Hak Hee Kim

Deep learning with domain adaptation for accelerated projection-reconstruction MR

The radial k‐space trajectory is a well‐established sampling trajectory used in conjunction with magnetic resonance imaging. However, the radial k‐space trajectory requires a large number of radial lines for high‐resolution reconstruction. Increasing the number of radial lines causes longer acquisition time, making it more difficult for routine clinical use. On the other hand, if we reduce the number of radial lines, streaking artifact patterns are unavoidable. To solve this problem, we propose a novel deep learning approach with domain adaptation to restore high‐resolution MR images from under‐sampled k‐space data.

Inflammatory pseudotumor (myoblastic tumor) of the breast: a case report and review of the literature.

Inflammatory pseudotumor of the breast is a very rare cause of breast mass. To our knowledge, only a few cases have been described in the English literature. In this case, the lesion appeared on mammography as a round high‐density mass with ill‐defined margins and on sonography as an irregular mass of complex echogenicity with ill‐defined margins and an echogenic rim. The lesion was resected with no evidence of local recurrence after 3 years.

Comparison of mammography, ultrasound, and MRI in size assessment of ductal carcinoma in situ with histopathologic correlation:

Background The ability to accurately assess tumor size in ductal carcinoma in situ (DCIS) is an important clinical issue when selecting the appropriate treatment plan. Purpose To compare the accuracy of using mammography, ultrasound (US), and magnetic resonance imaging (MRI) to assess DCIS tumor size based on imaging and histopathological findings. Material and Methods Fifty-six patients with DCIS were included. Mammography, US, and MRI were reviewed, and the accuracy of the measured tumor sizes were compared with the imaging and histopathological parameters. Results If visible, tumor measurements demonstrated high reliability with the pathologically determined size, with the best results obtained using US (ku2009=u20090.851) followed by mammography (ku2009=u20090.815) and MRI (ku2009=u20090.738). Tumor size assessment was significantly more accurate when the lesion was shown as a mass on US (Pu2009=u20090.003) or MRI (Pu2009<u20090.001) with minimal and mild background parenchymal enhancement (Pu2009=u20090.016) on MRI. When mammography was used to assess tumor size, the tumors with positive estrogen receptor status and luminal A subtype demonstrated a significantly more accurate tumor size. Conclusion The combination of US and MRI, in addition to mammography, has an important role in assessing the exact tumor extent of DCIS.

Comparison study of reconstruction algorithms for prototype digital breast tomosynthesis using various breast phantoms

AbstractDigital breast tomosynthesis (DBT) is a recently developed system for three-dimensional imaging that offers the potential to reduce the false positives of mammography by preventing tissue overlap. Many qualitative evaluations of digital breast tomosynthesis were previously performed by using a phantom with an unrealistic model and with heterogeneous background and noise, which is not representative of real breasts. The purpose of the present work was to compare reconstruction algorithms for DBT by using various breast phantoms; validation was also performed by using patient images. DBT was performed by using a prototype unit that was optimized for very low exposures and rapid readout. Three algorithms were compared: a back-projection (BP) algorithm, a filtered BP (FBP) algorithm, and an iterative expectation maximization (EM) algorithm. To compare the algorithms, three types of breast phantoms (homogeneous background phantom, heterogeneous background phantom, and anthropomorphic breast phantom) were evaluated, and clinical images were also reconstructed by using the different reconstruction algorithms. The in-plane image quality was evaluated based on the line profile and the contrast-to-noise ratio (CNR), and out-of-plane artifacts were evaluated by means of the artifact spread function (ASF). Parenchymal texture features of contrast and homogeneity were computed based on reconstructed images of an anthropomorphic breast phantom. The clinical images were studied to validate the effect of reconstruction algorithms. The results showed that the CNRs of masses reconstructed by using the EM algorithm were slightly higher than those obtained by using the BP algorithm, whereas the FBP algorithm yielded much lower CNR due to its high fluctuations of background noise. The FBP algorithm provides the best conspicuity for larger calcifications by enhancing their contrast and sharpness more than the other algorithms; however, in the case of small-size and low-contrast microcalcifications, the FBP reduced detectability due to its increased noise. The EM algorithm yielded high conspicuity for both microcalcifications and masses and yielded better ASFs in terms of the full width at half maximum. The higher contrast and lower homogeneity in terms of texture analysis were shown in FBP algorithm than in other algorithms. The patient images using the EM algorithm resulted in high visibility of low-contrast mass with clear border. In this study, we compared three reconstruction algorithms by using various kinds of breast phantoms and patient cases. Future work using these algorithms and considering the type of the breast and the acquisition techniques used (e.g., angular range, dose distribution) should include the use of actual patients or patient-like phantoms to increase the potential for practical applications.n

