onc brain

About · curated by Nick Boehling, MD · @nb2276

    ASCO Annual Meeting 2026

    19 studies Subscribe via RSS →

    DeLLphi-304 (Tarlatamab CNS Outcomes)

    ForRelapsed SCLC, ≥1 brain metastasis, post-platinum (2L)

    TL;DRCNS PFS HR 0.54 ITT and 0.40 in brain-met subgroup; CNS CR 15% vs 5% favoring tarlatamab over chemo in 2L SCLC.

    vs leading data
    • RT relevance: intracranial control here is salvage atop prior brain RT, not evidence for upfront RT omission
    • Parent DeLLphi-304 established OS benefit vs chemo in 2L SCLC; this extends the signal intracranially

    Systemic Palliative Phase 3 RCT Caveats dominate ASCO Annual Meeting 2026

    DeLLphi-304 (Tarlatamab CNS Outcomes)
    ArmMedian CNS PFS, moHR (95% CI)
    Tarlatamab (n=254)NE (13.7, NE)0.54 (0.39, 0.75)
    Chemotherapy (n=255)7.2 (5.6, NE)n/a
    +2 more figures
    DeLLphi-304 (Tarlatamab CNS Outcomes)
    ArmMedian CNS PFS, moHR (95% CI)
    Tarlatamab (n=67)6.5 (4.3, 13.7)0.40 (0.24, 0.66)
    Chemotherapy (n=56)4.2 (2.9, 5.5)n/a
    DeLLphi-304 (Tarlatamab CNS Outcomes)
    EndpointTarlatamab (n=67)Chemo (n=56)
    Best CNS CR14.9% (10)5.4% (3)
    CNS DCR77.6% (52)71.4% (40)
    Median dur. CNS DC, mo8.25.2
    7 details 2 trials watching
    Methods
    • 🔍 Post hoc intracranial-efficacy analysis of DeLLphi-304 (randomized phase 3, tarlatamab vs chemo, 2L SCLC)
    • 🔍 Two readouts, different methods: ITT CNS PFS by RECIST/investigator; brain-met subgroup by mRANO-BM/BICR
    • 🔍 >70% of brain-met pts had prior CNS-directed Rx (likely prior brain RT)
    • 🔍 Data cutoff Jan 29 2025; median f/u 11.4mo (tarlatamab) / 11.5mo (chemo)
    Results
    • 📊 Other brain-met CNS signals (not in the tables)
      • CNS tumor shrinkage ≥30%: 56.3% (9/16) vs 38.5% (5/13)
      • Ongoing CNS response at cutoff: 5 (50%) vs 0
      • Median duration of CNS CR: NE vs 3.6mo
    Critique
    • ⚠️ Post-hoc endpoints, not pre-specified primary; brain-met subgroup HR from unstratified Cox, hypothesis-generating
    • ⚠️ ITT median CNS PFS not estimable for tarlatamab; CNS events concentrated in the brain-met subgroup (67 vs 56)
    Open questions
    • Does CNS activity let brain RT be deferred or sequenced later in SCLC?
      not yet Tarlatamab for SCLC Brain Metastases Phase 2
      n=35 · primary completion 2029-02 · single-arm intracranial efficacy in active brain mets
    • Durability of intracranial responses with longer follow-up
      not yet Tarlatamab for SCLC Brain Metastases Phase 2
      n=35 · primary completion 2029-02 · phase 2 intracranial efficacy, primary f/u 2029
    • Activity in CNS-naive vs previously irradiated brain mets
    📚 Sources · 🐦 1 tweet

    ctDNA in Nonoperative Management (Rectal Cancer)

    ForStage I-III MSS rectal adenoca in nonoperative mgmt after cCR/nCR

    TL;DRLocal-regrowth sensitivity only 41%; a negative ctDNA can't exclude regrowth, so it can't replace endoscopy/MRI surveillance in watch-and-wait.

    Curative Retrospective Early signal ASCO Annual Meeting 2026

    ctDNA in Nonoperative Management (Rectal Cancer)
    9 details 2 trials watching
    Methods
    • 🔍 Retrospective single-center INTERCEPT cohort: n=110 MSS stage I-III rectal adenoca in NOM after cCR/nCR; tumor-informed Signatera ctDNA
    • 🔍 NAT was TNT in 104/110 pts; median f/u 25mo
    Results
    • 📊 Events on NOM: local regrowth 23/110 (21%), distant mets 12/110 (11%)
    • 📊 Per-sample diagnostic accuracy of ctDNA (n=669 draws)
      MetricLocal regrowthDistant mets
      Sensitivity41% (12/29)74% (31/42)
      Specificity94% (480/509)97% (611/627)
      Accuracy91% (492/538)96% (642/669)
    • 📊 Ever-positive longitudinal ctDNA → worse 2yr survival vs persistently negative
      • Local regrowth-free: log-rank p=0.0002
      • Distant metastasis-free: log-rank p<0.0001
    • 📊 First post-NAT ctDNA positivity (≤180d) also predicted worse RFS (p=0.0006) / MFS (p<0.0001); evaluable subset tiny (n=12)
    • 📊 At salvage surgery, ctDNA-positive regrowths showed more advanced ypT3-4: 75% (6/8) vs 21% (3/14), p=0.01
    Critique
    • ⚠️ Low local-regrowth sensitivity means a negative ctDNA can't exclude regrowth — endoscopy/MRI surveillance stays mandatory in organ preservation
    • ⚠️ No external validation cohort; single-institution, retrospective — prognostic association, hypothesis-generating for surveillance algorithms
    Open questions
    📚 Sources · 🐦 1 tweet

