onc brain

About · curated by Nick Boehling, MD · @nb2276

2026-05-29 ASCO Annual Meeting 2026

digest generated 2026-08-11

Wait or Treat (NCT05236946): upfront cranial RT cut intracranial progression (sub-HR 0.35, 0.21-0.59) but 2y OS was 48% vs 60% for delayed RT.
Only CNS data today, and it cuts against reflexive upfront cranial RT in asymptomatic BM from EGFR/ALK+ NSCLC: a real intracranial control gain (2y 21.7% vs 50% progression) with OS numerically worse (HR 1.45, per attendee report). Timing, not omission, and dose/technique/target volume are absent from source.

CNS

Challenges SOC

Wait or Treat (NCT05236946) NCT05236946

ForMetastatic EGFR/ALK+ NSCLC, asymptomatic measurable brain mets, ECOG 0-2

Intracranial progression-free survival local control

sub-HR 0.35

95% CI 0.21-0.59, p<0.001; 2y icPD 21.7% vs 50%

TL;DRUpfront cranial RT cut intracranial progression (sub-HR 0.35, 0.21-0.59, p<0.001) but 2y OS favored delayed RT, 48% vs 60%.

Why it mattersRadiation oncology

The intracranial win is real (2y progression 21.7% vs 50%, sub-HR 0.35) yet does not convert: 2y OS 48% upfront vs 60% delayed, HR 1.45. With necrosis reported ~6% upfront vs none delayed, and RT dose/technique unstated in source, this moves the timing decision toward deferral with MRI q3m surveillance.

Monday clinic

In treatment-naive EGFR or ALK-driven metastatic NSCLC with asymptomatic measurable brain mets starting a TKI, this supports deferring cranial RT with q3m MRI rather than treating reflexively; it says nothing about symptomatic mets, large or dominant lesions, or oncogene-negative disease.

The longer read
Wait or Treat (NCT05236946)
+2 more figures
Trial schema, NCT05236946. R 1:1, upfront cranial RT (n=105) vs delayed cranial RT (n=103).
Trial schema, NCT05236946. R 1:1, upfront cranial RT (n=105) vs delayed cranial RT (n=103).
Wait or Treat (NCT05236946)
10 details

Phase III open-label RCT, single-centre (Tata Memorial, Mumbai), N=208 randomised 1:1 to upfront (n=105) vs delayed cranial RT, both on TKI plus chemotherapy. Stratified by GPA (0-2 vs >2) and synchronous vs metachronous BM. Median follow-up 30.6 mo (28.7, 36).

Metastatic NSCLC with an EGFR or ALK alteration, ECOG PS 0-2, radiologically measurable brain metastases that were completely asymptomatic. Number, size, and location of lesions are not reported in source.

Upfront arm received cranial RT at diagnosis; the delayed arm received it at intracranial progression or patient's wish, so both arms are RT-exposed and the question is timing, not omission. Dose, fractionation, and technique (WBRT vs SRS) are not reported in source, which is the main barrier to transferring this result.

Primary: intracranial PFS. Secondary: OS, PFS, toxicity, ORR, neurocognition, PROM. Surveillance was MRI brain q3m for the first year, then q6m, which is what makes a delayed strategy safe to run.

Primary endpoint met. Survival ran the other way: 2y OS 48% upfront vs 60% delayed, OS HR 1.45, reported by attendees rather than captured in the slide OCR.

TimepointUpfront RT (n=105)Delayed RT (n=103)
1-year8.7% (2.9%, 14.5%)25.7% (16.8%, 34.7%)
2-years21.7% (12.6%, 30.8%)50% (39.2%, 60.9%)
Sub-HR (95% CI)0.35 (0.21, 0.59), p<0.001ref

Radiation necrosis ~6% in the upfront arm and none in the delayed arm per attendee reports, and described as less severe when delayed. Full toxicity tables, neurocognition, and PROM were secondary endpoints not reported in the source.

asymptomatic, radiologically measurable brain mets in EGFR/ALK-driven metastatic NSCLC on TKI plus chemotherapy with q3m MRI available
Does not represent symptomatic brain mets, oncogene-negative NSCLC, or settings without reliable serial MRI surveillance.

The intracranial magnitude sits alongside the older WBRT-era data (QUARTZ, and the historic SRS-vs-WBRT trials) in showing that cranial RT controls the brain without buying survival. What is new is testing it where a CNS-penetrant TKI is the competing intracranial therapy, a setting those trials predate entirely.

Single-centre and open-label, and the RT prescription is absent from the source, so a reader cannot tell whether the necrosis signal reflects WBRT, SRS technique, or concurrent TKI exposure. Median follow-up of 30.6 mo is short relative to expected survival in this population, and the OS comparison was a secondary endpoint, not powered.

The result splits the two things RT is usually credited with: it clearly buys intracranial control, and it clearly does not buy time. Whether the OS direction is a real cost of upfront RT or noise in an underpowered secondary is the open question, and it decides whether deferral is merely non-inferior or actually preferred.

CONSORT flow
Randomized 208
Upfront cranial RT + TKI/chemo
allocated 105
2y icPD 21.7%
Delayed cranial RT + TKI/chemo
allocated 103
2y icPD 50%

Randomised phase 3, prespecified intracranial PFS met, but the survival signal runs opposite the intracranial win. Directly contests reflex upfront cranial RT.

  • Does the OS direction hold with longer follow-up?
  • Was cranial RT whole-brain or stereotactic, at what dose?
  • Neurocognition and PROM outcomes by RT timing

Sourced from @OncLive, @DrRiyazShah, @StephenVLiu

📚 Sources · 🐦 3 tweets