ISSN: 2168-9784
Journal of Medical Diagnostic Methods
Make the best use of Scientific Research and information from our 700+ peer reviewed, Open Access Journals that operates with the help of 50,000+ Editorial Board Members and esteemed reviewers and 1000+ Scientific associations in Medical, Clinical, Pharmaceutical, Engineering, Technology and Management Fields.
 
Meet Inspiring Speakers and Experts at our 3000+ Global Conferenceseries Events with over 600+ Conferences, 1200+ Symposiums and 1200+ Workshops on
Medical, Pharma, Engineering, Science, Technology and Business
 

False Positive Value of Computer-Aided Detection in Full Field Digital Mammography in Diagnosis of Breast Malignancy

Fatma Mohamed Awad*
Assistant Professor, Cairo University, Radiology, Egypt
Corresponding Author : Fatma Mohamed Awad
Assistant Professor
Cairo University, Egypt
Tel: 01144180999
E-mail: ftm_wd@yahoo.co.uk
Received: October 13, 2015; Accepted: November 30, 2015; Published: December 29, 2015
Citation: Awad FM (2015) False Positive Value of Computer-Aided Detection in Full Field Digital Mammography in Detection of Breast Malignancy. J Med Diagn Meth 4:1000.192. doi:10.4172/2168-9784.1000.192
Copyright:© 2015 Awad FM. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Related article at Pubmed, Scholar Google

Visit for more related articles at Journal of Medical Diagnostic Methods

Abstract

Objective: The aim of this study is to estimate the false-positive rate of computer-aided detection (CAD) with fullfield digital mammography (FFDM) in the detection of breast cancer.

Materials and Methods: This study was conducted in the period from February, 2013 until September, 2014. It included 659 patients who presented to the radiology department for mammography. Full-field digital mammography (FFDM), three-dimensional breast tomosynthesis and computer-aided detection (CAD) were done for all patients. The patients had breast ultrasound +/- MR Mammography, with special attention to the sites of CAD suspicious lesions. Follow up protocol was done for the selected patients.

Results: From the 659 patients included in the study, 56 patients, with 70 lesions, had false positive findings of malignancy on CAD, giving a false positive rate of 8%. These cases had suspicious lesions on CAD not detected on further investigations. The results of the follow up of all patients showed no newly developed suspicious lesions.

Conclusion: The false positive rate of a CAD system in detection of breast malignancy was introduced in this study. As per the available data, example the effect of breast density on false positive CAD scores helped the radiologists in our hospital to dismiss most of the false CAD marks with more confidence with no need for further investigation. However, if CAD marks were hard to dismiss, further investigation by ultrasound +/- MR mammography was advisable.