    ARACOG (AFT-47)

    ForAdvanced prostate cancer (mHSPC, nmCRPC, mCRPC) on AR pathway inhibitor

    TL;DRWorst-domain cognitive decline (MCCD) -15.8% (DAR) vs -36.1% (ENZ) at 24wk, p=0.009; darolutamide spared cognition vs enzalutamide.

    vs leading data
    • Directionally consistent with darolutamide's lower blood-brain-barrier penetration vs enzalutamide

    Systemic Palliative Phase 2 trial Confirmatory ASCO Annual Meeting 2026

    ARACOG (AFT-47)
    ArmMCCD domainMedian changep
    Darolutamide (N=48)PALFAM-15.8P=0.009
    Enzalutamide (N=47)SWM-36.1P=0.009
    +1 more figure
    ARACOG (AFT-47)
    6 details 3 trials watching
    Methods
    • 🔍 Randomized open-label phase II, N=111; advanced PC across mHSPC, nmCRPC, mCRPC, stratified by age (<65 / 65-80 / >80)
    • 🔍 1° EP: % change baseline→24wk in Maximally Changed Cognitive Domain (MCCD), 5 remote CANTAB modules
    Critique
    • ⚠️ Open-label: pts self-perform CANTAB tasks unblinded; expectation/effort bias not excluded
    • ⚠️ MCCD is each arm's maximally-changed domain (derived metric, different domain per arm), not a validated clinical cognition endpoint
    • ⚠️ Small N (~48 vs 47 analyzed), single 24-wk readout; durability of the gap untested
    • ⚠️ Crossover permitted; crossover pts scored at time of crossover within randomized arm
    Open questions
    📚 Sources · 🐦 2 tweets

    A-DREAM

    FormHSPC, deep responders (PSA<0.2) after ≥18mo ADT + ≥12mo ARPI

    TL;DR41% treatment-free with eugonadal testosterone recovery at 18mo (1° EP met, one-sided p=0.0249) after interrupting ADT+ARPI in deep-responding mHSPC.

    vs leading data
    • Tests full ADT+ARPI interruption in the ARPI era, unlike legacy intermittent-ADT trials; no randomized OS comparator here

    Systemic Palliative Phase 2 trial Early signal ASCO Annual Meeting 2026

    A-DREAM
    Milestone at 18mon (%)
    Eugonadal T recovery52 (66.7%)
    Treatment-free45 (57.7%)
    Treatment-free + T recovery (1° EP)32 (41.0%)
    +2 more figures
    A-DREAM
    EndpointEvents/TotalMedian
    rPFS15/78NE
    OS4/78NE
    A-DREAM
    8 details 3 trials watching
    Methods
    • 🔍 Single-arm phase 2 (ALLIANCE A-DREAM); N=78 eligible, enrolled 07/2022-03/2024
    • 🔍 Eligible: mHSPC (64.9% low-volume), PSA<0.2 after 18-24mo ADT + ≥12mo ARPI; interrupt both
    • 🔍 RT-exposed cohort (applicability)
      • Prior local RT to prostate: 40 (51.3%)
      • Metastasis-directed RT: 23 (29.5%)
      • RT as non-protocol salvage during interruption: 4 (5.1%)
    Results
    • 📐 1° EP statistically met: 80% CI 33.1-48.9%, one-sided p 0.0249
    • 📊 Time to eugonadal T: median 9.0mo (95% CI 8.4-12.4); 67.9% recovered
    • 📊 Disposition at 26.9mo median f/u
      • Still off treatment: 30 (38.5%)
      • Resumed ADT+ARPI per protocol: 29 (37.2%)
      • Resumed before meeting criteria: 5 (6.4%)
      • Started non-protocol treatment: 7 (9.0%)
      • Death before next treatment: 1 (1.3%, MI)
    Critique
    • ⚠️ 4 of 29 who resumed ADT+ARPI per protocol later progressed (2 rPD, 1 PDu, 1 PET), stopped original ARPI
    • ⚠️ Single-arm, no continuous-therapy control; durability of 41% off-tx beyond 2yr and survival impact unknown
    Open questions
    📚 Sources · 🐦 1 tweet

    NRG Clinico-Transcriptomic Risk Stratification (Abstract 5000)

    ForNCCN high-risk/very-high-risk localized prostate cancer

    TL;DR22-gene genomic classifier reclassifies ~¼ of NCCN ≥HR pts, refining who gets abiraterone intensification atop definitive RT+ADT in localized high-risk prostate.

    vs leading data
    • AAP-benefit magnitude is imported from STAMPEDE M0 (Attard 2022); not prospectively tested by GC subgroup here

