Volume 14, No. 2, December 2010 목차 CONTENTS 원저자화율차이로인해왜곡된영상으로부터금속바늘의위치결정김은주 김대홍 87 ORIGINAL CONTRIBUTIONS Determining the Location of Metallic Needle

Transcription

1 ISSN

2 Volume 14, No. 2, December 2010 목차 CONTENTS 원저자화율차이로인해왜곡된영상으로부터금속바늘의위치결정김은주 김대홍 87 ORIGINAL CONTRIBUTIONS Determining the Location of Metallic Needle from MR Images Distorted by Susceptibility Difference 87 EJ Kim, DH Kim 삼중음성유방암의자기공명영상소견 : 비삼중음성유방암과의비교최재정 김성헌 차은숙 강봉주 이지혜 95 MRI Findings of Triple Negative Breast Cancer: A Comparison with Non-Triple Negative Breast Cancer 95 이소연 정승희 임현우 송병주 JJ Choi, SH Kim, ES Cha, BJ Kang, JH Lee, SY Lee, SH Jeong, HW Yim, BJ Song 정상노인및경도인지장애및알츠하이머성 103 치매환자에서의한국인뇌구조영상표준판개발김민지 장건호 이학영 김선미 류창우신원철 이수열 Development of a Korean Standard Structural Brain Template in Cognitive Normals and Patients with Mild Cognitive Impairment and Alzheimer s Disease MJ Kim, GH Jahng, HY Lee, SM Kim, CW Ryu, WC Shin, SY Lee 103 활성자극파라다임 fmri에서저주파요동성분분석나성민 박현정 장용민 115 Low Frequency Fluctuation Component 115 Analysis in Active Stimulation fmri Paradigm SM Na, HJ Park, YM Chang 유방암환자에서추가병변평가를위한 3 테슬러유방자기공명영상의임상적경험이지혜 김성헌 강봉주 최재정 이아원 121 Clinical Experience of 3T Breast MRI in Detecting the Additional Lesions in Breast Cancer Patients 121 JH Lee, SH Kim, BJ Kang, JJ Choi, AW Lee 뇌피질이형성증의 3차원뇌표면연출영상황승배 곽효성 이상용 진공용 126 Three-Dimensional Brain Surface Rendering Imaging of Cortical Dysplasia 126 한영민 정경호 SB Hwang, HS Kwak, SY Lee, GY Jin, YM Han, GH Chung

3 증례보고뼈주위지방종의자기공명영상소견 : 증례보고배형주 홍석주 김예림 강은영 김학준류영준 정원용 134 CASE REPORTS MR Imaging Findings of Parosteal Lipoma: Case Report HJ Bae, SJ Hong, YL Kim, EY Kang, HJ Kim, YJ Ryu, WY Jung 134 췌장중복낭의자기공명영상소견과수소자기공명분광법 : 증례보고박성희 김미영 서창해 이건영최석진 조재영 139 MR Images and 1 H MR Spectroscopy of Enteric Duplication Cyst of the Pancreas in an Adult SH Park, MY Kim, CH Suh, KY Lee, SJ Choi, JY Cho 세남아에서발생한췌장의비기능성악성신경내분비종양 : 증례보고임세웅 이영환 최시성 조현선 145 Non-Functioning, Malignant Pancreatic Neuroendocrine Tumor in a 16-Year-old Boy: A Case Report SW Lim, YH Lee, SS Choi, HS Cho 145 일차성간림프종 : 자기공명영상과병리소견의연관김한나 신유리 나성은 정은선 오순남최준일 정승은 이영준 151 Primary Hepatic Lymphoma: MR Imaging and Pathologic Correlation HN Kim, YR Shin, SE Rha, ES Jung, SN Oh, JI Choi, SE Jung, YJ Lee 151 논문외 대한자기공명의과학회지제14권색인 156 대한자기공명의과학회회원카드 159 대한자기공명의과학회회원가입안내 160 대한자기공명의과학회입회원서 161 대한자기공명의과학회지윤리규정및투고규정 162 저작권이양에관한동의서 167

4 Determining the Location of Metallic Needle from MR Images Distorted by Susceptibility Difference Eunju Kim 1, Daehong Kim 2 Purpose : To calculate the appearance of the image distortion from metallic artifacts and to determine the location of a metallic needle from a distorted MR image. Materials and Methods : To examine metal artifacts, an infinite metal cylinder in a strong magnetic field are assumed. The cylinder s axis leaned toward the magnetic field along some arbitrary angle. The Laplace equation for this situation was solved to investigate the magnetic field distortion, and the simulation was performed to evaluation the image artifact caused by both readout and slice-selection gradient field. Using the result of the calculation, the exact locations of the metal cylinder were calculated from acquired images. Results : The distances between the center and the folded point are measured from images and calculated. Percentage errors between the measured and calculated distance were less than 5%, except for one case. Conclusion : The simulation was successfully performed when the metal cylinder was skewed at an arbitrary tilted angle relative to the main magnetic field. This method will make it possible to monitor and guide both biopsy and surgery with real time MRI. Index words : Magnetic resonance imaging (MRI) Metallic artifacts Simulations of image distortion Introduction Various metallic devices have been used for implantations or percutaneous procedures, but these metallic implants cause severe magnetic resonance imaging (MRI) image distortions called susceptibility artifacts. Metal-induced artifacts arise from the susceptibility difference between the body tissue and the metal, and the artifacts can be strong enough to deteriorate the whole image (1, 2). A number of methods have been introduced to overcome metallic artifact image distortions. The view angle tilting (VAT) method suggested by Cho et al. is one of the most promising methods to correct the artifact and proved to be useful in medical practice (3, 4). However, the JKSMRM 14:87-94(2010) 1 Fusion Technology Medical Device Team, Korea Food & Drug Administration 2 Molecular Imaging and Therapy Branch, National Cancer Center Received; October 4, 2010, revised; December 14, 2010, accepted; December 15, 2010 Corresponding author : Daehong Kim, Ph.D., Molecular Imaging and Therapy Branch, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si Gyeonggi-do , Korea. Tel Fax dkim@ncc.re.kr - 87-

