Background
CLDN18 is an actionable biomarker in gastrointestinal (GI) cancers, primarily assessed by immunohistochemistry (IHC), although interpretation may be limited by tumour heterogeneity and variable membranous staining. Transcriptomic profiling may offer a quantitative complementary approach for CLDN18 assessment.
Methods
CLDN18 expression was evaluated across 748 GI tumour samples and correlated with clinically relevant biomarkers, including IHC-based PD-L1 and HER2, MMR/MSI status, TMB, and MUC1 gene co-expression, using Exacta testing. IHC–mRNA concordance was assessed in 78 paired samples comprising esophagogastric (n=34), pancreatic (n=12), CRC (n=11), gastric (n=10), biliary/ampullary (n=8), and other GI malignancies (n=3). CLDN18 IHC positivity was defined as ≥75% moderate-to-strong membranous staining. CLDN18 mRNA positivity was defined using a threshold of log₂FC >2 with FDR <0.05.
Results
Transcriptomic profiling showed strong concordance with CLDN18 IHC, with 75.9% sensitivity, 81.6% specificity, 71.0% PPV, 85.1% NPV, and 79.5% overall concordance. CLDN18 mRNA-positivity significantly correlated with CLDN18 IHC-positivity (Fisher’s exact p=7.49×10⁻⁷; OR=13.97). Notably, nine IHC-negative samples showed elevated CLDN18 transcript expression, highlighting the relevance of transcriptomic profiling as a complementary approach.
Across 748 samples, CLDN18 was upregulated in 23.5% of tumours and was significantly associated with PD-L1 positivity, dMMR/MSI-H status, and MUC1 upregulation. No significant associations were observed with HER2 or TMB.
Association of CLDN18 Expression with Immuno-Molecular Biomarkers
|
Biomarker |
CLDN18-positive |
CLDN18-negative |
Significance |
|
PD-L1 |
35.7% |
17.1% |
p=6.56×10⁻⁶; OR=2.69 |
|
dMMR/MSI-H |
4.4% |
0.9% |
p=0.012; OR=5.23 |
|
MUC1 |
51.7% |
13.3% |
p=5.75×10⁻²⁴; OR=6.99 |
Conclusion
Integrated RNA- and protein-based assessment demonstrated high concordance while identifying additional CLDN18-positive tumours not detected by IHC. CLDN18 upregulation was significantly associated with PD-L1 positivity, dMMR/MSI-H status, and MUC1 upregulation, suggesting enrichment of a distinct immuno-molecular subset, potentially representing MUC1-driven tumours linked to chemoresistance and limited targetable options. These findings support transcriptomic profiling as a potentially valuable adjunct to IHC for refining CLDN18 biomarker assessment in GI malignancies.