« PreviousContinue »
METHOD FOR USING A BONE
DENSITOMETRY SYSTEM, WITH DUAL-
This invention relates to a procedure for the use of an osteodensitometry system by X-rays.
It will be remembered that osteodensitometry by X-rays is 1Q a technique for measuring bone masses and densities starting from radiographic acquisitions made at several energies.
Two energies are usually used, called "high energy" and "low energy" respectively.
Various configurations of a measurement system (X-ray 15 source and detector) could be used to make the measurements.
A distinction is made between three families of systems depending on the type of radiation used:
pencil beam systems that use an X-ray source collimated 20
through a hole and a X-ray monodetector that is also
fan beam systems that use an X-ray source collimated through a slit and a linear X-ray detector, and 25
cone beam systems that use a two-dimensional X-ray detector.
More particularly, the invention relates to the implementation of osteodensitometry systems by dual energy cone beam X-radiation. 30
STATE OF PRIOR ART
The methodological principles of osteodensitometry by dual energy X-radiation and the main technical solutions used at the moment are defined in the following two docu- 35 ments which should be referred to.
 Technical principles of Dual Energy X-Ray Absorptiometry", G. M. Blake and I. Fogelman, Seminars in Nuclear Medicine, Vol XXVII, No. 3, July 1997, pages 210 to 228 and 40
 "The Evaluation of Osteoporosis: Dual Energy X-Ray Absorptiometry and Ultrasound in Clinical Practice", Second Edition, G. M Blake, H. W Wahner and I. Fogelman, Martin Dunitz Editor, 1999, ISBN 1-85317-472-6.
In particular, refer to chapters 3, 4 and 5 in document  that describes principles for the measurement of bone densities using dual energies and known systems for making these measurements.
Commercially available systems for making diagnoses 50 and therapeutic monitoring of osteoporosis in relevant sites (spinal column, hips, forearm, whole body), are of the pencil beam and fan beam type. These systems use collimators that limit parasite radiation, and particularly scattered radiation. Furthermore, the geometry of these acquisition systems 55 limits the geometric deformations related to the size of the detector, which in particular would result from the conicity of the X-ray beam.
Furthermore, these commercially available systems use a scanning system that uses a high energy measurement and a go low energy measurement in sequence, for each position of the acquisition system. This guarantees perfect consistency between the parts of the patient that are observed in the two measurements.
Commercially available cone beam systems are only 65 designed to make bone density measurements of peripheral parts such as the fingers, toes, hands, forearms and heels. It
is more difficult to implement these cone beam systems than it is to use the other two families of systems.
Scattered radiation generated by interaction between incident X-radiation and the part of the body being studied is important and cannot be ignored. Furthermore, conicity makes the measurement dependent on the position of the part of the body considered in the image field. Furthermore, during the time in which the two-dimensional sensor is being read, the patient may have moved between the high energy measurement and the low energy measurement.
"Cone beam" type systems that are commercially available at the present time are limited to an analysis of the extremities of the human body such as the hands, forearms and heels, because these areas are not very thick and their dimensions are small. Consequently, parasite radiation is limited and positioning may be consolidated, for example by using an attachment system for a forearm or a heel.
On the other hand, if this type of cone beam systems is used for larger parts of the anatomy, and particularly the spinal cord or hips, areas that are used mainly for diagnosis and monitoring of osteoporosis, it would be necessary to take account of parasite phenomena together with the significant conicity of X-radiation and adapt to potential movements of the patient.
Osteodensitometry systems for two-dimensional areas are already known, comprising a source capable of supplying X-ray flux with at least two energies, a converter screen that can transform X-rays into visible light photons, an optical image correction system and a CCD camera, and are described in the following documents which should be referred to:
 U.S. Pat. No. 5,150,394, Sep. 22, 1992, "Dual Energy System for Quantitative Radiographic Imaging", (Andrew Karellas), and
 International application published on Nov. 14, 1996, publication No. WO 96/35372, "A System for Quantitative Radiographic Imaging", (Andrew Karellas).
PRESENTATION OF THE INVENTION
The purpose of this invention is a process for the use of an osteodensitometry system like the system described in documents  and , this process being capable of increasing the precision and reproducibility of bone density measurements in different parts of the anatomy of a patient, starting from two-dimensional radiographics of these parts of the anatomy acquired at more than one energy.
For this purpose, the invention takes account of the patient's movements between two measurements taken at different energies.
This invention does this using matching between an image acquired at high energy and an image acquired at low energy, and also (preferably) a means of helping the patient to position himself.
Specifically, the purpose of this invention is a process for the use of an osteodensitometry system by X-rays with at least two energies, with a cone beam, this system comprising an X-ray source that can provide a cone beam of X-rays at at least one first energy called the high energy, and at a second energy called the low energy and lower than the first energy, a two dimensional X-ray detector and electronic image processor for processing images supplied by this detector, this process being characterised in that the low energy image is acquired from one part of the anatomy of a patient, and to take account of the patient's movements when making bone density measurements, the high energy
image of this part of the anatomy is acquired and these images acquired at low energy and high energy are matched before creating the bone density map for the part of the anatomy.