5 Eunju Kim et al images obtained by the VAT method are convolved with a square function, which is related to sliceselection gradient, so they are blurred. Recently, in order to compensate this image blurring, the additional z-phase encoding technique called as SEMAC(Slice Encoding for Metal Artifact Correction) was reported (5). When imaging patients with metallic devices, such as implants or biopsy needles, image degradation and geometric mis-registration must be solved (3, 6, 8, 9). In particular, it is important to know the exact location of the metallic needle for image-guided biopsy. A few investigators have successfully calculated the amount of shift caused by metallic artifacts (1, 7). Ludeke et al. (1) calculated the shift when the metal cylinder was perpendicular to the main magnetic field (1). Ladd et al. (7) studied the case where the metal cylinder is at an arbitrary angle with respect to the magnetic field, but they did not consider the slice distortion caused by the slice-selection gradient. If the slice-selection gradient is not taken into account with the readout gradient, then the calculated artifact is different from the real artifact. To explore the metal artifact, we assumed an infinite, non-magnetic metallic cylinder in a strong magnetic field. The cylinder s axis leaned toward the main magnetic field along some arbitrary angle. We solved the Laplace equation for these situations to investigate the magnetic field distortion (7), and then we successively simulated the effect of the readout gradient and slice gradient magnetic field. As a result, the artifact s shape and size were generated, and the location of the cylinder s center was calculated and determined. We compared the results of the calculated images with the real images. The real images were acquired on a 1.5 T clinical MRI system (GE HORIZON). The phantom was made of a stainless steel rod immersed in a bottle containing doped water. The goal of our new method is to exactly calculate the location of the cylinder s center, not to reduce or remove the artifact, so this method can be applied during surgery with a real-time open magnet MRI system for successfully guiding needles. Materials and Methods Calculation of Field Distortion Ladd et al. calculated the field distortion caused by an infinite cylinder that leans along a strong, homogeneous magnetic field, B 0, which is illustrated in Fig. 1, where the permeability inside and outside the cylinder are μ i and μ e, respectively, and the radius of the cylinder is R. For convenience, the axis of cylinder is parallel to y-axis, and the tilt angle between the magnetic field and the plane perpendicular to the cylinder s axis is θ. The external magnetic field can be expressed as B 0 = y^b 0 sin θ+z^b 0 cos θ, [1] where x^, y^, z^, are unit vectors for each direction. The resulting fields are B μ e -μ i μ e -μ i i = z^b 0 cos θ 1- +y^b 0 sin θ1- μ e +μ i μ e μ i Fig. 1. An infinite metallic cylinder is put in a strong homogeneous magnetic field. The angle between the main magnetic field and the plane perpendicular to the metal cylinder is θ. The right hand coordinate system is used. The axis of the cylinder is parallel to y-axis and the main magnetic field Bo lies in y-z plane. B μ z 2 - x 2 e -μ i e = z^b 0 cos θ 1- R μ 2 (z 2 +x 2 ) 2 e +μ i μ e -μ i 2xz - x^b 0 cos θ R 2 y^b 0 sin θ, [2] μ (z 2 +x 2 ) 2 e +μ i where B i and B e denote the magnetic field inside and outside cylinder, respectively. The relative permeability μcan be written as a - 88-

6 Determining the Location of Metallic Needle from MR Images Distorted by Susceptibility Difference function of susceptibility χ, as follows: μ=1+χ. For MRI χ 1, so we can approximate μ e -μ i χ =, [3] μ e +μ i 2 χ=χ e -χ i Computer Simulations of the Distortion Distortions were simulated using Mathematica (Wolfram Research, version 3.0). When a slice-selection gradient is applied, the slice s shape is distorted by the susceptibility difference, so we simulated a distorted shape caused by the slice selection gradient and by the superposition of distortions caused by the readout gradient. First, for the effect of slice gradient, we assumed that an infinitely thin slice was selected by the slice selection gradient: G ss =-z^g ss sin θ+y^g ss cos θ, [4] whose direction was perpendicular to the main magnetic field. A frequency = (B 0 +yg ss ) to select the slice at y=z tan θresults in an actual slice at r(x, y, z) defined by Using Eq. [2], Eq. [4] and Eq. [5], the selected slice at exterior of the cylinder is given as χ B 0 z 2 - x 2 y=z tan θ+ R 2. [6] 2 (z 2 +x 2 ) 2 G ss Because we are not concerned with the inner metal, no calculation is needed for the interior of the cylinder. Using Mathematic, this situation is plotted in Fig. 2a. Next, for the distorted slice we applied a readout gradient. In the readout gradient, the points move along the direction of the readout gradient which is assumed to be applied in the same direction of B0, that is, G r = z^g r cos θ+y^g r sin θ. [7] We also calculated this shift in a similar way we used in the case of a slice-selection gradient. The pixel r will be displayed at position r -. r=r - χ B 0 z 2 - x 2 -r =-cos θ R 2 [8] 2 (z 2 +x 2 ) 2 From Eq. [8], we can calculate the shift for each direction, as follows: χ B 0 z 2 - x 2 z=-cos 2 θ R 2, 2 (z 2 +x 2 ) 2 G r G r B e +(xx^+yy^+zz^) G ss=b 0 +(z tan θy^+zz^) G ss. [5] Fig. 2. Simulated slices were plotted with Mathematica. The tilted angle of the metal cylinder is 30. (a). The distorted slice is selected by the slice selection gradient. The direction of the slice gradient is perpendicular to main magnetic field. (b). When the readout gradient is applied to the slice in (a), each pixel in the slice is shifted along the readout gradient direction. The direction of the readout gradient is parallel to the main magnetic field. (c). The top view of (b) is plotted where the point f represents the folded point and p represents the piled point

7 Eunju Kim et al χ z 2 - x 2 y=-cos θsin θ R 2. [9] 2 (z 2 +x 2 ) 2 Each of the points on the surface of Eq. [6] will be shifted by the readout gradient calculated with Eq. [9]. As a result of Eq. [6] and Eq. [9], the components of the distorted surface are expressed as χ z-component: Z=z-cos 2 θ 2 G r R 2 x-component: X=x χ B 0 y-component: Y=z tan θ+ 2 R 2 χ z 2 - x 2 - cos θsin θ R 2 [10] 2 (z 2 +x 2 ) 2 The simulated image in Fig. 2 is relatively similar to the real image generated from MR scanner. There were regions of condensed lines that appeared as bright as regions on the real image. To understand the shape of the susceptibility artifact, we watched the brightest points. As shown in Fig. 2c, the brightness is caused when a slice is folded at one point (i.e., the folded point, f ) or when the slice is piled up around a point (i.e., the piled point, p ). The tangent line at the folded point is parallel to the direction of slice selection gradient, so the folded point can be calculated by simple differentiation. Initially, due to symmetry, the folded point will be on the z-axis, so all x s in Eq. [10] can be set to zero. Then, we find the point where the incline of the tangent line is -1/tan θ, as follows: dy 1 1 dy dz B 0 G r B 0 G ss B 0 G r z 2 - x 2 (z 2 +x 2 ) 2 z 2 - x 2 (z 2 +x 2 ) 2 dz tan θ+ cos θsin θ - G r G ss χb 0 R 2 z -3 = = dz 1 dz 1+cos 2 θ G χb 0 R 2 z -3 r In the case of x 2 =z 2 in Eq. [10], the shift by susceptibility artifact does not occur and those points of x 2 =z 2 in Eq. [10] stay stationary under the susceptibility artifact. So, other points around those stationary points are shift and piled up at those stationary points. So, the piled point will lie on the plans in Eq. [13]. x=z, x=-z, y=z tan θ [13] Since Eq. [12] relates to the folded point, we can plot the distance from the cylinder s center to the folded point as a function of the metallic cylinder s tilt angle relative to the main magnetic field (Fig. 3). Piled points are located on the matching lines of all plans in Eq. [13]. By using these points, we can calculate the exact location of the cylinder s center in the image without any complementary devices. The simulated images in Fig. 4 show the slice shape caused the by the magnetic field distortion, which varied as a function of tilt angle (0, 15, 30, 45, 60, and 75 ). We confirmed that there were large changes in artifact size that were caused by varying the tilt angle. This result corresponded to that shown in Fig. 3. All of these findings indicate that image distortion is not proportional to tilt angle. As θapproached 90, the artifact decreased in size and eventually disappeared. Phantom and Image A stainless-steel cylindrical rod was chosen for the phantom imaging. The stainless steel rod was inserted in a water container doped with CuCl 2. The rod s susceptibility was not available because the rod s constituents were not published, so the rod s susceptibility was calculated from the image (θ=0) using Eq. [12] and Eq. [13] as χ= The 1 = -. [11] tan θ The result is tan θ 1 z 3 = χb 0 R 2 - cos 2 θ. G ss The distance from the zero-point to the folded point is tan θ 1 1 r 3 = χb 0 R 2 -. [12] cos θ G ss G r G r Fig. 3. The distance from the cylinder center to the folded point as a function of the metallic cylinder s tilt angle relative to the main magnetic field. X-axis is plotted versus radians. The distance was not proportional to the tilt angle

