PD-L1 Combined Positive Score Concordance and Reproducibility in Head and Neck Squamous Cell Carcinoma: A Systematic Review and Meta-analysis
In brief
PD-L1 scores matched across tumor specimens in 70% of paired samples
Across five studies, paired primary and nodal, recurrent, or metastatic specimens matched exactly in combined positive score category about 70% of the time. Three other studies found assay platforms agreed 96% on whether scores met the threshold of 1 or higher, but these results are not directly comparable; limited or remote samples can still miss a positive result.
- Journal
- Laboratory investigation; a journal of technical methods and pathology (Q1)
- Published
- 19 November 2026
- Study design
- Systematic review of cohort studies
- Evidence level
- Level 2, Moderate (CEBM 2a)
- Authors
- Chenglu Ruan, Lin Wang
- PMID
- 43057578
- DOI
- 10.1016/j.labinv.2026.106182
Why clinicians should know about it
- Picked for Pathology and Forensic Medicine (top studies of the week, 22 November 2026): PD-L1 assay concordance, reader reproducibility, and specimen variability
Abstract
PURPOSE: Programmed death-ligand 1 (PD-L1) combined positive score (CPS) guides pembrolizumab use in recurrent or metastatic head and neck squamous cell carcinoma, but results may vary with the assay workflow, reader, and specimen. We evaluated the analytical concordance and reproducibility of CPS testing and separated technical, observer, and sampling effects. MATERIALS AND METHODS: PubMed and Europe PMC were searched from inception through August 28, 2026. Original human studies comparing PD-L1 assays, platforms, protocols, observers, digital methods, or paired specimens were eligible. Case-level agreement proportions were pooled only within prespecified, clinically homogeneous domains by logit random-effects models with Paule-Mandel variance estimates and modified Hartung-Knapp confidence intervals. RESULTS: Twenty-seven studies were included; eight contributed independent effects to two separate meta-analyses. Three studies of assay or platform classification at CPS >=1 yielded pooled agreement of 95.6% (95% CI, 83.4%-99.0%). Five studies of exact agreement across CPS <1, CPS 1-19, and CPS >=20 in paired primary and nodal, recurrent, or metastatic specimens yielded pooled agreement of 70.3% (95% CI, 62.0%-77.4%). These estimates were not directly comparable because the study sets and category definitions differed. Observer agreement improved after indication-specific training. Small biopsies and cytology cell blocks had high positive predictive value but limited negative predictive value. CONCLUSIONS: Validated assay-platform substitutions can be concordant at CPS >=1, but the evidence is based on few independent cohorts and does not establish universal interchangeability. Specimen-related variability remains clinically important. Laboratories should validate the complete workflow, monitor category-level reader agreement, and consider additional tissue when a negative or borderline result is obtained from limited or clinically remote material.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.