What is the 22C3 immunohistochemistry assay and its role in assessing PD‑L1 expression for pembrolizumab eligibility?

What is the 22C3 Assay?

The 22C3 assay is an FDA-approved companion diagnostic immunohistochemistry (IHC) test that detects PD-L1 protein expression in tumor tissue to determine patient eligibility for pembrolizumab (Keytruda) immunotherapy across multiple cancer types.

Technical Specifications

The PD-L1 IHC 22C3 pharmDx assay uses a mouse monoclonal antibody (clone 22C3) that targets the extracellular domain of the PD-L1 protein 1. The assay was originally developed by Dako for use on the Autostainer Link 48 (ASL48) platform and has been analytically validated to demonstrate high sensitivity, specificity, and reproducibility 2.

Scoring Methods

The 22C3 assay employs two distinct scoring systems depending on the cancer type:

Tumor Proportion Score (TPS) - Used primarily in non-small cell lung cancer (NSCLC):

• Measures the percentage of viable tumor cells showing partial or complete membrane staining
• Clinically relevant cutoffs: <1%, 1-49%, and ≥50% 3, 4
• Required for first-line pembrolizumab monotherapy in NSCLC with TPS ≥50% 3

Combined Positive Score (CPS) - Used in multiple other cancer types:

• Calculates: (number of PD-L1 staining tumor cells + lymphocytes + macrophages) ÷ (total viable tumor cells) × 100
• Evaluates both tumor cells and tumor-associated immune cells in the tumor and surrounding stroma 1, 5
• Clinically relevant cutoffs vary by indication:
  • Triple-negative breast cancer: CPS ≥10 1
  • Cervical cancer: CPS ≥1 5
  • Head and neck squamous cell carcinoma: CPS ≥1 and ≥20 6
  • Gastroesophageal cancer: Variable cutoffs 7

FDA-Approved Indications

The 22C3 assay serves as a companion diagnostic (required testing) for:

• First-line pembrolizumab monotherapy in NSCLC with TPS ≥50% 3, 4
• Second-line pembrolizumab in NSCLC with TPS ≥1% 3
• Pembrolizumab plus chemotherapy in triple-negative breast cancer with CPS ≥10 1
• Pembrolizumab in cervical cancer with CPS ≥1 5

Critical Clinical Considerations

Assay-Specific Requirements

PD-L1 testing is highly assay-specific - different immunotherapy agents require different PD-L1 assays 1, 8. For example:

• Pembrolizumab requires the 22C3 assay
• Nivolumab uses the 28-8 assay
• Atezolizumab uses the SP142 assay
• Durvalumab uses the SP263 assay

When selecting an immune checkpoint inhibitor, you must obtain the specific assay used in the trial that supported that agent's approval 1. The 22C3, 28-8, and SP263 assays show high analytical concordance, but the SP142 assay stains fewer cells overall and is not interchangeable 3.

Laboratory-Developed Tests (LDTs)

While the FDA-approved 22C3 pharmDx kit is the gold standard, validated laboratory-developed tests using the 22C3 antibody concentrate on alternative platforms (Ventana BenchMark ULTRA, BenchMark XT) have demonstrated high concordance with the companion diagnostic 9, 10, 11, 12. These LDTs expand testing access but require rigorous validation:

• Intraclass correlation coefficient ≥0.88-0.95 across tumor types 12
• Overall percentage agreement 90-97% at clinically relevant cutoffs 12
• However, achieving diagnostic accuracy can be challenging, with only 40-75% of laboratories meeting acceptance criteria in validation studies 11

Specimen Considerations

Fresh tissue is strongly preferred over archival specimens for 22C3 testing. The glycosylated extracellular domain epitope recognized by 22C3 degrades over time, leading to:

• Significant reduction in positive staining in archival samples >3 years old 13
• Loss of signal intensity that can affect clinical scoring 13
• This degradation does not occur with antibodies targeting intracellular domains 13

Previously cut slides <3 months old may be acceptable 3, but fresh tissue blocks are optimal when available.

For patients with metastatic disease, there is no significant difference in CPS between primary tumors and metastatic sites in head and neck squamous cell carcinoma 14, suggesting either site can be tested. However, obtaining adequate tumor tissue with sufficient viable tumor cells is essential for accurate scoring.

Interobserver Variability

Pathologist agreement varies substantially by scoring method:

• Near-perfect agreement (κ = 0.652-1.0) when dichotomizing as positive vs. negative 14
• Fair agreement (κ = 0.318-0.371) when stratifying by therapeutic groups 14
• Slight agreement (κ = 0.025-0.117) for continuous CPS scores 14

This highlights the importance of using validated cutoffs and experienced pathologists for clinical decision-making.