WO2012066494A2

WO2012066494A2 - Method and apparatus for compensating intra-fractional motion

Info

Publication number: WO2012066494A2
Application number: PCT/IB2011/055143
Authority: WO
Inventors: Jens-Christoph Georgi; Michael Johannes Eble; Irene Torres Espallardo
Original assignee: Koninklijke Philips Electronics N.V.; Philips Intellectual Property & Standards Gmbh
Priority date: 2010-11-19
Filing date: 2011-11-17
Publication date: 2012-05-24
Also published as: WO2012066494A3

Abstract

The present invention proposes to couple easy to measure, robust external markers to a real internal tumor motion in a breathing model. By such a connection between internal motion and external markers, a reliable prediction of tumor position based on external markers alone is possible. Suitable models have already been developed and could be used and adapted right away. In a learning step the model is trained based on patient specific data containing both, the external marker characteristics and the internal motion, recorded synchronously. The internal motion can be determined by means of the motion estimation part of the local motion compensation algorithm.

Description

Method and apparatus for compensating intra-fractional motion

FIELD OF THE INVENTION

The present invention relates to a method, apparatus, imaging device and computer program product for compensating intra-fractional motion in radiation therapy by means of motion estimation from a data acquisition or imaging device that acquires images of a moving object. Examples of suitable imaging devices comprise a magnetic resonance unit (MR), a computer tomography unit (CT), an ultra-sound unit (US), a positron emission tomography device (PET), a single photon emitting computer tomography (SPECT), or any combination thereof. BACKGROUND OF THE INVENTION

Currently radiotherapy planning is based predominantly on X-ray CT imaging to outline target structures and organs at risk. In addition biological information gained from PET imaging gets used increasingly. The clinical target volume is extended by general safety margins to account for not yet visible tumor spread, for inter-fractional uncertainties in patient positioning, and also for intra-fractional motion of the tumor. However, the position of the target (e.g. tumor) varies during treatment due to motion-inducing actions, such as respiration, for example. The related movements can be divided into two parts: 1) motion- inducing action during irradiation causes intra-fractional movements and 2) change of mean target position between two treatment fractions causes inter-fractional movements. Intra- fractional motion is especially important for radiotherapy of lung tumors, but also tumors in the liver region are affected by breathing motion. In the lung mean amplitudes of motion in the lower lobes have been investigated to be on the magnitude of more than 1 cm in the cranial-caudal direction.

Solutions to account for or to compensate for intra-fractional motion are highly desired. A reliable, robust and precise method would improve therapy outcome and reduce side effects. Especially for the application of ion beam radiation, solutions for this problem are desirable.

To solve the problem of irradiating moving targets technical approaches have been made recently for lung and prostate tumors to control multi-leaf collimators (MLC) dynamically, based on predictions of the tumor location made by a breathing model. Examples are described in Poulsen, P. R. et al.: "Dynamic Multileaf Collimator Tracking of Respiratory Target Motion Based On a Single Kilovoltage Imager During Arc

Radiotherapy", Int J Radiat Oncol Biol Phys, 2010, 77, 600-607, in Poulsen, P. R. et al: Implementation of a new method for dynamic multileaf collimator tracking of prostate motion in arc radiotherapy using a single kV imager, Int J Radiat Oncol Biol Phys, 2010, 76, 914-923, and in Beddar A.S. et al.: Correlation between internal fiducial tumor motion and external marker motion for liver tumors imaged with 4DCT, Int J Radiat Oncol Biol Phys, 2007, 67,630-638, and in Krauss A. et al: Electromagnetic real-time tumor position monitoring and dynamic multileaf collimator tracking using a Siemens 160 MLC: Geometric and dosimetric accuracy of an integrated system, Int J Radiat Oncol Biol Phys, 2010. The breathing model is fed by motion information of fiducial markers derived from e.g.

kilovoltage imaging systems and extrapolates the motion into the next half-second. The aim is to adapt the MLC according to the tumor's predicted motion in real-time and thereby get more dose to the target while sparing surrounding normal tissue.

