WO1998031423A1

WO1998031423A1 - Radiobiologically-based imaging systems for radiotherapy treatment planning

Info

Publication number: WO1998031423A1
Application number: PCT/GB1998/000110
Authority: WO
Inventors: Roger Graham Dale
Original assignee: Imperial College Innovations Limited
Priority date: 1997-01-16
Filing date: 1998-01-14
Publication date: 1998-07-23
Also published as: GB9700816D0

Abstract

A method and apparatus for assisting in the planning and implementation of radiotherapy treatment schedules, which provides for the display, in coded map form, of changes in radiobiological effectiveness of a radiotherapy treatment as a function of spatial position within a patient's body, as a result of changes in the treatment parameters to be used on the body. The degrees of change in clinical effectiveness are shown in a colour coded pattern superimposed upon a grey-scale image of the patient.

Description

RADIOBIOLOGICALLY-BASED IMAGING SYSTEMS FOR RADIOTHERAPY TREATMENT PLANNING

The present invention relates to methods and apparatus for assisting in the planning and implementation of radiotherapy treatment schedules, including teletherapy, brachytherapy or combinations of both.

The planning and implementation of radiotherapy treatments for individual patients is a complex activity which is particularly susceptible of computer assistance. Radiotherapy treatments typically require the delivery of predetermined quantities or doses of radiation to a target site or sites within a patient (eg. tumour sites), maximising the dose delivered to the target site while minimizing the dose delivered to normal tissue or organs surrounding the target site.

This is achieved by using, for example, multiple beams of radiation which are directed toward the patient's body from different angles thereby concentrating the dose where the different beams overlap, ie. at the target site. In order to further limit the toxicity of the treatment to normal tissue, the radiotherapy treatment is delivered in a number of relatively small doses of radiation, termed fractions, over a longer period of perhaps several weeks. Calculation of the dose delivered to the target site, and to the rest of the patient's body is very complex, depending upon many factors such as beam shapes, angles of incidence to the patient, depth and type of tissue or other medium through which the beams may travel affecting attenuation and scatter etc. For brachytherapy schedules, additional factors need to be considered, such as source strengths, source loading patterns, application times, etc. In existing systems, mathematical algorithms have been used to construct the geometrical shapes of the beams which are used to treat the patients, and these are summed to compute isodose distributions within the patient which may be displayed by a computer as a map of isodose contours. A clinician may plan a treatment schedule by modifying the beams, eg. in terms of number, size, intensity and angle of incidence, to determine what effect this has on the isodose map.

Further levels of sophistication have been introduced, in which the isodose maps are directly overlaid onto the output of an imaging system, such as CT or MRI. This enables the generation of an image of a portion of the patient's body, which can be used by the planning system to allow for physical density difference between various tissues identified on the CT or MRI image.

Such existing planning systems, however, only take into consideration the physical distribution of the radiation dose within the patient tissue. As indicated above, it is known that the response to radiation of individual tissues or organs is not solely dependent upon the total physical dose which they receive; it is also critically dependent upon the manner in which the radiation is delivered, such as the number of fractions into which the total dose is divided, the dose per individual fraction, the time gap between the fractions, dose rate during the fraction, etc.

Further developments in radiotherapy treatment planning have attempted to take into account quantifiable radiobiological parameters which determine the response of a particular organ to irradiation, and to use biophysically-based mathematical models to examine how radiation response in specific tissues will alter as the pattern of radiation delivery is changed.

Knowing the physical dose and the pattern of radiation delivery and the tissue-specific radiobiological parameters means that a biological, rather than simply physical, dose may be ascribed to each irradiated organ. The biological dose, hereinafter referred to as the biologically effective dose (BED) is a measure of the amount of biological damage sustained by an irradiated tissue. For tumours, the BED is related to the cure probability and should be made as high as possible, while for critical normal tissues, the BED is related to the complication probability and should be made as low as possible.

It is an object of the present invention to provide a radiotherapy treatment planning system which assists a clinician in determining how to improve (or maintain) the local tumour control or the cure probability of a tumour-affected region, while at the same time maintaining (or reducing) the complication probability of a normal region, or in any other way improving the balance between tumour control and normal tissue complications.

