS lymphoma trial

For patients with primary central-nervous-system lymphoma who responded to initial chemotherapy, consolidating that response with high-dose chemotherapy and an autologous stem-cell transplant kept the cancer at bay longer than a gentler, non-transplant regimen, according to the MATRix/IELSG43 trial in The Lancet [s1]. Three-year progression-free survival was 78% with transplant versus 51% without, a hazard ratio of 0·43 [s1]. But the benefit came with more toxicity and more treatment-related deaths, a trade-off the trial documents plainly [s1].

The result carries weight partly because of how it was funded. This was an academic, cooperative-group trial run across European university hospitals and paid for largely by government research agencies and a cancer foundation, not by a drug company — the largest randomised trial in this rare and aggressive brain cancer to date [s1].

What the trial tested

Primary CNS lymphoma is a B-cell cancer confined to the brain, spinal cord, and eyes, treated first with intensive chemotherapy and then with a consolidation step meant to deepen and hold the remission [s1]. High-dose chemotherapy with an autologous stem-cell transplant — using a patient's own harvested stem cells to rescue the marrow after otherwise marrow-lethal drug doses — is one option, but whether it is worth its risks compared with standard chemoimmunotherapy had remained uncertain [s1].

MATRix/IELSG43 enrolled untreated, immunocompetent patients with B-cell primary CNS lymphoma across 56 university and academic hospitals in five European countries [s1]. All received four cycles of MATRix induction — rituximab, high-dose methotrexate, high-dose cytarabine, and thiotepa — and those reaching at least a partial response were randomly assigned 1:1 to two cycles of non-myeloablative R-DeVIC chemoimmunotherapy or to thiotepa-based high-dose chemotherapy with autologous transplant [s1]. The primary endpoint was progression-free survival [s1].

What it found

Between July 16, 2014, and Aug 31, 2019, 368 patients were enrolled; 346 (94%) started induction, and 230 who responded were randomly assigned, of whom 229 were analysed — 115 to R-DeVIC and 114 to transplant [s1]. After a median follow-up of 45·3 months, three-year progression-free survival was 78% (95% CI 69–85) in the transplant group and 51% (41–60) in the R-DeVIC group, with a hazard ratio of 0·43 (95% CI 0·27–0·68; p=0·0003) [s1]. The trial also reported superior overall survival with transplant [s1].

The cost side was real. The mean number of adverse events per patient was 9·3 with R-DeVIC and 14·6 with transplant [s1]. Fatal serious adverse events after consolidation occurred in two patients in the R-DeVIC group — both acute myeloid leukaemia — and in five patients in the transplant group, from infections (four) and pulmonary embolism (one); all but the embolism were judged possibly treatment-related [s1].

How to read it

This is a clear positive result on a hard endpoint, with an effect size — a 57% reduction in the hazard of progression or death — that is large and unlikely to be a statistical fluke given the confidence interval well below 1 [s1]. It should firm up transplant as the preferred consolidation for fit patients who tolerate induction [s1].

Two qualifications frame who that applies to. First, only patients who both completed MATRix induction and reached at least a partial response were randomised — 229 of the 368 enrolled — so the finding speaks to fit responders, not to every patient who walks in with this diagnosis [s1]. Second, the toxicity is not incidental: five treatment-related deaths in the transplant arm is the price of the progression-free survival gain, and it is the kind of number that belongs in any conversation about the choice [s1]. Deciding how aggressively to consolidate turns on fitness and on what a given cancer stage and response imply for the odds.

Why it matters

Because this trial was academic and independent, it could ask whether an intensive, toxic, but non-proprietary strategy is worth it — a question a company selling a single drug has little incentive to pose [s1]. For a rare cancer with few randomised trials, a well-powered head-to-head that measures both the benefit and its cost is exactly the evidence clinicians have lacked. It sits alongside other independent consolidation and transplant comparisons in blood cancer, where the harder question is usually not whether more treatment helps but whether its harms are acceptable.

What to watch

The open questions are which patients can safely be spared transplant — whether molecular or imaging markers of a deep remission could identify people who do just as well with the gentler regimen — and how the transplant's late effects on cognition and second cancers weigh against its progression-free survival advantage over longer follow-up [s1]. Confirmation in other health systems and populations would strengthen the case further [s1][s2].

This article describes research and is not medical advice. Decisions about treating primary CNS lymphoma are for patients and their treating clinicians.

Sources

• High-dose chemotherapy followed by autologous stem-cell transplantation versus non-myeloablative consolidation in primary CNS lymphoma (MATRix/IELSG43): a randomised phase 3 trial — The Lancet, 22 July 2026
• High-dose Chemotherapy and ASCT or Consolidating Conventional Chemotherapy in Primary CNS Lymphoma (MATRix, NCT02531841) — ClinicalTrials.gov, U.S. National Library of Medicine, last updated 25 August 2015