However, to compensate for the intra-fractional tumor motion, prediction of motion of moving targets in the next -600 ms is desired. Such a prediction should be precise, robust, easy to implement, and not associated with additional dose burden or implantation of fiducial markers. SUMMARY OF THE INVENTION

It is an object of the present invention to provide an intra-fractional motion prediction which is precise, robust, easy to implement, and not associated with additional dose burden or implantation of fiducial markers.

This object is achieved by an apparatus as claimed in claim 1, an imaging device as claimed in claim 7, a method as claimed in claim 12 and a computer program product as claimed in claim 13.

Accordingly, a precise, robust and easy implementable model is provided for predicting internal target motion based on external markers. Easy-to-measure external surrogate markers, like e.g. breathing air flow, motion of markers on the patient's chest, or signals from a chest belt, are coupled to the real internal tumor motion derived from PET imaging that (if at all) might be known only from a gated CT study. The use of PET imaging based internal motion characteristics relative to a gated CT saves the patient significant dose burden and furthermore allows to measure the tumor motion over a longer duration, without collapsing this information into a view gates only. According to a first aspect, the internal motion may be determined by a motion estimation part of a local motion estimation algorithm. This allows obtaining a three- dimensional motion vector for specifying the internal motion.

According to a second aspect which can be combined with the above first aspect, the external indication comprises at least one of a breathing air flow, a motion of markers on said target object, signals from a chest belt, and a thorax surface laser tracking system. These external indications can be easily measured and correlated with the internal motion.

According to a third aspect which can be combined with the above first and second aspects, self-learning may be applied based e.g. on neuronal networking to obtain the motion model. Thus, an object-individual learning process can be implemented.

According to a fourth aspect which can be combined with any one of the above first to third aspects, the motion compensation mechanism may be adapted to control radiation beams of said imaging device or any therapeutic radiation device by adjusting at least one of a multi-leaf collimator, a gantry magnet and a patient table. Thereby, a straight forward and robust mechanism for compensating intra-fractional motion can be provided.

According to a fifth aspect which can be combined with any one of the above first and fourth aspects, the motion compensation mechanism may be adapted to control gating of a follow-up imaging display. Thus, advanced gating of subsequent evaluations is allowed.

According to a sixth aspect which can be combined with any one of the above first to fifth aspects, the motion compensation mechanism may be controlled in real-time. This allows real-time compensation of intra-fractional motion during the imaging process.

The apparatus may be implemented as a chip module, chip set, circuit or software-controlled processor, and may be provided in a network node (i.e. wireless node) or in a station management entity for the wireless mesh network.

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following drawings:

Fig. 1 shows a schematic block diagram of an imaging device with intra- fractional motion compensation according to a first embodiment; Fig. 2 shows schematic diagrams indicating the principle of local motion estimation; and

Fig. 3 shows a schematic flow diagram of an intra- fractional motion

compensation procedure according to a second embodiment.

DETAILED DESCRIPTION OF EMBODIMENTS

In the following embodiments, PET based tumor motion estimation from a local motion compensation algorithm is used as an input to train a motion model for prediction of intra-fractional tumor motion caused by breathing. The derived breathing motion model can subsequently be used for real-time beam adjustment in external photon radiotherapy or hadrontherapy (protons, neutrons or positive ions) or for improved gating in follow up imaging studies.

According to some embodiments, robust external markers are linked to the real internal tumor motion (derived via the local motion estimation from PET imaging) in the breathing motion model. The obtained connection between internal motion and external markers allows reliable prediction of tumor position based on the external markers alone.

Fig. 1 shows a schematic block diagram of an imaging device (ID) 10 (such as a PET unit) with a patient internal motion estimation functionality which may be

implemented as a hardware circuit or as a software routine which controls a processor responsible for motion compensation. The imaging device 10 is adapted to image an internal target (e.g. a tumor or an organ at risk) of a target object (T) 20 which may be a body of a living human or animal. The motion compensation functionality comprises a marker extraction function or marker extractor 30, a motion model generation function or motion model generator (MMG) 40 and a motion compensation function or motion compensator (MC) 50 and may be integrated in the imaging device 10 or provided as a separate unit. The marker extractor 30, motion model generator 40 and motion compensator 50 may be implemented as separate or combined hardware circuits or as software routines controlling a processor or computing device which may be arranged separately or integrated in the imaging device 10. It is noted that a separate therapy unit of a therapy delivery system may be provided, which is not shown in Fig. 1. The therapy delivery system may as well be controlled to some extend by the motion compensator 50.

