onc brain

Single-arm multi-stage feasibility and safety trial of durvalumab added to 5-fluorouracil/mitomycin C chemoradiotherapy in muscle-invasive bladder cancer. N=55 enrolled, 54 treated. A stage 1 expansion opened to node-positive pts, the first 6 with N+ disease.

Median age 69 (IQR 62-76), 45 (82%) male. cT2 in 44 (80%), cT3 in 11 (20%), N+ in 7 (13%) (N1 4, N2 3). 41 (75%) had prior neoadjuvant chemotherapy, so this is largely a post-NAC bladder-preservation population.

Durvalumab before, during and for 12 months after chemoRT, with 5-fluorouracil and mitomycin C as the radiosensitising backbone. 33/54 (61%) completed planned treatment; 21/54 (39%) discontinued early.

55Gy in 20 fractions to bladder; node-positive pts received 46Gy in 20 fractions to nodes alongside the bladder dose. This is the standard UK hypofractionated schedule, not an escalated or de-escalated variant.

Primary: disease-free survival rate at 12 months post chemoRT, read against a prespecified GO/NO-GO framework (GO ≥75%, contextual 60-75%, NO-GO <60%).

40/50 (80%) disease free at 12 months, 95% CI 0.67 to 0.89, clearing the GO threshold. 2 pts pending and 3 not evaluable. Investigators report very high bladder preservation and survival versus their previous trial data; those comparator figures are not given in the source.

The benchmark is the team's own BC2001 trial (James, JCO 2016), whose chemotherapy randomisation gave DFS HR 0.78 (0.60-1.02), stratified logrank p=0.07 on the slide shown. Comparison is to that historic control experience, not to a contemporaneous arm.

A 12-month DFS read is short for a preservation strategy where salvage cystectomy and late nodal relapse both fall outside the window. The evaluable denominator dropped from 55 to 50, and with 2 still pending the point estimate can move relative to a 75% threshold it clears by 5 points.

This clears a feasibility gate, not an efficacy one: the design question it answers is whether a randomised trial of durvalumab plus chemoRT is worth running, and the answer is yes. The 39% early discontinuation is the finding that should shape that trial, since a 12-month adjuvant tail that four in ten pts do not finish may not be the schedule worth randomising.

Single-arm feasibility/safety trial, N=55, 12-mo endpoint benchmarked against historic in-house CRT data. No randomised comparator; hard maturity gate applies.

Alliance single-arm phase 2, N=75 planned, 78 eligible enrolled 07/2022-03/2024. Median follow-up 26.9 months. Monitoring PSA and testosterone q3mo, scans at least q6mo (q3mo with rising PSA), QOL q6mo.

mHSPC on ADT + ARPI with PSA <0.2 stable or falling after 18-24 months of ADT and at least 12 months of ARPI. Median age 70 (49-90), 84.6% White, 64.9% low volume by CHAARTED, 35.1% high volume. Median PSA prior to starting ADT+ARPI 18.99 ng/mL.

Not an intervention, but the cohort was RT-heavy at baseline: 40 (51.3%) had radiation as local therapy, 31 (39.7%) had no local therapy, 7 (9.0%) prostatectomy +/- RT. 23 (29.5%) had radiation to metastatic sites. Dose and fractionation not reported in source.

Primary: treatment-free with eugonadal testosterone (T ≥150 ng/dL) at 18 months. Secondary: time to eugonadal T, duration off treatment, QOL. Exploratory: rPFS, TTNT, OS (CSS/NCSS), cost. Re-initiation triggered by PSA ≥5 ng/mL, PCWG3 radiographic change, or prostate-cancer-related symptoms.

Primary endpoint met: 32/78 (41.0%), 80% CI 33.1-48.9%, one-sided p 0.0249. Testosterone recovered in 52 (66.7%) by 18 months with median time to eugonadal T 9.0 months (95% CI 8.4-12.4); 45 (57.7%) stayed treatment-free for 18 months. rPFS 15/78 events, median NE; OS 4/78 events, median NE.

The 18-month primary readout is short for a de-intensification question whose real risk is late: 4 of 29 pts who resumed per protocol later progressed and discontinued the original ARPI, so re-treatment did not uniformly restore control. Deaths (4) included non-prostate causes (myelodysplastic syndrome, influenza, myocardial infarction), and with no continuous-treatment arm the OS curve carries no comparative information.

Intermittent ADT has been tested before in metastatic disease (SWOG 9346 could not exclude an OS decrement with intermittent therapy in mHSPC), but A-DREAM asks the question in the modern ARPI-intensified era with a molecularly deeper response gate, which is why the cohort's off-treatment rate looks better than the older intermittent literature suggests. That comparison is a rationale, not a result: A-DREAM has no randomised arm.

The number that matters clinically is not 41% but its two components: testosterone recovery is the rate limiter (66.7% recovered, median 9.0 months), while disease control off treatment was more common (57.7% treatment-free at 18 months). A trial that de-intensifies successfully on disease grounds can still miss its endpoint on gonadal recovery in an older cohort, and that distinction determines who to counsel.

Single-arm phase 2, no continuous-treatment control, 18mo primary readout in a deeply selected responder cohort. Interruption safety needs a randomised comparator.

Single-arm expansion cohort of the CHRYSALIS-2 program, n=49, 1L atypical EGFR-mutated advanced NSCLC. Median follow-up 31.3 mo. No randomised comparator arm.

