Performance of computer-aided diagnosis in the interpretation of lesions on breast sonography

123

Purpose-To retrospectively investigate the effect of using a custom-designed computer classifier on radiologists' sensitivity and specificity for discriminating malignant masses from benign masses on three-dimensional (3D) volumetric ultrasonographic (US) images, with histologic analysis serving as the reference standard.Materials and Methods-Informed consent and institutional review board approval were obtained. Our data set contained 3D US volumetric images obtained in 101 women (average age, 51 years; age range, 25-86 years) with 101 biopsy-proved breast masses (45 benign, 56 malignant). A computer algorithm was designed to automatically delineate mass boundaries and extract features on the basis of segmented mass shapes and margins. A computer classifier was used to merge features into a malignancy score. Five experienced radiologists participated as readers. Each radiologist read cases first without computer-aided diagnosis (CAD) and immediately thereafter with CAD. Observers' malignancy rating data were analyzed with the receiver operating characteristic (ROC) curve.Results-Without CAD, the five radiologists had an average area under the ROC curve (A z ) of 0.83 (range, 0.81-0.87). With CAD, the average A z increased significantly (P = .006) to 0.90 (range, 0.86-0.93). When a 2% likelihood of malignancy was used as the threshold for biopsy recommendation, the average sensitivity of radiologists increased from 96% to 98% with CAD, while the average specificity for this data set decreased from 22% to 19%. If a biopsy recommendation threshold could be chosen such that sensitivity would be maintained at 96%, specificity would increase to 45% with CAD.Conclusion-Use of a computer algorithm may improve radiologists' accuracy in distinguishing malignant from benign breast masses on 3D US volumetric images. © RSNA, 2007 Address correspondence to B.S. (berki@umich.edu).. Author contributions: Guarantors of integrity of entire study, B.S., H.P.C.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, B.S., H.P.C., M.A.R., L.M.H.; clinical studies, M.A.R., C.P.; experimental studies, B.S., M.A.R., M.A.H., C.P., J.B., A.V.N., C.B.; statistical analysis, B.S., H.P.C., L.M.H.; and manuscript editing, B.S., H.P.C., M.A.R., M.A.H.Supplemental material: http://radiology.rsnajnls.org/cgi/content/full/2423051464/DC1 Authors stated no financial relationship to disclose. NIH Public Access Author ManuscriptRadiology. Author manuscript; available in PMC 2010 January 2. NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author ManuscriptIn current clinical practice, the rate of positive biopsy results for breast cancer is about 15%-30% (1-3). To reduce patient anxiety and morbidity and to decrease health care costs, it is desirable to reduce the number of benign biopsy results without missing malignant lesions.Resul...

Section: Discussionsupporting

confidence: 87%

Section: Discussionmentioning

confidence: 99%

Malignant and Benign Breast Masses on 3D US Volumetric Images: Effect of Computer-aided Diagnosis on Radiologist Accuracy

Sahiner¹,

Chan²,

Roubidoux³

et al. 2007

123

“…In addition, several observer studies of all three of these modalities have revealed the potential usefulness of CAD in clinical settings (36)(37)(38)(39)(40). Thus, in this study, our aim was to investigate whether our CAD method can be extended from diagnostic to prognostic tasks to ultimately yield MR imagingbased prognostic markers.…”

Section: Discussionmentioning

confidence: 99%

Cancerous Breast Lesions on Dynamic Contrast-enhanced MR Images: Computerized Characterization for Image-based Prognostic Markers

Bhooshan¹,

Giger²,

Jansen³

et al. 2010

177

155

Purpose:To assess the performance of computer-extracted dynamic contrast material-enhanced (DCE) magnetic resonance (MR) imaging kinetic and morphologic features in the differentiation of invasive versus noninvasive breast lesions and metastatic versus nonmetastatic breast lesions. Materials and Methods:In this institutional review board-approved HIPAA-compliant study, in which the requirement for informed patient consent was waived, breast MR images were retrospectively collected. The images had been obtained with a 1.5-T MR unit by using a gadodiamide-enhanced T1-weighted spoiled gradient-recalled acquisition in the steady state sequence. The breast MR imaging database contained 132 benign, 71 ductal carcinoma in situ (DCIS), and 150 invasive ductal carcinoma (IDC) lesions. Fifty-four IDC lesions were associated with metastasis-positive lymph nodes (LNs), and 64 IDC lesions were associated with negative LNs. Lesion segmentation and extraction of morphologic and kinetic features were automatically performed by a laboratory-developed computer workstation. Features were fi rst selected by using stepwise linear discriminant analysis and then merged by using Bayesian neural networks. Lesion classifi cation performance was assessed with receiver operating characteristic analysis. Results: Conclusion:Computer-aided diagnosis of breast DCE MR imagingdepicted lesions was extended from the task of discriminating between malignant and benign lesions to the prognostic tasks of distinguishing between noninvasive and invasive lesions and discriminating between metastatic and nonmetastatic lesions, yielding MR imaging-based prognostic markers.q RSNA, 2010 Supplemental material: http://radiology.rsna.org/lookup/ suppl

“…Various groups (40)(41)(42)(43)(44) have assessed the use of expert versus community radiologists; however, a BI-RADS descriptor-based CAD system such as that described herein has not, to our knowledge, been tested on community radiologists, who often read a smaller volume of mammograms or have fewer years of experience in breast imaging. Both of these characteristics have been associated with increased variability and false-positive rates (45)(46)(47)(48)(49)(50).…”

Section: Participating Physicians and Lesion Descriptionmentioning

confidence: 99%

Computer-aided Classification of Breast Masses: Performance and Interobserver Variability of Expert Radiologists versus Residents

Singh¹,

Maxwell²,

Baker³

et al. 2011

Purpose:To evaluate the interobserver variability in descriptions of breast masses by dedicated breast imagers and radiology residents and determine how any differences in lesion description affect the performance of a computer-aided diagnosis (CAD) computer classifi cation system. Materials and Methods:Institutional review board approval was obtained for this HIPAA-compliant study, and the requirement to obtain informed consent was waived. Images of 50 breast lesions were individually interpreted by seven dedicated breast imagers and 10 radiology residents, yielding 850 lesion interpretations. Lesions were described with use of 11 descriptors from the Breast Imaging Reporting and Data System, and interobserver variability was calculated with the Cohen k statistic. Those 11 features were selected, along with patient age, and merged together by a linear discriminant analysis (LDA) classifi cation model trained by using 1005 previously existing cases. Variability in the recommendations of the computer model for different observers was also calculated with the Cohen k statistic. Results:A signifi cant difference was observed for six lesion features, and radiology residents had greater interobserver variability in their selection of fi ve of the six features than did dedicated breast imagers. The LDA model accurately classifi ed lesions for both sets of observers (area under the receiver operating characteristic curve = 0.94 for residents and 0.96 for dedicated imagers). Sensitivity was maintained at 100% for residents and improved from 98% to 100% for dedicated breast imagers. For residents, the computer model could potentially improve the specifi city from 20% to 40% ( P , .01) and the k value from 0.09 to 0.53 ( P , .001). For dedicated breast imagers, the computer model could increase the specifi city from 34% to 43% ( P = .16) and the k value from 0.21 to 0.61 ( P , .001). Conclusion:Among fi ndings showing a signifi cant difference, there was greater interobserver variability in lesion descriptions among residents; however, an LDA model using data from either dedicated breast imagers or residents yielded a consistently high performance in the differentiation of benign from malignant breast lesions, demonstrating potential for improving specifi city and decreasing interobserver variability in biopsy recommendations.q RSNA, 2010 1