Why PD-L1 Expression Matters: PD-L1 expression measured by immunohistochemistry (IHC) is the primary biomarker used to select NSCLC patients for immune checkpoint inhibitor therapy. Patients with high PD-L1 tumor proportion scores (TPS) are prioritized for pembrolizumab monotherapy, while lower PD-L1 scores influence combination regimen selection.
The Fixation Problem: Tissue fixation is required to preserve cellular morphology and antigenicity for IHC, but different fixatives interact differently with the PD-L1 protein. Formalin fixation followed by paraffin embedding (FFPE) is the gold standard, but many cytology specimens are fixed in alcohol - a fundamentally different chemical environment.
The Sampling Method Problem: NSCLC is frequently diagnosed from small specimens - fine-needle aspirates (FNA) or core biopsies obtained bronchoscopically or via CT guidance. These small samples may not represent the full spatial heterogeneity of PD-L1 expression compared to surgical resection specimens.
Study Design: This prospective study compared PD-L1 TPS across three matched specimen types - alcohol-fixed FNA, FFPE core biopsy, and FFPE surgical resection - from the same 57 NSCLC patients, enabling direct head-to-head comparison while controlling for inter-patient variability.
Patient Population: 57 NSCLC patients with available matched specimens from all three collection methods were prospectively enrolled. This design ensured that differences in PD-L1 staining reflected specimen type and fixation rather than true biological differences between patients.
Specimen Types: Alcohol-fixed FNA specimens were processed as cell blocks. Core needle biopsies and surgical resection specimens were processed as FFPE tissue blocks. All specimens underwent identical PD-L1 IHC staining using the 22C3 clone (the FDA-approved companion diagnostic for pembrolizumab).
Scoring Method: Two pathologists independently scored PD-L1 TPS on each specimen type, blinded to results from other specimen types. TPS was calculated as the percentage of viable tumor cells showing complete or partial membrane staining at any intensity.
Statistical Analysis: Intraclass correlation coefficients (ICC) and Bland-Altman plots assessed agreement between specimen types. Concordance rates for TPS cutoffs (1%, 50%) clinically relevant to treatment decisions were calculated.
Reduced TPS in Alcohol-Fixed FNAs: Alcohol-fixed FNA specimens showed systematically lower PD-L1 TPS compared to matched FFPE core biopsies and resection specimens. This difference was statistically significant and clinically meaningful, suggesting alcohol fixation attenuates PD-L1 antigenicity.
False-Negative Risk: In several cases, patients who would have been classified as PD-L1 positive (TPS >=1% or >=50%) based on their surgical specimen were misclassified as PD-L1 negative based on the alcohol-fixed FNA alone. This represents a potential false-negative scenario with direct therapeutic consequences.
Core Biopsy vs. Resection Concordance: FFPE core biopsies showed good concordance with surgical resection specimens at both the 1% and 50% TPS thresholds. This confirms that when proper FFPE fixation is used, small biopsy specimens can reliably reflect the surgical specimen PD-L1 score.
Inter-Observer Agreement: Pathologist agreement was high for FFPE specimens (ICC > 0.85) but modestly lower for alcohol-fixed cell blocks, where staining quality and tumor cellularity variations made scoring more challenging.
Treatment Assignment Errors: Since pembrolizumab monotherapy requires PD-L1 TPS >=50%, a patient with true TPS of 55% who receives an alcohol-fixed FNA scoring 35% could be incorrectly assigned to combination chemotherapy-immunotherapy rather than monotherapy, exposing them to unnecessary chemotherapy toxicity.
Guidance for Pathology Labs: These findings argue that pathology laboratories should avoid using alcohol-fixed cell blocks as the primary specimen for PD-L1 testing whenever FFPE material is available. Institutional protocols should specify preferred specimen types for PD-L1 IHC.
Re-Testing Recommendation: When a patient is diagnosed on an alcohol-fixed FNA with low or negative PD-L1, pathologists and oncologists should consider requesting a repeat biopsy with FFPE processing before making final treatment decisions if feasible.
Regulatory Implications: PD-L1 IHC companion diagnostics are validated primarily on FFPE tissue. The use of alternative fixatives and specimen types is technically off-label. This study provides prospective data supporting more explicit regulatory guidance on specimen type requirements.
Intratumoral PD-L1 Variability: Even among well-matched specimens from the same tumor, some variability in PD-L1 TPS was observed between different sampling locations. This reflects the known spatial heterogeneity of PD-L1 expression, which can vary by tumor zone, necrosis areas, and immune infiltration patterns.
Hotspot vs. Average Scoring: Pathologists tend to score the highest-expression region they can identify (hotspot approach) in small specimens, while resection specimens allow sampling of multiple areas. This may partially explain discordance between small biopsy and resection TPS at threshold boundaries.
Necrosis and Tumor Cellularity: FNA specimens with high necrosis fractions or low tumor cellularity were more likely to produce unreliable PD-L1 scores regardless of fixation. Pre-analytical quality metrics (such as minimum 100 viable tumor cells) should be enforced before reporting TPS.
Multiple Biopsy Strategy: For tumors with heterogeneous appearance on imaging, taking core biopsies from multiple sites (e.g., both central and peripheral zones) may provide a more representative PD-L1 TPS average, particularly for borderline cases near the 1% or 50% thresholds.
Sample Size: With 57 patients, the study is adequately powered to detect systematic fixation effects but may be underpowered to fully characterize all subgroup differences by histology, tumor stage, or driver mutation status. Larger prospective cohorts would provide more granular insights.
Single PD-L1 Clone Tested: The study used the 22C3 clone exclusively. Other commonly used clones (28-8, SP142, SP263) may interact differently with alcohol fixation. Multi-clone comparison studies on alcohol-fixed specimens are warranted.
Lack of Treatment Outcome Data: The study demonstrates fixation-related PD-L1 score differences but does not link these discrepancies to actual treatment assignment or clinical outcomes. A follow-up study tracking whether false-negative FNA results led to suboptimal treatment choices would provide clinical impact data.
Alternative Fixation Protocols: Investigation of whether modified fixation protocols, such as pre-treatment antigen retrieval optimization for alcohol-fixed specimens, could mitigate PD-L1 signal attenuation would be a practical next step for laboratories that primarily receive cytology specimens.