Computer-aided Classification of Breast Masses: Performance and Interobserver Variability of Expert Radiologists versus Residents
RadiologyPublished 23 October 2010
Swatee Singh, J Wingelaar Maxwell, Jay A. Baker, Jennifer L. Nicholas, Joseph Y. Lo
Citations46
SJR quartileQ1
SJR score3.86
SNIP3.55
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
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 specificity and decreasing interobserver variability in biopsy recommendations.
Abstract
Among findings showing a significant 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 specificity and decreasing interobserver variability in biopsy recommendations.
Keywords
Computer ScienceMedicine
BiometricsThe Measurement of Observer Agreement for Categorical Data
78,934 Citations1977J. Richard Landis, Gary G. Koch
A general statistical methodology for the analysis of multivariate categorical data arising from observer reliability studies is presented and tests for interobserver bias are presented in terms of first-order marginal homogeneity and measures of interob server agreement are developed as generalized kappa-type statistics.
New England Journal of MedicineDiagnostic Performance of Digital versus Film Mammography for Breast-Cancer Screening
1,942 Citations2005Etta D. Pisano, Constantine Gatsonis +10 more
RadiologySolid breast nodules: use of sonography to distinguish between benign and malignant lesions.
1,763 Citations1995A. Thomas Stavros, David Thickman +4 more
Sonography can be used to accurately classify some solid lesions as benign, allowing imaging follow-up rather than biopsy, and this distinction could be definite enough to obviate biopsy.
New England Journal of MedicineTen-Year Risk of False Positive Screening Mammograms and Clinical Breast Examinations
1,055 Citations1998Joann G. Elmore, Mary B. Barton +4 more
Over 10 years, one third of women screened had an abnormal test result that required additional evaluation, even though no breast cancer was present, and Physicians should educate women about the risk of a false positive result from a screening test for breast cancer.
New England Journal of MedicineVariability in Radiologists' Interpretations of Mammograms
598 Citations1994Joann G. Elmore, Carolyn K. Wells +3 more
Efforts to improve accuracy and reduce variability in interpretation may increase the effectiveness of mammography in detecting early breast cancers.
RadiologyPeriodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases.
553 Citations1991Edward A. Sickles
The author prospectively evaluated the value of periodic mammographic surveillance among 3,184 consecutive cases of nonpalpable, probably benign breast lesions detected with mammography, finding that decreasing the number of biopsies of benign lesions and thereby substantially reducing costs may help overcome a major barrier to widespread use of mammographic screening.
RadiologyBI-RADS Lexicon for US and Mammography: Interobserver Variability and Positive Predictive Value
503 Citations2006Elizabeth Lazarus, Martha B. Mainiero +3 more
Interobserver agreement with the new BI-RADS terminology is good and validates the US lexicon and subcategories 4a, 4b, and 4c are useful in predicting the likelihood of malignancy.
Obstetrical & Gynecological SurveyDiagnostic Performance of Digital versus Film Mammography for Breast-Cancer Screening
482 Citations2006Etta D. Pisano, Constantine Gatsonis +10 more
RadiologyBenign versus Malignant Solid Breast Masses: US Differentiation
407 Citations1999Guita Rahbar, Angela Sie +7 more
The data confirm that certain US features can help differentiate benign from malignant masses, however, practice and interpreter variability should be further explored before these criteria are generally applied to defer biopsy of solid masses.
Academic RadiologyImproving breast cancer diagnosis with computer-aided diagnosis
312 Citations1999Yulei Jiang, Robert M. Nishikawa +4 more
CAD can be used to improve radiologists' performance in breast cancer diagnosis by using receiver operating characteristic (ROC) analysis and by comparing biopsy recommendations.
RadiologyDiagnosis of Breast Cancer: Contribution of US as an Adjunct to Mammography
297 Citations1999Harmine M. Zonderland, Emile G. Coerkamp +3 more
The use of US as an adjunct to mammography resulted in an increase in diagnostic accuracy and its contribution to the diagnosis of breast cancer in this study was 7.4%.
