📬 Lung/Head and Neck/Thymoma Intelligence Digest — Week 2026-W38

MAVERICK ends the PCI era in SCLC: MRI surveillance alone improves cognitive failure-free survival with no survival cost

In the MAVERICK trial, MRI surveillance alone improved cognitive failure-free survival versus PCI plus MRI (HR 0.60, 90% CI 0.46–0.78; p=0.0005), with 6-month CFFS rates of 38% vs 17% and 12-month rates of 17% vs 6%, and a median cognitive failure-free survival of 37.8 months versus 16.5 months. Brain metastases occurred more frequently with MRI surveillance alone (30% vs 15% with PCI), but this did not translate into differences in PFS or OS; preliminary OS was similar (HR 0.90, 90% CI 0.67–1.20), and grade 3–5 treatment-related AEs were lower with MRI alone (0.8% vs 7.9%). The trial included 68–69% limited-stage SCLC, with the remainder extensive-stage. Investigators caution that MAVERICK did not show equivalence without PCI — it showed notable clinical benefit with an MRI surveillance strategy omitting PCI. The open question is what MRI surveillance interval — q3 months or q6 months — is required to make omission safe, and who still needs PCI when MRI access is limited.

Discussed by: @StephenVLiu, @Lung_Cancers, @HHorinouchi, @StephenVLiu, @BalazsHalmosMD, @jennifermarksmd, @diegoadiazg, @Latinamd, @Lung_Cancers, @StephenVLiu, @NaglaAKarimMD

→ View representative tweet

Sources cited: Conference abstract

DeLLphi-305: tarlatamab plus durvalumab is the first BiTE to improve overall survival as first-line maintenance in ES-SCLC

The phase 3 DeLLphi-305 study showed a statistically significant and clinically meaningful overall survival benefit for tarlatamab plus durvalumab versus durvalumab alone as first-line maintenance in extensive-stage SCLC, making this the first phase 3 BiTE study to show an OS benefit in first-line ES-SCLC maintenance; PFS and ORR were also positive and no new safety signals were identified (NCT06211036). Patients had no progression after first-line induction with durvalumab plus platinum chemotherapy and etoposide before randomization. Full numerical data have not been released alongside the press release, and clinicians note that tarlatamab is not for all patients — it is expensive, complex to administer, requires monitoring, and is associated with significant toxicity. Tarlatamab was approved in the US for accelerated use on May 16, 2024, yet patients globally still cannot access it in most countries outside of a trial, sharpening the question of what infrastructure is required to deliver it safely at scale and what options remain in second line once it moves into maintenance.

Discussed by: @RManochakian, @M_Torasawa, @EladSharonMD, @EladSharonMD, @HHorinouchi, @StephenVLiu, @GlopesMd, @g_mountzios, @phsiao4, @pistoto2017, @Latinamd, @ChandrakanthMv

→ View representative tweet

Sources cited: AstraZeneca press release

The Lancet's Therapeutics in Lung Cancer Series lands with three reviews spanning early-stage, advanced NSCLC and SCLC

The Lancet and The Lancet Respiratory Medicine launched a Therapeutics in Lung Cancer Series on 7 September at ERS2026, comprising three free-access review papers covering early-stage NSCLC, advanced NSCLC, and SCLC. The early-stage NSCLC review addresses current standard of care with immunotherapy and targeted therapies and open questions including neoadjuvant versus perioperative approaches, immunotherapy in borderline resectable disease, and how to increase pCR rates. The advanced NSCLC paper, published in Lancet Respiratory Medicine and associated with Hendriks and Reck, covers new biomarkers and treatment strategies in targeted therapy. The unresolved question across the Series is whether the three reviews converge on a single position for the perioperative question — neoadjuvant alone versus full perioperative immunotherapy — or leave it open.

Discussed by: @Latinamd, @dplanchard, @dplanchard, @lancetrespirmed.bsky.social

→ View representative tweet

Sources cited: ScienceDirect · The Lancet

FDA grants first-line accelerated approval to sevabertinib in HER2 TKD-mutant NSCLC, crowding the frontline HER2 field

On September 9, 2026, the FDA granted accelerated approval to sevabertinib (Hyrnuo) for adults with locally advanced or metastatic non-squamous NSCLC harboring HER2 (ERBB2) TKD activating mutations, expanding its prior indication to treatment-naïve patients. In the single-arm SOHO-01 trial, 69 untreated patients achieved an ORR of 75% (95% CI 64–85), with 73% of responders maintaining responses for at least six months and 38% for at least 12 months; adverse events include diarrhea, LFT abnormalities, and pneumonitis. The approval joins zongertinib in first line, with trastuzumab deruxtecan expected to follow based on DESTINY-Lung04, though oncologists note the accelerated approval rests on a single-arm response rate in 69 patients and is not yet proof of survival benefit or superiority. A toxicity distinction has been noted: sevabertinib targets wild-type EGFR more extensively than zongertinib — having been originally developed as an EGFR inhibitor — and causes more mucocutaneous and GI toxicity, a difference that may inform the frontline choice alongside CNS activity and sequencing strategy.

