GB2543731A

GB2543731A - Improvements in dosimetry techniques for radiotherapy

Info

Publication number: GB2543731A
Application number: GB1510288.2A
Authority: GB
Inventors: Sayeed Abdul
Original assignee: Elekta Ltd
Current assignee: Elekta Ltd
Priority date: 2015-06-12
Filing date: 2015-06-12
Publication date: 2017-05-03
Also published as: US20160361569A1; EP3103520A1; EP3103520B1; GB201510288D0; CN106237523B; US10441819B2; CN106237523A

Abstract

The invention provides radiotherapy apparatus comprising a source or therapeutic radiation, a detector and a controller. The controller is adapted to receive dose delivery instructions and patient data including patient geometry, causes the apparatus to perform the delivery instructions and receives the output of the detector, and is adapted to combine this data with the patient data to calculate an first dose distribution; the controller then causes the apparatus to run further delivery instructions, receives a further output from the detector and calculates a second dose distribution which is then compared to the first. The apparatus may issue an alert if the difference between the first and second dose distributions exceeds a predetermined threshold value, or may cause treatment to cease. The first dose distribution may be carried out on a different day to the second. The analysis may involve back-projection of stored radiation fluence. The patient data may be derived from a CT or MRI scan. The detector may be a portal imager or EPID.

Description

Improvements in Dosimetry Techniques for Radiotherapy

FIELD OF THE INVENTION

The present invention seeks to improve upon existing techniques for dosimetry in radiotherapy treatment. It aims to move towards providing a form of real-time in-vivo dosimetry.

BACKGROUND ART

Radiotherapy apparatus delivers a beam of high-energy radiation (typically in the MeV range, apt to damage tumour cells) which is directed towards a tumour site (or other lesion) in a collimated and controlled manner. Thus, the lateral extent of the beam is limited by collimating elements so as to match a pattern determined in a predetermined treatment plan, such as the external profile of the tumour or a subsection of it. The direction of the beam is also varied, so that the tumour is irradiated from multiple directions, thereby reducing the dose delivered to tissue surrounding the tumour site. The treatment is also delivered in "fractions", i.e. individual fractional doses delivered at intervals of (for example) a day, which add up to a total dose to be delivered; delivering the dose in fractions alleviates the side-effects on the healthy tissue surrounding the tumour site.

Typically, a treatment plan will be drawn up prior to delivery of the first fraction, which will detail the beam shapes, directions, and intensity/duration of a number of beam segments that together will form the first fraction. These beam segments are designed to, collectively, deposit a three-dimensional dose distribution in the tumour which corresponds to that prescribed by a clinician, and which both generally minimises the dose delivered to non-tumour areas, and remains within upper dose limits for certain designated sensitive areas of the patient. This is a challenging problem, and the treatment plan is generally arrived at by an iterative process carried out on a computational device.

It is now fairly common for radiotherapy apparatus to have a detector, usually attached to the gantry, located opposite the high-energy radiation source and positioned so as to detect the radiation beam after it has passed through the patient. Such "portal imagers" usually comprise a flat-panel detector (in the form of an electronic portal image detector, or "EPID") which can create an image of the treatment beam as attenuated by the patient; from this image and from a priori knowledge of the beam that was delivered, information as to the distribution of radiation fluences passing through the patient can be arrived at. Once the sequence of instantaneous fluence patterns have been brought together after a fraction is complete, they can be used in conjunction with anatomical information acquired prior to or during treatment (for example a planning CT scan or intrafraction MR image(s)) to estimate the three-dimensional dose pattern or dose equivalent that was delivered during that fraction. A significant computational effort is involved, so the calculation is usually done after the fraction is complete, using all of the observed fluence patterns from the fraction, to allow a post-fraction QA check of the dose delivery.

SUMMARY OF THE INVENTION

This process of creating a post-fraction check of the delivered dose is a useful form of quality assurance. However, if there has been a significant error in the treatment planning and/or delivery processes then this QA check will only bring this to light after the incorrect dose has already been delivered. A preferable form of QA would be a real-time check on the instantaneous dose actually being delivered, offering the option to suspend the delivery if an error was evident.