RadiologyNonpalpable breast lesions: recommendations for biopsy based on suspicion of carcinoma at mammography.
291 Citations1988F M Hall, J M Storella +2 more
In the management of suspicious nonpalpable mammographic findings, the rate of carcinoma for lesions at biopsy can approximate 40%.
RadiologyNonpalpable, probably benign lesions: role of follow-up mammography.
234 Citations1992Ximena Varas, Félix Leborgne +1 more
Of 21,855 consecutive women prospectively studied with mammography, physical examination, and high-resolution ultrasonography when appropriate, 558 received a diagnosis of nonpalpable, probably benign lesions; nine patients ultimately proved to have carcinoma; two had noninvasive carcinoma, and two had axillary node metastases.
JNCI Journal of the National Cancer InstituteAccuracy of Screening Mammography Interpretation by Characteristics of Radiologists
216 Citations2004W. E. Barlow, C. Chi +6 more
No evidence that greater volume or experience at interpreting mammograms is associated with better performance is found, however, they may affect sensitivity and specificity, possibly by determining the threshold for calling a mammogram positive.
RadiologyBreast cancer: prediction with artificial neural network based on BI-RADS standardized lexicon.
216 Citations1995Jay A. Baker, Phyllis J. Kornguth +3 more
The BI-RADS lexicon provides a standardized language between mammographers and an ANN that can improve the PPV of breast biopsy and was significantly greater than the specificity of radiologists.
RadiologyDoes Training in the Breast Imaging Reporting and Data System (BI-RADS) Improve Biopsy Recommendations or Feature Analysis Agreement with Experienced Breast Imagers at Mammography?
205 Citations2002Wendie A. Berg, Carl J. D’Orsi +5 more
BI-RADS training resulted in improved agreement with the consensus of experienced breast imagers for feature analysis and final assessment, and it is important that trainees showed improved rates of recommending biopsy for malignant lesions.
RadiologyBreast Imaging Reporting and Data System Lexicon for US: Interobserver Agreement for Assessment of Breast Masses
203 Citations2009Nouf Abdullah, Benoı̂t Mesurolle +2 more
Medical PhysicsCADx of mammographic masses and clustered microcalcifications: A review
201 Citations2009Matthias Elter, Alexander Horsch
A comprehensive review of the state of the art of CADx approaches is presented, restricted to the two most important types of mammographic lesions: masses and clustered microcalcifications, and focuses on articles published in international journals.
Medical PhysicsImprovement of mammographic mass characterization using spiculation measures and morphological features
187 Citations2001Berkman Sahiner, Heang‐Ping Chan +3 more
The results indicate that combining texture features with morphological features extracted from automatically segmented mass boundaries will be an effective approach for computer-aided characterization of mammographic masses.
Ultrasound in Medicine & BiologyBreast cancer diagnosis using self-organizing map for sonography
181 Citations2000Dar‐Ren Chen, Ruey‐Feng Chang +1 more
This computer-aided diagnosis system can provide a useful tool and its high negative predictive value could potentially help avert benign biopsies.
IEEE Transactions on Medical ImagingComputer-aided characterization of mammographic masses: accuracy of mass segmentation and its effects on characterization
174 Citations2001Berkman Sahiner, Nicholas Petrick +5 more
The automated segmentation method was quantitatively compared with manual segmentation by two expert radiologists using three similarity or distance measures on a data set of 100 masses to investigate the effect of the segmentation stage on the overall classification accuracy.
European Journal of RadiologyObserver variability of Breast Imaging Reporting and Data System (BI-RADS) for breast ultrasound
173 Citations2007Hye‐Jeong Lee, Eun‐Kyung Kim +5 more
Because inter- and intra-observer agreement with the BI-RADS lexicon for US is good, the use of BI-CAD lexicon can provide accurate and consistent description and assessment for breast US.
