Programmed Cell Death Ligand 1 Expression in Esophageal Squamous Cell Carcinoma: A Comparative Analysis of Three Different Assays
Esophageal cancer is one of the most aggressive and lethal malignancies, with a 5-year survival rate of less than 20%. Among the two main pathological subtypes of esophageal cancer, esophageal squamous cell carcinoma (ESCC) is the predominant form in China, accounting for over 95% of cases. Recent advancements in cancer immunotherapy, particularly the discovery of drugs targeting specific immune checkpoints, have opened new avenues for treatment. Immunotherapy has shown promising results in squamous cell carcinoma, suggesting it could be a viable strategy for ESCC treatment in the future.
The expression of programmed cell death-ligand 1 (PD-L1) has been identified as a predictive diagnostic marker for selecting patients who may benefit from anti-PD-1 axis drugs. Currently, several qualitative detection antibodies are available to evaluate PD-L1 expression. The PD-L1 immunohistochemistry (IHC) 22C3 PharmDx kit, developed by Agilent Technologies, is the only companion diagnostic test approved by the USA Food and Drug Administration for pembrolizumab in non-small cell lung cancer (NSCLC). Additionally, the Dako antibody 28-8 and the SP263 PharmDx assay, developed for the Ventana BenchMark platform, have been approved for NSCLC treatment decisions. However, there is no widely accepted antibody for PD-L1 detection in ESCC, which affects the consistency of immunotherapy studies. Therefore, comparing the analytical performance and comparability of these methods in ESCC is crucial for standardizing immunotherapy for this cancer type.
This study aimed to compare the PD-L1 expression results of IHC 22C3 with those of Dako’s 28-8 and Ventana’s SP263 in ESCC. The study included 324 consecutive patients who underwent curative esophagectomy with R0 resection for histologically verified ESCC between December 2005 and June 2013 at the National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences. Patients who received pre-operative chemotherapy/radiotherapy and those with distant metastasis were excluded. The pathological classification of the primary tumor and the degree of lymph node metastasis were assessed according to the American Joint Committee on Cancer (AJCC) Cancer Staging Manual (eighth edition). All ESCC tissue samples were stained with hematoxylin and eosin and confirmed by two pathologists independently.
Tissue microarrays (TMAs) were performed with three validated assays: staining for PD-L1 22C3 assay and 28-8 was performed on the Dako Autostainer Link 48 platform, while staining for PD-L1 SP263 assay was performed on the Ventana Benchmark XT Stainer. Two pathologists, blinded to the clinical data, assessed the samples independently, with disagreements resolved by a third experienced pathologist. The combined positive score (CPS) results were divided into three groups: less than 1, 1 to 49, and 50 to 100 positive cells.
Statistical analyses were performed using SPSS 18.0 software. The overall percent agreement (OPA), positive predictive value (PPV), negative predictive value (NPV), and area under the curve (AUC) were calculated to evaluate the clinical performance of the assays. The median age of the patients was 59 years, and 23.8% were women. Immunohistochemical staining in a representative ESCC sample from the TMAs showed different staining scores for the same tumor, with relatively high inter-observational consistency (agreement higher than 90%).
When comparing the 28-8 and 22C3 assays at the “1” cutoff, 77.5% (251/324) and 77.2% (250/324) of patients were negative (CPS less than 1), respectively; 19.1% (62/324) and 19.4% (63/324) had CPS values of 1 to 49, respectively; and 3.4% (11/324) and 3.1% (10/324) were strongly positive (CPS 50 or greater), respectively. The OPA between 28-8 and 22C3 was 93.5% at the “1” cutoff and 99.1% at the “50” cutoff. At the “1” cutoff, the sensitivity and specificity were 85.1% and 96.0%, with a PPV and NPV of 86.3% and 95.6%, respectively, and an AUC of 0.906. At the “50” cutoff, the sensitivity and specificity were 90.0% and 99.4%, with a PPV and NPV of 81.8% and 99.7%, respectively, and the AUC was 0.947.
For the SP263 and 22C3 assays, the OPA was 81.5% at the “1” cutoff, with a difference in 18.5% of the samples. The sensitivity and specificity were 79.7% and 82.0%, respectively, with a PPV and NPV of 56.7% and 93.2%, respectively, and an AUC of 0.809. At the “50” cutoff, the OPA was 96.0%, with a sensitivity of 80.0% and a specificity of 96.5%. The PPV and NPV were 42.1% and 99.3%, respectively, and the AUC was 0.882.
Overall, the 28-8 and 22C3 IHC staining scores were highly consistent. For SP263, the CPS tended to be higher than those of the other two assays, with strong and intense membrane staining. This is the first study to assess the concordance of the PD-L1 IHC 22C3 assay with the SP263 and 28-8 assays in the TMA of ESCC. Previous studies in NSCLC suggested that the PD-L1 clone 28-8 and 22C3 displayed strong correlation across samples, and similar results were observed in the CPSs of ESCC when comparing 22C3 with 28-8 or SP263 at 1 and 50 cutoffs. Therefore, the detection of PD-L1 expression with the 28-8 assay may be appropriately used in place of the 22C3 assay for guiding therapy with anti-PD-1/PD-L1 in ESCC.
Medical centers often face challenges in running two different staining platforms simultaneously on limited samples due to the lack of multiple automatic detection systems. Additionally, different antibodies have different antigen epitopes, which affect the consistency of IHC staining intensity. Antibodies with high sensitivity used for PD-L1 detection may be beneficial in reducing false negatives in small biopsy specimens. The CPSs of the 28-8 assay were similar to those of 22C3, but the SP263 assay showed a higher rate of PD-L1-positive cases (both at 1 and 50 cutoffs) when compared with 22C3, possibly due to the high-intensity staining of SP263 on the cell membrane. The precise reason for this discrepancy is not yet clear, and the definition of positivity will depend on the clinical situation and the intended treatment.
Considering the tumor heterogeneity of esophageal cancer and lung cancer, the expression of PD-L1 in this study was affected by the use of TMAs rather than whole sections. Tumors may heterogeneously express PD-L1, which could lead to an overestimation or underestimation of the true PD-L1 levels. Although the results demonstrate that the 28-8 assay shows high agreement with the 22C3 assay, the efficacy of these antibodies in detecting PD-L1 expression for guiding therapy with PD1/PD-L1 inhibitors in ESCC still requires further research.
doi.org/10.1097/CM9.0000000000001642
Was this helpful?
0 / 0