Skip to content

Biomarkers

TL;DR

  • A biomarker is clinically useful only when a defined result, from a defined specimen and assay, changes a decision. Prognostic association alone is not predictive utility.
  • PD-L1 tumour proportion score is the principal routine immune-selection marker in metastatic LUSC, but it is spatially heterogeneous, assay-dependent, treatment-sensitive, and imperfect at both high and low values.
  • Broad DNA/RNA profiling remains justified despite the lower prevalence of classic targetable drivers in LUSC: it can reveal rare targets, misclassified lineage, resistance biology, and trial eligibility.
  • Tumour mutational burden (TMB), KEAP1/NFE2L2 status, blood inflammatory indices, and spatial immune features are biologically informative but generally should not independently override validated treatment pathways.
  • Circulating tumour DNA (ctDNA) is useful for advanced-disease genotyping and is strongly prognostic after curative treatment. A negative plasma result never proves absence of a tumour alteration or residual disease because shedding and assay sensitivity vary.
  • Serum proteins such as CYFRA 21-1, CEA, and squamous-cell-carcinoma antigen are not substitutes for tissue diagnosis, CT screening, response imaging, or recurrence assessment.

A decision-first framework

Decision Validated or useful input What the result can do What it cannot do
Immunotherapy alone vs combination PD-L1 TPS on validated IHC platform Enrich for benefit from single-agent PD-1/PD-L1 therapy in metastatic, driver-negative disease Guarantee response or exclude benefit at TPS 0%
Targeted therapy or trial Tissue or plasma DNA plus RNA/fusion testing Detect actionable or eligibility-defining alteration Exclude an alteration after a non-shedding negative plasma test
Confirm lineage Morphology plus p40, TTF-1 and context-specific IHC Establish squamous differentiation and exclude mimics Identify the organ of origin by p40 alone
Estimate recurrence risk Stage, pathology; investigationally postoperative ctDNA Refine prognosis and research stratification Yet routinely dictate omission or escalation of curative therapy in all settings
Assess response Symptoms, examination, imaging; selected serial ctDNA Integrate anatomic and molecular change Declare progression from one noisy analyte
Screen an asymptomatic person Validated low-dose CT eligibility pathway Reduce lung-cancer mortality in high-risk groups Be replaced by serum tumour-marker panels

Every report should record specimen site and date, tumour content, intervening treatment, assay name/version, analytes, quality limits, and whether a negative finding is technically informative.

PD-L1 immunohistochemistry

PD-L1 is both the most useful and most frequently overinterpreted LUSC biomarker. Clinical trials tied decisions to particular antibodies, platforms, scoring compartments, and thresholds. The Blueprint projects found broadly comparable tumour-cell staining for 22C3, 28-8, and SP263 in many samples, while SP142 stained fewer tumour cells; immune-cell scoring was less concordant ([PMID 27913228](https://pubmed.ncbi.nlm.nih.gov/27913228/){target="_blank" rel="noopener"}; [PMID 29800747](https://pubmed.ncbi.nlm.nih.gov/29800747/){target="_blank" rel="noopener"}). “PD-L1 positive” without the assay, score type, and percentage is therefore incomplete.

Reading a PD-L1 result

Report element Question to ask
TPS 0%, 1–49%, or ≥50% Was TPS actually used, or an immune-cell/combined score from another indication?
Assay clone/platform Is it analytically and clinically linked to the contemplated regimen?
Adequacy Were enough viable tumour cells present and was necrosis avoided?
Site Primary, node, distant metastasis, or cytology cell block?
Timing Pretreatment, after radiation, after chemotherapy, or at acquired resistance?
Heterogeneity Could a small biopsy miss a high- or low-expression region?

