Automatic computer-aided sacroiliac joint index analysis for bone scintigraphy

• Published in: Computer methods and programs in biomedicine Vol. 98; no. 1; pp. 15 - 26
• Main Authors: Huang, Jia-Yann, Tsai, Ming-Fong, Kao, Pan-Fu, Chen, Yung-Sheng
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ireland Ltd 01.04.2010; Elsevier
• Subjects: Adolescent; Adult; Aged; Algorithms; Bias; Biological and medical sciences; Bone and Bones - diagnostic imaging; Bone scintigraphy; Bones; Child; Computer programs; Diagnosis, Computer-Assisted; Female; Fuzzy Logic; Fuzzy sets; Humans; Image Interpretation, Computer-Assisted; Image Processing, Computer-Assisted; Internal Medicine; Linear Models; Male; Medical sciences; Middle Aged; Morphological operations; Nuclear medicine; Osteomyelitis - diagnostic imaging; Osteomyelitis - pathology; Other; Patients; Pelvis - diagnostic imaging; Pelvis - pathology; Physicians; Predictive Value of Tests; Radiography; Radionuclide Imaging; Radiotherapy Planning, Computer-Assisted; Radiotherapy. Instrumental treatment. Physiotherapy. Reeducation. Rehabilitation, orthophony, crenotherapy. Diet therapy and various other treatments (general aspects); Regression; Reproducibility of Results; Retrospective Studies; Sacroiliac Joint - diagnostic imaging; Sacroiliac joint-to-sacrum uptake ratios; Sacroiliitis; Surgical implants; Technology. Biomaterials. Equipments. Material. Instrumentation; Young Adult; Sacroiliac joint-to-sacrum uptake ratios; Fuzzy sets; Sacroiliitis; Bone scintigraphy; Morphological operations; Scintigraphy; Biomedical engineering
• ISSN: 0169-2607; 1872-7565
• DOI: 10.1016/j.cmpb.2009.07.010

Abstract Bone scintigraphy helps to detect sacroiliitis before the radiographic changes. Nuclear medicine physicians interpret lesion on sacroiliac joint by the aid of sacroiliac joint-to-sacrum uptake ratios (sacroiliac joint index, SII). Usually, the SII is measured by manually drawing regions of interest (ROI) over the sacroiliac joints and sacrum, which is a tedious, time consuming, and highly operator dependent procedure. In this approach, we developed an automatic SII measurement program based on fuzzy sets histogram thresholding, anatomy-based image segmentation method and pelvis reference points located by morphological operations. To validate the program, the results of automatic SII measurement from 33 patients were compared with the results of manual SII measurement. For each patient, six SIIs representing the upper, middle, and low portions of the right and left SIIs were obtained. Totally, 198 SIIs were included for linear regression, bias, and precision analyses. The coefficient of determination ( R 2 ) between our approach and 2 medical experts’ manual measurements are 0.923 and 0.917 respectively. The predictive performance (bias, precision) shows a good agreement result. From the result of this study, we expect that this automatic computer-aided approach may be applied to help nuclear medicine physicians in sacroiliitis interpretation.