Fully automated nipple detection in digital breast tomosynthesis

BACKGROUND AND OBJECTIVEnWe propose a nipple detection algorithm for use with digital breast tomosynthesis (DBT) images. DBT images have been developed to overcome the weaknesses of 2D mammograms for denser breasts by providing 3D breast images. The nipple location acts as an invaluable landmark in DBT images for aligning the right and left breasts and describing the relative location of any existing lesions.nnnMETHODSnNipples may be visible or invisible in a breast image, and therefore a nipple detection method must be able to detect the nipples for both cases. The detection method for visible nipples based on their shape is simple and highly efficient. However, it is difficult to detect invisible nipples because they do not have a prominent shape. Fibroglandular tissue in a breast is anatomically connected with the nipple. Thus, the nipple location can be detected by analyzing the location of such tissue. In this paper, we propose a method for detecting the location of both visible and invisible nipples using fibroglandular tissue and changes in the breast area.nnnRESULTSnOur algorithm was applied to 138 DBT images, and its nipple detection accuracy was evaluated based on the mean Euclidean distance. The results indicate that our proposed method achieves a mean Euclidean distance of 3.10±2.58mm.nnnCONCLUSIONSnThe nipple location can be a very important piece of information in the process of a DBT image registration. This paper presents a method for the automatic nipple detection in a DBT image. The extracted nipple location plays an essential role in classifying any existing lesions and comparing both the right and left breasts. Thus, the proposed method can help with computer-aided detection for a more efficient DBT image analysis.

Simplified computer-aided detection scheme of microcalcification clusters in digital breast tomosynthesis images

A computer-aided detection (CADe) algorithm for clustered microcalcifications (MCs) in reconstructed digital breast tomosynthesis (DBT) images is suggested. The MC-like objects were enhanced by a Hessian-based 3D calcification response function, and a signal-to-noise ratio (SNR) enhanced image was also generated to screen the MC clustering seed objects. A connected component segmentation method was used to detect the cluster seed objects, which were considered as potential clustering centers of MCs. Bounding cubes for the accepted clustering seed candidate were generated and the overlapping cubes were combined and examined. After the MC clustering and false-positive (FP) reduction step, the average number of FPs was estimated to be 0.87 per DBT volume with a sensitivity of 90.5%.A computer-aided detection (CADe) algorithm for clustered microcalcifications (MCs) in reconstructed digital breast tomosynthesis (DBT) images is suggested. The MC-like objects were enhanced by a Hessian-based 3D calcification response function, and a signal-to-noise ratio (SNR) enhanced image was also generated to screen the MC clustering seed objects. A connected component segmentation method was used to detect the cluster seed objects, which were considered as potential clustering centers of MCs. Bounding cubes for the accepted clustering seed candidate were generated and the overlapping cubes were combined and examined. After the MC clustering and false-positive (FP) reduction step, the average number of FPs was estimated to be 0.87 per DBT volume with a sensitivity of 90.5%.

Abstract P3-08-05: Association of mammographic density in high risk BRCA mutated tumors compared to average risk tumors and healthy controls: Analysis of Korean Hereditary Breast and Cancer Study (KOHBRA)