8 Determining the Location of Metallic Needle from MR Images Distorted by Susceptibility Difference length and diameter of the cylindrical rod were 20 cm and 1 mm, respectively. The phantom images were acquired with a conventional spin echo pulse sequence on a 1.5 T MRI system (General Electric Medical Systems, Milwaukee, WI). The matrix size was , and the field of view (FOV) was 16 cm. Bandwidth was +/-16 khz. Slice thickness was 1 mm. Repetition time (TR) and echo time (TE) were msec and 20.0 msec, respectively. The actual strength of the readout and slice gradients were gauss/cm and gauss/cm, respectively. The tilt angles of stainless still rod to main magnetic field were tried as 0, 8, 24, 32, 44, 56 for MR image. Comparison of Simulated and Acquired Images Using the piled points and folded point, the center of metallic cylinder was determined from MR images. Also, the distance between the center and the folded point was measured. The measured distances were compared to the calculated value from Eq. [12]. Firstly, the piled point were identified on the image as one of the brightest points, and then a straight line were drawn using x=z or x=-z in Eq. [13], which will connect the piled point with the center of the metal cylinder. Secondly, we identified the folded point, which is the other brightest point, and then the vertical line was drawn which connected the folded point and the center of metal cylinder. Finally, the overlapped point of two lines is the center of the metal cylinder. We measured the number of pixels from the center to the folded point, and we could measure the distance from the center to the folded point from images. Using Eq. [12], we could calculate the distance from the center to the folded point for each tilt angle and compared these values to the measured image values. Results We have a sound understanding of the artifact s shape from the simulated metallic artifact. The simulated images with various tilt angles in Fig. 4 coincide with the real images from MR scanner in Fig. 5. In the direction of read-out gradient, one can observed little changes with the distance between the brightest points in Fig. 4 and Fig. 5. On the other hand, in the direction of phase gradient the ranges of artifact decrease Fig. 4. Simulated slices are plotted for various tilt angles ((a)-(f)): 0, 15, 30, 45, 60, and 75. The shape of the magnetic field distortion varied as a function of tilt angle. Each slice was generated with the same parameters, except for the tilt angle

9 Eunju Kim et al Table 1. Calculated and Measured Distances from the Cylinder s Center to the Folded Point Tilt Angle(degree) Distance from calculation(cm) Distance from measurement(cm) Error(%) error = { measure - calculate /measure} 100 Fig. 5. Cross sectional images of the stainless steel cylindrical rod were acquired. Its diameter and length were 1 mm and 20 cm, respectively. The tilt angles ( ) measured during imaging were 0, 8, 24, 32, 44, 56, and 73 ((a)-(g)), respectively. The readout gradient direction was top-to-bottom. The artifact shapes did not increase in size along the readout encoding direction, which is similar to the results shown in Fig. 4. according to the tilt angle. As mentioned above, the location of the center can be determined with real image. The distance from the center to the folded point can be also measured from the real image. The distances measured from images were compared to the calculated value using Eq. [12] and those distances were in good agreement, as shown in Table 1. The calculated distances of Table 1 show that the distance does not make a big change according to the tilt angle. These results can be observed in Fig. 3 and both simulated and real images. Percentage errors in Table 1 were less than 5%, except for one case. Errors can occur because it is difficult to precisely select the folded and piled points. For a pixel size of cm, considerable error may result from mispositioning by one or two pixels. Despite this challenge, these results are in good agreement for simulated and acquired images. Discussion The shape of the metallic artifact was determined with simulation under the assumption that the cylinder was infinite in length and skewed in an arbitrary tilted angle relative to the main magnetic field. Artifact shape can be calculated with the method we presented here. Our method is applicable for a stainless steel cylinder as well as other less susceptible materials (e.g., aluminum or titanium). The calculated and measured distances were similar, but there was some percentage of error in the measurement: %. One reasonable explanation is that we did not perfectly mark the folded point and piled point on the images; mismarking by a few pixels can cause considerable - 92-

10 Determining the Location of Metallic Needle from MR Images Distorted by Susceptibility Difference errors. In practice, marking the folded point is easier than marking the piled point; then, using Eq. [12], a user can find the metal cylinder s location. Additionally, if we know the magnetic susceptibility of the implanted metal and the cylinder s tilt angle, we can calculate the radius of the implanted cylinder from Eq. [12]. Precisely knowing the cylinder s location is useful in medical practice. Artifact shape and the location of metal cylinder implanted to human body can be determined, if the metallic cylinder is assumed to be infinite. Magnetic field distortion cannot be calculated at the sharp end of metallic biopsy needles (8-10) because an exact solution of the Laplace equation at the needle s point has not been found (8-10). Recently, artifacts caused by arbitrarily-shaped metallic implants were simulated with numerical methods and complex computer simulation (11). However, there is no method for measuring the location of metallic implants using only images, so further research is needed to be able to measure the location and shape of metallic implants Acknowledgments This research was supported by the Pioneer Research Center Program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology ( ). References 1.Lü deke KM, Röschmann P, Tischler R. Susceptibility artefacts in NMR imaging. Magn Reson Imaging 1985;3: Schenck JF. The role of magnetic susceptibility in magnetic resonance imaging: MRI magnetic compatibility of the first and second kinds. Med Phys 1996;23: Butts K, Pauly JM, Daniel BL, Kee S, Norbash AM. Management of biopsy needle artifacts: techniques for RFrefocused MRI. J Magn Reson Imaging 1999;9: Cho ZH, Kim DJ, Kim YK. Total inhomogeneity correction including chemical shifts and susceptibility by view angle tilting. Med Phys 1988;15: Lu W, Pauly KB, Gold GE, Pauly JM, Hargreaves BA. SEMAC: Slice Encoding for Metal Artifact Correction in MRI. Magn Reson Med 2009;62: Arbogast-Ravier S, Gangi A, Choquet P, Brunot B, Constantinesco A. An in Vitro Study at Low Field for MR Guidance of a Biopsy Needle. Magn Reson Imaging 1995;13: Ladd ME, Erhart P, Debatin JF, Romanowski BJ, Boesiger P, McKinnon GC. Biopsy needle susceptibility artifacts. Magn Reson Med 1996;36: Lufkin R, Teresi L, Hanafee W. New Needle for MR-Guided Aspiration Cytology of the Head and Neck. AJR Am J Roentgenol 1987;149: Lufkin R, Teresi L, Chiu L, Hanafee W. A Technique for MR- Guided Needle Placement. AJR Am J Roentgenol 1988;151: Liu H, Martin AJ, Truwit CL. Interventional MRI at High- Field (1.5 T): Needle Artifacts. J Magn Reson Imaging 1998; 8: Balac S, Caloz G, Cathelineau G, Chauvel B, de Certaines JD. Integral method for Numerical Simulation of MRI Artifact Induced by Metallic Implants. Magn Reson Med 2001;45:

