What is CPS in PD-1 Therapy?
CPS (Combined Positive Score) is a biomarker scoring method that measures PD-L1 expression by counting all PD-L1-positive cells (tumor cells, lymphocytes, and macrophages) divided by total tumor cells, multiplied by 100—it is the preferred method over TPS for selecting patients who will benefit from PD-1 inhibitor therapy. 1
Definition and Calculation
CPS is calculated as:
- (Number of PD-L1+ tumor cells + PD-L1+ lymphocytes + PD-L1+ macrophages) / Total number of viable tumor cells × 100
This differs fundamentally from Tumor Proportion Score (TPS), which only counts the percentage of PD-L1-positive tumor cells 1, 2.
Why CPS is Superior to TPS
CPS has emerged as the superior biomarker because it captures immune cell PD-L1 expression in addition to tumor cell expression, making it more sensitive and prognostic for immunotherapy response. 1, 3
Key evidence supporting CPS superiority:
- In KEYNOTE-040 post hoc analysis, CPS showed equivalent performance to TPS at cutoffs ≥50, but CPS was significantly more sensitive than TPS at lower cutoffs (≥1) 1
- For gastric and gastroesophageal junction cancers, CPS is considered "a more prognostic biomarker for immunotherapy efficacy" than TPS 3
- CPS ≥1 prevalence in gastric cancer was 57.6% with reasonable response enrichment (odds ratio 2.8), while TPS ≥1% prevalence was only 12.5% with minimal enrichment (odds ratio 1.4) 2
Clinical Cutoff Values
The optimal CPS threshold varies by cancer type and treatment line, but CPS ≥1 is preferred for maximum sensitivity, with higher cutoffs (≥5 or ≥10) providing greater predictive value:
Head and Neck Cancer:
- CPS ≥1 is the preferred lower threshold to maximize sensitivity 1
- Both CPS ≥20 and ≥1 were associated with superior OS in KEYNOTE-048, leading to adoption of the lower cutoff 1
Gastroesophageal Cancers:
- CPS ≥5: Strong recommendation for nivolumab plus chemotherapy 3
- CPS ≥10: Used for pembrolizumab plus chemotherapy in esophageal squamous cell carcinoma with high-quality evidence 3
- CPS ≥1: May be used as threshold when TPS is unavailable (91% of patients in CheckMate 648 had CPS ≥1) 3
Evidence-Based Thresholds:
Meta-analysis data suggests CPS ≥5 provides optimal predictive value for OS and CPS ≥10 for PFS in gastroesophageal adenocarcinomas 4. Patients with CPS <5 may not derive significant benefit from ICI therapy 4.
Clinical Application Algorithm
For treatment decision-making, follow this hierarchy:
If CPS is available: Use CPS with cancer-specific cutoffs
If only TPS is available:
If neither available: Consider TMB testing (TMB ≥10 correlates with PD-1 inhibitor benefit) 1
Reproducibility and Practical Considerations
CPS demonstrates excellent reproducibility with 96.6% interpathologist agreement and 97.2% intrapathologist agreement 2. This makes it a robust companion diagnostic for pembrolizumab approval.
Important Caveats:
- Heterogeneity exists: PD-L1 expression shows high concordance (86.5%) between different blocks of the same specimen, but moderate concordance (78%) between primary and metastatic lesions 5
- Sample adequacy matters: Intratumoral heterogeneity, temporal changes, and sample size can impact assessment 1
- Positive inflammatory cells drive variability: PD-L1 heterogeneity relates mainly to mononuclear inflammatory cells, which has clinical implications 5
Relationship to Clinical Outcomes
Higher CPS values correlate with better immunotherapy outcomes:
- In ESCC meta-analysis, HR for OS was 0.60 (95% CI 0.51-0.70) for CPS ≥10 versus 0.83 (95% CI 0.69-1.00) for CPS <10 (p for interaction 0.009) 6
- Pembrolizumab plus chemotherapy in ESCC with CPS ≥10 showed HR 0.57 for OS and 0.53 for PFS with high-quality evidence 3
- CPS positivity correlates with lymph node metastasis (59.2%) and venous/lymphatic invasion (66.3%) 5
The key clinical message: CPS is not just a binary positive/negative marker—higher scores predict progressively better responses to PD-1 inhibitors, making it essential for risk-benefit discussions with patients.