Keywords
Full-field digital mammography; Computer-aided detection (CAD); False positive rate; Breast malignancy
Introduction
Screening mammography is recognized as the single most important tool for the detection of early stage, clinically occult breast cancer, and numerous studies have demonstrated that screening mammography reduces breast cancer mortality [1,2].
Several studies have shown that nonpalpable breast cancers detected at screening are often visible in retrospect on prior mammograms [3,4]. Most studies evaluated the usefulness of computer-aided detection (CAD) to reduce the false-negative interpretations of mammograms [5-7].
Some studies have explored radiologists’ experience in evaluating mammograms as factors affecting the efficacy of CAD as an interpretive aid. Slightly increased benefits from CAD have been observed with experienced radiologists compared to novice radiologists and residents [8].
Throughout many years since the introduction of CAD, there were relatively few studies reporting the frequency of false-positive computer-aided detection (CAD) marks [9].
The aim of this study is to estimate the false-positive rate of computer-aided detection (CAD) with full-field digital mammography (FFDM) in the detection of breast cancer.
Materials and Methods
Patient selection
This study was conducted in the period from February, 2013 until September, 2014. It included 659 patients who were referred to the radiology department for mammography.
Inclusion criteria: Patients included in the study (fifty-six patients) were those who had suspicious lesions on computer-aided detection (CAD), with CAD score more than 60%, but were proven by further investigation and follow up to be negative for malignancy.
Exclusion criteria: Patients excluded from the study were those who had suspicious lesions on computer-aided detection (CAD) proven by further investigation to be positive for malignancy.
The study was approved by the local ethics committee, and an informed written consent was obtained from each patient prior to mammography and MR Mammography, whenever the latter was done.
Full-Field Digital Mammography (FFDM)
The mammography examinations were obtained using (Selenia, Hologic Dimensions, Bedford, USA). Two craniocaudal and two mediolateral oblique views were done for each breast. No additional views were required. Three-dimensional breast tomosynthesis and CAD were generated for all patients on the Selenia workstation. The CAD system algorithms use a triangle to mark clusters of bright spots (suggestive of clustered micro calcifications) and an asterisk to mark areas of central density and radiating lines (suggestive of a spiculated mass or asymmetrical density) [10].
Interpretation of FFDM
All obtained FFDM and breast tomosynthesis images were examined on the Selenia workstation (SecurView version7.3). In each case, an initial interpretation and assessment was done, without viewing the CAD marks. The CAD marks were then viewed and the regions of interest were re-examined on FFDM and breast tomosynthesis images, if they were not appreciated initially. The CAD marks were then classified as easy to dismiss or hard to dismiss.
Data Collection
Whether the CAD detected abnormality had features of spiculated mass, asymmetrical density or micro calcifications was detected for each patient. The location of the assumed lesions within the breast was recorded in centimeters.
Breast density was assessed according to breast imaging-reporting and data system (BIRADS) as fatty breasts, breasts with scattered fibroglandular tissue, heterogeneously dense and extremely dense breasts [11]. The former two categories were recorded as non-dense breasts, while the latter two categories were reported as dense breasts.
Further Investigation
All patients had breast ultrasound done for them; as a part of the breast screening program in the hospital, using (Aplio 500, Toshiba Medical Systems, Tokyo, Japan) and (Voluson 7, GE health Care, South Carolina, USA). Special attention was given to the sites of CAD suspicious lesions. Two patients did ultrasound outside our hospital upon their request. However, they came for follow up mammography after two years with no newly developed lesions.
For the patients with hard to dismiss CAD marks, if ultrasound offered an explanation of the cause for the CAD mark, no further investigation at that time was recommended. If ultrasound wasn’t satisfactory regarding this aspect, MR Mammography was done. MRI was performed in fifteen patients.
MRI Technique
MR Mammography was obtained using a 1.5 Tesla MRI (Optima MR 450W, GE Healthcare, South Carolina, USA) using a bilateral phased-array breast coil. The sequences used were axial STIR and sagittal fat-suppressed T2-weighted (TR/TE, 3850/67.4 ms and 4664/99.8 ms, respectively; slice thickness, 5 mm; matrix, 512 × 512), sagittal T1-weighted (TR/TE, 542/13 ms; slice thickness, 5 mm; matrix, 512 × 512), DWIs and contrast-enhanced three-dimensional dynamic Water VIBRANT-Flex sequences (TR/TE, 7.1/3.3 ms; flip angle, 12°; slice thickness, 1.5 mm; matrix, 512×512). One precontrast sequence was followed by six post contrast sequences for dynamic contrastenhanced images. Gadopentetate Dimeglumine (Magnevist; Schering, Berlin, Germany) was used as a contrast medium. The contrast medium was given intravenously over 20 s by an automatic MRcompatible injector. The dose was 0.1 mmol/kg.
The DWI sequences were performed with a two-dimensional echoplanar imaging (EPI) sequence (TR/TE, 8700/63.2 ms; slice thickness, 5.5 mm; matrix, 256× 256) in the axial plane. The images were obtained with b values of 0 and 600 mm2/s. The ADC map images were created automatically by the system.
Follow Up
All patients were followed every six months for 2 years from the initial mammography with clinical examination and ultrasound. Follow up mammography was done every one or two years according to the patient’s age, as a part of the screening program.