    Curative Retrospective Caveats dominate ASCO Annual Meeting 2026

    NRG Clinico-Transcriptomic Risk Stratification (Abstract 5000)
    EndpointHR (95% CI)p
    MFS0.53 (0.44-0.64)<0.0001
    OS0.60 (0.48-0.73)<0.0001
    +2 more figures
    NRG Clinico-Transcriptomic Risk Stratification (Abstract 5000)
    NRG Clinico-Transcriptomic Risk Stratification (Abstract 5000)
    8 details 2 trials watching
    Methods
    • 🔍 Combines NCCN clinical risk with a 22-gene genomic classifier (GC) into one clinico-transcriptomic (CT) risk score
    • 💊 Proposed algorithm: CT HR (≤2 points) → RT+ADT; CT VHR (≥3 points) → RT+ADT + abiraterone
    • 🔍 GC bins: <0.6 lower, 0.6-0.85 HR, >0.85 VHR; extends prior VHR work (Spratt JCO 2018)
    Results
    • 📊 GC independently improves prognostic estimates beyond clinical variables for MFS, DM, and OS (p<0.001)
    • 📊 Clinical vs 22-gene biomarker risk concordance (% of NCCN ≥HR pts)
      Biomarker lowerBiomarker higher
      Clinical lower49%15%
      Clinical higher9%27%
    • 📊 Discordant pts are the decision target: biomarker can upstage a clinically-lower pt or downstage a clinically-higher one
    Critique
    • ⚠️ Prognostic ≠ predictive: GC stratifies risk but doesn't prove the high-GC group differentially benefits from added AAP
    • ⚠️ Classifier development is retrospective; GC-guided intensification not yet tested in a randomized trial
    Open questions
    📚 Sources · 🐦 2 tweets

    GU Precision Oncology Biomarkers Discussion (Lang)

    TL;DREducation session on genomic biomarkers in prostate cancer; Decipher vs clinical-risk discordant in 24% of high-risk pts, with pre-ARPI validation caveats.

    Trials discussed

    ENZAMETCHAARTED

    vs leading data
    • Radonc note: Decipher's validated role is localized RT/post-RP staging, not these mHSPC intensification reads

    ASCO Annual Meeting 2026

    GU Precision Oncology Biomarkers Discussion (Lang)
    +1 more figure
    GU Precision Oncology Biomarkers Discussion (Lang)
    8 details 3 trials watching
    Methods
    • 🔍 ASCO26 GU education session (J. Lang): how to apply genomic biomarkers in prostate cancer practice
    • 💊 2026 landscape framing: tissue-panel testing for HRR + tumor-suppressor loss is SOC; 'genomics is not genetics'
    Results
    • 📊 Decipher vs clinical risk discordant in 24% of events in ≥high-risk pts
    • 📊 ENZAMET: Decipher ≤0.85 pts had worse OS with ADT+enza+docetaxel vs ADT+enza
    Critique
    • ⚠️ Discordance cohort had tissue in only 427 of 3816; representative of the full population?
    • ⚠️ Those validation trials predate ARPIs; transfer to current ADT+ARPI SOC unproven
    • ⚠️ But docetaxel arm enriched for high-volume disease + older pts → selection bias, not a biomarker-negative signal
    Notes
    • 🔬 ENZAMET Decipher biomarker-cohort attrition
      • N=1,125 enrolled
      • 1,071 consented for biomarker analyses
      • 764 had gene-expression profiling (GEP)
      • 634 final biomarker cohort
      • 320 in ADT+ENZA+docetaxel cohort
    Open questions
    📚 Sources · 🐦 1 tweet

    ENZAMET + Decipher

    FormHSPC on ADT+enzalutamide, Decipher classifier-tested

    TL;DROS HR 0.75 for docetaxel triplet vs doublet in Decipher >0.85 (interaction p=0.04); docetaxel avoidable if Decipher ≤0.85 in mHSPC.

    vs leading data
    • Consistent with CHAARTED / STAMPEDE: docetaxel OS benefit concentrated in higher-risk disease; here genomics, not just metastatic volume, flags benefiters

    Systemic Palliative Phase 3 RCT Caveats dominate ASCO Annual Meeting 2026

    ENZAMET + Decipher
    Decipher subgroupDoce vs no-doce OS HR, unweightedDoce vs no-doce OS HR, IPTW
    ≤0.852.78 (1.49-5.21), p=0.0011.94 (0.95-3.96), p=0.07
    >0.851.13 (0.71-1.79), p=0.600.75 (0.43-1.33), p=0.33
    Interaction p0.020.04
    +1 more figure
    ENZAMET + Decipher
    7 details 2 trials watching
    Methods
    • 🔍 ENZAMET biomarker substudy; Decipher (DPMC) genomic classifier dichotomized at >0.85; docetaxel-evaluable cohort N=320
    Results
    • 📊 Decipher >0.85 marks a poor-prognosis subset on ADT+enza doublet; adding docetaxel appears to neutralize that genomic disadvantage
    • 📊 Clinical read: Decipher >0.85 → consider docetaxel intensification; ≤0.85 → docetaxel likely avoidable
    • 📊 Low-volume subgroup: triplet directionally favored in high-Decipher pts, but N small and CI uninformative; no reliable effect size in source
    Critique
    • ⚠️ Docetaxel use was not randomized in ENZAMET (clinician/patient choice); doce-vs-no-doce compared by propensity-score / IPTW weighting
    • ⚠️ Authors note residual imbalance after propensity scoring; the docetaxel-by-Decipher interaction is exploratory, not confirmatory
    • ⚠️ High-Decipher triplet benefit is directional: the docetaxel HR's 95% CI crosses 1; significance rests on the interaction term, not the subgroup HR
    Open questions
    📚 Sources · 🐦 1 tweet

    TALAPRO-3

    ForHRR-deficient metastatic prostate cancer, first-line, enza+ADT backbone

    TL;DR3-yr rPFS 77% vs 56% (HR 0.48, 0.36-0.65, p<0.001) adding talazoparib to enza+ADT in HRR-deficient metastatic prostate cancer.

    vs leading data
    • vs TALAPRO-2 (1L mCRPC): same tala+enza combo, moved one line earlier