Discussed by: @RManochakian, @StephenVLiu, @HHorinouchi, @OncBrothers, @Latinamd, @GlopesMd, @ozdogan_md, @Dr_ElvinaA, @RielyMD, @StephenVLiu

→ View representative tweet

Sources cited: FDA press release

Update: the ATLAS registry gives real-world osimertinib outcomes in uncommon EGFR mutations, and exon 20 crowding dominates the BTG Lung discussion

In the Italian ATLAS registry, osimertinib in uncommon EGFR-mutant NSCLC produced an ORR of 66%, median PFS of 18.3 months, and median OS of 34.5 months overall, with exon 19 delins showing superior ORR, PFS, and OS versus other uncommon EGFR mutations and mirroring classical ELREA outcomes. At BTG Lung 2026, the breadth of variants sharing a single label was a recurring point — uncommon EGFR mutations demand individualized treatment rather than a blanket approach, and leptomeningeal disease is now seen in 11% of patients with EGFR NSCLC in the modern era, with incidence rising as survival lengthens. The debate is how to generate evidence across many uncommon variants — through functional databases, international registries, or prospective real-world evidence — in a now crowded exon 20 space, and what minimum dataset a functional database should supply before a variant is treated as osimertinib-sensitive.

Discussed by: @HHorinouchi, @StephenVLiu, @OncLive, @StephenVLiu

→ View representative tweet

Sources cited: ESMO Open · OncLive

IASLC expert panel proposes a revised pulmonary carcinoid/NET classification with a pulmonary-specific Ki-67 threshold

An IASLC expert panel has proposed an updated pulmonary carcinoid/NET classification expected to be included in the forthcoming WHO thoracic blue book. The proposal retains "carcinoid" as the main diagnostic term for thoracic clinicians while moving toward the NET nomenclature used in other organs, adds Ki-67 as a diagnostic criterion with a pulmonary-specific 5% threshold, and formally recognizes rare carcinoid/NET G3 tumors — defined by higher proliferation but molecularly closer to NETs than to high-grade NECs. The open question is whether the 5% Ki-67 threshold and the formally recognized G3 carcinoid category will change systemic therapy choices or only pathology reporting, with DLL3 remaining relevant given reported tarlatamab activity in this reclassified group.

Discussed by: @jennifermarksmd, @StephenVLiu

→ View representative tweet

Sources cited: JTO

SWOG/NRG S1914 publishes in the Lancet: atezolizumab adds toxicity but no survival to SBRT in inoperable early-stage NSCLC

In high-risk T1–T3N0M0 early-stage NSCLC patients who were medically inoperable or declined surgery, adding peri-SBRT atezolizumab did not improve PFS or OS and increased toxicity; the trial was stopped early for futility, with OS HR 1.04 and grade ≥3 AEs 12% versus 3%. Clinicians contrasted S1914 with IMpower030, noting two negative phase 3 atezolizumab trials sending different messages: IMpower030 missed its statistical threshold despite a strong efficacy signal, while S1914 showed no survival gain and greater toxicity. The unresolved question is whether any rationale remains for adding checkpoint blockade to SBRT in medically inoperable early-stage NSCLC, or whether the field should move to different mechanisms entirely.

Discussed by: @HHorinouchi, @EladSharonMD, @ozdogan_md, @_ShankarSiva

→ View representative tweet

Sources cited: The Lancet

Two B7-H3 ADCs report identical hazard ratios in relapsed SCLC on the same day, and clinicians warn against cross-trial comparison

At WCLC 2026, two independent phase 3 B7-H3 ADC trials — TAISHAN-302 (n=451) and ARTEMIS-008 (n=461) — each reported HR 0.46 against topotecan in relapsed SCLC, reducing the risk of death by 54% in both cases. ARTEMIS-008 reported an OS gain of 8.2 months, more than doubled median PFS, higher response and disease control rates, and no new safety signals. Oncologists have been explicit that identical hazard ratios are not grounds for comparing the two drugs across trials. Analysts identify B7-H3, DLL3, and SEZ6 as three validated second-line SCLC targets, with Daiichi's B7-H3 agent expected to reach market first, though the question of what will differentiate these agents — payload toxicity, CNS activity, or dosing schedule — remains open.

