Skip to content

Open questions — lung squamous-cell carcinoma

Last curated: 2026-08-30. These are concrete questions exposed by the assembled evidence. Stable IDs are retained across updates; a resolved question moves to Closed with the closing evidence and date. Evidence-gap statements below were re-searched in PubMed and ClinicalTrials.gov on 2026-08-30.

How to read priority

Tier 1 questions could change practice or reorder the field and can be addressed with present methods. Tier 2 questions matter but depend on a new cohort, assay, earlier answer, or longer follow-up. Tiers are curator judgment.

Dots not yet connected

Each row joins two evidence bodies that coexist in this knowledge base but have not yet been adequately connected.

# Dot A Dot B Missing junction Powers
D1 TCGA found frequent pathway alteration and extraordinary genomic complexity ([PMID 22960745](https://pubmed.ncbi.nlm.nih.gov/22960745/){target="_blank" rel="noopener"}) Lung-MAP matched-target cohorts produced only 10 responses among 143 treated patients ([PMID 33125909](https://pubmed.ncbi.nlm.nih.gov/33125909/){target="_blank" rel="noopener"}) Patient-level analysis separating genomic presence, functional dependency, drug exposure, and resistance is missing OQ-1, OQ-2
D2 NFE2L2/KEAP1 alterations define redox biology and adverse outcome associations ([PMID 20534738](https://pubmed.ncbi.nlm.nih.gov/20534738/){target="_blank" rel="noopener"}; [PMID 39111731](https://pubmed.ncbi.nlm.nih.gov/39111731/){target="_blank" rel="noopener"}) A 2026-08-30 registry search found redox/metabolic trials, including recruiting DRP-104 study NCT07249372, but no validated redox-directed treatment or routine assay-selection rule A prospective, assay-locked, biomarker–drug trial with pharmacodynamic confirmation has not closed the chain OQ-2
D3 KEYNOTE-407 has a five-year survival tail of 18.4% ([PMID 36735893](https://pubmed.ncbi.nlm.nih.gov/36735893/){target="_blank" rel="noopener"}) PD-L1 is spatially variable and incomplete as a predictor ([PMID 30280635](https://pubmed.ncbi.nlm.nih.gov/30280635/){target="_blank" rel="noopener"}) A 2026-08-30 PubMed search found no prospective, locked pretreatment LUSC-specific signature that distinguishes durable benefit from primary resistance OQ-3
D4 Post-PD-(L)1 CTLA-4 rescue in Lung-MAP had 0% response in acquired-resistant and 7% in primary-resistant cohorts ([PMID 34429332](https://pubmed.ncbi.nlm.nih.gov/34429332/){target="_blank" rel="noopener"}) Modern subsequent-line evidence still rests heavily on pre-chemo-immunotherapy comparators Resistance-state biopsies have not been joined to randomized sequence selection OQ-4
D5 NELSON and NLST reduce overall lung-cancer mortality ([PMID 31995683](https://pubmed.ncbi.nlm.nih.gov/31995683/){target="_blank" rel="noopener"}; [PMID 21714641](https://pubmed.ncbi.nlm.nih.gov/21714641/){target="_blank" rel="noopener"}) Central CT-occult LUSC exists in bronchoscopy referral cohorts ([PMID 38443800](https://pubmed.ncbi.nlm.nih.gov/38443800/){target="_blank" rel="noopener"}) Trial datasets have not adequately reported LUSC-specific sensitivity, interval cancers, location, and stage shift OQ-5
D6 High-grade airway lesions can carry 3q amplification and progress ([PMID 20299530](https://pubmed.ncbi.nlm.nih.gov/20299530/){target="_blank" rel="noopener"}) Intensive autofluorescence surveillance has not shown an outcome advantage in low-grade lesions ([PMID 35236723](https://pubmed.ncbi.nlm.nih.gov/35236723/){target="_blank" rel="noopener"}) A 2026-08-30 PubMed search found mechanistic and chemoprevention studies but no externally validated clinical model that selects a small group for interception while sparing low-risk people serial bronchoscopy OQ-6
D7 CheckMate 816 achieves pCR without mandatory adjuvant immunotherapy ([PMID 35403841](https://pubmed.ncbi.nlm.nih.gov/35403841/){target="_blank" rel="noopener"}) Perioperative trials add a postoperative checkpoint phase ([PMID 39288781](https://pubmed.ncbi.nlm.nih.gov/39288781/){target="_blank" rel="noopener"}; [PMID 37870974](https://pubmed.ncbi.nlm.nih.gov/37870974/){target="_blank" rel="noopener"}) A 2026 multicentre retrospective pCR cohort found no DFS advantage for adjuvant therapy (HR 1.22, 95% CI 0.51–2.95), but no randomized component isolates its marginal benefit and toxicity ([PMID 42055888](https://pubmed.ncbi.nlm.nih.gov/42055888/){target="_blank" rel="noopener"}) OQ-7
D8 pCR and ctDNA clearance strongly associate with long-term outcome ([PMID 40454642](https://pubmed.ncbi.nlm.nih.gov/40454642/){target="_blank" rel="noopener"}) Neither is validated to direct escalation or de-escalation Response-guided randomization is missing, especially for LUSC OQ-7, OQ-8
D9 LUSC is disproportionately shaped by smoking and comorbidity Pivotal systemic trials largely enroll ECOG 0–1 populations There is no robust causal framework separating tumour-driven poor performance status from irreversible frailty for regimen selection OQ-9
D10 Stigma is associated with delayed help-seeking and impaired communication ([PMID 24769603](https://pubmed.ncbi.nlm.nih.gov/24769603/){target="_blank" rel="noopener"}; [PMID 26553030](https://pubmed.ncbi.nlm.nih.gov/26553030/){target="_blank" rel="noopener"}) A systematic review found 11 intervention studies, mostly pilots, and a clinician-training pilot improved provider-reported outcomes ([PMID 39197098](https://pubmed.ncbi.nlm.nih.gov/39197098/){target="_blank" rel="noopener"}; [PMID 38836528](https://pubmed.ncbi.nlm.nih.gov/38836528/){target="_blank" rel="noopener"}) As of 2026-08-30, trials NCT05456841, NCT06191939, NCT03750864, and NCT06286592 confirm an intervention pipeline, but no adequately powered result connects reduced stigma to diagnostic interval, treatment uptake, or survival OQ-10
D11 Financial toxicity changes over time and affects households ([PMID 33555936](https://pubmed.ncbi.nlm.nih.gov/33555936/){target="_blank" rel="noopener"}; [PMID 38775918](https://pubmed.ncbi.nlm.nih.gov/38775918/){target="_blank" rel="noopener"}) Trials report efficacy and adverse events far more consistently than time, travel, and household cost Comparative treatment value has not integrated financial and time toxicity into LUSC outcomes OQ-11
D12 Molecular screening succeeds technically in most Lung-MAP submissions ([PMID 33125909](https://pubmed.ncbi.nlm.nih.gov/33125909/){target="_blank" rel="noopener"}) Fewer than half of assigned patients registered to a substudy The operational causes of assignment-to-treatment attrition are not linked to redesign experiments OQ-12

