Colour Doppler sonography in breast tumours: an update

Area of Science:

• Medical Imaging
• Oncology
• Diagnostic Ultrasound

Background:

• Computer-assisted analysis of color Doppler ultrasound images offers quantitative assessment.
• Previous studies suggested high diagnostic performance for image-based quantification in breast lesion characterization.

Purpose of the Study:

• To reassess the diagnostic performance of computer-assisted color Doppler quantification (using mean color value [MCV] and color pixel density [CPD]) under routine clinical conditions.
• To evaluate the reproducibility and reliability of this quantitative method in breast cancer diagnosis.

Main Methods:

• A computer-assisted protocol was used to analyze color Doppler images from 44 patients with breast carcinoma and 49 with benign breast lesions.
• Quantitative parameters, including MCV and CPD, were assessed over two distinct periods by two different examiners.
• Diagnostic performance metrics (sensitivity and specificity) were calculated for each period.

Main Results:

• In the first period, MCV showed 92% sensitivity and 75% specificity; CPD showed 80% sensitivity and 81% specificity.
• In the second period, MCV sensitivity dropped to 58% (specificity 77%); CPD sensitivity decreased to 68% (specificity 71%).
• A notable decline in diagnostic performance was observed during the second study period.

Conclusions:

• The diagnostic performance of computer-assisted color Doppler quantification, specifically MCV and CPD, may decrease under routine clinical conditions.
• Variability in performance suggests challenges in maintaining consistent diagnostic accuracy with this method in everyday practice.
• Further research is needed to optimize and standardize computer-assisted ultrasound quantification for reliable breast lesion assessment.