Statistical Analysis
The recorded data were analyzed to determine the false positive rate of CAD. McNemar’s and Chi-square tests were used to compare proportions of variables. P-values <0.05 were considered statistically significant.
Results
From the 659 patients referred to the radiology department for mammography, 56 had false positive CAD readings. The ages of the patients varied between 40 and 68 years (mean age, 48 years).
The 56 patients had 70 lesions on CAD (10 patients had multiple lesions), (Table 1). False CAD marks were in the form of clustered micro calcifications (16 lesions), spiculated masses (19 lesions) (Figures 1 and 2) and asymmetrical breast density (35 lesions) (Figure 3).
50% of the lesions suggested by CAD (35 lesions) were easy to dismiss by the radiologist, while the other 50% were hard to dismiss, 90% of which had CAD scores 90% or more, requiring further investigation (Table 2).
The rates of false positive CAD marks were higher in patients with dense breasts (11%) than in patients with non-dense breasts (6.2%). From the patients with false CAD scores, 62.5% (35 patients) had dense breasts on mammography. 80% of the lesions were in the denser parts of the breasts (upper outer quadrants and retroareolar regions) (Tables 3-5).
The lesions with false positive CAD marks were focal fibroadenosis (20 lesions), fibrocystic changes (11 lesions), single breast cysts (6 lesions) and benign-looking masses (3 lesions; one of which was an enlarged intramammary lymph node). Fourteen of the false CAD marks showed no abnormalities on further investigation (Table 6).
The results of the follow up of all the patients showed no newly developed suspicious lesions.
Accordingly, from the 695 patients who had FFDM in the study period, 56 patients, with 70 lesions had false positive findings of malignancy on CAD, giving a false positive rate of 8%.
Discussion
The false positive rate of CAD causes radiologists to dismiss high numbers of CAD marks detected in a screening population [11]. With the false positive rate, a typical screening population of 1,000 women would generate 2,000 false positive marks, while detecting approximately five cancers [12].
Few studies illustrated the false positive rate of CAD. These studies used different CAD systems. Our study was the first to be performed using Selenia, Hologic Dimensions, Bedford, USA.
In our study, 50% of the lesions were considered hard to dismiss. One study pointed out that most false positive marks were easy to dismiss and didn’t affect the radiologists’ final diagnosis [13]. In another study [12], 12% and 16% (in two versions of the CAD system) of these marks were average, hard, or very hard to dismiss. The reason for the difference between our results and others could be that, in our study, most of these lesions (90%) had CAD scores equal to or more than 90% which was worrisome to the radiologist who preferred to go for further investigation to exclude breast lesions.
Yang et al. [14] found higher false positive CAD marks in non-dense breasts than in dense breasts. Conversely, Brem et al. [15] and Malich et al. [16] showed a trend towards lower false positive rates with lower breast densities. Our study results agree with Brem et al. and Malich et al. as 62.5% of the patients with false positive CAD results, had heterogeneously dense breasts, while the rest of the patients (37.5%) had fatty breasts and scattered fibroglandular tissue. In addition, from all the patients included in the study, those with false positive CAD were 11% of the patients with dense breasts and 6.2% from those with nondense breasts. The differences were statistically significant (P-value <0.05). Also, most (80%) of the false positive lesions were located in the denser parts of the breasts (Retroareolar regions and upper outer quadrants). Summation of fibroglandular tissue is the reason for the false positive CAD marks in dense breasts, giving false spiculated margins.
Most of the false positive appearances on CAD were in the form of spiculated masses and asymmetrical densities (77%). On further imaging, fibroadenosis and fibrocystic disease were present. Again, the reason for the false CAD readings was the summation of fibroglandular tissue shadows. In another patient, the lesion suspicious on CAD was an enlarged lymph node. In two patients, the lesions were benign-looking masses, likely fibro adenomas, as seen on ultrasound and MRI. In six patients, the suspicious lesions were seen on ultrasound as simple breast cysts.
Brem et al. [15] reported 63% and 58% specificities for CAD in cases with microcalcifications in non-dense and dense breasts, respectively. Yang et al. [17] reported 67% and 69% specificities in non-dense and dense breasts. In both studies, there was no statistical difference in specificity between non-dense and dense breasts. Mahoney et al. [12] had similar results with 57.4%, 70.5% and 69.2%, 82.1% specificities for v5.0 and v7.2 in non-dense and dense breasts, respectively. The specificity of CAD for microcalcifications, in our study, was 97.2% and 98.2% in dense and non-dense breasts, respectively. This difference wasn’t statistically significant.
The limitation of the study is that the assessment of the images was made by a single radiologist; who has 10 years’ experience in breast imaging and reporting.
Conclusion
The false positive rate of a CAD system in detection of breast malignancy was introduced in this study. As per the available data, example the effect of breast density on false positive CAD scores helped the radiologists in our hospital to dismiss most of the false CAD
marks with more confidence, with no need for further investigation. However, if CAD marks were hard to dismiss, further investigation by ultrasound +/- MR mammography was advisable.
References

















