Introducere



UNIVERSITY OF MEDICINE AND PHARMACY CRAIOVA

FACULTY OF MEDICINE

[pic]

The role of ultrasound elastography in diagnosis breast focal lesions

-ABSTRACT-

Scientific coordinator,

Professor Andrei Bondari, MD, PhD

PhD Student,

Ioana Andreea Gheonea

CRAIOVA

2011

Contents

Background 3

Material and methods 3

Results and discussions 7

Elastography study 7

Qualitative analysis 7

Quantitative analysis 8

Conclusions 10

References 12

Curriculum vitae 13

Key words

Ultrasound elastography, mammography, ultrasound, magnetic resonance imaging, immunohistochemistry, breast focal lesions, breast cancer

BACKGROUND

Breast cancer is an important cause of morbidity and mortality, representing the second cause of death by cancer in women. A study made in 2005 by the Society for Women's Health Research indicated breast cancer as the most important disease among women. According to World Health Organization more than 1.2 million people are annually diagnosed breast cancer and over 500.000 die from this disease. However, breast cancer mortality rate has declined since 1990 due to early diagnosis and treatment methods improvement.

Ultrasound elastography has recently been developed for clinical applications, allowing reconstruction of tissue elasticity distribution and directly revealing tissue physical properties. Ultrasound elastography estimates axial strength of tissues along the ultrasound. The method reveals physical tissue properties, characterizing the difference in hardness between pathological and normal tissue. Sonoelastography is a complementary method for breast nodules, increasing two-dimensional ultrasound specificity. This could reduce the number of benign biopsy results for non-suspect lesions, as well as following examinations. The validation of the elastography could probably represent the initial step to tactile imaging, visualizing and reconstructing tissue elasticity distribution on breast covered surface.

Magnetic resonance imaging is a complementary technique for mammography and ultrasound. Malignant lesions can be identified by studying the behavior after intravenous paramagnetic contrast. Recently, new MRI techniques have improved, with possible applications in breast cancer diagnosis, techniques that are still being evaluated (MR spectroscopy, diffusion-weighted images, perfusion, bold).

This paper is divided into two parts: a general part – state of knowledge and a personal part – with personal contributions.

In the first part, we tried to present the latest theoretical data from specialized literature on the topic. In the special part we presented the material and methods used and results and discussions arising from the study of the groups of patients.

MATERIAL AND METHOD

The study was conducted at the University of Medicine and Pharmacy Craiova, Department of Radiology and Medical Imaging, Center for Research in Gastroenterology and Hepatology and Medical Center Camen in Craiova during November 2006 and August 2011 with prospective data. The total number of patients with breast pathology included in the study was 345 at who were added 17 cases with normal breast, of which two men evaluated for other conditions. Mammary pathology was: breast cancer (n=95), fibroadenoma (n=87), fibrocystic mastopathy (n=79), breast cysts (n=60), lipomas (n= 14), papillomas (n=4), breast abscess (n=6).

More studies were conducted on the group of patients evaluated:

• Demographic study

• Imaging study

• Cytological and immunohistochemical study

• Elastography study

All patients included were suspected of breast tumors by at least one of the following: detection of self palpation lesions, palpation by the physician, evidence of suspicious injuries in females examined by mammography within breast cancer screening program, highlighting suspicious imaging mass (ultrasound, mammography, CT, MRI), breast pain and inflammatory phenomena. Data were collected through a structured form.

Imaging examinations included the performance of breast ductal ultrasound and elastography in all cases and selectively mammography and/or magnetic resonance, with an interval between examinations of up to 7 days. In cases with suspected breast cancer imaging was performed cytological examination or surgery with histopathology exam.

Before imaging explorations or minimally invasive and sampling of biological materials, all patients signed an informed consent after explaining the details and clarify any arising ambiguities.

Breast ultrasound and elastography were performed using a Hitachi 8500 EUB, Hitachi Preirus and Siemens Acuson Antares Preirus with multifrequency transducers (7-15 MHz), allowing adaptation to breast thickness, thus achieving optimization of spatial resolution and contrast, all breast plans being examined with maximum accuracy.

Conventional ultrasound examinations were performed with the patient supine or oblique, ipsilateral arm raised above his head to clear upper-outer quadrant, axillary extension and submammary groove. Longitudinal and transverse sections were made. Good contact between probe and skin and an even distribution of gel were aimed. Examination of axillary lymph node groups, subclavian and internal mammary artery by axial and transverse sections on vascular tracks was also present.

The examination included vascular ultrasound Doppler study, given that malignant lesions have multipolar, centripetal, distortional blood supply, neovascularization appearing even in tumor sized less than 5 mm.

Similar to the Power Doppler mode examination, examination by elastography interface presents on the right side of the screen gray mode image and to the left the elastography images. Elastography images were obtained by conventional ultrasound lesions evaluation. Next a region of interest (ROI) was manually set that included both breast lesions and subcutaneous tissue and pectoral muscles without coastal springs. ROI areas involved the inclusion of a sufficiently large area of surrounding tissues because elastography values are tightly related to all ROI structures compressibility.