Treatment-naive advanced NSCLC harbouring an atypical (uncommon) EGFR mutation. Baseline strata shown on the swimmer plot include age ≥65y, Asian ancestry and CNS involvement; per-stratum counts are not reported in source.

IV amivantamab (EGFR/MET bispecific) plus lazertinib (third-generation EGFR TKI). No chemotherapy backbone. Median duration of treatment 13.3 months (range <0.1-53.2), with 39% of 1L participants still on treatment beyond 2 years.

Median OS 41.0 mo (95% CI 27.7-NE), described by the presenters as roughly 3.5 years. The upper CI bound is not estimable, so the point estimate is anchored on the lower bound of 27.7 mo.

Reported only as consistent with prior reports, no new safety signals with longer follow-up. No grade 3+ rates, infusion-reaction rates or dermatologic AE rates appear in the source.

The atypical EGFR label pools biologically distinct alterations (G719X, S768I, L861Q and compound variants) whose single-agent TKI sensitivity differs; n=49 cannot resolve per-variant benefit, and the curator's read of no variant-outcome association is an absence of signal in a small cohort, not evidence of uniformity. No comparator, so the 41.0 mo median cannot be positioned against afatinib or osimertinib series in the same population.

Atypical EGFR is the part of the EGFR-mutant space where no regimen is settled, so a 41.0 mo median in 49 pts is meaningful as a benchmark even without randomisation. The practical question this leaves open is whether the bispecific's added toxicity and IV schedule are justified over an oral TKI alone, which this design cannot answer.

Single-arm n=49 expansion cohort, no randomised comparator against afatinib or osimertinib in atypical EGFR. Maturity gate: single-arm never reaches confirmatory.

Randomised phase II, N=224, asandeutertinib (n=111) vs osimertinib (n=113). Endpoints read by BICR. Follow-up duration, stratification factors and alpha allocation are not reported in the source slide.

First-line EGFR-mutated NSCLC with brain metastases. The source does not state lesion number, size, symptom status, or whether prior CNS radiotherapy was permitted, which is the eligibility gate an RT reader needs.

Intracranial ORR by BICR is the headline, with intracranial PFS and overall PFS reported alongside. No overall survival data in source.

Any TRAE 99.1% vs 95.6%; serious TRAE 10.8% vs 7.1%. The excess serious-event rate is small in absolute terms but sits on a phase II denominator, and no organ-level breakdown is given in source.

The intracranial separation (HR 0.46) is larger than the systemic one (HR 0.64), which points at CNS penetration rather than a general potency gain as the mechanism. For radiation oncology the question is whether that CNS margin is durable enough to defer SRS in asymptomatic pts; a phase II with unreached medians and no OS cannot answer it.

Phase II, conference-slide source only; no OS, borderline overall-PFS p, and no reported CNS-RT or prior-RT stratification. Needs phase III before displacing osimertinib.

Randomized, multicenter, open-label phase 3 (NCT06448312), 1:1, N=413. Stratified by histology (squamous vs non-squamous), PD-L1 TPS (1-49% vs ≥50%), and ECOG (0 vs 1). Median follow-up 10.5 months at this analysis.

Locally advanced stage IIIB/IIIC or metastatic stage IV NSCLC with no prior systemic antitumor therapy. Required PD-L1 TPS ≥1% by IHC 22C3 central lab, no sensitizing EGFR or ALK alteration, ECOG 0 or 1.

Sac-TMT 4 mg/kg Q2W plus pembrolizumab 400 mg versus pembrolizumab 400 mg Q6W alone, until progression or unacceptable toxicity. Pembro was capped at a maximum of 18 cycles in both arms. One pt in the pembro group never received assigned treatment.

Primary: PFS by BICR. Key secondary: OS. Other secondary: investigator-assessed PFS, ORR, DCR, DOR, safety. The updated efficacy boundary at the actual 194 PFS events was 0.0174 (2-sided).

PFS crossed its boundary at p<0.0001. OS was descriptive at the PFS interim, HR 0.55 (0.36, 0.85) with both medians not reached and only 33 vs 54 deaths.

The pembro-alone control performed as expected for an all-comers TPS ≥1% population, with the TPS ≥50% arm reaching 9.5 mo and the TPS 1-49% arm 4.3 mo, consistent with the KEYNOTE-042 signal that low-expressors gain least from single-agent IO. This is the setting where chemo-IO has historically been chosen, so the trial tests whether an ADC can occupy that slot instead of platinum doublet chemotherapy.

No toxicity data in the source, which is the decisive missing piece for a doublet whose entire premise is avoiding chemotherapy. The comparator is pembro monotherapy rather than the chemo-IO most oncologists actually give at TPS 1-49%, so the PFS gap partly measures a weak control rather than a strong experimental arm.

An HR of 0.35 with a median not reached is a large PFS effect, and the OS point estimate moves in the same direction, but neither settles whether sac-TMT plus pembro beats the regimen it would actually replace. The larger effect in TPS 1-49% (HR 0.28) than TPS ≥50% (HR 0.47) is the expected pattern when the control arm is weakest in the low-expressors, not evidence of a biomarker-selected ADC effect.

PFS interim of an open-label phase 3; OS descriptive at 10.5 mo f/u with both medians NR. No safety data in source, and no comparison vs pembro + chemo.