Rationale and Objectives
The aim of this study was to evaluate the performance of a computer-aided diagnosis (CAD) system for breast ultrasound to improve the characterization of breast lesions detected on ultrasound by junior and senior radiologists.
Materials and Methods
One hundred sixty ultrasound breast lesions were randomly reviewed blindly by four radiologists with different levels of expertise (from 20 years [radiologist A] to 4 months [radiologist D]), with and without the help of an ultrasound CAD system (B-CAD version 2). All lesions had been biopsied. Sensitivity and specificity with and without CAD were calculated for each radiologist for the following evaluation criteria: Breast Imaging Reporting and Data System category and the final diagnosis (benign or malignant). Intrinsic sensitivity, specificity, and predictive values of CAD alone were also calculated.
Results
CAD detected all cancers, and its use increased radiologists’ sensitivity scores when this was possible (with vs without CAD: radiologist A, 99% vs 99%; radiologist B, 96% vs 87%; radiologist C, 95% vs 88%; radiologist D, 91% vs 88%). Seven additional cancers were diagnosed. However, the low specificity of CAD (48%) decreased the specificity of radiologists, especially of the more experienced among them (with vs without CAD: radiologist A, 46% vs 70%; radiologist B, 58% vs 80%; radiologist C, 57% vs 69%; radiologist D, 71% vs 71%).
Conclusions
CAD for breast ultrasound appears to be a useful tool for improving the diagnosis of malignant lesions for junior radiologists. Nevertheless, its low specificity must be taken into account to limit biopsies of benign lesions.
In clinical practice, breast ultrasound is the main complementary examination to mammography, both for detecting breast lesions and for lesion characterization. This is the method of choice for differentiating cysts from solid lesions . Stavros et al were the first to describe specific ultrasound criteria for differentiating benign and malignant features of solid lesions.
Ultrasound is a readily available, nonradiating, inexpensive, well-tolerated technique that allows interventional procedures , but the main limitation is operator dependence . Thus, less experienced radiologists are at greater risk for misdiagnosing a cancer and for increasing the number of false-positives.
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Materials and methods
Study Population
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Radiologist Readers
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
CAD System
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Radiologists’ Analysis
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
B-CAD Analysis
Get Radiology Tree app to read full this article<
Statistical Analysis
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Results
Study Population Characteristics
Get Radiology Tree app to read full this article<
Table 1
Description of Lesions
Pathology Number of Lesions Benign lesions 83 Cysts with three apocrine 16 Fibroadenoma 39 Benign lesions (fine-needle aspiration) with five cystic 24 Papilloma 3 Lymph node 1 Malignant lesions 77 Ductal invasive carcinoma with two apocrine, one tubular, and one mucinous 63 Lobular invasive carcinoma 14 Total 160
Get Radiology Tree app to read full this article<
Performance of B-CAD
Get Radiology Tree app to read full this article<
Table 2
Sensitivity and Specificity of B-CAD Alone
B-CAD Histologic Results Total Malignant Benign BI-RADS category 4 or 5 77 43 120 BI-RADS category 2 or 3 0 40 40 Total 77 83 160 Sensitivity Specificity 100% 48%
BI-RADS, Breast Imaging Reporting and Data System.
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Performance of Radiologist A (20 Years of Experience)
Benign and malignant classification
Get Radiology Tree app to read full this article<
Table 3
Sensitivity and Specificity, with and without B-CAD, of Senior Radiologist A (20 Years of Experience) for the Classification of Lesions According to Benign or Malignant Nature
Radiologist A Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total Malignant 76 25 101 76 45 121 Benign 1 58 59 1 38 39 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 99% 70% 99% 46%
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Classification according to BI-RADS categories
Get Radiology Tree app to read full this article<
Table 4
Sensitivity and Specificity, with and without B-CAD, of Senior Radiologist A (20 Years of Experience) for the Classification of Lesions According to BI-RADS Categories
Radiologist A Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total BI-RADS category 4 or 5 77 43 120 77 57 134 BI-RADS category 2 or 3 0 40 40 0 26 26 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 100% 48% 100% 31%
BI-RADS, Breast Imaging Reporting and Data System.
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Performance of Radiologist B (5 Years of Experience)
Benign and malignant classification
Get Radiology Tree app to read full this article<
Table 5
Sensitivity and Specificity, with and without B-CAD, of Intermediate-level Radiologist B (5 Years of Experience) for the Classification of Lesions According to Benign or Malignant Nature and BI-RADS Categories
Radiologist B Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total BI-RADS category 4 or 5 67 17 84 74 35 109 BI-RADS category 2 or 3 10 66 76 3 48 51 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 87% 80% 96% 58%
BI-RADS, Breast Imaging Reporting and Data System.