The marker extractor 30 comprises a detecting function or detector for detecting an external marker or indicator of the target object 10 so as to derive an external motion signal or output. The external marker may be a breathing air flow, a motion of at least one marker attached to the target object (e.g. a marker on the patient's chest), an output from a chest belt, or any other indicator of a motion-inducing action of the target object 10.

In a learning step of the subsequent model generator 40 the model is trained based on patient specific data containing both, the external marker characteristics and the internal motion of the internal target, which may be recorded synchronously. The internal motion can be determined for example by means of a motion estimation part of a local motion compensation algorithm.

Fig. 2 shows schematic diagrams for explaining the principle of local motion estimation. A sequence of images of the lesion or target of interest over time t is

reconstructed with a high temporal resolution. Based on these images, the position of the center of activity is calculated over time t. As indicated in the right-hand diagrams of Fig. 2, waveforms of time-dependent translations Δχ(ΐ), Ay(t) and Δζ(ΐ) are generated from the reconstructed images. The result is a so called "motion vector", indicating the motion of the lesion over time.

The motion model generator 40 correlates the obtained internal motion with the external motion output of the marker extractor 30 to derive a model for predicting the internal motion (i.e. intra-fractional motion) based on the observed external motion. The results of the motion prediction model can then be used by the motion compensator 50 a motion compensation mechanism of the imaging device 10, e.g., to control beams in photon radiotherapy (e.g. by adjusting multi-leaf collimators in real-time) or for particle beams by adjusting the gantry magnets. It could also be used to adjust a patient table or for advanced gating in follow-up imaging studies.

Fig. 3 shows a schematic flow diagram of an intra-fractional motion compensation procedure according to a second embodiment. More specifically, a principal application flow for motion prediction based on external and internal (PET based) motion information is shown.

The procedure comprises an initial model preparation part (MP) with steps S100 to S120 and a subsequent model application part (MA) with steps S200 to S230.

Initially, in step SI 00 and as part of a PET/CT planning study, a dedicated acquisition of the target (without additional dose for the patient) is performed using parameters like the following: bed position, lesion centered in this bed position, 10 min acquisition time, list- mode data storage, etc. Parallel to the imaging study (PET), external surrogate markers are acquired simultaneously in an external marker acquisition (EMA) step and the data from imaging and external markers is recorded synchronously in a synchronous data collection (SDC) to allow later correlation. Then, in a data analysis (DA) step SI 10, the obtained internal and external motion data is analyzed and a local motion estimation algorithm is used to derive a 3D motion vector of the tumor or other internal target from the list mode data. In a patient individual learning step S120 a (breathing) motion model is derived as a base model (BM), e.g. by self-learning based on neuronal networking, which builds a link between derived external motion (observation of external marker motion) and measured internal target motion (e.g. motion vector). For this learning step the synchronously recorded data is used. After individualization the motion model allows to predict motion in the model application part based on input from the external motion markers only.

In step S200 of the application part, external marker data (EMD) is obtained and an individualized motion model (IM) is derived in step S210 based on the available motion base model (derived in the model preparation part) and additional model training (e.g. modification by self-learning etc.). The internal target motion is then estimated or predicted in step S220 to obtain a predicted motion (PM) based on the individualized motion model. Finally, various control or compensation mechanisms (multi-leaf collimators, patient table, magnets) are applied in step S230 to provide a compensation mechanism (CM) for compensating for the estimated intra-fractional target motion. E.g., the beam of the imaging device is directed to where the internal target (e.g. lesion or organ of interest) is expected to be.