Academic RadiologyAutomated computerized classification of malignant and benign masses on digitized mammograms
167 Citations1998Zhimin Huo, Maryellen L. Giger +4 more
Automated computerized classification schemes may be useful in helping radiologists distinguish between benign and malignant masses and thus reducing the number of unnecessary biopsies.
JNCI Journal of the National Cancer InstituteScreening Mammograms by Community Radiologists: Variability in False-Positive Rates
158 Citations2002J. G. Elmore
Community radiologists varied widely in their false-positive rates in screening mammograms; this variability range was reduced by half, but not eliminated, after statistical adjustment for patient, radiologist, and testing characteristics.
JNCI Journal of the National Cancer InstitutePhysician Predictors of Mammographic Accuracy
145 Citations2005Rebecca Smith‐Bindman, Philip Chu +8 more
Higher overall accuracy was associated with more experience and with a higher focus on screening mammography, and raising the annual volume requirements in the Mammography Quality Standards Act might improve the overall quality of screening Mammography in the United States.
Clinical ImagingClassification of breast ultrasound images using fractal feature
143 Citations2005Dar‐Ren Chen, Ruey‐Feng Chang +6 more
A computer-aided diagnosis (CAD) system based on the fractal analysis is proposed to classify the breast lesions into two classes: benign and malignant and the k-means classification method is used to classify benign tumors from malignant ones.
JNCI Journal of the National Cancer InstituteDecreasing Women's Anxieties After Abnormal Mammograms: A Controlled Trial
136 Citations2004Mary B. Barton, Debra S. Morley +5 more
Immediate reading of screening mammograms, but not an educational intervention targeting coping skills, was associated with less anxiety among women with false-positive mammograms 3 weeks after mammography.
RadiologyBreast Cancer: Effectiveness of Computer-aided Diagnosis—Observer Study with Independent Database of Mammograms
134 Citations2002Zhimin Huo, Maryellen L. Giger +2 more
Computer-aided diagnosis can potentially help radiologists improve their diagnostic accuracy in the task of differentiating between benign and malignant masses seen on mammograms.
RadiologyStandardized Abnormal Interpretation and Cancer Detection Ratios to Assess Reading Volume and Reader Performance in a Breast Screening Program
128 Citations2000Lisa Kan, Ivo A. Olivotto +3 more
A minimum of 2,500 interpretations per year is associated with lower abnormal interpretation rates and average or better cancer detection rates in a screening program.
RadiologyBreast Mass Lesions: Computer-aided Diagnosis Models with Mammographic and Sonographic Descriptors
116 Citations2007Jonathan L. Jesneck, Joseph Y. Lo +1 more
Results showed that combining mammographic and sonographic descriptors in a CAD model can result in high classification and generalization performance.
RadiologyInduced costs of low-cost screening mammography.
111 Citations1988Dvora Cyrlak
Telephone follow-up in a recent low-cost ($50) mammography screening project in Orange County, California, was done to determine the detection rate and the induced costs of detection of small breast cancers, finding that at least five cancers were detected per 1,000 women screened.
Artificial Intelligence in MedicineSelf-organizing map for cluster analysis of a breast cancer database
109 Citations2003Mia K. Markey, Joseph Y. Lo +2 more
A self-organizing map was used to identify clusters in a large, heterogeneous computer-aided diagnosis database based on mammographic findings (BI-RADS) and patient age and the clusters showed logical separation of clinical subtypes such as architectural distortions, masses, and calcifications.
Medical PhysicsCharacterization of mammographic masses based on level set segmentation with new image features and patient information
106 Citations2007Jiazheng Shi, Berkman Sahiner +10 more
An automated method for mammographic mass segmentation is developed and new image based features in combination with patient information are explored in order to improve the performance of mass characterization.