Intratumour and intertumour heterogeneity are common; biopsy and resection specimens, or primary and metastatic sites, can disagree ([PMID 29049375](https://pubmed.ncbi.nlm.nih.gov/29049375/){target="_blank" rel="noopener"}; [PMID 34422362](https://pubmed.ncbi.nlm.nih.gov/34422362/){target="_blank" rel="noopener"}; [PMID 32030214](https://pubmed.ncbi.nlm.nih.gov/32030214/){target="_blank" rel="noopener"}). A result near a treatment threshold deserves particular attention to sampling and technical quality.

Clinical predictive value is probabilistic. KEYNOTE-407 improved survival across PD-L1 strata when pembrolizumab was added to chemotherapy ([PMID 30280635](https://pubmed.ncbi.nlm.nih.gov/30280635/){target="_blank" rel="noopener"}). CheckMate 017 found nivolumab benefit over docetaxel in previously treated squamous NSCLC regardless of PD-L1 expression ([PMID 26028407](https://pubmed.ncbi.nlm.nih.gov/26028407/){target="_blank" rel="noopener"}). Conversely, PD-L1-high monotherapy trials enrich for benefit but still contain primary progressors ([PMID 27718847](https://pubmed.ncbi.nlm.nih.gov/27718847/){target="_blank" rel="noopener"}).

Genomic profiling

The Cancer Genome Atlas demonstrated that LUSC is genomically complex, with recurrent alterations across TP53, CDKN2A, oxidative-stress, PI3K, squamous-differentiation, and chromatin pathways ([PMID 22960745](https://pubmed.ncbi.nlm.nih.gov/22960745/){target="_blank" rel="noopener"}). Most are not currently matched to an approved LUSC therapy. That distinction—altered versus actionable—is essential.

Finding class Current role Common error
Established sensitizing driver Use a guideline-linked targeted pathway where valid Assuming a squamous label makes testing unnecessary
FGFR alteration Trial eligibility; actionability depends on exact variant and context Treating amplification, mutation, and fusion as interchangeable
DDR2 mutation Research/clinical-trial hypothesis Equating preclinical sensitivity with proven benefit
PI3K-pathway alteration Trial stratification; may be subclonal Calling pathway activation a drug indication
KEAP1/NFE2L2 alteration Adverse biology and trial stratification Withholding checkpoint therapy solely on retrospective association
TP53 mutation Common lineage context; limited stand-alone actionability Treating prevalence as predictive value
High TMB Selected historical trial contexts and research Assuming all mutations generate presented, clonal neoantigens

Rare targetable findings can signal genuine LUSC, mixed histology, or a diagnostic reclassification. Never-smoker status, unusually young age, scant biopsy, or atypical morphology should lower the threshold for broad testing, not substitute for it.

Tissue and plasma are complementary

Tissue preserves morphology, tumour context, copy-number structure, and RNA; plasma samples DNA shed from multiple sites and can return quickly. In prospective studies, plasma sequencing shortened time to treatment and found clinically relevant alterations, while tissue retained unique findings ([PMID 35392653](https://pubmed.ncbi.nlm.nih.gov/35392653/){target="_blank" rel="noopener"}; [PMID 36201167](https://pubmed.ncbi.nlm.nih.gov/36201167/){target="_blank" rel="noopener"}). Prospective matched-therapy work shows feasibility but also the attrition between alteration detection and actual treatment ([PMID 30496436](https://pubmed.ncbi.nlm.nih.gov/30496436/){target="_blank" rel="noopener"}).

Plasma result Interpretation Next action
Actionable alteration detected at credible allele fraction Often sufficient to act if the assay and alteration are validated Correlate with phenotype; obtain tissue if diagnosis or transformation remains uncertain
No alteration detected, adequate sample Uninformative if tumour shedding is low Reflex to tissue; do not label “wild type”
Multiple low-fraction variants typical of haematopoiesis Possible clonal haematopoiesis Compare leukocyte DNA or use expert molecular review
Discordant plasma and tissue May reflect heterogeneity, timing, or analytic difference Review specimen dates, treatment interval, coverage, and variant details
Very high ctDNA burden Often adverse prognosis Do not intensify therapy solely from burden without validated rule

Different body fluids and tumour tissue yield different genomic profiles, reinforcing that “liquid biopsy” is not a unitary specimen ([PMID 30767431](https://pubmed.ncbi.nlm.nih.gov/30767431/){target="_blank" rel="noopener"}).