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Classification according to BI-RADS categories
Get Radiology Tree app to read full this article<
Performance of Radiologist C (1 Year of Experience)
Benign and malignant classification
Get Radiology Tree app to read full this article<
Table 6
Sensitivity and Specificity, with and without B-CAD, of Junior Radiologist C (1 Year of Experience) for the Classification of Lesions According to Benign or Malignant Nature
Radiologist C Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total Malignant 68 26 94 73 36 109 Benign 9 57 66 4 47 51 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 88% 69% 95% 57%
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Classification according to BI-RADS categories
Get Radiology Tree app to read full this article<
Table 7
Sensitivity and Specificity, with and without B-CAD, of Junior Radiologist C (1 Year of Experience) for the Classification of Lesions According to BI-RADS Categories
Radiologist C Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total BI-RADS category 4 or 5 75 55 130 73 38 111 BI-RADS category 2 or 3 2 28 30 4 45 49 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 97% 34% 95% 54%
BI-RADS, Breast Imaging Reporting and Data System.
Get Radiology Tree app to read full this article<
Performance of Radiologist D (4 Months of Experience)
Benign and malignant classification
Get Radiology Tree app to read full this article<
Table 8
Sensitivity and Specificity, with and without B-CAD, of Junior Radiologist D (4 Months of Experience) for the Classification of Lesions According to Benign or Malignant Nature
Radiologist D Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total Malignant 68 24 92 68 24 92 Benign 9 59 68 9 59 68 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 88% 71% 91% 71%
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Classification according to BI-RADS categories
Get Radiology Tree app to read full this article<
Table 9
Sensitivity and Specificity, with and without B-CAD, of Junior Radiologist D (4 Months of Experience) for the Classification of Lesions According to BI-RADS Categories
Radiologist D Without B-CAD With B-CAD Histologic Results Histologic Results Malignant Benign Total Malignant Benign Total BI-RADS category 4 or 5 73 34 107 73 34 107 BI-RADS category 2 or 3 4 49 53 4 49 53 Total 77 83 160 77 83 160 Sensitivity Specificity Sensitivity Specificity 95% 59% 95% 53%
BI-RADS, Breast Imaging Reporting and Data System.
Get Radiology Tree app to read full this article<
Discussion
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Study Biases and Limitations
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Future Prospects
Improvement of CAD system specificity
Get Radiology Tree app to read full this article<
Improved algorithm
Get Radiology Tree app to read full this article<
Training of radiologists
Get Radiology Tree app to read full this article<
Clinical and imaging data in the breast ultrasound CAD system
Get Radiology Tree app to read full this article<
Time optimization: Real-time CAD system
Get Radiology Tree app to read full this article<
Learning tool for breast ultrasound semiology
Get Radiology Tree app to read full this article<
Conclusions
Get Radiology Tree app to read full this article<
Acknowledgments
Get Radiology Tree app to read full this article<
References
1. Berg W.A.: Supplemental screening sonography in dense breasts. Radiol Clin North Am 2004; 42: pp. 845-851.
2. Corsetti V., Ferrari A., Ghirardi M., et. al.: Role of ultrasonography in detecting mammographically occult breast carcinoma in women with dense breasts. Radiol Medica 2006; 111: pp. 440-448.
3. Corsetti V., Houssami N., Ferrari A., et. al.: Breast screening with ultrasound in women with mammography-negative dense breasts: evidence on incremental cancer detection and false positives, and associated cost. Eur J Cancer 2008; 44: pp. 539-544.
4. Jackson V.P., Reynolds H.E., Hawes D.R.: Sonography of the breast. Semin Ultrasound CT MR 1996; 17: pp. 460-475.
5. Stavros A.T., Thickman D., Rapp C.L., et. al.: Solid breast nodules: use of sonography to distinguish between benign and malignant lesions. Radiology 1995; 196: pp. 123-134.
6. Skaane P., Engedal K., Skjennald A.: Interobserver variation in the interpretation of breast imaging. Comparison of mammography, ultrasonography, and both combined in the interpretation of palpable noncalcified breast masses. Acta Radiol 1997; 38: pp. 497-502.
7. Baker J.A., Kornguth P.J., Soo M.S., et. al.: Sonography of solid breast lesions: observer variability of lesion description and assessment. AJR Am J Roentgenol 1999; 172: pp. 1621-1625.
8. Baker J.A., Soo M.S.: Breast US: assessment of technical quality and image interpretation. Radiology 2002; 223: pp. 229-238.
9. Eadie LH, Taylor P, Gibson AP. A systematic review of computer-assisted diagnosis in diagnostic cancer imaging. Eur J Radiol. In press.
10. Balleyguier C., Kinkel K., Fermanian J., et. al.: Computer-aided detection (CAD) in mammography: does it help the junior or the senior radiologist?. Eur J Radiol 2005; 54: pp. 90-96.
