NSCLC (Part 2): Stage IV, Targeted and Systemic Therapy

Continued from: NSCLC (Part 1): Biology, Screening, Staging, Early-Stage and Stage III

Stage IV, general principles

• Always test for ALL drivers + PD-L1 BEFORE 1L in advanced non-squamous (and broadly in squamous if non/light smoker).
• For classic actionable drivers (EGFR ex19del/L858R, ALK, ROS1, RET, BRAF V600E, MET ex14, NTRK, HER2) give targeted therapy first regardless of PD-L1; exceptions: KRAS G12C follows 1L chemo-IO, and EGFR exon 20 insertions get 1L amivantamab + carbo/pemetrexed. IO/chemoIO inferior in driver+ disease and ↑ toxicity if TKI given after IO (esp. osimertinib pneumonitis after PD-1).
• Oligometastatic disease (single extrathoracic met M1b per AJCC, or ≤3 to 5 ablatable sites):
  • Local consolidative tx (SBRT/surgery) + systemic tx improved PFS/OS in small phase II trials (Gomez), but NRG-LU002 was negative; remains a selective, multidisciplinary option.
  • Gomez et al. (randomized phase II, ≤3 mets): mPFS 14.2 vs 4.4 mo; mOS 41.2 vs 17.0 mo.

Targeted therapy by driver

• 4 EGFR (erbB) family receptors: HER1/erbB1 (EGFR), HER2/erbB2, HER3/erbB3, HER4/erbB4. Activation → ↑ proliferation, angiogenesis, metastasis and ↓ apoptosis.
• Activating mutations: 10 to 30% of NSCLC; associated with female sex, East Asian race, never/light smoking, adenocarcinoma. Detect on tissue or ctDNA.
• Classic sensitizing mutations (~90%): exon 19 in-frame deletions (most commonly E746_A750del) and exon 21 L858R substitution.
• EGFR TKI class AEs: rash, diarrhea, paronychia (wild-type EGFR in skin/GI), generally favorable tolerability vs chemo (especially osimertinib), but not uniformly lower grade ≥3 toxicity (e.g., afatinib comparable to chemo in LUX-Lung 3).

• Frequency: ~15% of US adenos; ~50% of Asian non-smokers.
• FLAURA: 1L osimertinib (3rd-gen) > 1st-gen TKI. mPFS 18.9 vs 10.2 mo (HR 0.46); mOS 38.6 vs 31.8 mo (HR 0.80); ↓ CNS progression (6% vs 15%); ↓ grade ≥3 AEs.
• FLAURA2 (NEJM 2023 + final OS WCLC 2025): osimertinib + platinum/pemetrexed × 4 → osi + peme maintenance.
  • Efficacy: mPFS 25.5 vs 16.7 mo (HR 0.62); mOS 47.5 vs 37.6 mo (HR 0.77).
  • Best for: high-risk pts, CNS disease, L858R, TP53 co-mutation, brain mets.
• MARIPOSA (NEJM 2024): amivantamab (EGFR-MET bispecific) + lazertinib vs osimertinib 1L.
  • Efficacy: mPFS 23.7 vs 16.6 mo (HR 0.70); mOS HR 0.75 (3-yr OS 60% vs 51%).
  • FDA approval: Aug 2024. Toxicity: ↑ VTE, rash, infusion rxn, paronychia. Give VTE prophylaxis × first 4 mo.
  • SC amivantamab + hyaluronidase (Rybrevant Faspro): FDA Dec 17, 2025 across all amivantamab indications. PALOMA-3 (Leighl JCO 2024): SC vs IV amivantamab + lazertinib, non-inferior PK; markedly less infusion reaction (any-grade 13% vs 66%); 12-mo OS 65% vs 51% (HR 0.62). Administration time from hours to ~5 min.
• Historical combination: RELAY (PMID 31591063), ramucirumab (anti-VEGFR2) + erlotinib → mPFS 19.4 vs 12.4 mo (HR 0.59); brain mets excluded.
• Resistance to osimertinib (heterogeneous, no single dominant mechanism): EGFR C797S, MET amp, HER2 amp, SCLC transformation (RB1/TP53 loss), squamous transformation, off-target fusions. Resistance generally develops within ~2 yrs.
• After osimertinib progression:
  • Standard: platinum-doublet chemo. Continuing a first-gen TKI (gefitinib) with chemo is not beneficial (IMPRESS). Continuing osimertinib with platinum/pemetrexed after 1L osimertinib progression improved PFS (COMPEL: 8.4 vs 4.4 mo, HR 0.43); standard remains platinum-doublet chemo.
  • MARIPOSA-2: amivantamab + chemo (+/- lazertinib), improved PFS over chemo alone. FDA-approved 2L as amivantamab + carbo/pemetrexed (Sep 2024).
  • Datopotamab deruxtecan (Datroway), TROP2 ADC. FDA accelerated approval Jun 23, 2025 for EGFR-mut NSCLC after EGFR TKI + platinum chemo (pooled TROPION-Lung01/05; ORR 45%, mDOR 6.5 mo). Broader unselected indication NOT approved (TROPION-Lung01 missed OS); label is EGFR-mut-specific.
  • Patritumab deruxtecan (HER3-DXd), HER3 ADC. FDA complete response letter Jun 2024 (manufacturing inspection findings); HERTHENA-Lung02 phase 3 met PFS primary endpoint (HR 0.77) without OS benefit, and the BLA was voluntarily withdrawn May 2025. Not currently advancing in NSCLC.
  • SCLC transformation → treat as SCLC (platinum/etoposide); more common with inactivated RB1 and TP53.
• 1st/2nd-gen TKI progression with T790M: osimertinib restores response (AURA3: PFS 10.1 vs 4.4 mo, HR 0.30; RR 71% vs 31%; fewer AEs). T790M was found in ~half of resistant cases (analogous to T315I in CML).
• Afatinib (2nd-gen, irreversible): approved for nonresistant EGFR mutations including uncommon variants; more diarrhea/rash from targeting wild-type EGFR/HER2/HER4.

