We have developed quantitative descriptors to provide an objective means of noninvasive identification of cancerous breast lesions. These descriptors include quantitative acoustic features assessed using spectrum analysis of ultrasonic radiofrequency (rf) echo signals and morphometric properties related to lesion shape. Acoustic features include measures of echogenicity, heterogeneity and shadowing, computed by generating spectral-parameter images of the lesion and surrounding tissue. Spectral-parameter values are derived from rf echo signals at each pixel using a sliding-window Fourier analysis. We derive quantitative acoustic features from spectral-parameter maps of the lesion and adjacent areas. We quantify morphometric features by geometric and fractal analysis of traced lesion boundaries. Initial results on biopsy-proven cases show that although a single parameter cannot reliably discriminate cancerous from noncancerous breast lesions, multi-feature analysis provides excellent discrimination for this data set. We have processed data for 130 biopsy-proven patients, acquired during routine ultrasonic examinations at three clinical sites and produced an area under the receiver-operating-characteristics (ROC) curve of 0.947±0.045. Among the quantitative descriptors, lesion-margin definition, spiculation and border irregularity are the most useful; some additional morphometric features (such as border irregularity) also are particularly effective in lesion classification. Our findings are consistent with many of the BI-RADS (Breast Imaging Reporting and Data System) breast-lesion-classification criteria in use today.
American Cancer Society. in Cancer Facts & Figures2009, p. 6 (2009).
2.
PisaniPParkinDMFerlayJ. Estimates of the worldwide mortality from eighteen major cancers in 1985. Implications for prevention and projections of future burdenInt J Cancer55, 891–903 (1993).
3.
RiesLKosaryLHankeyBMillerBEdwardsB. SEER Cancer Statistics Rev, 1975–1996 (National Cancer Institute, Bethesda, MD1999).
4.
ViciniFAKestinLChenP, Limited-field radiation therapy in the management of early-stage breast cancerJ National Cancer Institute95, 1205–1210 (2003).
5.
TohnoECosgroveDOSloaneJP, Ultrasound Diagnosis of Breast Diseases (Churchill Livingstone, London, UK1994).
6.
StavrosATThickmanDRappCL, Solid breast nodules ruse of sonography to distinguish between benign and malignant lesionsRadiology196, 123–134 (1995).
7.
Cole-BeugletCSorianoRZKortzABGoldbergBB, Ultrasound analysis of 104 primary breast carcinomas classified according to histologic typeRadiology147, 191–196 (1983).
8.
Advanced Technology Laboratories, Summary of Atl-Sponsored Breast Ultrasound Study and Preliminary Findings (ATL publications, 1994).
9.
KolbTMLichyJNewhouseJH, Occult cancer in women with dense breasts: detection with screening US—diagnostic yield and tumor characteristicsRadiology207, 191–199 (1998).
10.
SkaanePEngedalK. Analysis of sonographic features in the differentiation of fibroadenoma and invasive ductal carcinomaAJR170, 109–114 (1998).
11.
HuberSDanesJZunaI, Relevance of sonographic B-mode criteria and computer-aided ultrasonic tissue characterization in differential/diagnosis of solid breast massesUltrasound Med Biol26, 1243–1252 (2000).
12.
American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) Atlas (American College of Radiology, Reston, VA2003).
13.
MandelsonEBBergNAMerrittCRB, Towards a standardized breast ultrasound lexicon bi-rads: ultrasoundSemin Roentgenol36, 217–25 (2001).
14.
DonohueKDHuangLBurksTForsbergFPiccoliCW, Tissue classification with generalized spectrum parametersUltrasound Med Biol27, 1505–1514 (2001).
15.
ShankarPMDumaneVAReidJM, Classification of ultrasonic b-mode images of breast masses using nakagami distributionIEEE Trans Ultrason Ferroelec Freq Contr48, 569–580 (2001).
16.
ShankarPMDumaneVAPiccoliCW, Computer-aided classification of breast masses in ultrasonic B-scans using a multiparameter approachIEEE Trans Ultrason Ferroelec Freq Contr50, 1002–1009 (2003).
17.
DrukkerKGigerMLMendelsonEB, Computerized analysis of shadowing on breast ultrasound for improved lesion detection. Med Phys30, 1833–42 (2003).
18.
JooSYangYSMoonWKKimHC, Computer-aided diagnosis of solid breast nodules: use of an artificial neural network based on multiple sonographic featuresIEEE Trans Med Imag23, 1292–1300 (2004).
19.
ChangRFWuWJMoonWKChenDR, Improvement in breast tumor discrimination by support vector machines and speckle-emphasis texture analysisUltrasound Med Biol29, 679–686 (2003).
20.
ZhengYGreenleafJFGisvoldJJ, Reduction of breast biopsies with a modified self-normalizing mapIEEE Trans Neural Networks8, 1386–1396 (1997).
21.
LefebvreFMeunierMThibaultFLaugierPBergerG. Computerized ultrasound B-scan characterization of breast nodulesUltrasound Med Biol26, 1421–1428 (2000).
22.
AndréMGalperinMOjeda-FournierH, Performance of a method to aid breast ultrasound interpretation using image processing and case-based reasoningMed Phys36, 2778–2778 (2009).
23.
AndréMGalperinMControG, Diagnostic performance of a computer-aided image analysis system for breast ultrasound, Acoust Imag28, pp. 267–277 (2007).
24.
AndréMGalperinMControGOmidN. Optimization of a breast mass classifier for computer-aided ultrasound analysisAcoust Imag28, 341–348 (2007).