11 Eunju Kim et al 대한자기공명의과학회지 14:87-94(2010) 자화율차이로인해왜곡된영상으로부터금속바늘의위치결정 1 식품의약품안전청융합기기팀 2 국립암센터분자영상치료연구과 김은주 1 김대홍 2 목적 : 금속에의한영상왜곡에대한정확한계산하고영상으로부터의금속물질의위치결정한다. 대상및방법 : 주자기장과일정각도를이루는무한히긴비자성금속실린더에대한라플라스방정식을풀고, 이결과를이용하여절편선택경사자계와주파수부호화경사자계에의한영상에왜곡을계산한다. 계산결과를바탕으로하여왜곡된영상으로부터원통형보철물의위치를계산한다. 결과 : Folded point와금속실린더의중심사이의거리를영상으로부터측정하여계산결과와비교한다. 측정결과와계산결과간의퍼센트오차는한경우를제외하고 5% 이내였다. 결론 : 금속실린더가자기장하에있을때, 영상의왜곡을시뮬레이션하였고, 이기술은생검술또는외과수술등을자기공명영상법을이용여실시간모니터링하는데적용할수있을것으로기대한다. 통신저자 : 김대홍, ( ) 경기도고양시일산동구일산로 323, 국립암센터분자영상치료연구과 Tel Fax dkim@ncc.re.kr - 94-

12 MRI Findings of Triple Negative Breast Cancer: A Comparison with Non-Triple Negative Breast Cancer Jae Jeong Choi 1, Sung Hun Kim 1, Eun Suk Cha 1, Bong Joo Kang 1, Ji Hye Lee 1, So-Yeon Lee 1, Seung Hee Jeong 2, Hyeon Woo Yim 2, Byung Joo Song 3 Purpose : To evaluate the magnetic resonance imaging (MRI) and clinicopathological features of triple negative breast cancer, and compare them with those of non-triple negative breast cancer. Materials and Methods : This study included 231 pathologically confirmed breast cancers from January 2007 to May We retrospectively reviewed the MRI findings according to the Breast Imaging Reporting and Data System (BI-RADS) lexicon: mass or non-mass type, mass shape, mass margin, non-mass distribution, and enhancement pattern. Histologic type, histologic grade, and the results for epidermal growth factor receptor, p53, and Ki 67 were reviewed. Results : Of 231 patients, 43(18.6%) were triple negative breast cancer. Forty triple negative breast cancers (93.0%) were mass-type lesion on MRI. A round or oval or lobular shape (p=0.006) and rim enhancement (p=0.004) were significantly more in triple negative breast cancer than non- triple negative breast cancer. In contrast, irregular shape (p=0.006) and spiculated margins (p=0.032) were significantly more in non-triple negative breast cancer. Old age (p=0.019), high histologic grade (p<0.0001), EGFR positivity (p<0.0001), p53 overexpression (p=0.038), and Ki 67 expression (<0.0001) were significantly associated with the triple negative breast cancer. Conclusion : MRI finding may be helpful for differentiation between triple negative and non-triple negative breast cancer. Index words : Breast cancer Estrogen receptor Progesterone receptor HER2 Magnetic resonance imaging (MRI) JKSMRM 14:95-102(2010) 1 Department of Radiology, College of Medicine, The Catholic University of Korea 2 Clinical Research Coordinating Center, College of Medicine, The Catholic Medical Center 3 Department of Surgery, College of Medicine, The Catholic University of Korea Received; September 27, 2010, revised; November 17, 2010, accepted; December 1, 2010 Corresponding author : Sung Hun Kim, M.D., Department of Radiology, College of Medicine, Seoul St. Mary s Hospital, The Catholic University of Korea, 505 Banpo-dong, Seocho-gu, Seoul , Korea. Tel Fax rad-ksh@catholic.ac.kr - 95-

13 Jae Jeong Choi et al Introduction Triple negative (TN) breast cancer is a subtype that is negative for the three main receptors for breast cancer, namely estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor type 2 (HER2). TN breast cancer has been considered to have a clinical feature with aggressive behavior and poor prognosis because there is no specific therapeutic target for the tumor. Previous literatures reported that TN breast cancer had a pathologic entity with a high histologic grade, and an overexpression of molecular factors such as epidermal growth factor receptor (EGFR), and Ki 67 (1-4). However, there are a few reports describing MR imaging features of TN breast cancer (5, 6). We hypothesized that comparison of MRI features between TN breast cancer and non-tn breast cancer could yield additional information for pretreatment planning and assessment of prognosis. Therefore, the purpose of our study was to evaluate the MRI findings of TN breast cancer and to compare the findings with those of non- TN breast cancer. We also compared the MRI findings with the clinicopathological results. Materials and Methods Patients This study consisted of 302 consecutive patients who were confirmed to have breast cancer and who underwent breast MRI in our institution from January of 2007 to May of Seventy one patients were excluded from the analysis for the following reasons; 42 patients underwent MRI after the excision of breast cancer, 24 patients received neo-adjuvant chemotherapy, two patients had cancers that were not visible on MRI scans, one patient had recurrent breast cancer after surgery, one patient had breast sarcoma and one patient had no information available on the ER status of the breast cancer. Two hundred-thirty one breast cancers in 231 patients were enrolled in this study. In patients with two or more pathologicallyconfirmed tumors, including three patients with bilateral breast cancers, the largest one was selected. Total 43 TN breast cancers and 188 non-tn breast cancers were included in this study. The age of the patients ranged from 31 to 86 years (mean age, 53.2 years). The size of the breast cancer ranged from 0.3 to 11.0 cm (mean size, 2.76 cm). This study was approved by the institutional review board of our institution. Imaging Acquisition MR images were acquired with a 1.5T scanner (Signa; GE Medical Systems, Milwaukee, WI, U.S.A. and Achieva; Philips Medical system, Best, the Netherlands) using a breast coil. MRI with the Signa scanner was performed using the following sequences; sagittal, fat-suppressed, fast spin-echo T2-weighted imaging, and axial or sagittal, fat-suppressed, fat-spoiled gradient-echo T1-weighted imaging (TR/TE=6.2/3.1, flip angle of 10, 2.6 mm section thickness, and an acquisition time of 1 min 31 minutes) obtained before and 91, 182, 273, 364 and 455 sec after the rapid bolus injection of 0.2 mmol/kg body weight of Gd-DPTA (Magnevist, Schering, Berlin, Germany). MRI with the Achieva scanner was performed using the following sequences; sagittal, fat-suppressed, fast spin-echo T2- weighted imaging and axial fat-suppressed, fat-spoiled gradient-echo T1-weighted imaging obtained before and 91, 182, 273, 364 and 455 sec after the rapid bolus injection of the same contrast agent. Interpretations of MR Image Findings MRI findings for TN breast cancers and non-tn breast cancers were reviewed by the consensus of two radiologists (with 4 and 9 years of breast MR imaging experience). The morphology and enhancement of the lesions were described according to the BI-RADS lexicon. The lesions were divided into mass or nonmass types. Mass type lesions were assessed for size, shape, margins, and enhancement pattern. Non-mass type lesions were assessed for distribution and enhancement pattern. We evaluated the enhancement pattern on 2 min post-contrast MR images. Pathological Analysis We reviewed the size, histologic type, and histologic grade of the TN and non-tn breast cancers. Immunohistochemistry was performed to assess the expression of the following molecular markers; ER, PR, HER2, p53, Ki 67, and epidermal growth factor receptor (EGFR). ER and PR positivity was defined as the presence of 10% or more positively stained nuclei - 96-

