Novel image analysis approach for quantifying expression of nuclear proteins assessed by immunohistochemistry: application to measurement of oestrogen and progesterone receptor levels in breast cancer

2008

Automated Analysis of Hormone Receptors in Breast Cancer

Sample size: 743 publication 10 minutes Evidence: high

Author Information

Author(s): Elton Rexhepaj, Donal J Brennan, Peter Holloway, Elaine W Kay, Amanda H McCann, Goran Landberg, Michael J Duffy, Karin Jirstrom, William M Gallagher

Primary Institution: University College Dublin

Hypothesis

Can an automated image analysis approach improve the quantification of oestrogen and progesterone receptor levels in breast cancer compared to manual methods?

Conclusion

The study demonstrates that an automated approach for quantifying ER and PR levels in breast cancer is effective and may provide better prognostic indicators than current manual methods.

Supporting Evidence

The automated algorithm showed an excellent correlation with manual analysis (Spearman's ρ = 0.9).
7% positive tumor cells for ER and 5% for PR were identified as optimal thresholds for predicting treatment response.
The study involved 743 breast cancer patients, providing a robust sample size for analysis.
Automated analysis reduced intra-observer variability compared to manual scoring.
The algorithm was validated by a histopathologist on 18 representative images.

Takeaway

Researchers created a computer program to help doctors measure important proteins in breast cancer more accurately, which could help patients get better treatment.

Methodology

The study used two cohorts of breast cancer patients and developed an automated algorithm to analyze digital images of tissue samples for ER and PR expression.