Academic RadiologyComputerized detection and classification of cancer on breast ultrasound1
106 Citations2004Karen Drukker, Maryellen L. Giger +2 more
The results show promising performance of the computerized lesion detection and classification method, and indicate the potential of such a system for clinical breast ultrasound.
RadiologyArtificial neural network: improving the quality of breast biopsy recommendations.
104 Citations1996Jay A. Baker, Phyllis J. Kornguth +2 more
Use of an ANN with radiologists' descriptions of abnormal findings may improve interpretation of mammographic abnormalities.
Radiologic Clinics of North AmericaScreening mammography: proven benefit, continued controversy
101 Citations2002Cheng‐Han Lee
Evidence supports beginning regular screening at age 40 in women at average risk and suggests that the screening interval should be yearly, especially in younger women, while the impact of double reading and computer-aided detection in the interpretation of screening mammograms warrants further evaluation.
American Journal of RoentgenologyLocalization and needle aspiration of breast lesions: complications in 370 cases.
98 Citations1991Mark A. Helvie, Debra M. Ikeda +1 more
It is concluded that wire localizations and imaging-guided aspirations are generally well tolerated procedures, however, vasovagal reactions are frequent enough to warrant close observation of patients and Radiologists and breast-imaging personnel should be able to recognize and treat vasovageal reactions.
Academic RadiologyPerformance of computer-aided diagnosis in the interpretation of lesions on breast sonography
96 Citations2004Karla Horsch, Maryellen L. Giger +2 more
Computer analysis of ultrasound images of breast lesions has been shown to improve the diagnostic accuracy of radiologists in the task of distinguishing between malignant and benign breast lesions and in recommending cases for biopsy.
Seminars in Ultrasound CT and MRIComputerized analysis of images in the detection and diagnosis of breast cancer
84 Citations2004Maryellen L. Giger
This chapter reviews CAD in breast cancer detection and diagnosis, including examples of image analyses, multi-modality approaches (i.e., special-view diagnostic mammography, ultrasound, and MRI), and means of communicating the computer output to the human.
Medical PhysicsAnalysis of temporal changes of mammographic features: Computer‐aided classification of malignant and benign breast masses
75 Citations2001Lubomir M. Hadjiiski, Berkman Sahiner +4 more
A new classification scheme was developed to classify mammographic masses as malignant and benign by using interval change information and stepwise feature selection and linear discriminant analysis classification were used to select and merge the most useful features.
Medical PhysicsOptimized approach to decision fusion of heterogeneous data for breast cancer diagnosis
58 Citations2006Jonathan L. Jesneck, L. W. Nolte +3 more
Decision fusion directly optimized clinically significant performance measures, such as AUC and pAUC, and sometimes outperformed two well-known machine-learning techniques when applied to two different breast cancer data sets.
RadiologySolid Breast Masses: Classification with Computer-aided Analysis of Continuous US Images Obtained with Probe Compression
56 Citations2005Woo Kyung Moon, Ruey‐Feng Chang +3 more
RadiologyRobustness of Computerized Lesion Detection and Classification Scheme across Different Breast US Platforms
55 Citations2005Karen Drukker, Maryellen L. Giger +1 more
The results indicate that the performance of this fully automated computerized lesion detection and classification method, which demonstrated robustness over the different US equipment used, is promising.
RadiologyBreast US Computer-aided Diagnosis Workstation: Performance with a Large Clinical Diagnostic Population
52 Citations2008Karen Drukker, Nicholas P. Gruszauskas +2 more
Current levels of computer performance warrant a clinical evaluation of the potential of US CAD to aid radiologists in lesion work-up recommendations.
BMC Health Services ResearchAssociation between Radiologists' Experience and Accuracy in Interpreting Screening Mammograms
51 Citations2008Eduard Molins, Francesc Macià +3 more
Among radiologists who read routinely, volume is not associated with better performance when interpreting screening mammograms, although specificity decreased in radiologists not routinely reading mammograms.