We have realised that once the first fraction is complete and has been verified in this way, there exists a body of data which is representative of a sequential dose development during delivery of the treatment for that patient and that treatment plan. This is based on individual or groups of EPID images and/or the individual fluences derived from the EPID images and the delivered beam. The individual fluences observed during the first fraction can therefore be used to calculate a developing in-vivo dose during the delivery which can be used as a comparator during subsequent fractions. This will not help during the first fraction, which can be verified by other means such as in the usual manner post-fraction, but it will allow real-time dosimetry verification for the second and subsequent fractions after a successful first fraction.

The present invention therefore provides a radiotherapy apparatus comprising a source of therapeutic radiation for delivering a dose to a patient, a detector for the therapeutic radiation after attenuation by the patient, and a control unit adapted to: i. receive a sequence of dose delivery instructions; ii. receive data describing the patient and characterising at least the patient geometry iii. cause the apparatus to perform the dose delivery instructions and acquire information from the detector as to a delivered radiation fluence; iv. after commencement of the sequence of instructions, perform a first computational analysis based on the acquired information as to the delivered radiation fluence in conjunction with the patient data in order to produce a first dose distribution; v. subsequent to completion of the first computational analysis, cause the apparatus to perform the dose delivery instructions a further time; and vi. during the further performance of the dose delivery instructions, a. acquire the output of the detector; b. perform a second computational analysis based on the acquired output in conjunction with the patient data to produce a subsequent dose distribution, and c. compare the subsequent dose distribution to the first dose distribution.

The control unit can react to the result of the comparison in a number of ways. A simple alert signal if the comparison shows a significant difference between the subsequent dose distribution and the first dose distribution will enable the operators of the apparatus to intervene and take appropriate action. Alternatively, the discovery of a significant difference could prompt the cessation of the treatment fraction. A difference can be considered significant if it exceeds a predetermined threshold, for example. It may also be possible for the apparatus to compute a corrective action or a damage-mitigating action, especially where the comparison reveals an under-dose.

In general, steps (i) to (iii) will take place on a first calendar day and steps (v) and (vi) will take place on a later calendar day. This allows time for the computational analysis of step (iv), which may take place during delivery of the first fraction, or subsequently on either day, or (often) overnight.

The source will often include sensors to measure its radiation output, in which case the delivered radiation can be determined by observing a difference between an output beam as measured by the sensors and an attenuated beam as measured by the detector.

The data describing the patient ideally includes at least data defining an external surface of the patient, and can also include data defining at least one of tissue types and tissue densities within the patient. Where it includes just an external surface of the patient, the computational analysis can comprise a back-projection of the stored radiation fluence information through a body that consists of a standard material (such as water) having a shape corresponding to the external surface of the patient, thereby to determine a dose distribution. Such an analysis is straightforward to compute and can therefore be done during the further performance. Where sufficient computing resources are available, the data describing the patient can be in greater detail and the computational analysis can comprise back-projection of the stored radiation fluence information through the defined tissue type information.

The data describing the patient can be derived from one of a CT scan and an MRI scan of the patient in a generally known manner.

The analysis is preferably carried out on a semi-continuous basis during the subsequent performance. Thus, the control apparatus is preferably arranged to perform the first computational analysis taking into account the acquired information resulting from a subset of the steps within the sequence of dose delivery instructions, and perform the second computational analysis during the further performance of the dose delivery instructions and after a corresponding subset of the steps within the sequence of dose delivery instructions. A first subset may (for example) comprise steps 1 to 'n' of the sequence of instructions. The control apparatus is ideally arranged to perform multiple pairs of computational analyses after different subsets of the steps within the sequence of dose delivery instructions, such as after a second subset which may comprise steps 1 to'm' of the sequence of instructions.

