How to Agree on a CTC: Evaluating the Consensus in Circulating Tumor Cell Scoring

Area of Science:

• Oncology
• Biotechnology
• Medical Imaging

Background:

• Accurate enumeration and characterization of circulating tumor cells (CTCs) are essential for clinical decision-making.
• Current CTC detection technologies yield a wide range of reported values, hindering inter-institutional comparability.
• Standardization is needed to ensure the reliability of CTC counts in clinical practice.

Purpose of the Study:

• To introduce an open-source CTC scoring tool to improve reviewer consistency and consensus in CTC identification.
• To assess the concordance of CTC identification among multiple reviewers using standardized images.
• To explore the potential of deep learning for automated CTC enumeration.

Main Methods:

• Development and application of an open-source CTC scoring tool.
• Evaluation of reviewer concordance using 100 images from two different CTC detection platforms.
• Assessment of inter-reviewer variability with 15 reviewers and an expert panel.
• Demonstration of a deep learning model for CTC identification on the same image set.

Main Results:

• Significant variability was observed among reviewers in classifying objects as CTCs.
• The study highlights the need for automated or computer-assisted CTC recognition to reduce subjectivity.
• A deep learning approach demonstrated potential for accurate and reproducible CTC enumeration.

Conclusions:

• The developed open-source tool facilitates comparison and consensus among CTC scorers.
• Automated methods, such as deep learning, are promising for enhancing the accuracy and consistency of CTC analysis.
• Standardized CTC enumeration is critical for reliable clinical application and decision support.