Again, the above steps S100 to S120 and S200 to S230 may be implemented as software routines controlling a processor or computing device which may be arranged separately or integrated in the imaging device. The steps may be performed by software code portions of a computer program product for a computer when the product is run on the computer.

The above embodiments could be implemented as (a) dedicated module(s) in a radiotherapy planning software. Furthermore, they could be used to control MLCs in realtime to adjust the beam according to the presumed lesion position. Due to the sharp Bragg peak this application is of special importance in particle radiotherapy, as well.

The embodiments could also be used for improved gated acquisitions in follow-up imaging studies (as already mentioned, a first imaging study would always be needed to train the model).

It is noted that the present invention is not restricted to the above embodiments and implementations. Rather, the proposed estimation and compensation of intra-fractional motion may be used in any PET, SPECT, PET/CT, SPECT/CT, PET/MR, SPECT/MR, MRI systems. The therapeutic unit is not necessarily controlled by an imaging device in combination with the motion compensation mechanism, but the motion compensation algorithm or mechanism, once trained for a specific patient, can itself control the therapeutic device, only based on the external motion markers.

In summary, the present invention proposes to couple easy-to-measure, robust external markers to a real internal tumor motion in a breathing model. By such a connection between internal motion and external markers, a reliable prediction of tumor position based on external markers alone is possible. Suitable models have already been developed and could be used and adapted right away. In a learning step the model is trained based on patient specific data containing both, the external marker characteristics and the internal motion, recorded synchronously. The internal motion can be determined by means of the motion estimation part of the local motion compensation algorithm.

Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor, sensing unit or other unit may fulfil the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

Any reference signs in the claims should not be construed as limiting the scope thereof.

Claims

CLAIMS:

1. An apparatus for compensating motion of a target object (20), said apparatus comprising:

- a detector (30) for detecting motion of said target object based on at least one external indicator of said target object (20);

a correlator (40) for correlating said detected external motion with an internal motion to obtain a motion model of said target object (20), said internal motion being derived from a synchronized sequence of medical images of an internal target portion of said target object (20); and

a motion compensator (50) for predicting internal motion of said internal target portion based on said motion model and said external detected motion, and for generating a motion compensation output in accordance with said predicted motion.

2. The apparatus according to claim 1, wherein said correlator (40) is adapted to determine said internal motion by a motion estimation part of a local motion compensation algorithm.

3. The apparatus according to claim 1, wherein said external indication comprises at least one of a breathing air flow, a motion of markers on said target object, and signals from a chest belt.

4. The apparatus according to claim 1, wherein said correlator (40) is adapted to record data of said internal and external motions synchronously for later correlation.

5. The apparatus according to claim 1, wherein said correlator (40) is adapted to extract a three-dimensional motion vector of said internal target portion by applying a local motion estimation algorithm.

6. The apparatus according to claim 1, wherein said correlator (40) is adapted to apply self-learning based on neuronal networking to obtain said motion model.

7. An imaging device for generating a medical image of an internal target portion of a target object, said imaging device comprising an apparatus according to claim 1, wherein said compensation output is supplied to a motion compensation mechanism of said imaging device or any other device used for localized therapeutic purposes of said target object.

8. The device according to claim 7, wherein said motion compensation mechanism is adapted to control radiation beams of said imaging device by adjusting at least one of a multi-leaf collimator, a gantry magnet and a patient table.

9. The device according to claim 7, wherein said motion compensation mechanism is adapted to control gating of a follow-up imaging display.

10. The device according to claim 7, wherein said apparatus is adapted to control said motion compensation mechanism in real-time.

The device according to claim 7, wherein said device comprises a multileaf

12. A method of compensating motion of a target object (20), said method comprising:

detecting motion of said target object based on at least one external indication of said target object (20);

correlating said detected external motion with an internal motion to obtain a motion model of said target object (20), said internal motion being derived from a synchronized sequence of medical images of an internal target portion of said target object (20);

- predicting internal motion of said internal target portion based on said motion model and said external detected motion; and

generating a motion compensation output in accordance with said predicted motion.

13. A computer program product comprising code means for producing the steps of claim 12 when run on a computing device.