GB2416101A - Imaging internal structures and correcting for motion artifacts - Google Patents

Abstract

A system for imaging the internal structure of a volume which varies periodically due to, for example, the breathing motion of the patient. The arrangement utilises a CT scanner to produce a series of projected images. The dataset is split into phase-correlated data sets, which are compared to each other and the relative transformation determined. This transformation then applied in reverse to bring the data sets back to a single phase. The whole data set is then used to reconstruct an image of that phase, thus eliminating image artifacts due to motion.

Description

Imaging Internal Structures

FIELD OF THE INVENTION

The present invention relates to the imaging of internal structures of a volume. It is particularly (but not exclusively) relevant to the analysis of CT images to eliminate artefacts and inaccuracies caused by a patient's breathing cycle.

BACKGROUND ART

Existing computed tomography (CT) scanners rely on a radiation source and a detector that rotate around the patient and observe the attenuation of the beam as it passes through the patient from a variety of directions. From this data, a three dimensional representation of the internal structure of the patient is computed.

These fall into two distinct groups. First, the simple CT scanner directs a narrow "fan" beam through the patient. Each complete rotation results in a two dimensional image in the form of a "slice" or section through the patient. The scanner (or the patient) is then indexed along the axis of rotation and a further slice is scanned. These slices can then be assembled into a three dimensional image.

Second, a cone beam CT scanner directs a divergent beam toward the patient to produce (at any selected time) a two dimensional projection of the entire region of interest. As the beam rotates, projections are acquired from different directions and a three-dimensional image can be constructed.

Cone beam CT is in general advantageous as compared to simple CT since its resolution is the same in all directions. Simple CT has a high resolution in the plane of each slice, but the resolution perpendicular to this is limited by the index distance.

Both techniques assume that the patient is static. This is an invalid assumption for a living patient, as some parts such as the heart, lungs and diaphragm will inevitably be moving. As the patient breathes, structures around the lungs and diaphragm will move and this presents difficulties in obtaining good quality scans. In simple CT, artefacts in the image arise; as the slice is indexed along the patient, the breathing cycle can move structures into and out of the slice being scanned at that time. Periodic artefacts thus develop in the reconstructed volume. In cone beam CT, some artefacts result from the reconstruction process, but the main problem is a blurring of the image in the form of an averaging process.

In the treatment of lung tumours (for example) it is important to know the position of the tumour and its movement as the breathing cycle progresses.

The time-averaged information derived from cone beam CT is inadequate, as this cannot distinguish between a large diffuse tumour that dwells in one area and a small dense tumour that briefly passes an area.

Hitherto, when performing CT scans of the thorax, patients have been asked to control their breathing in accordance with an external stimulus, or a proxy for the phase of the breathing cycle has been detected. Examples of proxies that have been used include the local temperature around the nostril, and the dimensions of the thorax. These have proved to be of assistance but generally unsatisfactory.

SUMMARY OF THE INVENTION

The present invention therefore provides an apparatus for imaging the internal structure of a volume exhibiting a periodic time-dependent internal variation, comprising a source of penetrating radiation and a two dimensional detector for that radiation, the source and the detector being arranged to produce a series of projected images of the volume, a reconstruction means arranged to derive information as to the three dimensional structure in the volume from selected images of the series, and a processing means arranged to group images with similar phase from the series for use by the reconstruction means, pass the thus defined groups of images to the reconstruction means for production of a plurality of information sets representing the three dimensional structure of the volume in different phases, for at least one pair of information sets, identify the transformation between a first information set and a second information set, and pass to the reconstruction means data corresponding to at least the first and second information sets, subject to a transformation to a common phase.

The present invention further provides a corresponding method, i.e. one of imaging the internal structure of a volume exhibiting a periodic timedependent internal variation, comprising obtaining a series of projected images of the volume via a source of penetrating radiation and a two dimensional detector for that radiation, grouping images with similar phase from the series for use by the reconstruction means, deriving information as to the three dimensional structure in the volume from each of the thus defined groups of images for production of a plurality of information sets representing the three dimensional structure of the volume in different phases, for at least one pair of information sets, identifying the transformation between a first information set and a second information set, and deriving information as to the three dimensional structure in the volume from data corresponding to at least the first and second information sets, subject to a transformation to a common phase.

Finally, the present invention provides a software module for assisting with the imaging of an internal structure of a volume exhibiting a periodic time -4 - dependent internal variation, the module being adapted to obtain a series of projected images of the volume from a source of penetrating radiation and a two dimensional detector for that radiation, group images with similar phase from the series for use by the reconstruction means, derive information as to the three dimensional structure in the volume from each of the thus defined groups of images for production of a plurality of information sets representing the three dimensional structure of the volume in different phases, for at least one pair of information sets, identify the transformation between a first information set and a second information set, and derive information as to the three dimensional structure in the volume from data corresponding to at least the first and second information sets subject to a transformation to a common phase.

