Computerized segmentation of whole-body bone scintigrams and its use in automated diagnostics
Introduction
Whole-body scan or bone scintigraphy is a well known clinical routine investigation and one of the most frequent diagnostic procedures in nuclear medicine [1]. Indications for bone scintigraphy include benign and malignant diseases, infections, degenerative changes [2]. Bone scintigraphy has high sensitivity and the changes of the bone metabolism are seen earlier than changes in bone structure detected on radiographs [1].
The investigator’s role is to evaluate the image, which is of poor resolution due to the physical limitations of gamma camera. There are approximately 158 bones visible on both anterior and posterior whole-body scans [3]. Poor image resolution and the number of bones to inspect make the evaluation of images difficult. Some research on automating the process of counting the bone lesions has been done, but only a few studies attempted to automatically segment individual bones prior to the computerized evaluation of bone scans [4], [5], [6].
First attempts to automate scintigraphy in diagnostics for thyroid structure and function were made in 1973 [7]. Most of the research on automatic localization of bones has been done at the former Institute of medical information science at the University of Hildesheim in Germany from 1994 to 1996. The main contribution was made by Berning [5] and Bernauer [4] who developed semantic representation of the skeleton and evaluation of the images. Benneke [6] has realized their ideas in 1996.
Yin and Chiu [8] tried to find lesions using a fuzzy system. Their preprocessing of scintigrams includes rough segmentation of six parts with fixed ratios of the whole skeleton. Those parts are rigid and not specific enough to localize a specific bone. Their approach for locating abnormalities in bone scintigraphy is limited to point-like lesions with high uptake.
When dealing with lesion detection other authors like Noguchi [3] have been using intensity thresholding and manual lesion counting or manual bone ROI (region of interest) labelling. Those procedures are only sufficient for more obvious pathologies whereas new emerging pathological regions are overlooked.
Section snippets
Aim and our approach
In everyday practice, when a bone is observed, it is diagnosed by the expert physician according to several possible pathologies (lesions, malignom, metastasis, degenerative changes, inflammation, other pathologies, no pathologies). Some pathologies are obvious and could be found even by a less experienced observer, but most are not and sometimes missed even by a specialists. Therefore this process can be supported by using some machine learning classifier [9] which produces independent
Patients and images
Retrospective review of 467 consecutive, non-selected scintigraphic images from 461 different patients who visited University Medical Centre in Ljubljana from October 2003 to June 2004 was performed. Images were not preselected, so the study included standard distribution of patients coming to examination in 9 months. The 19% of the images were diagnosed as normal, which means no pathology was detected on the image. The 57% of the images were diagnosed with slight pathology, 20% with strong
Segmentation
Approximately half of the available images were used for tuning rule parameters to optimize the recognition of the reference points and another half to test it. All 246 patients examined from October 2003 to March 2004 were used as the tuning set and 221 patients examined from April 2004 to June 2004 were used as the test set. In the tuning set there were various non-osseous uptakes in 38.9% of the images, 47.5% images with the visible injection point and 6.8% images of adolescents with the
Discussion
The testing showed encouraging results since the detection of proposed reference points gave excellent results for all bone regions but the extremities, which was expected.
Special attention has been paid to the images with partial skeletons since this is often the case in clinical routine (in our study 18% of the images were partial and no particular problem appeared in detecting) and a robust segmentation algorithm should not fail on such images. As expected, the detection of ribs showed to be
Conclusion
The presented computer-aided system for bone scintigraphy is a step forward in automating routine medical procedures. Some standard image-processing algorithms were tailored and used in combination to achieve the best reference point detection accuracy on scintigraphic images which have very low resolution. Because of poorer image resolution compared to conventional radiography, the presence of artifacts and pathologies necessitate that algorithms use as much background knowledge on anatomy and
Acknowledgement
This work was supported by the Slovenian Ministry of Higher Education, Science and Technology through the research programme P2-0209. Special thanks to nuclear medicine specialist Jure Fettich at the University Medical Centre in Ljubljana for his help and support.
References (20)
- et al.
Analysing and improving the diagnosis of ischaemic heart disease with machine learning
Artif. Intelligence Med.
(1999) - et al.
DeQuant: a flexible multiresolution restoration framework
Signal Process.
(2004) - et al.
Bone Scintigraphy in Clinical Routine
Radiol. Oncol.
(2001) - et al.
When to use bone scintigraphy. It can reveal things other studies cannot
Postgraduate Med.
(1998) - et al.
Percentage of the positive area of bone metastasis is an independent predictor of disease death in advanced prostate cancer
Br. J. Cancer
(2003) - Bernauer J. Zur Semantischen Rekonstruktion Medizinischer Begriffssysteme, Habilitationsschrift, Institut für...
- K.-C. Berning, Zur Automatischen Befundung und Interpretation von Ganzkörper-Skelettszintigrammen, Ph.D. thesis,...
- A. Benneke, Konzeption und Realisierung Eines Semi-Automatischen Befundungssystems in Java und Anbindung an ein...
- et al.
Validation of a rapid computerized method of measuring 99mTc pertechnetate uptake for routine assessment of thyroid structure and function
J. Clin. Endocrinol. Metab.
(1973) - et al.
A computer-aided diagnosis for locating abnormalities in bone scintigraphy by a fuzzy system with a three-step minimization approach
IEEE Trans. Med. Imaging
(2004)
Cited by (24)
Assessment of a software for semi-automatically calculating the bone scan index on bone scintigraphy scans
2021, Clinical ImagingCitation Excerpt :One way to avoid limitations related to subjective methods is to use a computer system to classify scintigraphy data according to user-defined criteria and suggest versions of image interpretation. The development of computer-aided diagnosis (CAD) systems began in the 1960s,18 and CAD systems for bone scintigraphy have been developed.19–21 Some of them have already been integrated into current commercial software packages for 99mTc-MDP bone scintigraphy scans, such as Exini bone (EXINI Diagnostics AB, Lund, Sweden) and BONENAVI® (FUJIFILM Toyama Chemical Co., Ltd., Japan).22–25
Artificial intelligence for bone cancer imaging
2021, Bone Cancer: Bone Sarcomas and Bone Metastases - From Bench to BedsideAutomatic computer-aided sacroiliac joint index analysis for bone scintigraphy
2010, Computer Methods and Programs in BiomedicineSegmentation of kidney from ultrasound B-mode images with texture-based classification
2006, Computer Methods and Programs in BiomedicineAn automatic fine-grained skeleton segmentation method for whole-body bone scintigraphy using atlas-based registration
2022, International Journal of Computer Assisted Radiology and SurgeryBone metastasis segmentation based on Improved U-NET algorithm
2021, Journal of Physics: Conference Series