The pressure applied was set between 1 and 6 score (score 1 indicating the lowest pressure and the 6 the highest). We applied a maximum pressure of 3 or 4, because too much pressure can cause compression even in hard lesions, which can lead to false information. Elastography images were evaluated according to the Tsukuba score of elasticity developed by Itoh and Ueno.

Hitachi elastography system is set so that very hard areas appear dark blue and still, with the decrease in hardness, the colors go through shades of green, yellow and red. Regarding the Siemens system, tissue elasticity was visualized in color and monochrome mode, hard areas being displayed in red or black and elastic areas in blue or white.

Lesions were classified using a five grade score: score 1 for lesions with identical elasticity with surrounding breast tissue, score 2 for lesions with alternating hard and elastic areas. Score 3 was used for lesions with peripheral elasticity and central hardness. Score 4 was used in entirely hard lesions, but with surrounding elastic tissue and score 5 for entirely non-compressible injuries as well as for adjacent tissue areas. Scores 1 to 3 are used for benign lesions, while malignant lesions are characterized by scores 4-5. Also, regarding ultrasound performed with Hitachi module, we calculated the fat-to-lesion-ratio (FLR).

Ductal ultrasound was added to elastography examination. The latter were performed with the patient in sitting position, facing the examiner, with hands on hips to examine the upper part of the breast and with hands on their hands the inferior region.

The probe was positioned perpendicular to the skin and radial on the breast, with one edge placed near the areola, and the other to the periphery. Transducer was oriented so that the areola to appear to the left of the image.

This was rotated around the areola, and when a ductal image was identified rotation was stopped and we performed lateral movement back and forth to assess the entire duct and its branches. A second movement was made for the analysis of peripheral described areas. So, clockwise and anti-clockwise lobe movements assured lobe analysis, with tracking and highlighting all ducts pathological changes.

Computed analysis of the images

Each elastography film was analyzed using a computerized image processing java tool (ImageJ) developed at the National Institutes of Health, Bethesda, Maryland for which a special plug-in dynamic analysis was created in the IT Department of UMF Craiova.

To minimize bias induced by the examining physician, all digital post-processing analysis were performed without knowledge of clinical and laboratory data of patients. For each US elastography film of 10 seconds each (about 125 personnel) have automatically been selected only color frames (containing information elastography) for which the histograms were calculated.

Final numerical value assigned to each examination is composed of the average individual values of the histograms ​​of each elastography frame.

Statistic analysis

Statistic analysis was performed using MedCalc Software 9, 2008, Mariakerke, Belgium.

For each imaging method were calculated sensibility, specificity, positive predictive value (VPP) and negative predictive value (VPN).

Our study also included the calculation of correlation Spearman and Kendall Tau index.

RESULTS AND DISCUSSION

ELASTOGRAPHY STUDY

➢ Qualitative analysis

Lesions were classified using a five grade score: score 1 for lesions with identical elasticity with surrounding breast tissue, score 2 for lesions with alternating hard and elastic zones. Score 3 was used for lesions with peripheral elasticity and central hard area. Score 4 was used in entirely hard lesions, but with surrounding elastic tissue and score 5 for entirely non-compressible lesions, as well as adjacent tissue areas.

The analysis was conducted separately for patients to whom elastography was performed using Siemens and Hitachi EUB 8500 devices.

In order to estimate elastography score value using Hitachi EUB 8500 ultrasound in the differential diagnosis of focal lesions of the breast we realized the ROC analysis that revealed a value of 0.918 (95%, 0.874 - 0.950) (P3. Thus, for this value the sensitivity and specificity obtained were 81.5% and 92.5%.

For estimating elastography score value using Siemens Acuson Antares ultrasound in the differential diagnosis of focal breast lesions we realized the ROC analysis that revealed a value of 0.934 (95%, from 0.874 to 0.971) (P3. Thus, for this value the sensitivity and specificity obtained were 83.3% and 91.1%.

Also, for estimating the elastography score value on an even larger number of patients we developed statistical analysis across the entire group consisting of patients examined with both ultrasound systems. Thus, ROC analysis revealed a value of 0.923 (95%, 0.890 to 0.949) (P 3. Thus, for this value the sensitivity and specificity obtained were 82.1% and 92.0%. Out of the total of malignant lesions evaluated by the two elastography systems, one lesion had score 1, six lesions had score 2 and ten lesions had score 3.

➢ Quantitative analysis

o Fat-to-lesion-ratio (FLR)

Quantitative FLR analysis using ultrasound Hitachi 8500 and Hitachi EUB Preirus systems was performed on a number of 224 patients with benign and malignant focal breast lesions. The average FLR value for benign lesions was 2.28 and 7.49 for malignant ones.

For estimating FLR value used in the differential diagnosis of focal lesions of the breast we realized the ROC analysis that revealed a value of 0.922 (95%, 0.879 to 0.954) (P 4.25. Thus, for this value the sensitivity and specificity obtained were 75.4% and 93.1%.

o Qualitative analysis (elastography score) - quantitative analysis (FLR) correlation

Correlation between qualitative analysis methods (elastography score) and static quantitative (FLR) analysis was performed using Spearman and Kendall’s Tau coefficients. Spearman correlation coefficient between the FLR values ​ and elastography score was 0.827 (95% CI 0.781 to 0.864, p ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download