    Systemic Palliative Phase 3 RCT Practice-changing ASCO Annual Meeting 2026

    TALAPRO-3
    SubgrouprPFS HR (95% CI)Placebo median rPFS
    ITT0.48 (0.36-0.65)45.8 mo
    BRCA0.37 (0.22-0.61)35.1 mo
    Non-BRCA0.57 (0.39-0.82)NC (40.5-NC)
    8 details 4 trials watching
    Methods
    • 🔍 Phase 3 double-blind: talazoparib vs placebo added to enza+ADT; HRR-deficiency required for entry (~300 pts/arm)
    • 🔍 Castration-sensitive setting (enza+ADT backbone, long control rPFS) — extends PARP+ARSI from mCRPC into 1L mHSPC
    CONSORT flow
    Randomized 599
    ↓
    Talazoparib + enzalutamide
    allocated 300
    Placebo + enzalutamide
    allocated 299
    Results
    • 📊 1° EP rPFS, 3-yr landmark (ITT): 77% (67-85) vs 56% (50-62)
    • 📐 ITT events 67/300 (tala) vs 126/299 (placebo); tala median rPFS not reached
    • 📊 Non-BRCA subgroup 3-yr rPFS 76% (69-82) vs 60% (52-67)
    Critique
    • ⚠️ Benefit deepest in BRCA-altered, attenuated in non-BRCA; HRR-wide result driven by the BRCA subset
    • ⚠️ rPFS is a surrogate; OS not reported in source, so survival benefit remains unproven
    • ⚠️ Applies only to HRR-deficient pts; requires upfront HRR/BRCA testing to act on
    Open questions
    📚 Sources · 🐦 1 tweet

    RAD-IO

    ForMuscle-invasive bladder cancer, cT2-T3, 13% node-positive, median age 69

    TL;DR12-mo DFS 80% (40/50, 95% CI 0.67-0.89), clearing the ≥75% GO bar for durvalumab added to 55Gy/20fr chemoRT in MIBC bladder preservation.

    vs leading data
    • Benchmark BC2001 (JCO 2016): chemo-vs-no-chemo DFS HR 0.78 (0.60-1.02), p=0.07; RAD-IO's historical reference, not a within-trial comparison.

    Combined Curative Phase 2 trial Early signal ASCO Annual Meeting 2026

    RAD-IO
    +1 more figure
    Durvalumab cycles: 33/54 (61%) completed planned treatment, 21 (39%) discontinued early
    Durvalumab cycles: 33/54 (61%) completed planned treatment, 21 (39%) discontinued early
    8 details 5 trials watching
    Methods
    • 🔍 Single-arm multi-stage feasibility/safety trial, N=55 MIBC; no randomised comparator, benchmarked vs prior CRT data.
    • 💊 Durvalumab neoadjuvant + concurrent + 12mo adjuvant, added to 5FU + mitomycin C chemoRT.
    • 🔍 Baseline (N=55)
      • Age 69 (IQR 62-76)
      • Male 45 (82%), female 10 (18%)
      • cT2 44 (80%), cT3 11 (20%)
      • N+ 7 (13%) (N1 4, N2 3)
      • Prior neoadjuvant chemo 41 (75%)
    Results
    • 📊 1° EP met: 12-mo DFS rate 40/50 (80%), 95% CI 0.67-0.89; cleared pre-set ≥75% GO threshold (NO-GO <60%).
    • 📐 RT: 55Gy/20fr to bladder; node-positive pts add 46Gy/20fr elective nodal (stage 1 expansion, first 6 N+ pts).
    Critique
    • ⚠️ Safety: AEs reported 'in line with component parts' (chemoRT + durvalumab); no G3+ rates given in source.
    • ⚠️ DFS denominator 50: 2 pending + 3 not evaluable excluded; 'very high bladder preservation' claimed but no numeric rate in source.
    • ⚠️ 12-mo DFS is short f/u for curative bladder preservation; OS and durable preservation immature.
    Open questions
    📚 Sources · 🐦 3 tweets

    MIBC Post-pCR Management Review (Guercio)

    TL;DRPost-pCR MIBC: sandwich IO regimens continue planned adjuvant regardless of response; cisplatin-chemo alone → surveillance; pCR strongly prognostic (SWOG 8710 85% 5-yr OS).