• Resistant to classic EGFR TKIs (osi, gefitinib, erlotinib). If de novo (untreated), preferred 1L is amivantamab + carboplatin/pemetrexed (PAPILLON), not chemo alone with the exon 20 agent reserved for progression.
• PAPILLON (NEJM 2023): 1L amivantamab + carbo/peme vs carbo/peme.
  • Efficacy: mPFS 11.4 vs 6.7 mo (HR 0.40); ORR 73% vs 47%.
  • FDA approval: Mar 2024, preferred 1L. Amivantamab is an EGFR-MET bispecific antibody (CHRYSALIS: ORR 40%, mPFS 8.3 mo; AEs infusion reactions 66%, paronychia 45%).
• Mobocertinib (oral exon 20 TKI, CHRYSALIS-era ORR ~25 to 28%): WITHDRAWN (failed phase 3 EXCLAIM-2).
• Zipalertinib: next-gen exon 20 TKI; NDA accepted by FDA Apr 2026.
• Sunvozertinib (Zegfrovy), next-gen EGFR ex20-selective TKI. FDA accelerated approval Jul 2, 2025 after prior platinum (WU-KONG6/1B; ORR ~46%, intracranial activity). Oral once-daily; distinct from withdrawn mobocertinib.

• Afatinib FDA-approved; osimertinib also active. NOT for exon 20 insertions.