When a junction study appears, update the row and sharpen or close the powered question.

Tier 1 — highest-value directions

OQ-1. What distinguishes a genomic alteration from a therapeutic dependency in LUSC?

TCGA documented a mutation- and copy-number-rich tumour, yet pathway prevalence has repeatedly failed to predict drug response ([PMID 22960745](https://pubmed.ncbi.nlm.nih.gov/22960745/){target="_blank" rel="noopener"}). FGFR1 amplification was common in an early cohort ([PMID 21160078](https://pubmed.ncbi.nlm.nih.gov/21160078/){target="_blank" rel="noopener"}), but dovitinib and AZD4547 response rates remained low ([PMID 27315356](https://pubmed.ncbi.nlm.nih.gov/27315356/){target="_blank" rel="noopener"}; [PMID 31195180](https://pubmed.ncbi.nlm.nih.gov/31195180/){target="_blank" rel="noopener"}). The needed study combines quantitative copy number, expression, phosphoprotein state, spatial clones, pharmacokinetics, on-treatment target modulation, and outcome—not another binary amplification basket. → molecular landscape, failed and frontier targets

OQ-2. Can redox-state LUSC be therapeutically attacked?

NFE2L2/KEAP1 is among the most distinctive LUSC pathways, with protein and genomic studies associating its state with adverse outcomes ([PMID 20534738](https://pubmed.ncbi.nlm.nih.gov/20534738/){target="_blank" rel="noopener"}; [PMID 39111731](https://pubmed.ncbi.nlm.nih.gov/39111731/){target="_blank" rel="noopener"}). A 2026-08-30 registry search found active metabolic/redox studies, including NCT07249372, but no analytically validated clinical classifier or randomized intervention with demonstrated target engagement. A study must predefine NFE2L2 mutation, KEAP1 loss, expression, and functional redox state rather than treating them as interchangeable. → molecular landscape, biomarkers

OQ-3. Who will become a durable immunotherapy survivor before treatment begins?