    Trials discussed

    SWOG 8710ABACUSCHECKMATE-274KEYNOTE-905VOLGANIAGARAVESPERKEYNOTE-B15

    Combined Curative ASCO Annual Meeting 2026

    OS HR 0.50, periop EV+pembro vs RC alone (KEYNOTE-905)
    OS HR 0.50, periop EV+pembro vs RC alone (KEYNOTE-905)
    +1 more figure
    MIBC Post-pCR Management Review (Guercio)
    9 details 3 trials watching
    Methods
    • 🔍 pCR is a primary endpoint in signal-seeking phase 2; co-primary (not standalone registrational) in phase 3
    • 💊 Perioperative regimens by cisplatin-eligibility (IO 'sandwich': neoadjuvant → cystectomy → adjuvant)
      TrialCisplatinComparisonKey result
      KEYNOTE-905IneligiblePeriop EVP vs RC aloneOS HR 0.50
      VOLGAIneligiblePreop EV + periop durva/tremeEFS improved (press release)
      NIAGARAEligiblePreop Gem/Cis + periop durva24-mo OS 82.2% vs 75.2%
      VESPEREligibleddMVAC vs Gem/Cis5-yr OS 66% vs 57%
      KEYNOTE-B15EligiblePeriop EVP vs Gem/CisOS improved (FDA review)
    • 🔍 High-risk path after neoadjuvant chemo (≥ypT2 +N1) indicates adjuvant therapy (CHECKMATE-274 setting)
    Results
    • 📊 pCR (pT0N0) after neoadjuvant chemo is a strong positive prognostic marker
    • 📐 SWOG 8710: 85% 5-yr OS if pT0
    • 📐 Meta-analysis pooled RR 0.19 for RFS (pCR vs residual disease)
    • 📊 Post-pCR management hinges on the regimen
      • Sandwich IO (NIAGARA, periop EVP): continue planned adjuvant regardless of path response
      • NIAGARA: resume durvalumab ×8 mo post-cystectomy
      • Cisplatin chemo alone: surveillance after pCR is standard
    Critique
    • ⚠️ Not all pts in perioperative trials completed adjuvant therapy (toxicity); real-world adjuvant contribution uncertain
    • ⚠️ Scope is neoadjuvant → cystectomy only; trimodality/bladder-preservation not addressed (response there is clinical, not pathologic)
    Open questions
    📚 Sources · 🐦 1 tweet

    PREPEC

    ForImplant reconstruction after skin/nipple-sparing mastectomy

    TL;DRPre-pectoral implant improved 24mo BREAST-Q chest well-being +4.8 pts (1.0-8.7, p=0.01) vs sub-pectoral, but more unplanned implant loss (21% vs 15%).

    vs leading data
    • Prior pre- vs sub-pectoral data is mostly retrospective; this is a rare randomized PRO comparison, but the primary is a PRO, not a hard clinical-outcome endpoint

    Surgery Phase 3 RCT Confirmatory ASCO Annual Meeting 2026

    PREPEC
    Arm24mo well-being (chest), LS mean (95% CI)
    Pre-pectoral (N=191)79.2 (75.5-82.8)
    Sub-pectoral (N=189)74.3 (70.7-78.0)
    Difference4.8 (1.0-8.7), p=0.01
    +2 more figures
    PREPEC
    ArmUnplanned implant loss/replacement at 24mo, % (n/N)
    Pre-pectoral21.1% (41/194)
    Sub-pectoral14.5% (27/186)
    Adjusted difference5.7% (-2.4 to 13.8)
    PREPEC
    7 details 4 trials watching
    Methods
    • 🔍 International randomized trial (OPBC-02): pre- vs sub-pectoral implant-based reconstruction after skin- or nipple-sparing mastectomy, therapeutic or risk-reducing setting
    • 🔍 1° endpoint was patient-reported: BREAST-Q physical well-being (chest) at 24mo, 0-100 scale (higher = better)
    • 🔍 Safety reported by actual (as-treated) positioning while the PRO primary used assigned arms, implying some intraoperative crossover between groups
    CONSORT flow
    Randomized 380
    ↓
    Pre-pectoral IBBR
    allocated 191
    Sub-pectoral IBBR
    allocated 189
    Results
    • 📐 The adjusted between-arm implant-loss CI crosses zero, so the safety detriment is directional, not statistically conclusive
    Critique
    • ⚠️ Modest magnitude: 4.8-pt mean difference on a 0-100 PRO scale; lower 95% CI bound 1.0 leaves per-patient clinical relevance uncertain despite p=0.01
    • ⚠️ Safety tradeoff: pre-pectoral had more unplanned implant/expander loss or replacement, which authors call inconsistent with their non-inferiority hypothesis
    • ⚠️ Open-label by necessity (positioning can't be blinded); a subjective PRO primary under an unblinded design risks reporting bias
    Open questions
    📚 Sources · 🐦 1 tweet

    ROADS

    ForResected brain metastasis >2 cm

    TL;DRSurg bed recurrence 1% (GammaTile brachy) vs 12% (post-op SRS) in resected brain mets >2cm; 2-yr OS 62% vs 36%.

    vs leading data
    • Post-op cavity SRS is current SOC for resected brain mets (Mahajan Lancet Oncol 2017; N107C); ROADS positions intraoperative brachy as the comparator

    Radiation Palliative Challenges SOC ASCO Annual Meeting 2026

    ROADS
    EndpointGammaTilePost-op SRS
    Time to surg bed recurrenceNR17 mo
    Surg bed recurrence-free survivalNR11 mo
    2-yr OS62%36%
    7 details 4 trials watching
    Methods
    • 🔍 Randomized, resected brain metastasis >2cm, N=230; GammaTile (Cs-131 intraoperative brachytherapy) vs post-op cavity SRS
    • 🔍 GammaTile = Cs-131 collagen-tile implant placed at resection: immediate bed dosing, single session, no post-op SRS planning delay
    Results
    • 📊 Surg bed recurrence near-eliminated: 1% GammaTile vs 12% post-op SRS
    • 📊 Toxicity / competing-risk by arm
    Critique
    • ⚠️ Higher LMD with GammaTile (10% vs 3%) is the key tradeoff against its local-control gain
    • ⚠️ 2-yr OS gap (62% vs 36%) implausibly large for a cavity-local therapy; suspect baseline systemic-burden imbalance or small N
    • ⚠️ Open-label by design (implant vs SRS unblindable); abstract-only ASCO 2026, not yet peer-reviewed
    Open questions
    📚 Sources · 🐦 1 tweet

    CHRYSALIS-2

    ForTreatment-naïve atypical EGFR-mutant advanced NSCLC

    TL;DRmOS 41.0 mo (95% CI 27.7-NE) with 1L amivantamab + lazertinib in treatment-naïve atypical EGFR-mutant NSCLC; single-arm, n=49.