• Biology: most commonly EML4-ALK from a chromosome 2p inversion. Profile: younger non/light smokers, adenocarcinoma. Detect by IHC, FISH, or NGS.
• 1L preferred, lorlatinib (CROWN 5-yr update, JCO 2024).
  • Efficacy: mPFS not reached vs 9.1 mo crizotinib (HR 0.19); 5-yr PFS 60% vs 8%; time to intracranial progression NR vs 16.4 mo; ORR 76% vs 58%.
  • Toxicity: CNS effects (anxiety, mood/personality change, cognitive), hyperlipidemia, edema. Frequently dose-modify, rarely discontinue.
  • Note: longest single-agent PFS reported in any solid tumor.
• Alternatives:
  • Alectinib (ALEX: mPFS 34.8 vs 10.9 mo vs crizotinib, HR 0.43). No ROS1 activity.
  • Brigatinib (ALTA-1L: PFS HR 0.49, intracranial RR 78% vs 29%). Pneumonitis risk on loading, start 90 mg × 7 d then 180 mg.
  • Ceritinib (ASCEND-4: 1L PFS 16.6 vs 8.1 mo vs chemo, HR 0.55).
  • Ensartinib (Ensacove), eXalt3; FDA Dec 18, 2024 for 1L ALK+ NSCLC.
  • Crizotinib: obsolete in 1L (PROFILE 1014: PFS 10.9 vs 7.0 mo vs chemo, HR 0.45).
• Resistance: ALK 2nd-site mutations (~30% post-crizotinib, mostly non-G1202R).
  • G1202R: toughest; uniquely targetable by lorlatinib.
  • After 1L crizotinib, switch to a later-generation ALK TKI (alectinib, brigatinib, ceritinib, lorlatinib). After 1L lorlatinib progression, routine switch to an earlier ALK TKI is not established; use platinum/pemetrexed or resistance-directed options (local therapy plus continued lorlatinib for oligoprogression).

• Biology: ROS1 on chromosome 6, sequence homology to ALK; younger nonsmokers, adenocarcinoma.
• 1L preferred, repotrectinib (NEJM 2024, TRIDENT-1).
  • Efficacy (TKI-naive): ORR 79%, mDOR 34 mo, mPFS ~36 mo. CNS active. Post-crizotinib ORR 38%.
  • FDA approval: Nov 2023.
• Alternatives:
  • Entrectinib: ORR 67%, mPFS 15.7 mo; CNS-active.
  • Crizotinib: ORR 72%, mPFS 19.2 mo, but poor CNS penetration (historical).
  • Ceritinib: option.
  • Taletrectinib (Ibtrozi): FDA Jun 11, 2025 for ROS1+ NSCLC, TKI-naive or pretreated (TRUST-I/II: ORR 85 to 90% TKI-naive, 52 to 62% post-TKI); CNS active; 600 mg daily.
  • Zidesamtinib (Jideytro): FDA Jul 22, 2026 for ROS1+ NSCLC after ≥1 prior ROS1 TKI (ARROS-1: ORR 44%).
• Pearl: not all ALK inhibitors treat ROS1, alectinib has NO ROS1 activity despite ALK potency.

• Biology: KRAS mutated in ~30% of cancers overall; 3 RAS isoforms (KRAS, NRAS, HRAS), KRAS ~84% of RAS-driven disease. G12C (glycine-to-cysteine at codon 12) is the MC KRAS mutation in lung adeno; covalent inhibitors bind the mutant cysteine and extend into the switch II pocket.
• 1L: chemo-IO (KRAS G12C inhibitors NOT yet first-line; frontline trials disappointing).
• 2L (post chemo-IO):
  • Sotorasib (Lumakras), CodeBreaK100 phase II ORR 37%, mPFS 6.8 mo; CodeBreaK 200 mPFS 5.6 mo, modestly better than docetaxel.
  • Adagrasib (Krazati), KRYSTAL-1 ORR 43%; KRYSTAL-12 mPFS 5.5 vs 3.8 mo; grade ≥3 TRAE similar to docetaxel (47% vs 46%), different profile (more GI/hepatic).
• Resistance: switch-II pocket mutations (Y96D confers resistance to both sotorasib and adagrasib), acquired non-G12C KRAS (G12D, G12V, G13D), RTK rebound.

• ~50% of BRAF-mutant lung adeno carry the actionable V600E (sensitive to BRAF-i + MEK-i).
• Dabrafenib + trametinib: 1L ORR 64%, mPFS 10.9 mo (PMID 28919011); 2L ORR 63%, mPFS 9.7 mo (PMID 27283860).
• Encorafenib + binimetinib (PHAROS, 2023): newer FDA-approved option.
• Toxicity: pyrexia is the classic combo toxicity, manage with hold/dose mod, NSAIDs.