14 MRI Findings of Triple Negative Breast Cancer in ten high-power fields. The intensity of HER2 membrane staining was scored as 0, 1+, 2+ or 3+. Tumors with 2+ or 3+ scores were classified as positive for HER2 overexpression, whereas tumors with scores of 0 or 1+ were negative for HER2 overexpression. Among 231 patients, assessment of EGFR was performed in 204 patients, p53 in 132 patients, and Ki 67 in 228 patients. EGFR was considered as positive if membrane staining was observed. Ki 67 expression level of >=15% was considered as expression. Statistical Analysis Continuous variables are shown as means±standard deviation and categorical variables are presented as frequencies and percentages. The differences between the imaging findings for the TN cancer and non-tn cancer were compared using the unpaired t-test, Chisquare test and Fisher s exact test. In addition, logistic regression analysis was performed to assess the contribution of the major risk factors. Statistical significance was established at a p-value<0.05. Statistical analyses were performed using SAS version 9.1 (SAS Institute Inc., Cary, NC, USA). Results Forty three patients (18.6%) had TN breast cancer, Table 1. Characteristics of the 43 Triple Negative Breast Cancer and 188 Non Triple Negative Breast Cancer Groups Characteristics Triple Negative Breast Cancers (n=43) Non-Triple Negative Breast Cancers (n=188) P value Mean Age (year) Tumor Size (cm) Mean < 2 cm 16 (37.2%) 78 (41.5%) >=2 cm 27 (62.8%) 110 (58.5%) Histologic Grade < Grade 1 3 (7.0%) 40 (21.3%) Grade 2 10 (23.2%) 73 (38.8%) Grade 3 26 (60.5%) 38 (20.2%) Not Available 4 (9.3%) 37 (19.7%) Histologic Type Invasive ductal carcinoma 32 (74.4%) 151 (80.3%) Invasive lobular carcinoma 2 (4.7%) 3 (1.6%) Medullary carcinoma 5 (11.6%) 2 (1.1%) Mucinous carcinoma 0 2 (1.1%) Papillary carcinoma 0 2 (1.1%) Metaplastic carcinoma 1 (2.3%) 1 (0.5%) Mixed invasive carcinoma 0 2 (1.1%) Ductal carcinoma in situ 3 (7.0%) 25 (13.2%) EGFR Negative 17 (39.5%) 136 (72.3%) < Positive 20 (46.5%) 31 (16.5%) Not Available 6 (14.0%) 21 (11.2%) p53 Negative 13 (30.2%) 73 (38.8%) Positive 14 (32.6%) 32 (17.0%) Not Available 16 (37.2%) 83 (44.2%) Ki 67 Mean < < 15% 33 (76.7%) 76 (40.4%) < >= 15% 10 (23.3%) 109 (58.0%) Not Available 0 3 (1.6%) - 97-

15 Jae Jeong Choi et al and 188 patients (81.4%) had non-tn breast cancer. The characteristics of TN breast cancer and non-tn breast cancer are summarized in Table 1. The mean age was significantly older for the TN cancer group compared with the non-tn cancer group (p=0.019). There was no significant difference in the tumor size between two groups. TN breast cancers included invasive ductal carcinomas, medullary carcinomas, invasive lobular carcinomas, metaplastic carcinoma and ductal carcinoma in situ. Non-TN breast cancers included invasive ductal carcinoma, invasive lobular carcinoma, medullary carcinoma, mucinous carcinoma, papillary carcinoma, mixed carcinoma, metaplastic carcinoma and ductal carcinoma in situ. There were no significant differences in the histologic types between TN breast cancer and non-tn breast cancer groups. TN breast cancers were more likely to have high histologic grade tumors (p<0.0001). For the TN breast cancers, 26 breast cancers (67%) were grade 3. For the non-tn breast cancers, 38 breast cancers (25%) were grade 3. Compared with non-tn breast cancer, TN breast cancer was associated with EGFR positivity (p<0.0001), p53 overexpression (p=0.038), and Ki 67 expression (p<0.0001). Based on the multivariate analysis, a high histologic grade (odds ratio of grade 3 vs. grade 1, 9.1; 95% CI, ), p53 overexpression (odds ratio, 2.5; 95% CI, ), EGFR positivity (odds ratio, 5.2; 95% CI, ), and Ki 67 expression (odds ratio of < 15% vs. >= 15%, 4.9; 95% CI, ) were associated with the risk of TN breast cancer (Table 2). MRI Findings for TN Breast Cancer and non-tn Breast Cancer TN breast cancers (93.0%) were more likely to show mass-type than non-tn breast cancers (80.8%). However, there was no significant difference in the lesion type between TN breast cancers and non-tn breast cancers (p=0.055). Among mass-type lesions, 33 TN breast cancers (82.5%) showed round or oval or lobular shape. Irregular shape was found in 7 TN breast cancers (17.5%) and 69 non-tn breast cancers (45.4%). Irregular shape was significantly lesser in TN breast cancers than non-tn breast cancers (p=0.006) (Table 3). Smooth margins were more frequently found in TN breast cancers (37.5%) than non-tn breast cancers (23.0%). Spiculated margins were more frequently found in non-tn breast cancers (31.6%) than TN breast cancers (12.5%) (p=0.032). 25 TN breast cancers (62.5%) showed rim enhancement, which was significantly more than TN breast cancers (33.6%) (p=0.004). 59.9% of non-tn breast cancers showed heterogeneous enhancement (Fig. 1 and Fig. 2). There was no significant difference in the findings of nonmass lesions between TN breast cancers and non-tn breast cancers. Based on the multivariate analysis, the oval, round and lobular shapes (hazard ratio 3.92; 95% CI, ) and the rim enhancement (hazard ratio, 3.43; 95% CI, ) were associated with the risk of TN breast Table 2. Odds Ratios (95% CI) for the Triple Negative Breast Cancer Associated with Patient and Tumor Characteristics Triple Negative Cancers Odds Ratio (95% CI) Age (year) >= ( ) < (reference) Lesion Type on MRI Mass 3.16 ( ) Non-mass 1.0 (reference) Mass Shape Oval, Round, Lobular 3.92 ( ) Irregular 1.0 (reference) Mass Margin Smooth 2.01 ( ) Irregular, Spiculated 1.0 (reference) Mass Enhancement Homogeneous 1.4 ( ) Heterogeneous 1.0 (reference) Rim 3.43 ( ) Histologic Grade Grade (reference) Grade ( ) Grade ( ) p53 Negative 1.0 (reference) Positive 2.46 ( ) EGFR Negative 1.0 (reference) Positive 5.16 ( ) Ki 67 < 15% 1.0 (reference) >=15% 4.86 ( ) - 98-