Tumour mutational burden

Smoking-associated LUSC frequently carries high mutational burden, supplying a plausible neoantigen substrate. In CheckMate 227, nivolumab plus ipilimumab improved PFS versus chemotherapy among NSCLC classified as high TMB in a prespecified analysis ([PMID 29658845](https://pubmed.ncbi.nlm.nih.gov/29658845/){target="_blank" rel="noopener"}). Later survival analyses and other programs complicated its role, and TMB was not required for the regimen's broader survival finding ([PMID 31562796](https://pubmed.ncbi.nlm.nih.gov/31562796/){target="_blank" rel="noopener"}).

Blood TMB in MYSTIC showed potential predictive stratification but depends on circulating tumour fraction, panel size, calibration, and threshold ([PMID 33355200](https://pubmed.ncbi.nlm.nih.gov/33355200/){target="_blank" rel="noopener"}). Tissue TMB is also sensitive to panel design, germline filtering, mutation types counted, and purity. TMB should be reported as an assay-specific quantitative result, not a universal biological constant.

Why high TMB can fail:

  • mutations may be subclonal rather than shared by all cancer cells;
  • altered peptides may not be expressed, processed, or presented;
  • HLA loss and antigen-presentation defects can hide neoantigens;
  • T-cell exclusion or exhaustion can dominate antigen abundance;
  • KEAP1/NFE2L2-linked metabolic states may create immune resistance;
  • immunosuppression, steroids, antibiotics, or comorbidity may alter host response.

ctDNA and molecular residual disease

Phylogenetic ctDNA work in TRACERx showed that plasma can track clonal and subclonal evolution in early lung cancer ([PMID 28445469](https://pubmed.ncbi.nlm.nih.gov/28445469/){target="_blank" rel="noopener"}). In localized lung cancer, postoperative molecular residual disease detection can precede radiographic relapse ([PMID 28899864](https://pubmed.ncbi.nlm.nih.gov/28899864/){target="_blank" rel="noopener"}). Later TRACERx analyses linked ctDNA patterns to metastatic dissemination and the biology of relapse ([PMID 37055640](https://pubmed.ncbi.nlm.nih.gov/37055640/){target="_blank" rel="noopener"}).

Prospective cohorts consistently show strong prognostic separation:

Study context Key contribution Limitation
Longitudinal resected NSCLC Dynamic ctDNA status stratified recurrence risk and explored adjuvant-chemotherapy benefit ([PMID 34799585](https://pubmed.ncbi.nlm.nih.gov/34799585/){target="_blank" rel="noopener"}) Nonrandomized treatment interaction
LUNGCA-1 Perioperative tumour-informed MRD detection in a multicentre cohort ([PMID 34844976](https://pubmed.ncbi.nlm.nih.gov/34844976/){target="_blank" rel="noopener"}) Assay timing and sensitivity constrain negative predictive value
Tumour-naïve presurgical assay Tested clinical utility without bespoke tumour sequencing ([PMID 38992468](https://pubmed.ncbi.nlm.nih.gov/38992468/){target="_blank" rel="noopener"}) Broader applicability can trade off sensitivity/specificity
Pilot pre/postoperative sampling Postoperative positivity was more prognostic than preoperative detection ([PMID 33209612](https://pubmed.ncbi.nlm.nih.gov/33209612/){target="_blank" rel="noopener"}) Small cohort and exploratory thresholds

The crucial distinction is prognostic versus treatment-predictive validation. A positive postoperative test identifies high risk; it does not by itself prove that a particular escalation improves survival. A negative result is reassuring only within the assay's sensitivity, sampled volume, timing, and tumour-shedding phenotype. LUSC may shed more readily than some indolent adenocarcinomas, but small or anatomically sequestered disease can still be invisible.

