Computer-Aided Diagnostics Helps Accurately Determine Different Expression Levels of Claudin-18.2 in Gastric Cancer

Area of Science:

• Oncology
• Pathology
• Medical Imaging

Background:

• Claudin-18.2 expression assessment via immunohistochemistry (IHC) is crucial for zolbetuximab targeted therapy in gastric cancers (GCs).
• Interobserver variability in IHC scoring poses challenges for precise claudin-18.2 expression categorization.
• Computer-aided diagnosis (CAD) offers potential for enhanced diagnostic accuracy and reproducibility in pathology.

Purpose of the Study:

• To establish and validate a computer-aided analysis tool for quantifying claudin-18.2 positivity in gastric cancer.
• To improve the precision and reproducibility of claudin-18.2 expression assessment for patient stratification.

Main Methods:

• Development of a CAD tool for analyzing haematoxylin-3,3'-diaminobenzidine-stained tissue microarray (TMA) slides.
• The tool performs tumour tissue identification, cell segmentation, and membranous claudin-18.2 staining intensity estimation (clone 43-14A).
• Analysis of 2,248 cores from 417 primary resected GCs with available pathological data.

Main Results:

• No stained tumour cells were detected in 51.6% of TMA cores.
• Among positive cores, predominantly 1+ and 2+ staining was observed, with a minority of 3+ cells and uneven distribution.
• Utilizing the SPOTLIGHT threshold, 12% of cores and 2.5% of cases were identified as claudin-18.2 positive, with heterogeneity noted in tumour centres.

Conclusions:

• Computer-aided diagnostics accurately quantifies claudin-18.2 expression levels in gastric cancer.
• This tool enables precise determination of claudin-18.2 status for patient stratification in targeted therapy.
• Novel categorization of staining intensities may refine understanding of claudin-18.2 thresholds for patient selection.