📬 Lung/Head and Neck/Thymoma Intelligence Digest — Week 2026-W40

Merck and Daiichi Sankyo withdraw the I-DXd BLA in ES-SCLC, leaving relapsed patients without an approved ADC outside China

Daiichi Sankyo and Merck withdrew the BLA application for I-DXd in ES-SCLC ahead of the FDA's PDUFA process for accelerated approval, which had been expected to complete by October 2026. The stated regulatory rationale was that the IDeate-Lung01 data do not satisfy requirements needed to support an accelerated approval, with the company noting concern that I-DXd may not have been approved on phase 2 findings alone while the phase 3 IDeate-Lung02 study nears completion. Some clinicians argue the withdrawal should not be over-read — phase II was likely inadequate and the B7-H3 class still has tam-peli, ris-rez, and I-DXd — but the gap is compounded by lurbinectedin's withdrawal in the 2nd-line-and-beyond setting, leaving few options after tarlatamab.

Discussed by: @StephenVLiu, @drshieldsmd, @StephenVLiu, @LauraAlderMD, @TejasPatilMD, @SuyogCancer, @lungoncdoc, @phsiao4

→ View representative tweet

Sources cited: Conference abstract · Merck press release

Randomized ASTRO26 trial (NRG-CC009) makes fractionated SRS the standard for SCLC brain metastases over hippocampal-avoidance WBRT

The randomized NRG-CC009 trial presented at ASTRO26 compared fractionated SRS versus HA-WBRT in SCLC brain metastases, finding no difference in time to cognitive failure or neurologic death, a higher OS with SRS (median 17 vs 9 months), and more adverse events with HA-WBRT. Radiation oncologists are calling SRS the standard of care for SCLC with brain metastases. The open question is whether this survival advantage holds outside a trial population, or whether the miliary pattern of SCLC brain disease still makes whole-brain treatment the safer default in many patients.

Discussed by: @OncoAlert, @Sushilberiwal, @daakhsaab

→ View representative tweet

ESMO26 lung Presidential lineup lands: DeLLphi-305, KRASCENDO 1 and SAFFRON all read out positive by press release

Three phase 3 lung cancer late-breaking abstracts hold ESMO26 Presidential slots: DeLLphi-305 (LBA3), KRASCENDO 1 (LBA5), and SAFFRON (LBA6). KRASCENDO 1 compared divarasib with sotorasib or adagrasib in previously treated KRAS G12C-mutated NSCLC and met its primary and key secondary endpoints, with divarasib significantly improving PFS and OS; clinicians note that a next-generation KRAS inhibitor with greater activity than the modest benefit of current options is very welcome, though details are still needed. SAFFRON tested osimertinib plus savolitinib versus platinum–pemetrexed in EGFRm MET-overexpressed and/or -amplified advanced NSCLC post-osimertinib and met its primary PFS and key secondary OS endpoints, while DeLLphi-305 tested tarlatamab plus durvalumab versus durvalumab as 1L maintenance following durvalumab, platinum, and etoposide in ES-SCLC and met its primary endpoint with significantly improved OS, PFS, and ORR.

Discussed by: @M_Torasawa, @tnewsomdavis, @HHorinouchi, @Latinamd, @DrRiyazShah, @potsstrongem

→ View representative tweet

Sources cited: AstraZeneca press release · AstraZeneca press release

DLL3 IHC predicts tarlatamab response and time on treatment in the largest validation series to date

In 138 SCLC patients, a CLIA DLL3 IHC assay predicted response and long-term outcomes with tarlatamab, while dynamic ctDNA/CTC monitoring revealed resistance mechanisms — described as the largest series to validate DLL3 to date. DLL3 IHC expression ranged from 0% to 100% (medians of 70% and 75% in two datasets); patients with expression above 75% had a significantly longer median time on tarlatamab treatment (10.1 months vs 2.9 months) versus those with ≤75% expression, and patients achieving a partial response had higher median DLL3 expression (80%) than those with progressive disease (40%). Clinicians frame the central question as whether higher DLL3 expression reflects a stronger BiTE effect, with biomarker selection becoming more pressing as B7-H3 ADCs widen the available options.