25.
LizziFLGreenebaumMFeleppaEJElbaumMColemanDJ, Theoretical framework for spectrum analysis in ultrasonic tissue characterizationJ Acoust Soc Am73, 1366–1373 (1983).
26.
AlamSKLizziFLFeleppaEJLiuTKaliszA, Ultrasonic multi-feature analysis procedures for breast lesion classification, Proc SPIE 3982, Medical Imaging 2000: Ultrasonic Imaging and Signal Processing, pp. 196–201, Kirk-ShungK.InsanaMichael F. Eds. (2000).
27.
AlamSKLizziFLFeleppaEJLiuTKaliszA. Computer aided diagnosis of breast lesions using a multi-feature analysis procedureProc SPIE 4687, Medical Imaging 2002: Ultrasonic Imaging and Signal Processing, pp. 296–303, InsanaMichael F.WalkerW, Eds (2002).
28.
FeleppaEJLizziFLColemanDJYaremkoMM, Diagnostic spectrum analysis in ophthalmology: a physical perspectiveUltrasound Med Biol12, 623–31 (1986).
29.
LizziFLAstorMFeleppaEJShaoMKaliszA. Statistical framework for ultrasonic spectral parameter imagingUltrasound Med Biol23, 1371–82 (1997).
30.
LizziFLAlamSKMikaelianSLeePFeleppaEJ, On the statistics of ultrasonic spectral parametersUltrasound Med Biol32, 1671–1685 (2006).
31.
AlamSKLizziFLMikaelianSFeleppaEJLeeP. The effect of gate functions and planar calibration on ultrasonic spectral parameters, IEEE Trans Ultrason, Ferroelect, Freq Contr (2010) (submitted).
32.
BamberJC, Ultrasonic properties of tissues, Ultrasound in Medicine, DuckFABakerACStarrittHC Eds., Ch. 4, pp. 57–88 (Inst of Physics Pub, Bristol, England1998).
33.
MastTD, Empirical relationships between acoustic parameters in human soft tissuesAcoust Res Lett1, 37–42 (2000).
34.
LizziFLFeleppaEJAstorMKaliszA. Statistics of ultrasonic spectral parameters for prostate and liver examinationsIEEE Trans Ultrason, Ferroelect Freq Contr44, 935–942 (1997).
35.
YaoWZhaoBZhaoYWangWQianG. Ultrasonographic texture analysis of parenchymatous organs by the four-neighborhood-pixels algorithmJ Ultrasound Med20, 465–471 (2001).
36.
HurstHEBlackRPSimaikaYM, Long Term Storage: An Experimental Study (Constable, London1965).
37.
FederJ. Fractals (Plenum Press, New York1988).
38.
GarraBSKrasnerBHHoriiSC, Improving the distinction between benign and malignant breast lesions: the value of sonographic texture analysisUltrasonic Imag13, 267–285 (1993).
39.
ChenD-RChangR-FHuangY-L., Breast cancer diagnosis using self organizing map for sonographyUltrasound Med Biol26, 405–411 (2000).
40.
GonzalezRCWoodsRE, Digital Image Processing (Addison-Wesley, Reading, MA1992).
41.
ChouY-HTiuC-MHungG-S, Stepwise logistic regression analysis of tumor contour features for breast ultrasound diagnosisUltrasound Med Biol27, 1493–1498 (2001).
42.
SicklesEA, Mammographic features of “early” breast cancerAJR143, 461–464 (1984).
43.
RichardsonLF, The problem of contiguity: An appendix of statistics of deadly quarrels, General Systems Yearbook6, pp. 139–187 (1961).
44.
KayeBH, Image analysis techniques for characterizing fractal structures, The Fractal Approach to Heterogeneous Chemistry, pp. 55–66, (John Wiley & Sons, New York1989).
45.
MandelbrotBB, How long is the coast of Britain? Statistical self-similarity and fractional dimensionScience155, 636–638 (1967).
46.
RussJC, The Image Processing Handbook (CRC Press, Boca Raton, FL1999).
47.
LoydCJ, Using small receiver operating characteristic curves to summarize and compare diagnostic systems. J Am Stat Assoc93, 1356–1364 (1998).
48.
PressSJWilsonS. Choosing between logistic regression and discriminant analysisJ Amer Stat Assoc73, 699–705 (1978).
49.
JarqueCMBeraAK, A test for normality of observations and regression residualsInt Stat Rev55, 163–172 (1987).
50.
PoharMBlasMTurkS. Comparison of logistic regression and linear discriminant analysis: a simulation studyMetodološki Zvezki1, 143–161 (2004).
51.
DaPonteJSShermanP. Classification of ultrasonic image texture by statistical discriminant analysis and neural networksComputerized Medical Imaging Graphics15, 3–9 (1991).
52.
KaplanSS, Clinical utility of bilateral whole-breast us in the evaluation of women with dense breast tissue. Radiology221, 641–649 (2001).
53.
MendelsonEB, Imaging the post-surgical breastSemin Ultrasound CTMR10, 154–170 (1989).
54.
TabàrLFagerbergGDuffySW, Update of the Swedish two-county program of mammographic screening of breast cancerRadiol Clin North Am30, pp. 187–210 (1992).
55.
PeerPGMvan DijckJAAMHendricksJHCLHollandRVerbeekALM, Age-dependent growth rate of primary breast cancerCancer71, 3547–3551 (1993).
