Human Epidermal Growth Factor Receptor 2 Quantification Using Computational Pathology to Identify Novel Biomarkers

Ansh Kapil1, Henrik Failmezger1, Yu Fu1

  • 1AstraZeneca Computational Pathology GmbH, Oncology Research and Development, Munich, Germany.

Abstract

Insights

Quantitative continuous scoring (QCS) identifies new biomarkers for trastuzumab deruxtecan (T-DXd) therapy in gastric cancer. These biomarkers predict patient survival, improving treatment selection for HER2-positive cancers.

Area of Science:

  • Oncology
  • Computational Pathology
  • Biomarker Discovery

Background:

  • Trastuzumab deruxtecan (T-DXd) is approved for HER2-positive gastric cancer (GC) and gastroesophageal junction adenocarcinoma (GEJA).
  • Current eligibility for T-DXd relies on conventional HER2 immunohistochemistry (IHC) scoring.
  • Quantitative continuous scoring (QCS) offers precise HER2 quantification for improved T-DXd treatment selection.

Purpose of the Study:

  • To identify novel biomarker signatures predictive of T-DXd response using HER2-QCS.
  • To validate these signatures in independent patient cohorts.
  • To explore QCS as a quantitative method for identifying patients who may benefit from T-DXd therapy.

Main Methods:

  • HER2-QCS scoring was applied to identify biomarker signatures in the DESTINY-Gastric01 cohort.
  • External validation of identified signatures was performed using the DESTINY-Gastric02 cohort.
  • Survival analysis using Cox regression and log-rank tests evaluated biomarker performance.

Main Results:

  • QCS-derived continuous spatial proximity score (cSPS) and density of HER2-positive cells in tumor epithelium (DPC) showed significant survival benefits.
  • Biomarker-positive patients demonstrated superior outcomes compared to biomarker-negative patients in the T-DXd arm (P < .0001).
  • These findings were consistently validated in the DESTINY-Gastric02 cohort (log-rank P < .0001).

Conclusions:

  • cSPS and DPC are significant predictors of T-DXd clinical outcomes in GC.
  • QCS presents a promising quantitative approach for identifying patients likely to benefit from T-DXd.
  • This study supports the potential integration of QCS into clinical practice for personalized GC treatment.