Related Experiment Videos
Predicting breast cancer response to neoadjuvant chemotherapy using pretreatment diffuse optical spectroscopic
William T Tran1,2, Mehrdad J Gangeh1, Lakshmanan Sannachi1,3
1Department of Radiation Oncology, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada.
British Journal of Cancer
|April 19, 2017
Summary
Texture analysis of diffuse optical spectroscopy (DOS) images before neoadjuvant chemotherapy (NAC) can predict breast cancer response. Pretreatment DOS texture features accurately differentiate between responders and non-responders, potentially guiding treatment decisions for locally advanced breast cancer (LABC).
Area of Science:
- Medical Imaging
- Biophysics
- Oncology
Background:
- Diffuse optical spectroscopy (DOS) shows promise for monitoring neoadjuvant chemotherapy (NAC) response in locally advanced breast cancer (LABC).
- Texture features from pretreatment DOS functional maps are evaluated for predicting treatment outcomes.
Purpose of the Study:
- To assess the predictive capability of texture features derived from pretreatment DOS parametric maps for LABC response to NAC.
- To identify specific texture features that can differentiate between responders and non-responders.
Main Methods:
- 37 LABC patients underwent pretreatment DOS imaging before NAC.
- Texture analyses were performed on parametric maps using grey-level co-occurrence matrices.
- Leave-one-out cross-validation with logistic regression, naive Bayes, and k-nearest neighbor classifiers was employed.
Main Results:
- Pretreatment DOS texture features effectively differentiated treatment response outcomes.
- HbO2 homogeneity showed the highest accuracy (87.8%) with 86.5% sensitivity and 89.0% specificity.
- Multivariate features, like Hb-contrast + HbO2-homogeneity, achieved 79.5% accuracy.
Conclusions:
- Pretreatment DOS texture features can predict breast cancer response to NAC.
- These findings suggest DOS texture analysis may help guide personalized treatment strategies for LABC.