In accordance with one aspect, the present invention provides an apparatus and method for displaying, in a coded map form, changes in clinical effectiveness of a radiotherapy treatment schedule as a function of spatial position within a subject (such as a patient's body or part thereof), resulting from changes in the treatment parameters.

According to a further aspect, the present invention provides an apparatus and method for enabling the optimisation of a radiotherapy treatment schedule by imaging the changes in biological effectiveness of a modified physical dose distribution compared with a reference physical dose distribution, comprising the steps of:

(a) providing image data representing a subject;

(b) identifying component parts of the subject, within the image, having - differing radiotherapy responses and ascribing a specific radiobiological parameter to each component part;

(c) defining a reference radiotherapy treatment schedule;

(d) computing a reference physical dose distribution within the subject; (e) using said specific radiobiological parameters to compute a reference biologically effective dose distribution within the subject;

(f) defining a modified radiotherapy treatment schedule;

(g) computing a modified physical dose distribution within the subject; (h) using said specific radiobiological parameters to compute a consequent change to the biologically effective dose distribution within the subject and imaging said change distribution.

Embodiments of the present invention will now be described by way of example and with reference to the accompanying drawings in which:

Figure 1 shows apparatus according to one aspect of the present invention; and

Figure 2 shows a flow diagram indicating the steps of the method carried out by the apparatus of figure 1.

It has been recognized that there are distinct advantages in representing, in a two- or three-dimensional image, the biological effect of a treatment schedule, rather than the mere distribution of the physical dose. Conversion from a physical dose to a biologically effective dose is a relatively straightforward computation, providing that the requisite computing power and clinical information is accessible from a database. However, the displaying of such information in a manner which is meaningful and useful to the clinician is not a straightforward task.

Because differing tissue types are associated with differing radiobiological parameters, two adjacent tissues which each receive the same physical dose may have quite different biologically effective doses. Displaying this type of effect, either on a computer screen or other output medium proves to be a difficult exercise. Whereas existing physical radiation dose contours (isodose contours) are continuous where they cross adjacent tissues, the corresponding biologically effective dose curves are notably discontinuous across different tissue boundaries and thus very difficult to display in a meaningful manner and difficult to interpret clinically.

A further problem is that the biologically effective dose map thus generated is reliant upon the accuracy and integrity of the stored data sets of the radiobiological parameters. Such data sets necessarily include a proportion of experimental uncertainty and a degree of interpretational judgement is often required in their application.

In the present invention, it has been determined that the system should use the best of current knowledge of radiobiological parameters, but should not prevent clinical value judgement by the radiotherapist. This allows the clinician to place appropriate emphasis on certain results, while overriding other results where his experience dictates that this is expedient and where existing modelling data is known to be less reliable, eg. in relation to volume effects.

It has also been determined that the computation and display of comparative measures, rather than absolute measures of biological effect, is more reliable and clinically safer. This approach allows clinicians to see how new or alternative treatment schedules compare with existing radiotherapy treatment schedules.

With reference to figure 1 , there is shown a computer system suitable for implementing the present invention. The computer system 10 comprises a suitably programmed computer with a processor 12 and associated memory device, together with an output device 14 such as a monitor display or printer, and an input device 16 which preferably includes a keyboard and mouse or similar analogue input device. These components are coupled together in a standard configuration known in the art.

Within the memory is a clinical database 20 which provides a look- up table or data set 22 of radiobiological parameters, a library of various treatment schedules 24,26 as determined by the clinician or as computed by the system, and set-up / control parameters 28. The computer system 10 may be coupled to, or equipped to receive data from, a scanning system 30 such as a CT scanner or MRI system for providing an image of a subject (such as a patient) to be treated. The image may be input to the system via an appropriate i/o interface 32 (or by a standard transferable floppy disk), and stored in the memory 13 as an image file 29. The system may also include an image analyzer 35 for identifying component parts of an image of a subject. Typically, the subject would be a human patient, and the image would be a two- or three-dimensional scan of the patient or part thereof, stored in an image file 29.

A simulation generator 33 may be provided for computer generating images of a subject in place of those provided by a scanning system 30.

With reference to figures 1 and 2, the steps of a preferred method carried out by the system 10 will now be described.

