501-1194 1-1 501-1194 1-1 501-1194 1-1 501-1194 1-1 501-1194 1-1 2009 2 9 2009 6 4 Automated measurement method of bone mineral density of vertebral bones and spinal curvature using sagittal plane in X-ray torso CT images Mingxu HAN,Tatsuro HAYASHI, Xiangrong ZHOU,Huayue CHEN,Takeshi HARA, Hiroshi FUJITA,Ryujiro YOKOYAMA,Masayuki KANEMATSU,,andHiroaki HOSHI Department of Intelligent Image Information, Graduate School of Medicine, Gifu University Gifu 501-1194, Japan Department of Anatomy, Graduate School of Medicine, Gifu University Gifu 501-1194, Japan Radiology Services, Gifu Univ. School of Medicine and University Hospital Gifu 501-1194, Japan Department of Radiology, Gifu Univ. School of Medicine and University Hospital Gifu 501-1194, Japan Department of Radiology, Division of Tumor Control, Graduate School of Medicine, Gifu University Gifu 501-1194, Japan Received on February 9, 2009. In final form on June 4, 2009 Abstract : The number of patients with osteoporosis is increasing every year in Japan. The bone-mineral-densitybmd and spinal curvature are two important factors related to the fractures and osteoporosis. Recently, multi-detector-row CT images are widely used in clinical medicine and may have the potential for osteoporosis diagnosis. The development of a computer-aided diagnosiscadsystem that can support the osteoporosis diagnosis is now required. This study proposes a method that can measure the distributions of BMD in each of vertebral trabecular bones and angles of spinal curvatures automatically on torso CT images. This method extracts the regions of the vertebral bones firstly and then generates a sagittal plane of the human spine. Finally, 4 corner points of each vertebral bone region are identified and used for measuring the BMD distributions and spinal curvatures. This method was applied to 20 CT cases and the results were compared with the gold standard that were generated by a medical expert. The errors of the average value of BMD have the mean value of 6.93 mg/cm 3 std.dev. : 6.82. The angles have the mean value of 4.62 degreesstd. dev. : 3.23in thoracic region and 4.40 degreesstd. dev. : 5.08in abdomen. These results showed a good performance of our purposed method and possibility for supporting the CAD of osteoporosis diagnosis. Key words : X-ray CT images, CAD, Osteoporosis, Vertebral body, Bone mineral density X CT MDCT MDCT X CT MDCT bone-minaral-density BMD BMD MDCT BMD BMD [1-5]Kobayashi [6] BMD MDCT MDCT BMD computer-aided diagnosis : CAD [7-13] [13, 14] [13] 52
Campbell-Kyureghyan [15] 4 Campbell-Kyureghyan [15] 4 BMD 4 4 BMD 2 3 CT BMD 4 20 CT 5 6 2002 5 2003 4 MDCT GE LightSpeed Ultra 12 bitfov320380 mm 512512 pixel 120 kv Auto ma 1.25 mm 0.6 mm sinc 0.6250.742 mm 20 15 5 54 Th1-L5 BMD CT 1 BMD 4 BMD [8][9] [10] 3.1.3.3 3.4.3.5 4 BMD CT CT CT [8] Fig.1 Flowchart of our study.the previous works were marked with the shadow. BMD [9] [16] Closing [17] [9] 1 BMD Vol.26 No.32009 53
ath1image before stretching a b bth1image after stretching Fig.2 Recognition results of skeleton.avertebral canal.b Vertebral body. cth6image after stretching dl1image after stretching Fig.4 Stretched images for three vertebral bodies. l :Normalized width of vertebral bodies, t : Angle. The corresponding points before and after image stretching were attached with a same number1 8inaandb. a b Fig.3 Reconstructing a median plane image of spine.aone slice of sagittal section in CT images.bmedian plane image of spine. CT 1 sagittal 3.2.1 CT 1 1 CT 1 Fig.3 CT 1 Th 1 1 Th1-L5 3.2.1. 3.2.2. BMD Step 1 n Step 2 p p n d Step 3 n p l l 1.5 d Step 4 n Step 3 360 degree Step 5 l 30 l d 1.5 Step 4 Step 5 l pixel t degree Th1 Th1Th6L1 1-8 54