KEYNOTE-407 doubled five-year OS from 9.7% to 18.4%, but most patients did not reach that landmark ([PMID 36735893](https://pubmed.ncbi.nlm.nih.gov/36735893/){target="_blank" rel="noopener"}). PD-L1 enriches some monotherapy populations but does not identify all combination benefit ([PMID 30280635](https://pubmed.ncbi.nlm.nih.gov/30280635/){target="_blank" rel="noopener"}). A locked, externally validated model should integrate spatial immune state, tumour genetics, HLA, blood markers, microbiome/exposures only where reproducible, and early kinetics—then demonstrate decision utility, not merely association. → immunotherapy, biomarkers

OQ-4. What is the best sequence after platinum–taxane–checkpoint therapy?

The dominant first-line regimen changed, while many later-line estimates still come from patients never exposed to chemo-immunotherapy. Empiric CTLA-4 rescue showed little activity after PD-(L)1 resistance ([PMID 34429332](https://pubmed.ncbi.nlm.nih.gov/34429332/){target="_blank" rel="noopener"}); docetaxel–ramucirumab has randomized pre-immunotherapy evidence ([PMID 24933332](https://pubmed.ncbi.nlm.nih.gov/24933332/){target="_blank" rel="noopener"}), and an ADC trial showed an unfavourable LUSC subgroup signal ([PMID 39250535](https://pubmed.ncbi.nlm.nih.gov/39250535/){target="_blank" rel="noopener"}). A pragmatic LUSC-specific platform should randomize by resistance phenotype and include symptom, function, and time-to-treatment-failure outcomes. → systemic therapy, clinical-trials landscape

OQ-5. How well does LDCT screening detect LUSC specifically?

NLST and NELSON establish overall lung-cancer mortality benefit ([PMID 21714641](https://pubmed.ncbi.nlm.nih.gov/21714641/){target="_blank" rel="noopener"}; [PMID 31995683](https://pubmed.ncbi.nlm.nih.gov/31995683/){target="_blank" rel="noopener"}), but central flat or endobronchial lesions can be CT-occult ([PMID 38443800](https://pubmed.ncbi.nlm.nih.gov/38443800/){target="_blank" rel="noopener"}). Re-analysis of trial and program images should report histology-specific sensitivity, interval cancers, central/peripheral location, stage distribution, and competing smoking-related mortality. This can be done before launching new invasive screening. → screening and early detection, central-airway biology

OQ-6. Can high-risk bronchial premalignancy be intercepted without over-treatment?

High-grade lesions and 3q gain predict progression in small longitudinal samples ([PMID 20299530](https://pubmed.ncbi.nlm.nih.gov/20299530/){target="_blank" rel="noopener"}), while repeated autofluorescence bronchoscopy in low-grade lesions did not improve outcome ([PMID 35236723](https://pubmed.ncbi.nlm.nih.gov/35236723/){target="_blank" rel="noopener"}). A prospective model should combine grade, persistence, copy-number state, field biomarkers, smoking trajectory, and airway imaging; only then should an interception therapy be tested against surveillance with invasive-procedure and quality-of-life harms captured. → central-airway biology, screening

OQ-7. Who needs postoperative immunotherapy after neoadjuvant chemo-immunotherapy?

CheckMate 816 improved EFS and pCR without a protocol-mandated adjuvant checkpoint phase ([PMID 35403841](https://pubmed.ncbi.nlm.nih.gov/35403841/){target="_blank" rel="noopener"}). KEYNOTE-671 and AEGEAN used perioperative designs and therefore cannot isolate the postoperative component ([PMID 39288781](https://pubmed.ncbi.nlm.nih.gov/39288781/){target="_blank" rel="noopener"}; [PMID 37870974](https://pubmed.ncbi.nlm.nih.gov/37870974/){target="_blank" rel="noopener"}). A response-stratified randomized discontinuation design could test adjuvant value after pCR, non-pCR, and ctDNA clearance, while measuring immune toxicity and treatment time. → early-stage and perioperative therapy

OQ-8. Can ctDNA safely direct escalation or de-escalation?

Presurgical clearance is strongly prognostic in CheckMate 816 follow-up, but association does not prove that altering therapy improves outcome ([PMID 40454642](https://pubmed.ncbi.nlm.nih.gov/40454642/){target="_blank" rel="noopener"}). Assay sensitivity in low-shedding disease, clonal haematopoiesis, optimal time points, and salvage lead time remain unresolved. The closing experiment is an assay-locked randomized strategy trial, not another retrospective survival curve. → biomarkers, early-stage therapy

OQ-9. How should evidence be generated for ECOG 2–3, frail, and multimorbid patients?

The population carrying LUSC commonly has COPD, vascular disease, frailty, and smoking-related organ impairment, while pivotal trials largely select fitter patients. The key distinction is whether functional decline is tumour-reversible or comorbidity-driven. Prospective pragmatic cohorts and randomized reduced-intensity strategies should use geriatric measures, competing risks, symptom response, hospitalization, and days at home—not performance status alone. → systemic therapy, red flags and safety

OQ-10. Can reducing lung-cancer stigma improve clinical outcomes?