RadiologyBreast Masses: Computer-aided Diagnosis with Serial Mammograms
48 Citations2006Lubomir M. Hadjiiski, Berkman Sahiner +13 more
CAD involving interval change analysis of preselected regions of interest can significantly improve radiologists' accuracy in classifying masses on digitized screen-film mammograms as malignant or benign.
RadiologyDifferences between Computer-aided Diagnosis of Breast Masses and That of Calcifications
47 Citations2002Mia K. Markey, Joseph Y. Lo +1 more
Comparing the performance of a computer-aided diagnosis (CAD) system for diagnosis of previously detected lesions, based on radiologist-extracted findings on masses and calcifications found the BP-ANN performed significantly better on masses than microcalcifications.
American Journal of RoentgenologyCross-Institutional Evaluation of BI-RADS Predictive Model for Mammographic Diagnosis of Breast Cancer
43 Citations2002Joseph Y. Lo, Mia K. Markey +2 more
The BI-RADS lexicon facilitated the cross-institutional test of a breast cancer prediction model and the model generalized reasonably well, but there were significant performance decreases.
Medical PhysicsBreast cancer CADx based on BI-RADS™ descriptors from two mammographic views
42 Citations2006Shalini Gupta, Priscilla F. Chyn +1 more
Some improvement in performance of BI-RADS based CADx algorithms may be achieved by combining information from two mammographic views.
Academic RadiologyMultimodality Computerized Diagnosis of Breast Lesions Using Mammography and Sonography1
41 Citations2005Karen Drukker, Karla Horsch +1 more
Computerized classification of cancer significantly improved when lesion features from both modalities were combined, and this results are encouraging and warrant further exploration of computerized methods for multimodality imaging.
Breast Cancer Research and TreatmentVariation in false-positive rates of mammography reading among 1067 radiologists: a population-based assessment
38 Citations2006Alai Tan, Daniel Freeman +2 more
Academic RadiologyEvaluation of Computer-aided Diagnosis on a Large Clinical Full-field Digital Mammographic Dataset
35 Citations2008Hui Li, Maryellen L. Giger +7 more
Results show that the computerized analysis methods for the diagnosis of breast mass lesions on FFDM are promising, and can potentially be used to aid clinicians in the diagnostic interpretation of FFDM.
Breast Cancer Research and TreatmentPredictors of interobserver agreement in breast imaging using the Breast Imaging Reporting and Data System
33 Citations2010Anna Liza Antonio, Catherine M. Crespi
The best multivariate model identified training, use of two-view mammograms and BI-RADS categories (masses, calcifications, and final assessments) as predictors of kappa, which may be useful in increasing reliability in mammogram interpretation.
Academic RadiologyMulti-modality CADx
30 Citations2009Berkman Sahiner, Heang‐Ping Chan +11 more
A well-trained CADx system that combines features extracted from mammograms and US images may have the potential to improve radiologists' performance in distinguishing malignant from benign breast masses and making decisions about biopsies.
Academic RadiologyComputer Aid for Decision to Biopsy Breast Masses on Mammography
30 Citations2005Anna O. Bilska‐Wolak, Carey E. Floyd +2 more
The LRb classifier shows promise as a potential aid in reducing the number of biopsies performed on benign masses in breast mass classification for mammography on a new, independent database of cases not used for algorithm development.
Archives of SurgeryComputer-Aided Diagnosis for Surgical Office-Based Breast Ultrasound
22 Citations2000Ruey‐Feng Chang
An accurate preoperative diagnosis can be routinely established for surgical office-based digital US of the breast and the high negative predictive rate by the CAD system can avert benign biopsies.
Academic RadiologyPotential Effect of Different Radiologist Reporting Methods on Studies Showing Benefit of CAD
16 Citations2008Karla Horsch, Maryellen L. Giger +1 more
Use of different types of reporting data in the computation of sensitivity and specificity may result in different conclusions concerning the benefit of computer-aided diagnosis, as well as the use of user-dependent thresholds.