In CheckMate 816 follow-up, presurgical ctDNA clearance was associated with five-year OS of 75.0% versus 52.6% without clearance ([PMID 40454642](https://pubmed.ncbi.nlm.nih.gov/40454642/){target="_blank" rel="noopener"}). This is a response association, not randomized proof that changing treatment based on clearance improves survival.

Host and inflammatory markers

Neutrophil-to-lymphocyte ratio (NLR), derived NLR, platelet-to-lymphocyte ratio, lactate dehydrogenase, albumin, weight loss, and composite scores repeatedly correlate with outcomes. Meta-analyses associate high baseline NLR with worse survival on checkpoint inhibitors ([PMID 36013536](https://pubmed.ncbi.nlm.nih.gov/36013536/){target="_blank" rel="noopener"}; [PMID 31304800](https://pubmed.ncbi.nlm.nih.gov/31304800/){target="_blank" rel="noopener"}); LDH plus derived NLR also has prognostic value ([PMID 35924149](https://pubmed.ncbi.nlm.nih.gov/35924149/){target="_blank" rel="noopener"}).

These measures are cheap but nonspecific. Infection, corticosteroids, smoking-related inflammation, obstruction, cachexia, and marrow effects can all change them. They mostly describe prognosis and host state; they do not prove that immunotherapy is ineffective or identify which alternative works better.

Serum tumour markers

CYFRA 21-1 and squamous-cell-carcinoma antigen may be elevated in LUSC; CEA, CA125, NSE, and others can also vary. Reference intervals differ by age, sex, assay, geography, renal function, and smoking, as illustrated by a large multicentre reference-interval study ([PMID 33982344](https://pubmed.ncbi.nlm.nih.gov/33982344/){target="_blank" rel="noopener"}). Their overlap with benign disease and other cancers prevents use as a stand-alone diagnostic test.

Proposed use Assessment
Population screening Not validated; low-dose CT has outcome evidence, serum panels do not
Establishing LUSC Inadequate; tissue morphology and immunophenotype are required
Choosing immunotherapy Inadequate; use validated PD-L1 and clinical context
Monitoring response At most adjunctive when clearly elevated at baseline; imaging and clinical state govern
Detecting recurrence A rise may prompt evaluation but cannot locate or confirm recurrence

Emerging composite biomarkers

The next useful biomarker is likely composite rather than singular: clonal mutation burden, antigen presentation, PD-L1 geography, tumour–immune spatial organization, circulating immune state, microbiome, and on-treatment ctDNA dynamics. The statistical danger is severe: many features, small cohorts, shifting cut-offs, and treatment-confounded validation. A credible biomarker needs locked methods, an external cohort, calibration, clinically relevant discrimination, and—ideally—a trial showing that biomarker-guided treatment improves outcomes.

Open questions

  1. Can a composite LUSC immune-resistance score outperform PD-L1 without becoming too complex to reproduce?
  2. Which ctDNA time point after resection best balances false reassurance against actionable lead time?
  3. Does randomized escalation for MRD-positive disease improve survival, and can MRD-negative patients safely de-escalate?
  4. How should KEAP1/NFE2L2 status alter trial design without prematurely denying standard immunotherapy?
  5. Can spatial profiling be reduced to a robust pathology assay usable on small bronchoscopic samples?