Select your language of interest to view the total content in your interested language
 
Share This Article
   
 
   
 
Relevant Topics
Disc Acute Coronary Syndrome
Disc Adrenocortical Carcinoma
Disc Advanced Parkinson Treatment
Disc Advances in Alzheimers Therapy
Disc Alzheimer Disease
Disc Alzheimers Medicine
Disc Alzheimers Products & Market Analysis
Disc Alzheimers Symptoms
Disc Anatomic Pathology
Disc Angina Pectoris
Disc Aortic Dissection
Disc Atrial Fibrillation
Disc Biomarkers in Clinical Trials
Disc Biomarkers in Oncology
Disc Biopsy
Disc Cancer Pathology
Disc Cardiomyopathy
Disc Cardiovascular Diseases
Disc Case Report
Disc Clinical Diagnostic Test
Disc Clinical Microbiology
Disc Clinical Trials Research
Disc Computed Tomography
Disc Computer Aided Diagnosis
Disc Congenital Heart Disease
Disc Coronary Artery Disease
Disc Degenerative Disorders
Disc Diagnosis and treatment of Infectious Disease
Disc Diagnosis of Tuberculosis
Disc Diagnostic Alzheimer
Disc Diagnostic Imaging
Disc Diagnostic Methods
Disc Diagnostic Procedure
Disc Diagnostic Therapeutics
Disc Differential Diagnosis
Disc Disease Diagnosis
Disc Electronic health record
Disc Endocarditis
Disc Genetic Diagnosis
Disc Genetic Screening
Disc Germ cell tumours
Disc Haemodialysis
Disc In vitro Diagnostics
Disc Industrial Microbiology
Disc Laboratory Diagnosis
Disc Magnetic Resonance Imaging
Disc Medical Case
Disc Medical Certificate
Disc Medical Clinical Studies
Disc Medical Diagnosis
Disc Medical Encounters
Disc Medical Equipments
Disc Medical Indication
Disc Medical Pathology
Disc Medical Screening
Disc Medical papers
Disc Microbiology and Immunology
Disc Microbiology and Immunology Research
Disc Microbiology and Pathology
Disc Molecular Biomarkers
Disc Molecular Diagnostic Market
Disc Molecular Tools for Earlier Diagnosis
Disc Mycology
Disc Oxidative Stress Biomarkers
Disc Parkinson
Disc Parkinsonism Diagnosis
Disc Parkinsonism Gene Therapy
Disc Parkinsonism Stages and Treatment
Disc Pathology Diagnosis
Disc Percutaneous Coronary Intervention
Disc Positron Emmision Tomography
Disc Prenatal Diagnosis
Disc Principles of Diagnosis
Disc Pulmonary Embolism
Disc Retrospective Diagnosis
Disc Salivary Biomarkers
Disc Serum Biomarkers
Disc Stem cell Treatment Parkinson
Disc Surgery
Disc Surgical Case Reports
Disc Surgical Pathology
Disc Telepathology
Disc Tumor Pathology
Disc Tumour Biomarkers
Disc Virus Diagnosis
Disc clinical Microbial Pathology
 