Thus, the present invention permits a more accurate reconstruction to be derived. A simple reconstruction of the entire untransformed dataset benefits from the use of a large volume of data, but suffers in that many of the sampled volumes are different. Likewise, reconstructions using partial or phase- correlated data sets avoid this problem but are less accurate in that there is correspondingly less data from which the reconstruction can be prepared. The invention avoids this choice and allows an accurate reconstruction using the whole dataset, or at least a larger proportion thereof.

The invention is particularly applicable to the imaging of a breathing patient. However, it is in principle applicable to the imaging of a volume exhibiting a periodic time-dependent internal variation. That volume is preferably a living organism, in which case the variation will usually be a natural variation exhibited by the organism, such as breathing.

Assuming that N groups of images are prepared, it will usually be necessary to identify (N-1) transformations, i.e. the transformations between one group and each of the remaining (N-1) such groups. The choice of group is not in principle limited, although there may be advantages in selecting the group so as to minimise the extent of the transformations involved, either in terms of the aggregate or the maximum. In this case, it may be preferable to reconstruct from one untransformed data set and (N-1) transformed data sets, or to reconstruct from N transformed data sets, being one null transformation and the (N-1) other transformations.

The reconstruction means can be unitary and invoked a plurality of times as required. Alternatively, the reconstruction means can contain a plurality of modules each dedicated to the task at hand, such as a module or modules arranged to accept phase-correlated data and a further module or modules arranged to accept transformed data. In this way, the reconstruction process can be tuned to the type of incoming data, if required.

It is of course preferred that the source and detector are rotateable relative to the volume, such that the series of projected images show the volume in different orientations. A cone-beam CT scanner is thus the preferred platform for the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

An embodiment of the present invention will now be described by way of example, with reference to the accompanying figures in which; Figure 1 is a flow diagram of a first embodiment of the present invention; and Figure 2 is a flow diagram of a second embodiment of the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Referring to figure 1, a complete projection image set 10 is prepared using standard cone beam CT techniques. This comprises a number of image sets 10a, Job, 10c etc. which are spaced in time. These image sets are then subjected to phase selection, which identifies the particular phase within the breathing cycle of each image 10a, Job, 10c etc. This can be done as, for example, set out in our earlier co-pending applications GB0327675.5 lodged on 28 November 2003, US 10/760,628 filed 20 January 2004 and International application PCT/GB04/000155 filed on 20 January 2004 all entitled "Imaging Internal Structures", the contents of which are all hereby incorporated by reference.

After the phase selection step, a number of image sets are produced, each being a sub-set of the complete projection image set 10. The sub-sets 12a, 12b, 12c each contain a number of projection images, each image within each set being of the patient at the same breathing phase, but each set showing images of a different breathing phase. In this case, there are three sub-sets but obviously this number can be chosen as required. In general, the greater the number of sub-sets, the fewer images there will be in each sub-set and hence the quality of the image may suffer. Likewise, the greater number of sub-sets there are, the closer in phase the images will be within each sub-set, and this will tend to improve the image quality. A balance needs to be struck.

Each phase-selected image set 12a, 12b, 12c is then reconstructed independently to produce corresponding reconstructed volumes 14a, 14b and 14c. One such reconstructed volume is selected, in this case 14c, and the remaining reconstructed volumes 14a and 14b are compared to that reconstructed volume 14c in order to determine the transformation that has occurred between 14a and 14b and between 14b and 14c respectively.

That transformation is then applied in reverse to the original phase selected image sets 12a, 12b to produce transformed image sets 12a' end 12b' and a complete reconstruction is then carried out based on the transformed image sets 12a' and 12b' and the original phase selected image set 12c. This produces a single reconstructed volume 16, which shows the patient in the phase of image set 12c, but reconstructed using the complete data set.

In practice, it may be simplest to apply the transformation to the dataset (effectively) in three dimensions as part of the conebeam reconstruction.

Referring to figure 2, an alternative scheme is shown. The same image set 10 is again subjected to phase selection to produce sub-sets 12a, 12b and 12c. These are again each reconstructed to produce reconstructed volumes 14a, 14b and 14c. Once again, the transformations as between 14a and 14c and as between 14b and 14c are determined, but in this case the transformations are applied to the reconstructed volumes 14a and 14b to produce transformed volumes 14a' end 14bi these transformed volume data sets are then averaged with the original reconstructed volume 14c to produce an overall reconstructed volume 18. Depending on the nature of the initial data, the first and second embodiments may yield better results.

It will of course be understood that many variations may be made to the above-described embodiment without departing from the scope of the present invention.