    vs leading data
    • Author frames this as extending the regimen's durable 1L benefit from common EGFR to atypical variants

    Systemic Palliative Early signal ASCO Annual Meeting 2026

    Overall survival: median 41.0 mo (95% CI 27.7-NE). Median follow-up 31.3 mo.
    Overall survival: median 41.0 mo (95% CI 27.7-NE). Median follow-up 31.3 mo.
    +1 more figure
    Median treatment duration 13.3 mo (range <0.1-53.2); 39% on treatment >2 yr.
    Median treatment duration 13.3 mo (range <0.1-53.2); 39% on treatment >2 yr.
    8 details 2 trials watching
    Methods
    • 🔍 Single-arm, n=49, median f/u 31.3 mo; IV amivantamab + lazertinib
    • 🔍 No clear association between EGFR variant subtype and outcome (author-stated)
    • 💊 Safety consistent with prior reports, no new signals at longer f/u
    Results
    • 📊 mOS 41.0 mo (95% CI 27.7-NE) in 1L atypical EGFR+ advanced NSCLC
    • 📊 Median treatment duration 13.3 mo (range <0.1-53.2)
    • 📊 39% of 1L pts remained on treatment >2 yr
    Critique
    • ⚠️ Single-arm, no randomised comparator vs standard 1L atypical-EGFR therapy (afatinib/osimertinib); 41 mo mOS is uncontrolled
    • ⚠️ Atypical EGFR is heterogeneous (e.g. G719X, L861Q, S768I, exon 20 ins); variant mix not detailed in source
    Open questions
    📚 Sources · 🐦 1 tweet

    ESAONA

    For1L EGFR-mutated NSCLC with brain metastases

    TL;DRIntracranial ORR 95.5% vs 79.6% (p=0.0004) and iPFS HR 0.46 favoring asandeutertinib over osimertinib in 1L EGFR-mut NSCLC with brain mets.

    vs leading data
    • Osimertinib (FLAURA-established 1L SOC) is the comparator; head-to-head challenge centered on CNS disease

    Systemic Palliative Phase 2 trial Challenges SOC ASCO Annual Meeting 2026

    7 details 4 trials watching
    Methods
    • 🔍 Randomized phase II, n=224 (asandeutertinib 111 vs osimertinib 113), 1L EGFR-mutated NSCLC with brain metastases
    • 🔍 RT-relevant: strong intracranial activity supports deferring upfront brain RT (SRS/WBRT) in favor of TKI; no arm tested RT directly
    CONSORT flow
    Randomized 224
    ↓
    Asandeutertinib
    allocated 111
    Osimertinib
    allocated 113
    Results
    • 📊 Efficacy by arm (BICR-assessed)
      EndpointAsandeutertinibOsimertinibHR / p
      Intracranial ORR95.5% (89.8-98.5)79.6% (71.0-86.6)p=0.0004
      Intracranial PFSNR17.5 mo (15.18-NA)HR 0.46, p=0.0020
      Overall PFSNR17.2 mo (15.18-19.55)HR 0.64, p=0.0473
    Critique
    • ⚠️ Safety: numerically more toxicity with asandeutertinib
      TRAEsAsandeutertinibOsimertinib
      Any TRAEs99.1%95.6%
      Serious TRAEs10.8%7.1%
    • ⚠️ Overall PFS edge borderline: HR 0.64, p=0.0473; benefit concentrated intracranially
    • ⚠️ Phase II, modest N; superiority over osimertinib needs phase III confirmation
    • ⚠️ Source is an ASCO 2026 LBA (LBA2007) presentation, not yet peer-reviewed
    Open questions
    📚 Sources · 🐦 1 tweet

    OptiTROP-Lung05 NCT06448312

    ForUntreated stage IIIB-IV NSCLC, PD-L1 TPS≥1%, EGFR/ALK wild-type

    TL;DRmPFS NR vs 5.7mo, HR 0.35, adding TROP2 ADC sac-TMT to pembro in 1L PD-L1+ NSCLC; OS immature (HR 0.55).

    vs leading data
    • Control mPFS 5.7mo aligns with KEYNOTE-042 1L pembro-mono, validating the comparator for the PD-L1≥50% population