According to one particular embodiment of the process 5 according to the invention, the following procedure is used to match the images acquired at high and low energies respectively;
sets of contours of bone areas in the part of the anatomy
are extracted from these images, 10 a search is made for an optimum plane transformation in order to match the set of contours for the image acquired at high energy with a set of contours for the image acquired at low energy (or vice versa according to one variant), this transformation is used to bring the image acquired at high energy into the coordinate system of the image acquired at low energy (or vice versa according to the above variant), and 2o these two images are combined to determine the bone
density measurement map. According to one preferred embodiment of the process according to the invention, a radioscopy plate, in other words a plate at a low X-ray dose, is made of the part of the 25 patient's anatomy, to help in positioning this area in the system before making the acquisitions at high and low energies.
Preferably, when the patient is examined for the first time, this low X-ray dose plate is used to retroact on the mechan- 30 ics of the system in order to position the part of the anatomy with respect to the predetermined coordinate system.
Preferably, when the patient is subjected to a check-up, the plate taken at a low X-ray dose will be used to put the part of the anatomy into exactly the same position that it was 35 in during the previous examination.
More explicitly, in the preferred embodiment, a plate with a low X-ray dose and a single energy is made of the part of the patient's anatomy before making acquisitions at low and high energies. The bone contours detected on this plate are 40 used to determine the geometric parameters either to centre the part of the anatomy in the acquisition field for a first examination, or to position the part of the anatomy with respect to the acquisition field in the same way as in the previous examination, for an examination after the first 45 examination. In both cases, this positioning is done by moving the patient with respect to the source-detector system, or the source-detector system with respect to the patient, by manual or automatic control.
According to one particular embodiment of the invention, 50 a plate is made at a low X-ray dose before making the high and low energy acquisitions, in order to match the irradiation dose by varying the X-ray flux, by varying the current applied to the X-ray source used and/or the voltage applied to this source. 55
According to another particular embodiment, a plate with a low X-ray dose is made before making the high and low energy acquisitions, in order to automatically position masks in order to limit the irradiation area.
The possibility of using a plate before the examination 60 (also called a prescan scout data plate) in order to "centre" the patient with is known in the field of tomography from U.S. Pat. No. 5,457,724 "Automatic field of view and patient centring determination from prescan scout data", Oct. 10, 1995. 65
In this patent, two single dimensional projections (fan beam) at 0° and 90° of a tomographic section of a patient are
acquired, before this section is reconstructed. The points corresponding to the edges of the patient are detected in the two projections, and the position of the centre of the area and the size of the tomographic acquisition are detected. These parameters are given to the operator and he can use them to move the patient to optimise centering of the patient for the tomographic acquisition. The purpose is to obtain the best possible image quality, since tomographic systems are designed such that the maximum attenuation occurs at the centre of the acquisition area and spectrum hardening corrections are made as a function of the size of the acquisition field.
This invention makes use of radiology (using a two dimensional detector) rather than tomography (using a fan beam detector in rotation). The final image is a two dimensional projection rather than a reconstructed section. Furthermore, in the preferred embodiment, a single prior plate is used rather than two plates at 90° from each other. Furthermore, one purpose of the invention is reproducibility of the measurement of the bone mass calculated from the image, rather than the quality of the image itself. Furthermore, in tomographic systems, the patient is not recentred automatically (he is recentred automatically in height but not in width, since lateral displacement of the table is not designed or is not necessary).
"Reproducibility" is the property of the measurement instrument to give the same measurement for different examinations on the same patient (assuming a constant bone density) and on the same anatomic site. In the case of a patient whose bone mass varies with time, for example under the influence of a disease or treatment, this reproducibility property can be used to quantify these variations of the bone mass.
Furthermore, in this invention, this "prescan scout data" can be used to automatically position masks to limit the irradiation area to the bone area, which is not possible in tomography otherwise truncated projections occur. In this case, the irradiation dose received by the patient is minimised.
Remember that image matching is used in several image processing domains, for example such as stereoscopic vision, analysis of image sequences, and the fusion of modal images or images with different natures.
Different approaches have been developed, for example matching between layers of grey levels or the extraction of primitives and matching between the primitives.
The following document provides further information about this subject:
 "Overview of Image Matching Techniques", C. Heipke, Proceedings of OEEPE Workshop on the Application of Digital Photogrammetric Workstation, Kolb O. Editor, OEEEPE Official publications, No. 33, pages 173 to 189, 1996.
Concerning positioning of the patient, the first positioning of the patient on these pencil beam or fan beam type systems is made using a laser pointer that identifies the area to be examined starting from external morphological observations. Scanning then starts. If the patient is correctly positioned, the examination continues but if the observation of the first acquired lines on the screen shows that the positioning is not good, the operator stops everything and makes a new positioning and then restarts the examination.
Further information about this subject is given in document , pages 198 to 200 for the spinal cord and pages 265 to 267 for the hip.
A recent study has shown that repositioning with pencil beam type systems was necessary in 50% of cases, and that