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 shows a schematic form of apparatus to which the present application can be applied;

Figure 2 shows the known comparison process;

Figure 3 shows a sample dose development according to the present invention; and

Figure 4 shows the comparison process according to the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Figure 1 schematically illustrates a radiotherapy apparatus to which the present invention can be applied. A patient 10 including a tumour or other lesion 12 is supported on a patient table 14. This is positionable in a generally known manner which allows the position of the patient 10 to be adjusted in up to 6 axes (three translational, three rotational), thus allowing the tumour 12 to be positioned at a location convenient to the apparatus. A gantry 16 is provided closely adjacent to the patient table 14, and is rotatable around a horizontal axis 18. As illustrated, the patient table 14 is arranged so that the horizontal axis 18 is generally aligned with the patient's cranio-cordal axis, but in practice the table may be rotated so that the patient is at whatever angle is convenient for the treatment that is intended. Rotation of this type can be useful in moving sensitive structures of the patient out of the field of treatment or into an area of the field that is more easily avoided. An arm (not visible) extends from the gantry 16 and carries a source of therapeutic radiation 20 which emits a beam 22 generally towards the horizontal axis 18. Collimators 24 are provided within the source structure, which are able to shape and direct the beam 22 into a chosen cross-section being a subset of its maximum aperture. One or more sensors 25 are also provided within the source, to measure the radiation output.

As shown in figure 1, the source 20 on the gantry 16 is at its uppermost position, thus directing the beam 22 vertically downwardly. As the gantry 16 rotates and carries the source 20 with it, the beam will approach the patient from a range of directions. Assuming that the collimators 24 are used to shape the beam cross-section to the shape of the tumour 12 when viewed along that direction (known as rotational conformal arc therapy) then the effect will be that the tumour 12 remains in the beam at all times whereas surrounding tissue is only briefly irradiated when the gantry 16 is at or passing through one of two particular rotational positions. Thus, this arrangement allows the dose delivered to the tumour to be maintained whilst delivering a significantly lower dose to surrounding tissue.

Other treatment modalities are possible, involving collimator movements that are more complex than the relatively simple rotational conformal arc therapy. In such treatments, the collimation may be adjusted along with the rotation speed of the gantry and the dose rate delivered by the source 20 in order to build up a three-dimensional dose distribution that is optimised around the tumour and the patient's anatomy. Such a dose distribution can ensure a sufficient dose is delivered into the tumour whilst observing maximum permitted doses to other areas, which may vary according to the nature of the tissue at those areas. For example, sensitive areas such as the spinal cord can be protected from radiation doses. Such a dose distribution is delivered by computing a "treatment plan" which consists of the necessary gantry movements, collimator movements, and dose rates which need to take place in order to deliver the distribution. The treatment plan is generally obtained by an iterative computation process which takes into account the dose distribution and the apparatus constraints in a generally known manner to arrive at a treatment plan which is individual to that patient.

The gantry 16 is usually set within a wall 19 so that the relatively bulky mechanisms for supporting the gantry and rotating it accurately can be concealed, allowing a cosmetically acceptable environment for the treatment room.

Also attached to the gantry 16 is an electronic portal imaging device (EPID) 26 in the form of a flat panel detector for the therapeutic radiation, usually using scintillation techniques to derive a digital image of the therapeutic beam. The EPID 26 is positioned on the gantry opposite the source 20 and thus the patient 10 lies between the source 20 and the EPID 26. As a result, the EPID 26 sees the therapeutic beam after attenuation by the patient 10, and thus with knowledge of the nature of the beam that was emitted, and the patient's anatomical information, a measure of the radiation absorbed by the patient can be determined. From the set of such doses obtained throughout the treatment, a three- dimensional representation of the actual dose delivered to the patient, can be computed once the treatment fraction is complete. This computation process takes some time, but is usually available prior to the next treatment fraction. It means that the dose that was actually delivered can be compared to the dose that was planned to be delivered, thus validating the treatment plan and confirming that the treatment fraction was successful. Generally, this is done after each fraction.

