Ultrasonic Elastography in Breast Cancer Diagnosis: Strain Ratio vs 5-point Scale
Received 12 April 2010; accepted 10 May 2010. published online 23 July 2010. Corrected Proof
Rationale and Objectives
The aim of this study was to develop a more reliable ultrasonic elastographic diagnostic method than a five-point scoring system by analyzing the difference in stiffness between benign and malignant breast lesions.
Materials and Methods
From January 2008 to April 2009, 559 solid lesions (415 benign, 144 malignant) in 437 consecutive patients (age range, 12–77 years) were examined using ultrasound elastography (UE). Final diagnosis was made on the basis of histopathologic findings. The strain ratios of the lesions were calculated. The area under the curve and cutoff point, both of which were obtained using receiver-operating characteristic curve analysis, were used to assess diagnostic performance. Diagnostic performance was further compared to that generated using a five-point scoring system with the z test. The sensitivity, specificity, and accuracy of these two evaluation systems were compared using McNemar's test.
Results
The strain ratios of benign lesions (mean, 1.83 ± 1.22) and malignant lesions (mean, 8.38 ± 7.65) were significantly different (P < .00001). When a cutoff point of 3.05 was introduced, UE had 92.4% sensitivity, 91.1% specificity, and 91.4% accuracy. The area under the curve for strain ratio–based elastographic analysis was 0.944, and the area under the curve for the five-point scoring system was 0.885. The diagnostic performance of strain ratio–based elastographic analysis was better than that of the five-point scoring system with UE (P < .05).
Conclusions
Strain ratio–based elastographic analysis can provide a new, more reliable diagnostic tool in comparison to a five-point scoring system for UE.
ABSTRACT