16 MRI Findings of Triple Negative Breast Cancer Fig. 1. MRI finding for triple negative breast cancer in 42 year-old woman. T1-weighted sagittal MR image with contrast enhancement shows a lobular mass with smooth margins and a rim enhancement pattern. Fig. 2. MRI finding for non-triple negative breast cancer in 46 year-old woman. T1-weighted sagittal MR image with contrast enhancement shows an irregular mass with spiculated margins and a heterogeneous enhancement pattern. cancer (Table 2). Discussion Triple negative breast cancer has been reported in 10-25% of all types of breast cancer. It occurs with a higher incidence in pre-menopausal African/American women (1-4). In our study, TN breast cancer accounted for 18.6% of all types of breast cancers. Patients with TN breast cancer were older than those with non-tn breast cancer. This result was similar to the findings obtained for the Japanese series (2). In our study, 93% of the TN breast cancers were mass-type lesions on MRI. Several studies have reported that TN breast cancer is more likely to exhibit mass on MRI and mammography (5-7). However, there was no significant difference in the lesion type between TN breast cancers and non-tn breast cancers in our study. This result may be due to the fact that 28 DCIS were included in our study. Non-mass type enhancement is a common feature of DCIS on MRI (8, 9). We found that the lobular shape, smooth margins, and rim enhancement are associated with TN breast cancers. Because TN breast cancers have an aggressive growing nature, they may reveal bulging shape with pushing borders. Rim enhancement may be due to tumor necrosis. Uemastu T, et al. reported that a very high intratumoral signal intensity on T2-weighted MR images is associated with intratumoral necrosis (5). In contrast, non-tn breast cancers were associated with spiculated margins, due to the desmoplastic reaction of the tumor. Metaplastic carcinoma and medullary carcinoma show a basal-like subtype, and they have higher incidences in TN breast cancer (10-15). Our study included five medullary carcinomas and one metaplastic carcinoma in 43 TN breast cancers. The frequency of the histologic types did not differ between the TN breast cancers and the non-tn breast cancers. This result may be due to the small number of TN breast cancers evaluated. In our study, EGFR, p53 and Ki 67 were overexpressed in TN breast cancer. We assumed that aggressiveness and rapid growing of TN breast cancer - 99-

17 Jae Jeong Choi et al Table 3. MRI findings for the 43 Triple Negative Breast Cancer and 188 Non Triple Negative Breast Cancer Groups MRI Findings Triple Negative Cancers (n=43) Non-Triple Negative Cancers (n=188) Lesion Type Mass 40 (93.0%) 152 (80.9%) Non-mass 03 (7.0%) 036 (19.1%) Mass Shape Round, Oval 16 (40.0%) 041 (27.0%) Lobular 17 (42.5%) 042 (27.6%) Irregular 07 (17.5%) 069 (45.4%) Mass Margins Smooth 15 (37.5%) 035 (23.0%) Irregular 20 (50%) 069 (45.4%) Spiculated 05 (12.5%) 048 (31.6%) Mass Enhancement Homogeneous 02 (5.0%) 010 (6.6%) Heterogeneous 13 (32.5%) 091 (59.9%) Rim 25 (62.5%) 051 (33.5%) Non-mass Distribution Focal 01 (33.3%) 005 (13.9%) Segmental 02 (66.7%) 028 (77.8%) Regional 00 ( 003 (8.3%) Non-mass Enhancement Homogeneous 00 ( 005 (13.9%) Heterogeneous 01 (33.3%) 015 (41.7%) Stippled, Punctuate 00 ( 002 (5.5%) Clumped 02 (66.7%) 014 (38.9%) P value related to overexpression of the markers. EGFR, a type of cell surface receptor, is associated with cell proliferation. EGFR has been considered as a potential therapeutic target in TN breast cancer (16). p53 is a tumor suppression gene that regulates cell proliferation and apoptosis (17). p53 overexpression is an indicator used to predict the response to anthracycline-based chemotherapy in breast cancer. They are associated with a poor prognosis (16-21). Ki 67 is a nuclear antigen that appears during the proliferative phase of the cell cycle. It is related to a high mitotic count and a high level of cell proliferation (22-24). Our study has limitations. First, a small number of patients were enrolled. Further validation in a larger study is warranted. Second, we used only immunohistochemistry to define HER2 status. We classified the HER2 2+ score as positive for HER2 overexpression, without considering the results of fluorescence in situ hybridization (FISH). However, the definition of HER2 status in TN breast cancer remains controversial. Third, inter-observer variability in the assessment of BI-RADS-based MRI findings was not considered in the present study. However, two radiologists reached a consensus in the evaluation of the MRI findings. In conclusion, TN breast cancer occurred in elderly women. On MRI, rim enhancing mass with round or oval or lobular shape is favorable to TN breast cancer rather than non-tn breast cancer. It may be due to aggressive histologic behavior of TN breast cancer. MRI finding may be helpful for planning treatment and prediction prognosis in triple negative and non-triple negative breast cancer patients. References 1.Reis-Filho JS, Tutt AN. Triple negative tumours: a critical review. Histopathology 2008;52: Iwase H, Kurebayashi J, Tsuda H, et al. Clinicopathological analyses of triple negative breast cancer using surveillance data from the Registration Committee of the Japanese Breast Cancer Society. Breast Cancer 17: Dent R, Trudeau M, Pritchard KI, et al. Triple-negative breast