• Capmatinib (GEOMETRY mono-1): 1L ORR 68%, mPFS 12.4 mo; 2L ORR 41%, PFS 5.4 mo.
• Tepotinib (VISION): ORR ~48 to 50%, mDOR 11.1 mo; watch peripheral edema and LFT elevation (~7%, onset ~30 d).
• Crizotinib: ORR 32%, mPFS 7.3 mo (historical).
• MET amplification (high-level GCN ≥10 in GEOMETRY mono-1): emerging biomarker; capmatinib has activity.
• Telisotuzumab vedotin (Emrelis), c-MET ADC. FDA accelerated approval May 14, 2025 for previously treated nonsquamous NSCLC with high c-MET protein overexpression (strong 3+ staining in ≥50% of tumor cells by VENTANA MET SP44); MET ex14 skipping or MET amplification alone does not qualify.

• Selpercatinib preferred (LIBRETTO-001 + LIBRETTO-431).
  • Efficacy: 1L (treatment-naive) ORR ~85% (LIBRETTO-001); the 17.5-mo mDOR was the pretreated cohort (ORR 64%); naive mDOR ~20 mo on update, mPFS ~24 mo; CNS active. LIBRETTO-431: selpercatinib > chemo +/- pembro (mPFS 24.8 vs 11.2 mo).
  • Toxicity: HTN, transaminitis.
• Pralsetinib (ARROW: ORR 65%): alternative; ↑ pneumonitis risk.
• Cabozantinib also inhibits RET (multikinase).

• Profile: most are exon 20 insertions in the TKD. ↑↑ brain mets at presentation.
• T-DXd (fam-trastuzumab deruxtecan) at 5.4 mg/kg q3wk (DESTINY-Lung02; DESTINY-Lung01 ORR 55%, mOS 17.8 mo): ORR 49 to 58%, mDOR 16.8 mo (5.4 mg/kg), mPFS ~10 mo.
  • FDA approval: 2L (after prior tx).
  • Critical AE: ILD ~12% (including grade 5), monitor with serial imaging.
• Zongertinib (Hernexeos), selective HER2 TKI. FDA accelerated approval Aug 2025 (2L), expanded to 1L Feb 2026 (Beamion LUNG-1: ORR 76% tx-naive).
• Sevabertinib (Hyrnuo): selective HER2 TKI (20 mg PO BID). FDA accelerated approval Nov 19, 2025 (previously treated), expanded to 1L Sep 9, 2026 for HER2 TKD-mutant nonsquamous NSCLC (SOHO-01: ORR 75% in tx-naive).
• Older option: T-DM1 (ado-trastuzumab emtansine), ORR 44%, mPFS 5 mo. Largely replaced by T-DXd.

• Larotrectinib (tumor-agnostic; ORR 75% across NTRK+ tumors), entrectinib (CNS-penetrant, ORR 57%, mPFS 11 mo), repotrectinib, tumor-agnostic FDA approvals.

• Zenocutuzumab (Bizengri), HER2/HER3 bispecific. FDA accelerated approval Dec 2024 for NRG1+ NSCLC or pancreatic ca after prior tx.

1L systemic, driver-negative metastatic NSCLC

• PD-L1 ≥50%:
  • Pembrolizumab monotherapy (KEYNOTE-024: mOS 30 vs 14 mo).
  • Atezolizumab mono (IMpower110), cemiplimab, alternatives.
  • Chemo-IO acceptable if high disease burden / symptomatic / urgent time-to-progression.
• PD-L1 1 to 49% or <1%: chemo-IO doublet preferred.
  • Squamous: pembrolizumab + carboplatin + paclitaxel/nab-paclitaxel (KEYNOTE-407).
  • Non-squamous: pembrolizumab + carboplatin + pemetrexed (KEYNOTE-189).
  • Ipi/nivo +/- 2 cycles chemo (CheckMate 9LA / 227): chemo-sparing IO option; less popular d/t toxicity.
  • Tremelimumab + durvalumab + chemo (POSEIDON): especially STK11/KEAP1 mutants where benefit may be enriched.
• Ivonescimab (PD-1/VEGF bispecific): HARMONi-2 in China, mPFS 11.1 vs 5.8 mo vs pembro mono in PD-L1+. NOT FDA-approved; BLA (+ chemo, EGFR-mut post-TKI, HARMONi) under FDA review, PDUFA Nov 14, 2026.