t45 degree135 degree 225 degree315 degree 4 3-6 1278 0 t090180270 degree l1 pixel t45135225315 degree l2 pixel t90180270360 degree l1 pixel d1 pixel 0 CT Gaussian Laplacian fscmg f femasktl fsc 1 2 G Gaussian M L Laplacian T Mask f CT fsc Gaussian fe Th1-L5 4 4 4 P1 P2 P1A1 * sin t 180 * π * ω1αc 3 P2A2 * sin 180 t π * ω2αc 4 ω1 360 s*2 5 ω2 180 s*2 360 6 s A1 A2 α C sc α 3 4d2 pixel P1 P2 3.4.3. 3.4.4. 4 4 d2 pixel 1 4 BMD [4] as45, C 0, bs30, C 0, 0 0 cs30, C 15, 0 Fig.5 Range of cortical bone at vertebral bodies. ds30, C 0, 10 es30, C 0, 10 Fig.6 Result image after edge-enhansement filtering. Fig.7 Examples of the pattern images.adetails of pattern,b pattern adjusted by parameter s,cpattern by parameter C,d,epattern by parameter. Vol.26 No.32009 55
a b Fig.8 Extraction results by pattern matching.aresult of thoracic vertebra, andbresult of lumbar vertebrae. BMD 3.3. 3 CT BMD [5] x y3.2382 1.1316 x BMDy CT 7 3.4. 4 Th1-Th12 Th1 2 Th12 2 L1-L5 L1 2 L5 2 2 2 3.4.2.max6 min15d1153.4.3.gaussianσ2.1 3.4.4.A d2 A16A24d25s3045 α1010c 015 1 4800 4 BMD 4 4 20 2.35 mm0.84 1.95 mm0.25 3.30 mm1.00 20 CT BMD 6.93 mgcm 3 6.8220 Th1-L1 BMD 20 CT 4.62 degree3.23 4.40 degree5.08 42.35mm 4 1.35 mm 2 CT 4 BMD BMD [9] BMD 4 BMD6.93mgcm 3 19 4 BMD Fig.9 BMDs for gold standard and measurement resultth1-l5. averaged BMD for gold standard, averaged BMD for result. 56
a b Fig.10 Curvature results of thoracic vertebrae and lumbar vertebrae.acurvature result of thoracic vertebrae andbcurvature result of lumbar vertebrae. : gold standard and, : measuring result, : subtraction between the gold standard and measured result. 5 degree 4 Pruijs [18] Morrissy [19] Cobb Pruijs 3 2.0 degree 7degreeMorrissy 4 2.4 degree 10 degree 20 40 37 92.5 10 degree 4 4 4 CT BMD 20 CT 6.93 mgcm 3 6.82 4.62 degree3.23 4.40 degree 5.08 [ 1 ] D. Weishaupt, M. E. Schweitzer, M. N. DiCuccio, et al : Relationships of cervical, thoracic, and lumbar bone mineral density by quantitative CT, J Comput Assist Tomogr., 251, 146-150, 2001. [ 2 ] N. Yoganandan, F. A. Pintar, B. D. Stemper, et al : Trabecular bone density of male human cervical and lumbar vertebrae, Bone, 392, 336-344, 2006. [ 3 ] W. W. Lu, Y. Zheng, A. Holmes, et al : Bone mineral density variations along the lumbosacral spine, Clin Orthop Relat Res., 378, 255 263, 2000. [ 4 ] X CT 464, 451-457, 2008 [ 5 ] X CT 454, 256- Vol.26 No.32009 57
266, 2007 [ 6 ] T. Kobayashi, N. Takeda, Y. Atsuta, et al : Flattening of sagittal spinal curvature as a predictor of vertebral fracture, Osteoporos Int., 19, 65-69, 2008. [ 7 ] X CT MI2008-14, 75-80, 2008 [ 8 ] X. Zhou, T. Hara, H. Fujita, et al: Automated segmentations of skin, soft-tissue, and skeleton from torso CT images, Proc. SPIE, Medical Imaging 2004, 5370, 1634-1639, 2004. [ 9 ] T. Hayashi, X. Zhou, T. Hara, et al : Automated segmentation of the skeleton in torso X-ray volumetric CT images, In : Proceedings of the 20 th International Congress and Exhibition of Computer Assisted Radiology and Surgery 2006, Springer, 522-523, 2006. [10] X CT MI2008-22, 35-40, 2008 [11] 3 CT MI2006-162, 19-22, 2007 [12] CT 27 JAMIT Annual Meeting 2008, B5-06, 2008 [13] CT MI2007-136, 401-404, 2008 [14] CT MI2007-144, 439-442, 2008 [15] N. Campbell-Kyureghyan, M. Jorgensen, D. Burr, et al : The prediction of lumbar spine geometry : Method development and validation, Clin Biomech., 205, 455-64, 2005. [16] X CT Medical Imaging Technology, 264, 261-271, 2008 [17] 199643 [18] J. E. H. Pruijs, M. A. P.E.Hageman,W.Keessen, et al : Variation in Cobb angle measurements in scoliosis, Skeletal Radiol., 237, 517-20, 1994. [19] R. T. Morrissy, G. S. Goldsmith, E. C. Hall, et al : Measurement of the Cobb angle on radiographs of patients who have scoliosis, Evaluation of intrinsic error. JBone Joint Surg., 723, 320-327, 1990. CT forz 0; zn ; z fory 0; ym ; y gy, zfcenter[z], y,z f CT g center[z] z M N CT 58