18 MRI Findings of Triple Negative Breast Cancer cancer: clinical features and patterns of recurrence. Clin Cancer Res 2007;13: Trivers KF, Lund MJ, Porter PL, et al. The epidemiology of triple-negative breast cancer, including race. Cancer Causes Control 2009;20: Uematsu T, Kasami M, Yuen S. Triple-negative breast cancer: correlation between MR imaging and pathologic findings. Radiology 2009;250: Chen JH, Agrawal G, Feig B, et al. Triple-negative breast cancer: MRI features in 29 patients. Ann Oncol 2007;18: Yang WT, Dryden M, Broglio K, et al. Mammographic features of triple receptor-negative primary breast cancers in young premenopausal women. Breast Cancer Res Treat 2008;111: Facius M, Renz DM, Neubauer H, et al. Characteristics of ductal carcinoma in situ in magnetic resonance imaging. Clin Imaging 2007;31: Rosen EL, Smith-Foley SA, DeMartini WB, Eby PR, Peacock SLehman CD. BI-RADS MRI enhancement characteristics of ductal carcinoma in situ. Breast J 2007;13: Weigelt B, Kreike B, Reis-Filho JS. Metaplastic breast carcinomas are basal-like breast cancers: a genomic profiling analysis. Breast Cancer Res Treat 2009;117: Reis-Filho JS, Milanezi F, Steele D, et al. Metaplastic breast carcinomas are basal-like tumours. Histopathology 2006;49: Leibl S, Moinfar F. Metaplastic breast carcinomas are negative for Her-2 but frequently express EGFR (Her-1): potential relevance to adjuvant treatment with EGFR tyrosine kinase inhibitors? J Clin Pathol 2005;58: Reis-Filho JS, Milanezi F, Carvalho S, et al. Metaplastic breast carcinomas exhibit EGFR, but not HER2, gene amplification and overexpression: immunohistochemical and chromogenic in situ hybridization analysis. Breast Cancer Res 2005;7: R Kuroda N, Fujishima N, Inoue K, et al. Basal-like carcinoma of the breast: further evidence of the possibility that most metaplastic carcinomas may be actually basal-like carcinomas. Med Mol Morphol 2008;41: Sasaki Y, Tsuda H. Clinicopathological characteristics of triple-negative breast cancers. Breast Cancer 2009;16: Corkery B, Crown J, Clynes M, O Donovan N. Epidermal growth factor receptor as a potential therapeutic target in triple-negative breast cancer. Ann Oncol 2009;20: Rolland P, Spendlove I, Madjd Z, et al. The p53 positive Bcl-2 negative phenotype is an independent marker of prognosis in breast cancer. Int J Cancer 2007;120: Viale G, Rotmensz N, Maisonneuve P, et al. Invasive ductal carcinoma of the breast with the triple-negative phenotype: prognostic implications of EGFR immunoreactivity. Breast Cancer Res Treat 2009;116: Biswas DK, Iglehart JD. Linkage between EGFR family receptors and nuclear factor kappab (NF-kappaB) signaling in breast cancer. J Cell Physiol 2006;209: Nogi H, Kobayashi T, Suzuki M, et al. EGFR as paradoxical predictor of chemosensitivity and outcome among triplenegative breast cancer. Oncol Rep 2009;21: Chae BJ, Bae JS, Lee A, et al. p53 as a specific prognostic factor in triple-negative breast cancer. Jpn J Clin Oncol 2009;39: Yamamoto Y, Ibusuki M, Nakano M, Kawasoe T, Hiki R, Iwase H. Clinical significance of basal-like subtype in triplenegative breast cancer. Breast Cancer 2009;16: Viale G, Regan MM, Mastropasqua MG, et al. Predictive value of tumor Ki-67 expression in two randomized trials of adjuvant chemoendocrine therapy for node-negative breast cancer. J Natl Cancer Inst 2008;100: Ding SL, Sheu LF, Yu JC, et al. Expression of estrogen receptor-alpha and Ki67 in relation to pathological and molecular features in early-onset infiltrating ductal carcinoma. J Biomed Sci 2004;11:

19 Jae Jeong Choi et al 대한자기공명의과학회지 14:95-102(2010) 삼중음성유방암의자기공명영상소견 : 비삼중음성유방암과의비교 1 가톨릭대학교의과대학영상의학과 2 가톨릭대학교의과대학임상연구지원센터 3 가톨릭대학교의과대학외과 최재정 1 김성헌 1 차은숙 1 강봉주 1 이지혜 1 이소연 1 정승희 2 임현우 2 송병주 3 목적 : 삼중음성유방암은국소재발및원격전이가흔하고예후가불량한유방암이다. 이의자기공명영상소견과임상적, 병리학적소견이비삼중음성유방암과차이가있는지알아보고자한다. 대상및방법 : 2007년부터 2008년까지수술로확진된 231명의유방암환자를대상으로하였다. 자기공명영상소견에서대표병변을 Breast Imaging Reporting and Data System (BI-RADS) 에따라후향적으로분석하였고삼중음성유방암의소견이비삼중음성유방암과차이가있는지알아보았다. 또한나이, 조직학적형태, 분화도, 표피성장인자수용체, p53, Ki 67의발현정도가두군간에차이가있는지분석하였다. 결과 : 총 231명중 43명 (18.6%) 이삼중음성유방암이었다. 삼중음성유방암중 40개 (93%) 가자기공명영상에서종괴병변이었다. 삼중음성유방암은비삼중음성유방암에비해원형, 난원형또는소엽성모양 (p=0.006), 변연조영증강 (p=0.004) 소견이많았다. 반면불규칙모양 (p=0.006) 과침상경계 (p=0.032) 는비삼중음성유방암에유의하게많았다. 고령 (p=0.019), 높은조직분화도 (p < ), 표피성장인자수용체양성 (p < ), p53 (p=0.038) 과 Ki 67 (< ) 과발현이삼중음성유방암과관련이있었다. 결론 : 자기공명영상소견은삼중음성유방암과비삼중음성유방암을구분하는데도움이된다. 통신저자 : 김성헌, ( ) 서울시서초구반포동 505, 가톨릭대학교의과대학서울성모병원영상의학과 Tel Fax rad-ksh@catholic.ac.kr

20 정상노인및경도인지장애및알츠하이머성치매환자에서의한국인뇌구조영상표준판개발 김민지 1, 2 장건호 2 이학영 3 김선미 2 류창우 2 신원철 3 이수열 1 목적 : 자기공명영상을이용한치매연구에서삼차원 T1강조뇌구조영상으로뇌기능을분석할경우복셀기반형태분석방법이이용된다. 그러나일반적으로 The Montreal Neurological Institute (MNI) 152 라는젊은서양성인뇌로만들어진표준판에정규화되고이는분석오차가생길수있어한국노인및치매환자의뇌를분석할경우부정확한결과를초래한다. 따라서, 본연구의목적은뇌분석을최적화하기위해한국노인및치매환자의뇌구조영상의표준판을개발하는데있다. 대상및방법 : 검사장비는 3테슬러를이용하였으며 8채널 SENSE (sensitivity encoding) 머리코일을이용하였다. 펄스열은삼차원 T1 강조터보경사자장타입으로뇌의해부학영상을획득하였다. 신경심리검사점수결과에따라정상노인 43명, 경도인지장애환자 44명, 알츠하이머성치매환자 36명의세그룹으로나누었다. 일반선형모델방정식을사용하여나이와성별및질환간의차이에의한인자를고려하였으며, 평가된인자는쌍일치접근방법으로한국노인과치매환자의뇌표준판을만들었다. 두표준판의 Talairach 기준점에따른평균거리와뇌실의거리를계산하였다. 또한뇌전체영역에서회백질과백질을확률이 50% 이상인복셀개수를세어회백질과백질의영역을계산하였다. 결과 : 최종뇌표준판은 MNI152뇌표준판과비교했을경우공간분해능이높았고, 평균거리와뇌실의크기에서차이가있었다. 회백질및백질의영역은본연구에서개발한뇌표준판의회백질과백질모두에서더적었고, 백질보다회백질에서더많은차이가있었다. 결론 : 본연구에서개발한한국뇌표준판은앞으로한국노인과치매환자의질환을분석하는연구에유용할것으로생각된다. 서론구조적자기공명영상 (Structural Magnetic Resonance Imaging, MRI) 과기능적자기공명영상 (functional Magnetic Resonance Imaging, fmri) 을이용한임상적용및인지과학연구가많이진행되고있다. 현재가장많이사용 되는대표적인뇌구조펄스열 (pulse sequence) 로삼차원 T1 강조 (3Dimension-T1 Weighted Imaging) 영상을얻을수있는 MPRAGE (Magnetization Prepared Rapid acquisition Gradient Echo) (1) 혹은 SPGR (Spoiled Gradient-Recalled) (2) 방법이있다. 삼차원 T1강조뇌구조영상을이용한특정영역을분석하기위하여관심영역 (Region of interest, ROI) 방법이사용되나, 이방법은주관적인판단 대한자기공명의과학회지 14: (2010) 1 경희대학교대학원생체의용공학과 2 경희대학교의과대학강동경희대학교병원영상의학과 3 경희대학교의과대학강동경희대학교병원신경과 * 본연구는보건복지가족부보건의료연구개발사업의지원에의하여이루어진것임 (A092125). 접수 : 2010 년 9 월 30 일, 수정 : 2010 년 11 월 29 일, 채택 : 2010 년 11 월 30 일통신저자 : 장건호, ( ) 서울시강동구상일동 149 번지, 경희대학교의과대학부속강동경희대학교병원영상의학과 Tel. (02) Fax. (02) ghjahng@gmail.com