Thus, figure 2 shows the comparison process in a conventional radiotherapy apparatus. The treatment plan comprises a series of individual sub-doses 28a, 28b, 28c, 28d, 28e, 28f; a specific treatment plan may include more sub-doses or fewer sub-doses as required. Together, these subdoses 28a-28f cumulatively deliver a total dose 30. During the treatment planning stage, this will have been compared to the dose prescribed by the physician and confirmed as being identical or within a set tolerance that is deemed acceptable. The apparatus then delivers the first treatment fraction, delivering sub-doses 32a, 32b, 32c, 32d, 32e, 32f. During this time, the EPID 26 monitors the radiation exiting the patient and records this. After delivery of the fraction, it combines this information with patient-specific anatomical information to determine the sub-doses delivered during the fraction. These sub-doses are summed to create a cumulative dose 34 and a comparison 36 is carried out with the planned dose 30 or the prescribed dose to confirm that any differences are within a set tolerance that is deemed acceptable.

If there are non-trivial differences between the planned or prescribed dose and the dose actually delivered, then an alert is raised to draw the attention of medical staff who can investigate why this has happened. Otherwise, the second fraction can then be delivered in a similar manner. Thus, sub-doses 38a to 38f will be delivered, and monitored throughout by the EPID 26. At the end of the second fraction, a comparison 42 is made between the total dose 40 delivered during that fraction and the planned dose 30 or the prescribed dose, again to confirm that any differences are within a set tolerance that is deemed acceptable. Similar alerts can be raised if a significant difference is observed, or the treatment can continue to a further fraction (if required) in the same way.

Clearly, if there is an error in the delivery of a fraction then this will not be detected until that fraction is complete. Whilst this is obviously better than nothing, it would be useful to be able to detect errors sooner than this. The present invention therefore uses the first fraction to ascertain the 3D dose building up during the delivery. This can be then used in all subsequent fractions as a basis to compare and determine any variations in the delivery. Once the first fraction is analysed, EPID dosimetry can be used as at present to determine if the fraction was delivered correctly. If this was, then because its build up characteristic has been recorded, this can be used to check subsequent fractions.

To illustrate the concept, imagine a single point dose being calculated per treatment plan step (or with respect to time) as shown in figure 3. Along the xax\s of figure 3 is the progress of the treatment fraction, measured in any convenient scale such as time, or treatment step, or the like. Along the /axis is the dose delivered to the point in question. The diamond-shaped points 44 illustrate the total dose delivered up to that point during that treatment fraction. As can be seen, this increases steadily with each sub-dose 46 until at the end of the treatment the planned total dose 48 has been delivered.

In a subsequent fraction (such as the second, third, fourth etc), the same process can be carried out to produce a further dose/progress profile, this time illustrated with square points 50. In this case, the square points 50 of the subsequent fraction closely follow the diamond points 44 of the first fraction, indicating that the subsequent fraction is developing in the same manner as the first fraction. This comparison can be made with an acceptable tolerance in mind. Thus, a comparison can be made after each sub-dose 46 of the second and subsequent fractions in order to determine that the fraction is progressing normally. This can be repeated for each point within the tumour, or within the patient, or within a region of interest, depending on the computational load involved and the computational power available. If a significant divergence is detected then an alert can be issued and/or the treatment fraction stopped or suspended. This should (in most cases) result in a lower accidental dose being delivered. It is notable that if the error is in the early part of the treatment and in relation to a location that is to receive a significant dose, then it may be possible to detect a developing error before the prescribed dose 48 has been delivered.

Figure 4 shows a version of figure 2, showing the additional comparisons provided by the present invention. Thus, in addition to the comparisons 36 and 42 carried out at the end of the treatment fraction, there are a series of comparisons 52a to 52f between the effects produced by the sub-doses 32x and 38x of the first and the subsequent fraction (respectively).

The dose itself (for each location) is calculated from knowledge of the fluence patterns obtained from the EPID and from knowledge of the patient anatomy. The ideal approach is to start from a known shape and internal structure of the patient (i.e. tissue types, tissure densities etc) and reconstruct the observed ray fluences through the patient in order to calculate the dose delivered at each point within the patient, using known techniques. This approach produces the most accurate representation of the dose pattern that was delivered. Alternatively, a less computationally demanding approach is to use only the patient outline and assume that the patient is composed of a uniform standard material, such as water. Without having to take account of internal variations in the patient tissue density, the dose pattern is much easier to compute, allowing it to be done several times within the time required to deliver a treatment fraction. The dose pattern produced in this way is less accurate, but (we have found) is sufficient for the purpose of comparison with other treatment fractions.