Stigma is associated with help-seeking delay and communication ([PMID 24769603](https://pubmed.ncbi.nlm.nih.gov/24769603/){target="_blank" rel="noopener"}; [PMID 26553030](https://pubmed.ncbi.nlm.nih.gov/26553030/){target="_blank" rel="noopener"}). Intervention evidence now includes small empirical studies and registered randomized work, but a 2024 systematic review judged most studies to be pilots and called for larger randomized evaluations ([PMID 39197098](https://pubmed.ncbi.nlm.nih.gov/39197098/){target="_blank" rel="noopener"}). Cluster-randomized service interventions should combine clinician language feedback, public messaging, cessation care, navigation, and patient-reported discrimination, then measure diagnostic intervals, treatment uptake, distress, and retention. → patient experience and advocacy

OQ-11. Do patient-weighted outcomes reorder treatment choices?

Survival and conventional toxicity tables omit time in clinic, caregiver work, financial strain, breathlessness, neuropathy, cognition, and days at home. Financial toxicity is longitudinal ([PMID 33555936](https://pubmed.ncbi.nlm.nih.gov/33555936/){target="_blank" rel="noopener"}), while direct LUSC PRO evidence shows group-level QoL can be measured ([PMID 31751163](https://pubmed.ncbi.nlm.nih.gov/31751163/){target="_blank" rel="noopener"}). A patient-partnered core outcome set and re-analysis of comparative trials could reveal whether regimen rankings change. → patient experience, systemic therapy

OQ-12. How can master protocols convert assignment into treatment and completion?

Lung-MAP returned genomic results for 90.9% of tissue submissions and assigned 83.9% of those with results, yet only 46.7% of assigned patients registered to a substudy ([PMID 33125909](https://pubmed.ncbi.nlm.nih.gov/33125909/){target="_blank" rel="noopener"}). The exact losses—clinical deterioration, slot availability, eligibility, travel, physician choice, patient preference, or assay delay—need authoritative classification and prospective workflow experiments. Platform success should include completed, interpretable cohorts and participant-reported burden, not screening volume alone. → clinical-trials landscape

Tier 2 — enabling and narrower questions

Biology and pathology

Staging and curative therapy

Advanced treatment and safety

  • OQ-21. Can any antibody–drug conjugate overcome the unfavourable squamous signal seen with datopotamab deruxtecan? Target abundance, internalization, payload sensitivity, lung toxicity, and histology must be co-developed rather than extrapolated from all-NSCLC results ([PMID 39250535](https://pubmed.ncbi.nlm.nih.gov/39250535/){target="_blank" rel="noopener"}). → failed and frontier targets
  • OQ-22. When does local treatment of oligoprogression extend immunotherapy benefit? The randomized oligometastatic evidence is small and predates current systemic therapy ([PMID 31067138](https://pubmed.ncbi.nlm.nih.gov/31067138/){target="_blank" rel="noopener"}). Modern trials need progression topology and time off systemic therapy. → systemic therapy
  • OQ-23. Can high-risk immune toxicity be predicted without denying beneficial therapy? Autoimmune disease, transplant, interstitial lung disease, frailty, and steroid need are underrepresented; registries need causal comparators and organ-specific outcomes. → red flags
  • OQ-24. What prevents histology-inappropriate pemetrexed or antiangiogenic use in practice? The efficacy interaction and pulmonary-haemorrhage concern are established, but real-world wrong-regimen frequency, causes, and electronic prevention controls are poorly quantified. → chemotherapy constraints

Equity and lived experience

  • OQ-25. What is the LUSC-specific global burden by stage and mortality? Current global models estimate incidence, while robust histology-specific mortality, stage, and survival data remain incomplete ([PMID 39914442](https://pubmed.ncbi.nlm.nih.gov/39914442/){target="_blank" rel="noopener"}). → epidemiology
  • OQ-26. Which patients are systematically absent from LUSC evidence? ECOG 3–4, low-resource, rural, non-English-speaking, uninsured, and caregiver-poor populations require explicit participation denominators and transportability analyses. → patient experience
  • OQ-27. What instrument captures central-airway fear and function? Existing lung-cancer PROs capture symptoms but may not represent haemoptysis fear, suffocation fear, and episodic obstruction; qualitative item generation should precede validation. → patient experience
  • OQ-28. Can early palliative-care benefit be reproduced across contemporary treatment and health-system contexts? The landmark trial is strong but single-centre and from an earlier treatment era ([PMID 20818875](https://pubmed.ncbi.nlm.nih.gov/20818875/){target="_blank" rel="noopener"}). Implementation trials should include caregiver and days-at-home outcomes. → systemic therapy

Closed

(None yet. Move questions here only when decisive evidence closes the stated design, with the closing citation and date.)