21 김민지외 에의해평가에오차가존재할수있다 (3). 복셀기반형태분석 (Voxel-Based Morphometry, VBM) 을통한삼차원T1강조뇌구조영상방법은회백질감소 (gray matter loss) 또는백질의신호강도를이용한특정부위의부피차이를관찰하는방법으로두뇌전체를분할 (Segmentation) 하여뇌의전체적인변화를객관적으로조사할수있다는장점을가지고있다 (4). VBM 분석을위해서는뇌표준판 (brain template) 에맞춰공간정규화 (Spatial normalization) 하는과정을통하여위치정보를얻는다 (5). 현재삼차원 T1강조영상뿐만아니라뇌확산강조 (diffusion-weighted imaging, DWI) 및확산텐서 (diffusion tensor imaging, DTI) 혹은동맥라벨링 (arterial spin labeling, ASL) 을이용한뇌관류 (cerebral perfusion imaging) 에서도 VBM 방법을이용한뇌변화연구가활발하게진행되고있다. 이때에도삼차원 T1강조영상을이용한뇌표준판을많이이용하게된다. 최근들어고령인구의급격한증가때문에 MRI를이용한치매환자에대한뇌연구가많이진행되고있다. 치매는나이가들어감에따라뇌가후천적인외상이나질병등외부적인요인에의해손상되거나파괴되어점진적인기억력장애와함께한가지이상의인지기능장애를말하며뇌의조직학적변화가나타난다 (6). 노인치매의종류에는여러가지가있는데그중에서알츠하이머성 (Alzheimer s disease) 은치매의 50% 를차지할정도의기억, 사고및행동에장애를초래하는뇌의진행성, 퇴행성질병이다. 노인및치매환자에대한뇌연구를할경우에도 VBM을통한분석이필요하다 (7, 8). 이는해마 (hippocampus) 혹은내후각내피질 (entorhinal cortex) 과같은조직의손상또는특질상취약한뇌영역을관찰하므로정상노인에서경도인지기능장애 (Mild Cognitive Impairment, MCI) 혹은알츠하이머성치매 (Alzheimer s disease, AD) 환자로전환됨을예측할수있다 (9). 그예로, 알츠하이머성치매환자에서 3차원뇌구조 T1 강조영상을이용한회백질감소 (7, 10), 알츠하이머성치매에서 Voxel-based DTI 분석 (11), 알츠하이머성치매에서 ASL를이용한 perfusion 분석 (12) 등이보고된바있다. 현재쥐 (13), 토끼 (14) 등의동물의뇌, 신생아 (15) 혹은젊은정상한국인 (16, 17) 혹은국내어린이 (18) 등을대상으로한표준판이개발되었으며계속해서임상에필요한여러표준판들이개발중에있다. 이들표준판을이용할경우각연구에서보고한것과같이개개의객체를표준좌표화하여군간비교를하거나개개인분석을할경우보다정밀하게분석을할수있는특징이있다. 따라서분석결과의신빙성을높이기위해서는연구하고자하는객체에대한특정표준판을만들어서사용하는것이바람직하다. 일반적으로VBM 분석을위해서많이사용되는뇌표준판은몬트리올신경과학연구소 (Montreal Neurological Institute, MNI) 152 에서개발한뇌표준판을사용하고있다. MNI 표준판을통해뇌영상을맞출때개개인의뇌를 Talairach 좌표로 표준좌표화하여이용하고있다 (19). MNI152 표준판은평균연령이27.8세 (± 5.1) 로젊은정상청년의뇌를이용하여만들어졌으며오른손잡이서양인의뇌를바탕으로하고있다 (20). 한국인노인과치매에대한뇌영상을분석할경우에도서양성인을기준으로한뇌표준판에맞추게된다. 하지만유럽인과일본인의뇌형태를비교했을때인종이나성별에따라서뇌반구의크기와모양이다르다고보고하였다 (21). 또한연령과성별에따른정상한국인뇌실의계측학적연구에서연령이높을수록뇌실의크기가크다고보고되었다 (22). 그러므로 MNI 뇌표준판로한국노인뇌를분석할경우, 인종에의한오차나연령과성별에의한형태적기능적인오차가생길수있어한국인의노화에따른뇌기능연구에부정확한결과를가져올수있다. 이때문에어린환자군혹은나이가많은노인환자군에서얻은영상을분석하기위해서는연구중심의표준화된특정뇌표준판이필요하다. 현재까지어린이뇌를바탕으로하거나 20대와 40대한국인정상남녀를대상으로한뇌표준판은많이보고되었으나, 아직까지한국노인이나치매환자를위한한국인뇌표준판에대한자료가부족한실정이다. 따라서본연구의목적은 50대에서 80대사이의한국정상노인 (Cognitive Normal), 경도인지장애 (Mild Cognitive Impairment) 환자, 알츠하이머성치매 (Alzheimer s disease, AD) 환자군의삼차원T1강조영상을각회백질 (Gray matter), 백질 (White matter) 뇌척수액 (Cerebrospinal Fluid) 의조직맵 (tissue maps) 으로분할한뒤특정뇌표준판을만들고 MNI152 표준판과비교하여한국노인및치매환자의뇌분석을최적화하기위한방법을연구하는데있다. 대상및방법대상본연구는연구를주관한기관의임상실험인증을받은후에실시하였으며, 모든피험자에서참여동의서를획득하고연구를진행하였다. 대상군은표준신경심리검사 (Seoul Neuropsychological Screening Battery, SNSB) 를수행하고 MRI뇌영상을얻었다. SNSB검사는주의집중능력, 언어및그와관련된기능, 시공간기능, 기억력및전두엽, 집행기능의 5가지인지영역을평가함으로써치매환자를진단하는신경심리검사이다. 피험자군의분류는 SNSB 검사를바탕으로이루어졌다. 본연구에참여한총피험자는 148명이었으며이중 25명은피험자의움직임이나 MRI뇌영상을판독한결과뇌에병변이있는경우로본연구에서배제되었다. 최종적으로정상노인 43명 ( 평균나이64.9세 ; 표준편차7.6세 ; 나이범위50-82세 ; 남자15 명, 여자28명 ), MCI환자44명 ( 평균나이67.7세 ; 표준편차7.69 세 ; 나이범위50-83세 ; 남자20명, 여자22명 ),AD환자36명 ( 평균나이72.7세 ; 표준편차9.2세 ; 나이범위53-87세 ; 남자7명, 여자29명 ) 이본연구에포함되었다. 이들을정리한내용이 Table 1에있다