11. Kuo W.J., Chang R.F., Moon W.K., et. al.: Computer-aided diagnosis of breast tumors with different US systems. Acad Radiol 2002; 9: pp. 793-799.
12. Chen C.M., Chou Y.H., Han K.C., et. al.: Breast lesions on sonograms: computer-aided diagnosis with nearly setting-independent features and artificial neural networks. Radiology 2003; 226: pp. 504-514.
13. Chen D., Chang R.F., Huang Y.L.: Breast cancer diagnosis using self-organizing map for sonography. Ultrasound Med Biol 2000; 26: pp. 405-411.
14. Chang R.F., Wu W.J., Moon W.K., et. al.: Automatic ultrasound segmentation and morphology based diagnosis of solid breast tumors. Breast Cancer Res Treat 2005; 89: pp. 179-185.
15. Huang Y.L., Chen D.R., Jiang Y.R., et. al.: Computer-aided diagnosis using morphological features for classifying breast lesions on ultrasound. Ultrasound Obstet Gynecol 2008; 32: pp. 565-572.
16. Chen D.R., Chang R.F., Huang Y.L.: Computer-aided diagnosis applied to US of solid breast nodules by using neural networks. Radiology 1999; 213: pp. 407-412.
17. Wu W.J., Moon W.K.: Ultrasound breast tumor image computer-aided diagnosis with texture and morphological features. Acad Radiol 2008; 15: pp. 873-880.
18. Shen W.C., Chang R.F., Moon W.K., et. al.: Breast ultrasound computer-aided diagnosis using BI-RADS features. Acad Radiol 2007; 14: pp. 928-939.
19. Shen W.C., Chang R.F., Moon W.K.: Computer aided classification system for breast ultrasound based on Breast Imaging Reporting and Data System (BI-RADS). Ultrasound Med Biol 2007; 33: pp. 1688-1698.
20. Chang R.F., Kuo W.J., Chen D.R., et. al.: Computer-aided diagnosis for surgical office-based breast ultrasound. Arch Surg 2000; 135: pp. 696-699.
21. Horsch K., Giger M.L., Vyborny C.J., et. al.: Performance of computer-aided diagnosis in the interpretation of lesions on breast sonography. Acad Radiol 2004; 11: pp. 272-280.
22. Sahiner B., Chan H.P., Roubidoux M.A., et. al.: Malignant and benign breast masses on 3D US volumetric images: effect of computer-aided diagnosis on radiologist accuracy. Radiology 2007; 242: pp. 716-724.
23. Drukker K., Gruszauskas N.P., Sennett C.A., et. al.: Breast US computer-aided diagnosis workstation: performance with a large clinical diagnostic population. Radiology 2008; 248: pp. 392-397.
24. American College of Radiology: BI-RADS: ultrasound.Breast Imaging Reporting and Data System: BI-RADS atlas.2003.American College of RadiologyReston, VA:
25. Drukker K., Giger M.L., Vyborny C.J., et. al.: Computerized detection and classification of cancer on breast ultrasound. Acad Radiol 2004; 11: pp. 526-535.
26. Wang Y., Jiang S., Wang H., et. al.: CAD algorithms for solid breast masses discrimination: evaluation of the accuracy and interobserver variability. Ultrasound Med Biol 2010; 36: pp. 1273-1281.
27. Horsch K., Giger M.L., Metz C.E.: Potential effect of different radiologist reporting methods on studies showing benefit of CAD. Acad Radiol 2008; 15: pp. 139-152.
28. Gruszauskas N.P., Drukker K., Giger M.L., et. al.: Performance of breast ultrasound computer-aided diagnosis: dependence on image selection. Acad Radiol 2008; 15: pp. 1234-1245.
29. Collins M.J., Hoffmeister J., Worrell S.W.: Computer-aided detection and diagnosis of breast cancer. Semin Ultrasound CT MR 2006; 27: pp. 351-355.
30. Giger M.L.: Computerized analysis of images in the detection and diagnosis of breast cancer. Semin Ultrasound CT MR 2004; 25: pp. 411-418.
31. Weinstein S.P., Conant E.F., Sehgal C.: Technical advances in breast ultrasound imaging. Semin Ultrasound CT MR 2006; 27: pp. 273-283.
32. Athanasiou A., Tardivon A., Ollivier L., et. al.: How to optimize breast ultrasound. Eur J Radiol 2009; 69: pp. 6-13.
33. Sickles E.A., Philpotts L.E., Parkinson B.T., et. al.: American College of Radiology/Society of Breast Imaging curriculum for resident and fellow education in breast imaging. J Am Coll Radiol 2006; 3: pp. 879-884.
34. Berg W.A., D’Orsi C.J., Jackson V.P., et. al.: Does training in the Breast Imaging Reporting and Data System (BI-RADS) improve biopsy recommendations or feature analysis agreement with experienced breast imagers at mammography?. Radiology 2002; 224: pp. 871-880.