    Systemic Palliative Phase 3 RCT Challenges SOC ASCO Annual Meeting 2026

    OptiTROP-Lung05
    ArmMedian PFS, moPFS eventsHR (95% CI)p
    Sac-TMT+PembroNR (13.6, NE)66 (31.7%)0.35 (0.26-0.47)<0.0001
    Pembro5.7 (4.3-7.0)128 (62.4%)n/an/a
    +2 more figures
    OptiTROP-Lung05
    PD-L1 TPSSac-TMT+Pembro mPFSPembro mPFSHR (95% CI)
    ≥50%NR (NE, NE)9.5 (6.9-13.8)0.47 (0.29-0.77)
    1-49%NR (11.1, NE)4.3 (2.9-5.5)0.28 (0.19-0.41)
    OptiTROP-Lung05
    ArmMedian OSOS eventsHR (95% CI)
    Sac-TMT+PembroNR (NE, NE)33 (15.9%)0.55 (0.36-0.85)
    PembroNR (NE, NE)54 (26.3%)n/a
    8 details 3 trials watching
    Methods
    • 💊 Sac-TMT (sacituzumab tirumotecan, SKB264/MK-2870): TROP2-directed ADC, here added to pembrolizumab as a chemo-free 1L doublet
    • 🔍 Phase 3, open-label, multicenter; N=413 randomized 1:1; 1° EP PFS by BICR (blinded central review)
    • 🔍 Eligibility: untreated stage IIIB/IIIC or IV NSCLC, PD-L1 TPS≥1%, EGFR/ALK wild-type, ECOG 0-1; stratified by histology, TPS, ECOG
    CONSORT flow
    Randomized 413
    ↓
    Sac-TMT + Pembro
    allocated 208
    Pembro
    allocated 205
    Results
    • 📊 ORR 70.2% vs 42.0% (sac-TMT+pembro vs pembro)
    • 📊 PFS benefit consistent across both PD-L1 strata, numerically larger in the TPS 1-49% subgroup than ≥50%
    Critique
    • ⚠️ Comparator caveat: control was pembro monotherapy
      • Pembro mono used as control across all enrolled TPS≥1%
      • For TPS 1-49%, real-world 1L standard is chemo-IO, not pembro mono
      • So relative benefit in the 1-49% stratum likely overstated vs current SOC
    • ⚠️ Open-label; PFS by BICR mitigates assessment bias, but ORR and investigator-PFS are unblinded
    • ⚠️ OS descriptive/immature at 10.5mo median f/u; HR 0.55 a favorable trend, not the prespecified final OS analysis
    Open questions
    📚 Sources · 🐦 2 tweets

    Neo-CRAG

    ForHigh-risk locally advanced gastric/EGJ adenoca, ≥cT3N2, D2 resection

    TL;DRAdding neoadj chemoRT (45Gy/25fx) to perioperative XELOX improved DFS (HR 0.75, mDFS 52.7 vs 24.4mo) and OS (HR 0.78, 67.5 vs 37.6mo) in high-risk LAGC.

    vs leading data
    • vs TOPGEAR and CRITICS, both null for adding RT to perioperative chemo: Neo-CRAG is the positive outlier

    Combined Curative Phase 3 RCT Challenges SOC ASCO Annual Meeting 2026

    Neo-CRAG
    8 details 3 trials watching
    Methods
    • 🔍 Phase 3 open-label RCT, N=620 (310/arm), 13 Chinese centers, ≥cT3N2 gastric/EGJ adenoca, 36.3% EGJ primary
    • 🔍 RT (CRT arm): 45 Gy/25 fx concurrent after chemo cycle 1, with dose-reduced oxaliplatin/capecitabine
    • 💊 Both arms: perioperative XELOX (3 cycles pre + 3 post) + D2 gastrectomy
    CONSORT flow
    Assessed / enrolled 620
    ↓
    Randomized 620
    ↓
    CRT (chemoRT)
    allocated 310
    CT (chemo)
    allocated 310
    Results
    • 📊 Efficacy, CRT vs CT (chemoRT added to perioperative XELOX)
      EndpointCRTCTHR (95% CI), p
      Median DFS (1°)52.7 mo24.4 mo0.750 (0.607-0.928), p=0.008
      3-yr DFS55.6%42.4%n/a
      Median OS67.5 mo37.6 mo0.781 (0.628-0.970), p=0.025
      5-yr OS50.1%44.2%n/a
    • 📊 Pathologic / locoregional signal (RT-attributable), CRT vs CT — comparison values omitted (cell value "0-2" not verified in source)
    Critique
    • ⚠️ Added toxicity with RT (G3+), CRT vs CT
      G3+ eventCRTCT
      Hematologic14.6%10.3%
      Postop complications9.0%7.6%
    • ⚠️ Non-FLOT backbone (XELOX); current perioperative SOC is FLOT, so whether RT adds to a stronger backbone is untested
    • ⚠️ Median OS gap (67.5 vs 37.6mo) much larger than HR 0.78 implies; check proportional-hazards assumption / tail stability
    Open questions
    📚 Sources · 🐦 1 tweet

    Living Longer, Living Better: GU Oncology Review (Rini)

    TL;DRDiscussant round-up on curing GU cancers without sacrificing QoL: bladder-preservation chemoRT ± durvalumab, EV-based sparing, ctDNA-guided adjuvant RCC selection.

    Trials discussed

    EV209EV309RAMPART

    vs leading data
    • EV309 is the RT-relevant readout: EV+pembro randomized directly against chemoradiotherapy for bladder preservation, with an OS endpoint

    Curative ASCO Annual Meeting 2026

    RT: full 55 Gy/20 fx 54/54 (100%), no extension/delay 47/54 (87%). Durvalumab completed 33/54 (61%).
    RT: full 55 Gy/20 fx 54/54 (100%), no extension/delay 47/54 (87%). Durvalumab completed 33/54 (61%).
    +1 more figure
    Living Longer, Living Better: GU Oncology Review (Rini)
    10 details 4 trials watching
    Methods
    • 💊 Adding durvalumab to mitomycin/5-FU chemoRT didn't compromise RT or chemo delivery (discussant read)
    • 🔍 EV + pembrolizumab bladder-sparing trials in MIBC
      EV209EV309
      Populationcystectomy-eligible cT2-4a N0ineligible/refusing cT2-4a N0
      N240390
      Design9 cycles EV + 1yr pembrorandomized vs chemoradiotherapy
      EndpointscCR rate, 2yr BIEFSBIEFS, OS
    • 🔍 RAMPART (adjuvant IO, RCC): non-clear cell subgroups underpowered, few pts/events, wide CIs, central path review ongoing
    • 🔍 ctDNA vs pathology for adjuvant RCC selection (Choueiri): pathologic features called a crude selection tool
    Results
    • 📊 ctDNA-negative 24-mo DFS 79.9% (pembro, n=344) vs 72.9% (placebo, n=332)
    • 📊 ctDNA-positive (n=30/arm) markedly worse: median DFS 10.1 and 6.4 mo, both arms poor
    Critique
    • ⚠️ Long-term bladder preservation, function, symptom burden, QoL + safety still need longer f/u (discussant)
    • ⚠️ Adjuvant IO has no proven benefit in non-clear cell RCC (discussant)
    • ⚠️ Only 30 ctDNA-positive pts/arm; hypothesis-generating, assay not yet a standard selection tool
    Notes
    • 🔬 Discussant session (Rini, ASCO 2026) framing the curative-intent vs QoL tension across GU; multiple trials, not one new result
    Open questions
    📚 Sources · 🐦 1 tweet

