Analytical Validation of the PreciseDx Digital Prognostic Breast Cancer Test in Early-Stage Breast Cancer

Area of Science:

• Digital pathology
• Artificial intelligence in medicine
• Oncology

Background:

• PreciseDx Breast (PDxBr) is a digital diagnostic test designed for early-stage breast cancer.
• It aims to predict the recurrence of invasive breast cancer (IBC) within a six-year timeframe.

Purpose of the Study:

• To perform analytical validation of the PDxBr test's automated annotation and digital histologic features.
• To assess the impact of methodologic variability on the composite risk score.
• To evaluate the precision, repeatability, and reproducibility of the assay.

Main Methods:

• Utilized hematoxylin and eosin-stained whole slide images of IBC.
• Employed an artificial intelligence-enabled morphology feature array to generate microanatomic features.
• Combined morphometric attributes with clinical data (age, tumor size, stage, lymph node status) for risk prediction.
• Conducted studies on precision, repeatability, reproducibility, and interference, including multi-operator and multi-day assessments.

Main Results:

• High precision was achieved in tumor segmentation (0.98), lymphocyte detection (0.91-0.93), and mitotic figure identification (0.84-0.85).
• Assay risk score variation for reproducibility and repeatability was less than 2%.
• No interference was observed from variations in staining or tumor thickness, indicating assay robustness.

Conclusions:

• The analytical validation confirms strong performance of the digital IBC risk assessment test.
• The assay accurately predicts recurrence in early-stage invasive breast cancer patients.
• PDxBr is a reliable tool for assessing recurrence risk in early-stage breast cancer.