30 Gy vs 40 Gy consolidative RT, limited-stage DLBCL
ForStage I-II DLBCL, CR/CRu after anthracycline-based chemo ± rituximab
90.5% vs 87.2%
HR 0.87, 95% CI 0.59-1.28; one-sided P = 0.0013 for non-inferiority
TL;DR5-yr PFS 90.5% (30 Gy) vs 87.2% (40 Gy), HR 0.87 (0.59-1.28), non-inferiority met in stage I-II DLBCL in CR after chemo.
The local-regional read backs the PFS headline: locoregional recurrence did not differ between 30 Gy and 40 Gy involved-site/field RT, though no rates are given in the source text. Toxicity, including second malignancies, was also comparable at 7.2 years median f/u, so the late-effect gain from de-escalation is inferred, not shown.
In stage I-II DLBCL in CR/CRu after anthracycline-based chemo with or without rituximab, this supports 30 Gy over 40 Gy as the consolidative dose; it does not extend to pts short of CR, and it does not test omitting RT.
The local-regional read backs the PFS headline: locoregional recurrence did not differ between 30 Gy and 40 Gy involved-site/field RT, though no rates are given in the source text. Toxicity, including second malignancies, was also comparable at 7.2 years median f/u, so the late-effect gain from de-escalation is inferred, not shown.
12 details 1 trial watching
Multicenter, open-label, non-inferiority phase 3, randomized 1:1. 432 pts enrolled April 14, 2010 to July 1, 2019; median follow-up 7.2 years. Non-inferiority margin not reported in source.
Newly diagnosed stage I-II DLBCL achieving CR/CRu after anthracycline-based chemo with/without rituximab. Age criterion printed as '=15 years' in the portal text. Bulk, IPI and PET status at response not reported in source.
30 Gy vs 40 Gy consolidative involved-site/field RT. Fraction size, technique and the split between involved-site and involved-field volumes are not reported in source.
Primary: 5-year PFS, tested for non-inferiority of 30 Gy. Also reported: OS, locoregional recurrence, acute and late toxicity including second malignancies.
Primary endpoint met (table). OS and locoregional recurrence showed no difference between arms; the source gives no rates or HRs for either.
| Endpoint | 30 Gy | 40 Gy | Effect |
|---|---|---|---|
| Randomized | n=216 | n=216 | 1:1 |
| ITT analysis | 212 | 213 | n/a |
| 5-yr PFS (95% CI) | 90.5% (86.5-94.7) | 87.2% (82.7-91.9) | HR 0.87 (0.59-1.28), one-sided P = 0.0013 for non-inferiority |
Acute and late toxicities, including second malignancies, were comparable between arms. No rates or grades reported in source, so no measured toxicity advantage for 30 Gy.
The source states no prior randomized trial compared these two doses in DLBCL specifically. The UK NHL dose-reduction RCT (Lowry et al.) asked a similar de-escalation question in aggressive NHL more broadly, and ILROG guidance already favors the lower consolidative dose after CR, so this result is concordant with that direction.
Enrollment spans 2010 to 2019 with rituximab optional and response graded as CR/CRu, so systemic therapy and response assessment may not match current PET-adapted practice. Mixed involved-site/field volumes are not broken out. Abstract only: no event counts, no margin.
With PFS near 90% in both arms and locoregional recurrence no different, the higher dose bought no measurable disease control in pts already in CR. What the trial does not settle is whether the lower dose reduces late effects (not shown at 7.2 years) and whether any RT is needed after a PET-negative response.
CONSORT flow
Randomised phase 3, prespecified non-inferiority primary met at 7.2-year median f/u; source calls 30 Gy the new standard. Caveats: margin and fractionation unreported, rituximab optional.
- Can consolidative RT be omitted after PET-negative CR in limited-stage DLBCL?
- Does 30 Gy reduce late toxicity or second malignancies with longer follow-up?
- Is a dose below 30 Gy adequate after CR to rituximab-containing chemo? active Dose-Reduced Consolidation Radiation Therapy in Patients With Diffuse Large B-cell Lymphoma Phase NAn=243 · primary completion 2025-07 · 30 Gy to 20 Gy consolidation RT after PET-neg R-chemo
📚 Sources · 📄 1 paper
Abstract
The longer read
The question here is narrow and well posed: once a pt with stage I-II DLBCL is in complete remission after anthracycline-based chemotherapy, does the consolidative dose need to be 40 Gy, or does 30 Gy do the same job. The trial answers it with a randomized comparison, a prespecified non-inferiority primary, and a median follow-up of 7.2 years, long enough that most DLBCL relapses would have declared themselves. The 5-year PFS point estimate actually sits on the side of the lower dose (90.5% vs 87.2%), and the HR of 0.87 has an upper confidence bound of 1.28. That bound is the number to hold onto, since the abstract does not print the non-inferiority margin. A reader cannot check how generous the margin was, only that the one-sided P of 0.0013 cleared it.
The direction is unsurprising. The UK randomized dose-reduction trial in NHL (Lowry et al.) already pointed toward lower doses being adequate in aggressive histologies, and ILROG guidance has favored the lower consolidative dose after CR for years. What this trial adds is a DLBCL-specific, CR-restricted randomized dataset, which the authors say did not previously exist. In centers that already use the lower dose it reads as confirmation with better evidence behind it; in centers still prescribing 40 Gy or above (the abstract cites doses from 30 to 55 Gy in use) it is the reason to stop.
Several design features should temper how far the result is stretched. Enrollment ran from 2010 to 2019 and rituximab was optional, so some fraction of the cohort received systemic therapy weaker than current standard. If anything that makes the finding more reassuring, since RT was carrying more of the load in those pts and the lower dose still held. Response is described as CR/CRu, a category from CT-era criteria, and the abstract does not say whether PET defined remission. A CT-defined CRu may include pts who would be PET-positive today, and the trial says nothing about what dose residual metabolically active disease needs. The target is given as involved-site/field, which implies a shift in volume definition during accrual; outcomes by volume are not reported.
The weakest part of the abstract is the toxicity claim. The stated rationale for de-escalation is less long-term harm, yet acute and late toxicities and second malignancies were comparable between arms. At 7.2 years that is too early for radiation-induced solid tumors or cardiac events to separate, so the benefit of the lower dose rests on dosimetric logic and not on trial data. Open-label design matters little for PFS in this setting, but no rates are given for OS, locoregional recurrence or any toxicity, so the secondary claims cannot be weighed.
The decision this moves is dose, not indication. It does not test RT omission after a PET-negative response, which is the live controversy in limited-stage DLBCL, and it does not inform doses below 30 Gy.