Discussed by: @BZhangMD, @GlopesMd, @BalazsHalmosMD, @mgonzalezvelMD

→ View representative tweet

Sources cited: bioRxiv

IMpower030 fails its primary endpoint despite a large numerical EFS difference, and perioperative IO's OS record is being tallied

In IMpower030, EFS numerically favoured atezolizumab at 62.8 versus 34.9 months (HR 0.77), but the primary endpoint was not statistically significant (P = 0.07) against a required boundary of P <0.04, meaning OS was not formally tested. Across six phase III perioperative IO trials, EFS was positive in 5 of 6; OS hazard ratios reported so far are 0.74 for KEYNOTE-671, 0.65 for RATIONALE-315, 0.85 for CheckMate 77T (interim, NS), 0.89 for AEGEAN (interim, NS), 0.62 for NEOTORCH (interim, NS), and 0.77 for IMpower030 (not formally tested). In the perioperative versus preoperative debate, both sides agree that no randomized trial has yet defined the contribution of adjuvant IO therapy, and the case against the adjuvant phase asks what it adds and what to do for patients who did not reach pCR.

Discussed by: @ChandrakanthMv, @ChandrakanthMv, @Latinamd, @Latinamd

→ View representative tweet

NCCN adds daraxonrasib for KRAS-mutant NSCLC at 200 mg, and clinicians immediately ask about access

The 9/2026 NCCN NSCLC guidelines added daraxonrasib off-label at 200 mg daily (instead of 300 mg) to consider in patients with KRAS-mutant NSCLC, listed in the emerging biomarkers table — not an FDA approval for lung cancer — alongside sevabertinib as a new first-line option for HER2-mutant NSCLC. In phase I/II data, post chemo-IO response rate was approximately 42% with DOR approximately 1 year and OS of 16 months, yet clinicians note there is no compassionate use expanded access program planned. At least one clinician reported insurance approving off-label daraxonrasib for an NSCLC patient outright without appeal, signalling that the NCCN listing is already having practical effect.

Discussed by: @TejasPatilMD, @StephenVLiu, @LauraAlderMD, @LuisRaezMD, @EmreCakirMD, @GlopesMd

→ View representative tweet

Two HER2 TKIs are now approved first-line in NSCLC and the resistance data are starting to separate them

Zongertinib and sevabertinib are both approved as first-line-plus treatment for advanced NSCLC with a HER2 TKD mutation; while efficacy is similar, they are distinct agents and clinicians are working to leverage those differences. In Beamion LUNG-1 with zongertinib, 8 paired tissue results from patients with progressive disease showed two acquired HER2 mutations — T862A and S783C — with T798I and C805S notably absent, and the acquired mutations were not detected in plasma, leading specialists to argue that re-biopsy at progression is mandatory. One perspective, citing Xiuning Le of UT MD Anderson, is that zongertinib is the go-to option for treatment-naive patients with a HER2 TKD mutation.

Discussed by: @StephenVLiu, @dplanchard, @StephenVLiu

→ View representative tweet

Sources cited: OncLive

GRAIL's Galleri faces the FDA's first MCED panel as NHS-Galleri misses its stage-shift endpoint

In the NHS-Galleri randomized trial of more than 142,000 people, adding a blood-based multicancer early detection test to usual screening did not reduce the incidence of Stage III/IV cancers — the study's primary endpoint — despite more Stage I/II diagnoses and fewer Stage IV cancers. The FDA panel packet gave head and neck performance for the first time; in PATHFINDER 2, head and neck cancer was the single largest contributor to the stage I–II cancers Galleri detected. A separate observation linked regional participation in a population-based MCED screening trial to a temporary increase in diagnostic delay rates among patients referred for suspected head and neck, lung, and upper gastrointestinal cancers. One view is that regulatory approval need not mean universal screening, and that informed patients and their physicians should examine the results and decide together.