Recommended Journals
Disc Diagnostic Techniques and Biomedical Analysis
Disc Alzheimers Disease & Parkinsonism
Disc Medical Reports & Case Studies
Disc Universal Surgery
Disc Rare Disorders: Diagnosis & Therapy
Disc Medical & Surgical Pathology
Disc Blood Research & Hematologic Diseases
Disc Imaging and Interventional Radiology
Disc Cardiovascular Diseases & Diagnosis
Disc Veterinary Science & Medical Diagnosis
Disc Medical Microbiology & Diagnosis
Disc Molecular Biomarkers & Diagnosis
  View More»
 
Recommended Conferences
Disc Radiology Congress
Jul 14-15, 2016 Cologne, Germany
Disc  Neurorehabilitation Conference
July 21-22, 2016 Brisbane, Australia
Disc Radiology and Imaging Summit  
Sep 19-20, 2016 Las Vegas, USA.
Disc Infectious Diseases and Diagnostic Microbiology   
Oct 3-5, 2016 Vancouver, Canada
Disc Radiography Conference              
Nov 17-18, 2016 Dubai, UAE
Disc Endoscopy Summit        
Nov 17-19, 2016 Atlanta, USA
View More»
 
Article Tools
Disc Export citation
Disc Share/Blog this article
 
Article usage
  Total views: 7547
  [From(publication date):
December-2015 - Jun 29, 2016]
  Breakdown by view type
  HTML page views : 7498
  PDF downloads :49
 
 

Post your comment

captcha   Reload  Can't read the image? click here to refresh

 
OMICS International Journals
 
Make the best use of Scientific Research and information from our 700 + peer reviewed, Open Access Journals
 
 
OMICS International Conferences 2016-17
 
Meet Inspiring Speakers and Experts at our 3000+ Global Annual Meetings
 
 

Contact Us

Agri, Food, Aqua and Veterinary Science Journals

Dr. Krish

agrifoodaquavet@omicsinc.com

1-702-714-7001 Extn: 9040

Clinical and Biochemistry Journals

Datta A

clinical_biochem@omicsinc.com

1-702-714-7001Extn: 9037

Business & Management Journals

Ronald

business@omicsinc.com

1-702-714-7001Extn: 9042

Chemical Engineering and Chemistry Journals

Gabriel Shaw

chemicaleng_chemistry@omicsinc.com

1-702-714-7001 Extn: 9040

Earth & Environmental Sciences

Katie Wilson

environmentalsci@omicsinc.com

1-702-714-7001Extn: 9042

Engineering Journals

James Franklin

engineering@omicsinc.com

1-702-714-7001Extn: 9042

General Science and Health care Journals

Andrea Jason

generalsci_healthcare@omicsinc.com

1-702-714-7001Extn: 9043

Genetics and Molecular Biology Journals

Anna Melissa

genetics_molbio@omicsinc.com

1-702-714-7001 Extn: 9006

Immunology & Microbiology Journals

David Gorantl

immuno_microbio@omicsinc.com

1-702-714-7001Extn: 9014

Informatics Journals

Stephanie Skinner

omics@omicsinc.com

1-702-714-7001Extn: 9039

Materials Science Journals

Rachle Green

materialsci@omicsinc.com

1-702-714-7001Extn: 9039

Mathematics and Physics Journals

Jim Willison

mathematics_physics@omicsinc.com

1-702-714-7001 Extn: 9042

Medical Journals

Nimmi Anna

medical@omicsinc.com

1-702-714-7001 Extn: 9038

Neuroscience & Psychology Journals

Nathan T

neuro_psychology@omicsinc.com

1-702-714-7001Extn: 9041

Pharmaceutical Sciences Journals

John Behannon

pharma@omicsinc.com

1-702-714-7001Extn: 9007

Social & Political Science Journals

Steve Harry

social_politicalsci@omicsinc.com

1-702-714-7001 Extn: 9042

 
© 2008-2016 OMICS International - Open Access Publisher. Best viewed in Mozilla Firefox | Google Chrome | Above IE 7.0 version