Thus, after each sub-dose or after a group of sub-doses, the cumulative dose delivered so far during the fraction can be calculated. For the first fraction, these can be stored for later use. For the second and subsequent fractions, these can be compared with the corresponding dose patterns computed using the data from the first fraction in order to provide a real-time confirmation that the fraction is proceeding as intended.

Thus, by allowing a simpler reconstruction (such as a water model of the patient) and by operating (at any one time) on only a subset of the sub-doses delivered during a fraction (instead of the entire fraction), the present invention reduces the computational load of this type of checking process to one that can be completed in real time as the treatment fraction progresses.

In some systems, the treatment plan for the second or subsequent fractions is adjusted to take account of variations in the size, shape, position of the tumour (etc). Corresponding variations could be made to the recorded data from the first fraction and the patient anatomical information in order to ensure that the comparison remains valid.

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

1. Radiotherapy apparatus comprising; a source of therapeutic radiation for delivering a dose to a patient, a detector for the therapeutic radiation after attenuation by the patient, a control unit adapted to: i. receive a sequence of dose delivery instructions; ii. receive data describing the patient and characterising at least the patient geometry; iii. cause the apparatus to perform the dose delivery instructions and acquire information from the detector as to a delivered radiation fluence; iv. after commencement of the sequence of instructions, perform a first computational analysis based on the acquired information as to the delivered radiation fluence in conjunction with the patient data in order to produce a first dose distribution; v. subsequent to completion of the first computational analysis, cause the apparatus to perform the dose delivery instructions a further time; and vi. during the further performance of the dose delivery instructions, a) acquire the output of the detector; b) perform a second computational analysis based on the acquired output in conjunction with the patient data to produce a subsequent dose distribution, and c) compare the subsequent dose distribution to the first dose distribution.

2. Radiotherapy apparatus according to claim 1 in which the control unit is adapted to issue an alert signal if the comparison shows a significant difference between the subsequent dose distribution and the first dose distribution.

3. Radiotherapy apparatus according to claim 2 in which a difference is significant if it exceeds a predetermined threshold.

4. Radiotherapy apparatus according to any one of the preceding claims in which steps i. to iii. take place on a first calendar day and steps v. and vi. take place on a later calendar day.

5. Radiotherapy apparatus according to any one of the preceding claims in which the source includes sensors to measure its radiation output, and the delivered dose is determined by a difference between an output beam measured by the sensors and an attenuated beam measured by the detector.

6. Radiotherapy apparatus according to any one of the preceding claims in which the data describing the patient includes data defining an external surface of the patient.

7. Radiotherapy apparatus according to claim 6 in which the computational analysis comprises back-projection of the stored radiation fluence information through a body consisting of a standard material having a shape corresponding to the external surface of the patient thereby to determine a dose distribution.

8. Radiotherapy apparatus according to claim 7 in which the standard material is water.

9. Radiotherapy apparatus according to any one of claims 1 to 6 in which the data describing the patient includes data defining at least one of tissue types and tissue densities within the patient.

10. Radiotherapy apparatus according to claim 9 in which the computational analysis comprises back-projection of the stored radiation fluence information through the defined tissue type information thereby to determine a dose distribution.

11. Radiotherapy apparatus according to any one of the preceding claims in which the data describing the patient is derived from one of a CT scan and an MRI scan of the patient.

12. Radiotherapy apparatus according to any one of the preceding claims in which the control apparatus is arranged to perform the first computational analysis taking into account the acquired information resulting from a subset of the steps within the sequence of dose delivery instructions, and perform the second computational analysis during the further performance of the dose delivery instructions and after a corresponding subset of the steps within the sequence of dose delivery instructions.

13. Radiotherapy apparatus according to claim 12 in which the control apparatus is arranged to perform multiple pairs of computational analyses after different subsets of the steps within the sequence of dose delivery instructions.

14. Radiotherapy apparatus substantially as herein described with reference to and/or as illustrated in the accompanying figures.