References

  1. Hirsch FR, McElhinny A, Stanforth D, Ranger-Moore J, Jansson M, Kulangara K, et al. PD-L1 Immunohistochemistry Assays for Lung Cancer: Results from Phase 1 of the Blueprint PD-L1 IHC Assay Comparison Project. J Thorac Oncol. 2017;12(2):208-222. PMID 27913228
  2. Tsao MS, Kerr KM, Kockx M, Beasley MB, Borczuk AC, Botling J, et al. PD-L1 Immunohistochemistry Comparability Study in Real-Life Clinical Samples: Results of Blueprint Phase 2 Project. J Thorac Oncol. 2018;13(9):1302-1311. PMID 29800747
  3. Nakamura S, Hayashi K, Imaoka Y, Kitamura Y, Akazawa Y, Tabata K, et al. Intratumoral heterogeneity of programmed cell death ligand-1 expression is common in lung cancer. PLoS One. 2017;12(10):e0186192. PMID 29049375
  4. Shen X, Wang Y, Jin Y, Zheng Q, Shen L, Chen Y, et al. PD-L1 expression in non-small cell lung cancer: heterogeneity by pathologic types, tissue sampling and metastasis. J Thorac Dis. 2021;13(7):4360-4370. PMID 34422362
  5. Saito Y, Horiuchi S, Morooka H, Ibi T, Takahashi N, Ikeya T, et al. Inter-tumor heterogeneity of PD-L1 expression in non-small cell lung cancer. J Thorac Dis. 2019;11(12):4982-4991. PMID 32030214
  6. Paz-Ares L, Luft A, Vicente D, Tafreshi A, Gümüş M, Mazières J, et al. Pembrolizumab plus Chemotherapy for Squamous Non-Small-Cell Lung Cancer. N Engl J Med. 2018;379(21):2040-2051. PMID 30280635
  7. Brahmer J, Reckamp KL, Baas P, Crinò L, Eberhardt WE, Poddubskaya E, et al. Nivolumab versus Docetaxel in Advanced Squamous-Cell Non-Small-Cell Lung Cancer. N Engl J Med. 2015;373(2):123-35. PMID 26028407
  8. Reck M, Rodríguez-Abreu D, Robinson AG, Hui R, Csőszi T, Fülöp A, et al. Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer. N Engl J Med. 2016;375(19):1823-1833. PMID 27718847
  9. Cancer Genome Atlas Research Network. Comprehensive genomic characterization of squamous cell lung cancers. Nature. 2012;489(7417):519-25. PMID 22960745
  10. Sugimoto A, Matsumoto S, Udagawa H, Itotani R, Usui Y, Umemura S, et al. A Large-Scale Prospective Concordance Study of Plasma- and Tissue-Based Next-Generation Targeted Sequencing for Advanced Non-Small Cell Lung Cancer (LC-SCRUM-Liquid). Clin Cancer Res. 2023;29(8):1506-1514. PMID 36201167
  11. Thompson JC, Aggarwal C, Wong J, Nimgaonkar V, Hwang WT, Andronov M, et al. Plasma Genotyping at the Time of Diagnostic Tissue Biopsy Decreases Time-to-Treatment in Patients With Advanced NSCLC-Results From a Prospective Pilot Study. JTO Clin Res Rep. 2022;3(4):100301. PMID 35392653
  12. Sabari JK, Offin M, Stephens D, Ni A, Lee A, Pavlakis N, et al. A Prospective Study of Circulating Tumor DNA to Guide Matched Targeted Therapy in Lung Cancers. J Natl Cancer Inst. 2019;111(6):575-583. PMID 30496436
  13. Wu Z, Yang Z, Li CS, Zhao W, Liang ZX, Dai Y, et al. Differences in the genomic profiles of cell-free DNA between plasma, sputum, urine, and tumor tissue in advanced NSCLC. Cancer Med. 2019;8(3):910-919. PMID 30767431
  14. Hellmann MD, Ciuleanu TE, Pluzanski A, Lee JS, Otterson GA, Audigier-Valette C, et al. Nivolumab plus Ipilimumab in Lung Cancer with a High Tumor Mutational Burden. N Engl J Med. 2018;378(22):2093-2104. PMID 29658845