Claims (30)

1. Apparatus for imaging the internal structure of a volume exhibiting a periodic time-dependent internal variation, comprising: a source of penetrating radiation and a two dimensional detector for that radiation, the source and the detector being arranged to produce a series of projected images of the volume; a reconstruction means arranged to derive information as to the three dimensional structure in the volume from selected images of the series; a processing means arranged to group images with similar phase from the series for use by the reconstruction means; pass the thus defined groups of images to the reconstruction means for production of a plurality of information sets representing the three dimensional structure of the volume in different phases; for at least one pair of information sets, identify the transformation between a first information set and a second information set; pass to the reconstruction means data corresponding to at least the first and second information sets, subject to a transformation to a common phase.

2. Apparatus according to claim 1 in which the processing means is arranged to create N groups of images, and identify (N-1) transformations being between one information set and each of the remaining (N-1) such sets.

3. Apparatus according to claim 2 in which the processing means is arranged, on passing the data to the reconstruction means on the second occasion, to pass the data with N transformations, consisting of one null transformation and the (N-1) transformations.

4. Apparatus according to any one of the preceding claims in which the reconstruction means is unitary in that it is arranged to accept data a plurality of times.

5. Apparatus according to any one of claims 1 to 3 in which the reconstruction means contains a plurality of modules, at least one being arranged to accept phase-correlated data and at least one other being arranged to accept transformed data.

6. Apparatus according to any one of the preceding claims, in which the source and detector are rotateable relative to the volume, such that the series of projected images show the volume in different orientations.

7. Apparatus according to any one of the preceding claims in which the volume is a living organism.

8. Apparatus according to claim 7 in which the variation is a natural variation exhibited by the organism.

9. Apparatus according to claim 8 in which the variation is caused by breathing.

10. A method of imaging the internal structure of a volume exhibiting a periodic time-dependent internal variation, comprising: obtaining a series of projected images of the volume via a source of penetrating radiation and a two dimensional detector for that radiation; grouping images with similar phase from the series for use by the reconstruction means; deriving information as to the three dimensional structure in the volume from each of the thus defined groups of images for production of a plurality of information sets representing the three dimensional structure of the volume in different phases; for at least one pair of information sets, identifying the transformation between a first information set and a second information set; deriving information as to the three dimensional structure in the volume from data corresponding to at least the first and second information sets, subject to a transformation to a common phase.

11. A method according to claim 10 in which N groups of images are created, and (N-1) transformations are identified, being between one information set and each of the remaining (N-1) such sets.

12. A method according to claim 11 in which the data is passed to the reconstruction means on the second occasion data with N transformations, being one null transformation and the (N-1) transformations.

13. A method according to any one of claims 10 to 12 in which a reconstruction means is provided, to which data is passed a plurality of times.

14. A method according to any one of claims 10 to 12 in which a plurality of the reconstruction means are provided, one to which phase-correlated data is directed and at least one further such means to which transformed data is directed.

15. A method according to any one of claims 10 to 14, in which the source and detector are rotated relative to the volume during image acquisition, such that the series of projected images show the volume in different orientations.

16. A method according to claim 15 in which the source of penetrating radiation and the two dimensional detector for that radiation comprise a cone beam CT scanner.

17. A method according to any one of claims 10 to 16 in which the volume is a living organism.

18. Apparatus according to claim 17 in which the variation is a natural variation exhibited by the organism.

19. Apparatus according to claim 18 in which the variation is caused by breathing.

20. A software module for assisting with the imaging of an internal structure of a volume exhibiting a periodic time-dependent internal variation, the module being adapted to: obtain a series of projected images of the volume from a source of penetrating radiation and a two dimensional detector for that radiation; group images with similar phase from the series for use by the reconstruction means; derive information as to the three dimensional structure in the volume from each of the thus defined groups of images for production of a plurality of information sets representing the three dimensional structure of the volume in different phases; for at least one pair of information sets, identify the transformation between a first information set and a second information set; derive information as to the three dimensional structure in the volume from data corresponding to at least the first and second information sets subject to a transformation to a common phase.

21. A module according to claim 20 in which N groups of images are created, and (N-1) transformations are identified, being between one information set and each of the remaining (N-1) such sets.

22. A module according to claim 21 in which the data is passed to the reconstruction means on the second occasion data with N transformations, being one null transformation and the (N-1) transformations.

23. A module according to any one of claims 20 to 22 in which a reconstruction means is provided, to which data is passed a plurality of times.

24. A module according to any one of claims 20 to 22 in which a plurality of the reconstruction means are provided, one to which phase-correlated data is directed and at least one further such means to which transformed data is directed.

25. A module according to any one of claims 20 to 24, in which the source and detector are rotated relative to the volume during image acquisition, such that the series of projected images show the volume in different orientations.

26. A module according to claim 25 in which the source of penetrating radiation and the two dimensional detector for that radiation comprise a cone beam CT scanner.

27. A module according to any one of claims 20 to 26 in which the volume is a living organism.

28. A module according to claim 27 in which the variation is a natural variation exhibited by the organism.

29. A module according to claim 28 in which the variation is caused by breathing.

30. Apparatus for imaging the internal structure of a volume substantially as described herein with reference to the accompanying figures.