• Cisplatin vs carboplatin: meta-analyses show ~1 mo OS edge for cisplatin in metastatic, but carboplatin preferred (tolerability, QoL).
• Histology-driven:
  • Pemetrexed, ONLY non-squamous. Scagliotti phase III: cis/pemetrexed vs cis/gemcitabine, ↑ OS in adeno (12.6 vs 10.9 mo) and large cell (10.4 vs 6.7 mo) but WORSE in squamous (9.4 vs 10.8 mo); cis/pemetrexed had fewer grade 3-4 cytopenias, febrile neutropenia, alopecia (more grade 3-4 nausea).
  • Bevacizumab, avoid in squamous (hemoptysis) and recent significant hemoptysis; cavitation or central location is a relative risk, assess individually; consider if not an IO candidate.
  • Maintenance pemetrexed (after 4 cycles cis/peme), mOS benefit in non-squamous (PARAMOUNT: mOS 13.9 vs 11.0 mo, HR 0.78).

• Single-agent chemo (grade 2C): pemetrexed (non-squam, if not used), docetaxel 75 mg/m2 q3wk, gemcitabine.
• Docetaxel +/- ramucirumab (REVEL): mOS 10.5 vs 9.1 mo, PFS 4.5 vs 3.0 mo, ORR 23% vs 14%; for robust PS.
• Docetaxel + nintedanib (LUME-Lung 1): adenocarcinoma only.
• Pemetrexed vs docetaxel: similar 1-yr OS (~29.7%), pemetrexed with less febrile neutropenia, infection, neutropenic-fever hospitalization.
• Datopotamab deruxtecan: NOT approved for driver-negative NSCLC (TROPION-Lung01 missed OS); approval is EGFR-mutant only (see EGFR section).
• Pembrolizumab + ramucirumab (Lung-MAP S1800A signal): phase 3 Pragmatica-Lung (S2302) negative, no OS benefit vs SOC (mOS 10.1 vs 9.3 mo, HR 0.99).
• Oligoprogression on PD-1 (1 to 2 sites): local SBRT/thermal ablation/surgery to progressing site(s) + continue PD-1 may extend benefit (limited data).
• IO rechallenge: reasonable months/years after last PD-1 dose if prior response.

High-yield board pearls

• Always know molecular profile + PD-L1 BEFORE 1L in advanced NSCLC.
• EGFR/ALK+ → TKI (NOT IO). IO before TKI ↑ pneumonitis risk.
• Sublobar OK for peripheral N0 tumors ≤2 cm (CALGB 140503, JCOG0802); note ≤2 cm = 7th-ed cT1a, now T1a (≤1 cm) to T1b (>1 to 2 cm) in AJCC 8th/9th.
• ADAURA, adj osimertinib × 3 yr in EGFR+ resected IB to IIIA (OS HR 0.49).
• ALINA, adj alectinib × 2 yr in ALK+ resected IB (≥4 cm) to IIIA (DFS HR 0.24).
• Perioperative IO (CheckMate 77T, KEYNOTE-671, AEGEAN), for resectable II to IIIA, EGFR/ALK WT.
• PACIFIC, durva × 1 yr after concurrent chemoRT for unresectable III.
• LAURA, osimertinib post-chemoRT in EGFR+ unresectable III.
• Lorlatinib, durable 5-yr PFS 60% in 1L ALK+ (CROWN).
• FLAURA2 (osi + chemo) and MARIPOSA (ami + lazertinib), both with OS benefit in 1L EGFR+; choose by tox tolerance and pt phenotype.
• T-DXd ILD risk, monitor.
• Pancoast, neoadj chemoRT → surgery → adj chemo. Look for Horner.
• Avoid pneumonectomy after induction chemoRT (INT 0139).
• PORT no longer routine after R0 N2 resection (Lung-ART).
• USPSTF 2021 screening: 50 to 80 yo, ≥20 PY, current or quit ≤15 yr.
• Pemetrexed non-squamous only; bevacizumab not in squamous (hemoptysis).
• New 2024 to 2026 approvals: zongertinib (HER2 TKI), datopotamab deruxtecan (TROP2 ADC, EGFR), sunvozertinib (EGFR ex20), telisotuzumab vedotin (c-MET ADC), zenocutuzumab (NRG1 fusions), ensartinib (1L ALK).