  15. Hellmann MD, Paz-Ares L, Bernabe Caro R, Zurawski B, Kim SW, Carcereny Costa E, et al. Nivolumab plus Ipilimumab in Advanced Non-Small-Cell Lung Cancer. N Engl J Med. 2019;381(21):2020-2031. PMID 31562796
  16. Si H, Kuziora M, Quinn KJ, Helman E, Ye J, Liu F, et al. A Blood-based Assay for Assessment of Tumor Mutational Burden in First-line Metastatic NSCLC Treatment: Results from the MYSTIC Study. Clin Cancer Res. 2021;27(6):1631-1640. PMID 33355200
  17. Abbosh C, Birkbak NJ, Wilson GA, Jamal-Hanjani M, Constantin T, Salari R, et al. Phylogenetic ctDNA analysis depicts early-stage lung cancer evolution. Nature. 2017;545(7655):446-451. PMID 28445469
  18. Chaudhuri AA, Chabon JJ, Lovejoy AF, Newman AM, Stehr H, Azad TD, et al. Early Detection of Molecular Residual Disease in Localized Lung Cancer by Circulating Tumor DNA Profiling. Cancer Discov. 2017;7(12):1394-1403. PMID 28899864
  19. Abbosh C, Frankell AM, Harrison T, Kisistok J, Garnett A, Johnson L, et al. Tracking early lung cancer metastatic dissemination in TRACERx using ctDNA. Nature. 2023;616(7957):553-562. PMID 37055640
  20. Qiu B, Guo W, Zhang F, Lv F, Ji Y, Peng Y, et al. Dynamic recurrence risk and adjuvant chemotherapy benefit prediction by ctDNA in resected NSCLC. Nat Commun. 2021;12(1):6770. PMID 34799585
  21. Xia L, Mei J, Kang R, Deng S, Chen Y, Yang Y, et al. Perioperative ctDNA-Based Molecular Residual Disease Detection for Non-Small Cell Lung Cancer: A Prospective Multicenter Cohort Study (LUNGCA-1). Clin Cancer Res. 2022;28(15):3308-3317. PMID 34844976
  22. Hong TH, Hwang S, Dasgupta A, Abbosh C, Hung T, Bredno J, et al. Clinical Utility of Tumor-Naïve Presurgical Circulating Tumor DNA Detection in Early-Stage NSCLC. J Thorac Oncol. 2024;19(11):1512-1524. PMID 38992468
  23. Ohara S, Suda K, Sakai K, Nishino M, Chiba M, Shimoji M, et al. Prognostic implications of preoperative versus postoperative circulating tumor DNA in surgically resected lung cancer patients: a pilot study. Transl Lung Cancer Res. 2020;9(5):1915-1923. PMID 33209612
  24. Forde PM, Spicer JD, Provencio M, Mitsudomi T, Awad MM, Wang C, et al. Overall Survival with Neoadjuvant Nivolumab plus Chemotherapy in Lung Cancer. N Engl J Med. 2025;393(8):741-752. PMID 40454642
  25. Platini H, Ferdinand E, Kohar K, Prayogo SA, Amirah S, Komariah M, et al. Neutrophil-to-Lymphocyte Ratio and Platelet-to-Lymphocyte Ratio as Prognostic Markers for Advanced Non-Small-Cell Lung Cancer Treated with Immunotherapy: A Systematic Review and Meta-Analysis. Medicina (Kaunas). 2022;58(8). PMID 36013536
  26. Xie X, Liu J, Yang H, Chen H, Zhou S, Lin H, et al. Prognostic Value of Baseline Neutrophil-to-Lymphocyte Ratio in Outcome of Immune Checkpoint Inhibitors. Cancer Invest. 2019;37(6):265-274. PMID 31304800
  27. Zhang Q, Gong X, Sun L, Miao L, Zhou Y. The Predictive Value of Pretreatment Lactate Dehydrogenase and Derived Neutrophil-to-Lymphocyte Ratio in Advanced Non-Small Cell Lung Cancer Patients Treated With PD-1/PD-L1 Inhibitors: A Meta-Analysis. Front Oncol. 2022;12:791496. PMID 35924149
  28. Li Y, Li M, Zhang Y, Zhou J, Jiang L, Yang C, et al. Age-stratified and gender-specific reference intervals of six tumor markers panel of lung cancer: A geographic-based multicenter study in China. J Clin Lab Anal. 2021;35(6):e23816. PMID 33982344