Discussed by: @GlopesMd, @GlopesMd, @Jdcramer

→ View representative tweet

Sources cited: Cancer Diagnostic Delay Rates Associated With a Population-Based Screening Trial Evaluati… (JAMA, 2026)

REACH-EGFR and SOLARA report early activity for two next-generation EGFR C797S TKIs at WCLC26

In REACH-EGFR, the mutant-selective EGFR TKI VRN110755 produced responses in all 6 patients with EGFR C797S, with ctDNA clearance and early signs of intracranial efficacy, and no discontinuations due to treatment-related adverse events among 67 patients. In SOLARA, the macrocyclic, CNS-penetrant EGFR TKI BH-30643 — which holds FDA fast-track designation — yielded a 45% response rate (18/40) in EGFR C797S with DCR 88%, 63% of patients still on therapy, 3% discontinuation from treatment-related adverse events, and 1% pneumonitis. Critics note that C797S accounts for approximately 7% of osimertinib-resistant patients, and more translational work is still needed for the vast majority of EGFR-resistant patients.

Discussed by: @StephenVLiu, @phsiao4, @TejasPatilMD

→ View representative tweet

COPERNICUS reports that COCOON-style supportive care cuts amivantamab-lazertinib toxicity versus MARIPOSA

Data from COPERNICUS show that subcutaneous amivantamab plus lazertinib with COCOON supportive care substantially reduced toxicity compared with published MARIPOSA results: paronychia fell from 50% to 35%, rash from 55% to 25%, VTE from 23% to 3%, and IRR from 55% to 3%, with discontinuation of amivantamab in the first year due to adverse events of special interest under 1% versus 9% in MARIPOSA. A new COCOON cohort focused on prevention of moderate to severe dermatologic adverse events with first-line subcutaneous amivantamab plus lazertinib is also being presented at ESMO26. In the first-line EGFR debate, combination regimens including amivantamab plus lazertinib are defended on the basis of survival benefit and CNS activity, with toxicity addressed through proactive mitigation strategies.

Discussed by: @StephenVLiu, @Latinamd, @HHorinouchi

→ View representative tweet

Tarlatamab's reduced monitoring window prompts oncologists to redesign outpatient delivery and debate prophylactic tocilizumab

The FDA approved reduced monitoring time for the first two tarlatamab doses — from 22–24 hours on cycle 1 day 1 and day 8 to 6–8 hours — lowering a key barrier to outpatient and community administration. In a real-world tarlatamab study, CRS occurred in 56% after C1D1 and 61% cumulatively through C1D8, with tocilizumab used in 33 instances, rates described as similar to those reported with myeloma bispecifics; in the MSK myeloma bispecific experience, low-dose prophylactic tocilizumab 4 mg/kg reduced CRS from approximately 59% to approximately 23% and grade 2 CRS from 21% to 2%, while cutting drug cost roughly in half versus 8 mg/kg. Evidence is also accumulating that the optimal tarlatamab dose may differ individually and possibly by histology of neuroendocrine neoplasm, in addition to DLL3 level and molecular subtype.

Discussed by: @LauraAlderMD, @BZhangMD, @jennifermarksmd, @BZhangMD

→ View representative tweet

Sources cited: Lung Cancer · Lung Cancer · ScienceDirect

ARMADA opens: first-in-human DLL3-targeted CAR T cells engineered to secrete IL-18 in DLL3+ cancers

ARMADA (NCT07783477), a first-in-human phase I trial developed at MSK, is now open and evaluates DLL3-targeted CAR T cells engineered to secrete IL-18 in DLL3+ cancers, including small cell, large cell, carcinoid, and other tumors; IL-18 armoring is designed to enhance CAR T-cell activity and help overcome immune suppression in solid tumors. Separately, off-the-shelf CAR T-cell data in lung cancer brain metastases showed both responders alive at 11+ and 13+ months from first dose with no GvHD — an early signal for allogeneic CAR T in this setting.

Discussed by: @g_mountzios, @EladSharonMD

→ View representative tweet