Discussed by: @ozdogan_md, @MorningGlorySci, @EladSharonMD, @ozdogan_md, @Medthority

→ View representative tweet

TAISHAN-302: the B7-H3 ADC tambotatug pelitecan cuts the risk of death 54% versus topotecan in relapsed SCLC

TAISHAN-302 randomized 451 patients 1:1 to tambotatug pelitecan (tam-peli, YL201) — a B7-H3 ADC with a topo-1 payload and DAR 8, given at 2 mg/kg q3w — versus standard topotecan. Median OS was 13.3 versus 9.4 months (HR 0.46; P<0.001), PFS was 7.4 versus 2.8 months (HR 0.29; P<0.001), and ORR was 59.1% versus 9.7%; grade ≥3 AEs were 55.4% versus 77.9%, and intracranial response was 32% versus 3% in patients with baseline brain lesions, with 71 patients still on tam-peli at approximately 9 months median follow-up versus 8 with topotecan. The trial was conducted in China only, open-label, with short follow-up, limiting generalizability; Roche holds development and commercial rights outside Greater China, making future European access a live question, and the debate concerns what global confirmatory data are needed before a B7-H3 ADC displaces topotecan in second-line SCLC.

Discussed by: @StephenVLiu, @DrRishabhOnco, @Dr_ElvinaA, @YGaritaonaindia, @LungCancerEu, @NaglaAKarimMD, @ashleygilstrapp

→ View representative tweet

Sources cited: Tambotatug Pelitecan in Small-Cell Lung Cancer after Platinum-Based Therapy (New England Journal of Medicine, 2026)

Update: the debate over ctDNA-guided switching reaches lung cancer, with APPLE's 22.0 vs 20.2 month PFS as the cautionary precedent

APPLE (EORTC LCG-1613) monitored plasma T790M on gefitinib and switched to osimertinib at molecular progression; median PFS was comparable to switching at radiographic progression (22.0 vs 20.2 months), and only 17% of patients ever switched on ctDNA before their scans changed. Clinicians at BTG Lung 2026 discussed whether ctDNA can support adaptive designs in first-line EGFR NSCLC, with attention to the US SHEDDER study. In unresectable stage III NSCLC, ctDNA-based MRD assessment has also been discussed as a means of identifying patients at high risk of relapse after chemoradiotherapy who may benefit from alternative or intensified treatment strategies. The unresolved question is what a positive ctDNA-guided trial in lung cancer would actually have to demonstrate, given that APPLE showed molecular-progression switching matched radiographic switching.

Discussed by: @maraantonoff, @StephenVLiu, @JTOonline

→ View representative tweet

Sources cited: JTO

FDA breakthrough designation for the FMISO hypoxia tracer to guide de-escalated radiation in HPV-positive throat cancer

The FDA has granted Breakthrough Therapy designation to JBS-003 (FMISO), a specialized imaging tracer, for the identification of tumor hypoxia to guide de-escalated radiation therapy in patients with HPV-positive oropharyngeal carcinoma. The designation positions FMISO as a tool to identify HPV-positive throat cancer patients who may safely receive less-intensive radiation treatment. The debate is whether hypoxia imaging should replace clinical risk factors as the gate for radiation de-escalation and what confirmatory outcome data are required before that shift is made.

Discussed by: @RielyMD, @CancerNetwrk

→ View representative tweet

Sources cited: Cancer Network

DeLLphi-309: extended-interval tarlatamab dosing trades response rate for convenience in relapsed SCLC

DeLLphi-309 evaluated three tarlatamab schedules in relapsed SCLC: BICR ORR was 39.8% with 10 mg Q2W, 31.0% with 20 mg Q3W, and 27.1% with 30 mg Q4W, with median PFS of 4.2, 4.1, and 2.7 months respectively; median duration of response was 8.3, 5.6, and not estimable months, with descriptive analysis only and no formal statistical testing. Median OS was 11.2 months with Q2W and not estimable at 9 months for Q3W/Q4W. CRS was mostly grade 1/2 but numerically higher with extended intervals; ICANS was higher in the Q4W arm, and one fatal event — cardiac arrhythmia/electrolytes — occurred on Q4W. The trade-off is whether the lower response rate with Q3W/Q4W is an acceptable price for fewer infusion visits, or whether Q2W should remain the default outside of access-limited settings.