Introduction: Mammographic density is a well-known risk factor of breast cancer as a whole. Nonetheless only few studies have examined the association of density among high risk breast cancer regarding BRCA mutation. We examined mammographic density of 2019 breast cancer patients and 2029 healthy controls, regarding risk factors and BRCA mutation status. Method: Total 2019 breast cancer patients diagnosed between 1980 to 2011 were divided into two groups- high versus average risk group. Women with 1)family history of breast/ovarian cancer or 2)younger than age 40 or 3)bilateral cases were considered high risk group and were participants of ‘Korean Hereditary Breast Cancer study’ (KOHBRA) whom undergone BRCA testing. Density of 2029 healthy women who took screening mammogram during the same period were analyzed for comparison. Density was measured of the unaffected contralateral CC view using computer-assisted method Cumulus by single observer (10% randomly selected, intra-class correlation coefficient=0.96). Percent density (PD, dense area/breast area, %) among three groups, association with BRCA mutation status and breast cancer subtypes were examined. Results and Discussion: Percent density (PD) was significantly higher in high risk group compared to average risk and controls in a consecutive manner. This finding was consistent after adjusting age and BMI (p* High mammographic density showed to be a significant risk factor throughout different subtypes. Among the 1066 high risk group, 81.5% (869) undergone BRCA testing and 70(6.6%) had BRCA1, 78(7.3%) had BRCA2 mutations without significant difference in density. Similar strong magnitude association of mammographic density was observed in both BRCA mutated/non-mutated tumors and among subtypes. The ongoing GWAS and whole exome analysis of this population-subset will give insight into the tumor etiology and how density could stratify breast cancer risk for personalized screening especially in high risk population. Citation Format: Jisun Kim, Jong Won Lee, Sung Won Park, Hee Jung Shin, Hak Hee Kim, Sae-Byul Lee, Jong Han Yu, Hee Jeong Kim, Beom Seok Koh, Byung Ho Son, Sei-Hyun Ahn. Association of mammographic density in high risk BRCA mutated tumors compared to average risk tumors and healthy controls: Analysis of Korean Hereditary Breast and Cancer Study (KOHBRA) [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-08-05.

Abstract P3-01-02: Correlation of Mammographic breast density and tumor characteristics in Korean breast cancer patients

Introduction: Western studies have demonstrated high breast density as a strong risk factor for breast cancer, it is poorly understood whether breast density affects the diverse phenotypes of breast cancer. We examined the association between various tumor characteristics and mammographic breast density in women with breast cancer. Methods: We conducted a cross-sectional analysis in 910 Korean women diagnosed with breast cancer to evaluate the associations between breast density and tumor size, lymph node status, lymphovascular invasion, histologic grade, estrogen receptor, progesterone receptor, HER2. Breast density was classified as fatty (percent density less than 50% by a computer-assisted thresholding program, named “Cumulus™”; n = 470) or dense (percent density 50% or more; n = 440) for the cancer-free breast at the time of operation. Logistic regression was used to examine whether the relationships were modified by adjustment for body mass index, age at diagnosis, age at first birth, menopausal status, history of breast-feeding, and breast cancer staging. Results: Total 910 patients were involved, the mean age and median age at the operation was 48 years old (range 20–82), and the mean percent density was 48.09 (SD = 9.62 %: normally distributed, Kolmogorov-Smirnov test p = 0.32). Crude analysis shows that tumor size over than 0.5cm were more likely to have dense breasts compared with women with a tumor size p = 0.001 for tumor sizes 0.6–1.0cm; OR = 2.02, 95% CI = 1.09–3.74, p = 0.03 for tumor sizes 1.1–1.5cm; OR = 1.8, 95% CI = 0.97–3.33, p = 0.06 for tumor sizes 1.6–2.0cm; and OR = 1.64, 95% CI = 0.92–2.94, p = 0.1 for tumor sizes 2.1cm or more). PD and histologic grade shows reverse association between histologic grade 1 and grade 2,3. Progesteron receptor positive patients tend to have more dense(OR = 1.27, 95% CI=0.97–1.66, p = 0.07) breast than receptor negative patients, although after adjustment of age the statistical significant disappeared. Percent density was not significantly associated with, ER ( p = 0.74), HER2 ( p = 0.72). Conclusion: These results suggest that breast density is associated with tumor size and histologic grade and progesterone receptor positivity. Additional studies are needed to address whether these associations are due to just density masking the detection of some tumors, biological causation, or both. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P3-01-02.

Imaging features of bilateral breast abnormalities

There are various-causes, broad-spectrum, heterogeneous groups with various biologic profiles, and imaging features of bilateral breast abnormalities. As imaging modalities continue to be developed, the ability to detect subtle or tiny abnormalities is improved in the contralateral breast of patients already diagnosed with unilateral breast disease, especially in breast cancer patients when using magnetic resonance imaging. Furthermore, some diseases involved bilateral breast, simultaneously. The purpose of this review is to describe imaging features of the bilateral breast abnormalities-common diseases and simultaneously involving diseases. In order to provide adequate treatment and to prevent misdiagnosis, a complete understanding of the imaging and clinical features of bilateral breast abnormalities of common diseases as well as those of simultaneously involving diseases is necessary.