    SPIN Score — Celiac Plexus Radiosurgery NCT03323489

    ForPancreatic cancer with intractable retroperitoneal/celiac pain

    TL;DRPain response 32%→89% across the 0-2 score; severe baseline pain and no prior neurotoxic chemo predict who responds to celiac SRS.

    vs leading data
    • Parent phase 2 overall pain response 53% (95% CI 42-64%); the score separates subgroups well above and below that

    Radiation Palliative Phase 2 trial Caveats dominate ASCO Annual Meeting 2026

    ROC AUC 0.716 (apparent), 0.714 (optimism-corrected; 500-iteration bootstrap).
    ROC AUC 0.716 (apparent), 0.714 (optimism-corrected; 500-iteration bootstrap).
    7 details 2 trials watching
    Methods
    • 🔍 Post-hoc analysis of 90 evaluable pts from the pivotal phase 2 trial (NCT03323489, Lancet Oncol 2024;25:1070-79)
    • 🔍 Pain response = ≥2-pt drop in average BPI-SF pain, baseline to 3 wks (per protocol)
    Results
    • 📊 Pain response by score (post-hoc derivation cohort)
      ScoreResponse raten
      032%31
      153%40
      289%19
    • 📐 Independent predictors on multivariate logistic regression
      PredictorORp
      Neurotoxic chemo exposure5.10.009
      Baseline pain intensity1.80.003
    • 📊 Response by baseline pain cutoff (>6 gave optimal discrimination)
      • >5: 61.5%
      • >6: 65.8%
      • >7: 83.3%
      • >8: 85.7%
    Critique
    • ⚠️ Caveats
      • Post-hoc exploratory score; internal bootstrap validation only, external validation still pending (per authors)
      • Small subgroups: 2-pt n=19, 0-pt n=31
      • Response measured at 3 wks only, durability of analgesia not assessed
      • RT dose / fractionation / target volume not reported in source
    Notes
    • Best responders have severe baseline pain plus no prior neurotoxic chemo, supporting earlier celiac SRS before chemo-induced neuropathy
    Open questions
    📚 Sources · 🐦 1 tweet

    Wait or Treat NCT05236946

    ForMetastatic EGFR/ALK+ NSCLC, asymptomatic brain mets, on TKI+chemo

    TL;DRUpfront cranial RT cut intracranial progression (sub-HR 0.35) but didn't improve OS; 2y OS 48% vs 60% favored delayed RT.

    vs leading data
    • First completed phase III RCT on timing of brain RT for asymptomatic BM in oncogene-mutated NSCLC

    Radiation Palliative Phase 3 RCT Confirmatory ASCO Annual Meeting 2026

    Wait or Treat
    Upfront RT (n=105)Delayed RT (n=103)
    Events2047
    1-yr IC progression8.7% (2.9-14.5)25.7% (16.8-34.7)
    2-yr IC progression21.7% (12.6-30.8)50% (39.2-60.9)
    Sub-HR0.35 (0.21-0.59)ref
    p<0.001
    +1 more figure
    Phase III open-label RCT, 1:1: upfront RT (n=105) vs delayed RT (n=103) + TKI/chemo
    Phase III open-label RCT, 1:1: upfront RT (n=105) vs delayed RT (n=103) + TKI/chemo
    8 details 5 trials watching
    Methods
    • 🔍 Phase III open-label RCT, N=208 (105 vs 103), 1:1; stratified by GPA score and synchronous vs metachronous BM
    • 🔍 Population: metastatic EGFR/ALK+ NSCLC, completely asymptomatic measurable BM, ECOG 0-2, on TKI + chemo
    • 🔍 Delayed arm: cranial RT withheld until intracranial progression or patient request
    • 🔍 Surveillance: MRI brain every 3 months for year 1, then every 6 months
    CONSORT flow
    Randomized 208
    ↓
    Upfront cranial RT + TKI/chemo
    allocated 105
    Delayed cranial RT + TKI/chemo
    allocated 103
    Results
    • 📊 OS: no upfront benefit (secondary EP, per @StephenVLiu)
      Upfront RTDelayed RT
      2-yr OS48%60%
      OS HR1.45 (favors delayed)ref
    Critique
    • ⚠️ Radiation necrosis ~6% with upfront RT; not observed in delayed arm (per @DrRiyazShah, @StephenVLiu)
    • ⚠️ Upfront RT improved intracranial control but not survival; pre-emptive cranial RT hard to justify in asymptomatic BM on CNS-active TKIs
    • ⚠️ RT modality, dose, fractionation (WBRT vs SRS) not reported in source, gating transferability
    Open questions
    📚 Sources · 🐦 3 tweets