Discussed by: @FernandoOnco, @Medthority, @Medthority, @HHorinouchi

→ View representative tweet

Sources cited: Conference abstract

ADAURA's 8-year OS landmark update keeps adjuvant osimertinib's benefit intact and reopens the adjuvant chemotherapy question

The 8-year OS update of ADAURA — described as the longest OS follow-up from any global adjuvant phase III trial in EGFR-mutated early-stage NSCLC — showed osimertinib OS HR 0.53 (95% CI 0.38–0.75) in the stage II–IIIA primary population and 0.52 (95% CI 0.39–0.71) in the stage IB–IIIA overall population, with 8-year landmark OS rates increased by 16 and 15 percentage points respectively, and 79% of patients alive at 8 years. The HR was 0.48 at 5 years and 0.52 at 8 years. A separate WCLC26 abstract reported that intercalating platinum chemotherapy with gefitinib improved DFS versus gefitinib alone (58.1 vs 48.9 months; HR 0.71), keeping the adjuvant chemotherapy question alive, with the OCEANiC trial cited as the prospective test of whether platinum chemotherapy is still required alongside adjuvant osimertinib.

Discussed by: @Tony_Calles, @TommyJohn00, @HHorinouchi, @dplanchard, @ChandrakanthMv

→ View representative tweet

Sources cited: Conference abstract

EVOKE-03 misses: sacituzumab govitecan plus pembrolizumab adds four months of PFS in PD-L1-high NSCLC without statistical significance

EVOKE-03/KEYNOTE-D46 reported median PFS of 11.8 versus 7.7 months (HR 0.81) in PD-L1 TPS ≥50% metastatic NSCLC with sacituzumab govitecan plus pembrolizumab versus pembrolizumab alone — a 4.1-month PFS gain — but the primary statistical endpoint was not met. The result was reported at the Presidential session alongside the two positive B7-H3 ADC readouts. The unresolved question is whether a 4.1-month PFS gain that misses statistical significance is enough to keep TROP2 ADC plus pembrolizumab combinations in frontline development for PD-L1-high NSCLC.

Discussed by: @Medthority, @Medthority, @ozdogan_md

→ View representative tweet

NAUTIKA1 tests divarasib before surgery in resectable KRAS G12C NSCLC, with 42% major pathologic response in 20 patients

Updated NAUTIKA1 data presented at WCLC26 (OA04.04) showed neoadjuvant divarasib produced a 42% major pathologic response rate and 16% pCR in resectable KRAS G12C-positive NSCLC, with 19 of 20 patients undergoing surgery and no adverse-event-related surgical delays. With only 20 patients, the result is proof of concept rather than practice-changing. The debate is whether a 42% major pathologic response rate from targeted monotherapy before surgery is sufficient to justify randomizing KRAS G12C patients away from perioperative chemoimmunotherapy.

Discussed by: @ChandrakanthMv, @OncoAlert

→ View representative tweet

PAPILLON final OS: amivantamab-chemotherapy gives the longest survival reported in EGFR exon 20 insertion NSCLC but misses in the ITT after 76% crossover

The final OS analysis of PAPILLON (PL03.03, NCT04538664) in first-line EGFR exon 20 insertion NSCLC showed median OS of 34.3 versus 27.9 months — the longest reported to date for this population — but the ITT OS HR was 0.87 (95% CI 0.66–1.14; P=0.307) at a median follow-up of 48 months, with 76% crossover; an IPCW-adjusted OS HR of 0.57 was reported to account for crossover. The amivantamab-chemotherapy arm also demonstrated superior PFS2 (HR 0.59), improved patient-reported outcomes, and improved time to symptomatic progression. The methodological question under discussion is whether IPCW-adjusted survival should carry the same weight in guidelines as an ITT result when a trial's crossover rate reaches 76%.

Discussed by: @StephenVLiu, @HHorinouchi, @NReguart, @drshieldsmd, @Tony_Calles, @jennifermarksmd, @OncologyMarwarD, @TommyJohn00

→ View representative tweet

Sources cited: Conference abstract

Update: REZILIENT3 reports its phase 3 numbers — zipalertinib plus chemotherapy nearly doubles PFS in first-line EGFR exon 20 insertion NSCLC

REZILIENT3 (PL03.04, NCT05973773) reported median PFS of 14.5 versus 8.5 months (HR 0.50; 95% CI 0.34–0.73) and ORR 65% versus 40.3% with zipalertinib plus chemotherapy versus chemotherapy alone in first-line EGFR exon 20 insertion NSCLC; OS is immature with 64% crossover, and a favorable OS trend was noted. 31% of patients in both arms had brain metastases, with an impressive benefit in those with CNS involvement; grade 3 or higher TRAEs were 80%, largely cytopenias, dose reduction and interruption were common, and 6% had ILD. With zipalertinib-chemotherapy, amivantamab-chemotherapy, and sunvozertinib all demonstrating benefit over chemotherapy in this setting, the question is which toxicity profile — early myelosuppression, lower-grade GI toxicity, or infusion-related reactions — should drive the first-line selection.

Discussed by: @HHorinouchi, @Tony_Calles, @LauraAlderMD, @DrRiyazShah, @drshieldsmd, @TommyJohn00, @KSamalaMD

→ View representative tweet

Sources cited: Conference abstract