An image of a subject is generated using techniques well known in the art, such as magnetic resonance imaging or CT scanning by the scanner 31 , which provides a two- or three-dimensional picture of the subject (step 100). Preferably, each image comprises a two-dimensional "slice" through the subject, or successive series of slices along an axis orthogonal to the planes of the slices. Each slice is divided up into a plurality of image elements, or pixels, each having an associated numeric value representing, for example, a grey-scale value. The image is saved to an image file 29.

Within the image, various component parts of the subject can be identified, using techniques well known in the art, as known organs or tissue types. Such identification may be carried out manually by the clinician using i/o device 16 or may be carried out automatically by pattern recognition systems such as image analyzer 30, or a combination of both.

Once component parts of the image are identified, the system uses the clinical database 20 to ascribe specific radiobiological parameters to each component part. Preferably this is performed by assigning each image element a tissue type identification code, and ascribing an appropriate radiobiological parameter, or a plurality of radiobiological parameters, thereto, which may also be stored as part of image file 29, or as an updated image file 29' (step 102).

The clinician then inputs into the system a reference radiotherapy treatment schedule (24). The schedule may include a number of parameters relating to the delivery of doses of radiation to the patient. The treatment may be an external, x-ray beam-based delivery regime (teletherapy) or an internal delivery regime using radioactive γ-emitting materials which are delivered through catheters or which are implanted (brachytherapy), or a combination of both. The brachytherapy regime can be low dose rate (LDR), medium dose rate (MDR) or high dose rate (HDR), pulse dose rate or the permanent implant. Other types of radiation delivery are also included in the present invention.

For example, for teletherapy treatments, the treatment schedule parameters may include field sizes, specifications of beam geometry and directions, the number of fractions into which the dose is divided, the dose per fraction, the intervening period between fractions, and, any other parameters which affect the physical and/or biologically effective dose delivered. For brachytherapy treatments, the treatment schedule parameters may include source geometry, source strengths, details of the temporal pattern of treatment delivery, details of dwell times, and any other parameters which affect the physical and/or biologically effective dose delivered.

This reference treatment schedule becomes the baseline against which other treatment schedules (which may include alterations to the reference treatment) will be compared, and is stored in memory as a reference treatment schedule file 40 (step 103).

The physical dose distribution in the subject is then computed using the treatment schedule parameters from file 40. Each image element is ascribed a physical dose value to form a matrix of values describing the physical dose distribution, which is stored as a physical dose map 41 (step

104).

Using the physical dose values so calculated, the system then uses the treatment schedule file 40 and subject image file 29 to convert the physical doses to biologically effective doses for each image element, by reference to an appropriate clinical data set 22 in the memory. In a preferred embodiment, the linear quadratic model is used to define the data set 22 in the clinical database 20. The conversion to BED preferably takes into account, for each image element, a number of factors such as relevant treatment schedule parameters identified above, together with radiobiological parameters such as the fractionation parameter c /β, the recovery constant μ, the potential growth rate of the tumour and normal structures, and any other parameters which are capable of being accommodated in current or future extensions of the radiobiological model. As a further possibility, the system may have an override function which enables the clinician to manually override parameters in the clinical database by entering alternative parameters where clinical judgement so dictates. The conversion to BED takes place, for each image element, for the tissue type within which the image element exists, as identified in the labelled image file 29'.

Any parts of the image which are not specifically identified as any particular organ or tissue type may assume a background type, and be given default "normal" tissue values. A matrix of values (BED_{i j k})_ref defining the resulting three-dimensional biologically effective dose distribution is thus stored in a BED map file 42 (step 105).

The clinician then defines an alternative, modified treatment schedule, for example by editing the reference treatment schedule file 40 to create a new ("modified") treatment schedule which is defined in file 45 (step 106). The modified treatment schedule may include changes to the dose and fractionation pattern, changes in the dose distribution by changing beam weighting factors or source dwell times and the like.

The system then performs the steps of computing the physical dose distribution resulting from the modified treatment schedule in the same manner as that described above for the reference treatment, and saves the result as a modified physical dose map 46 (step 107). The system then performs the steps of converting the physical doses to biologically effective doses for each image element in the same manner as that described above for the reference treatment, and saves the result as a modified biologically effective dose map 47 comprising a matrix of values (BED^,)^ (step 108).

By comparison between the reference BED map 42 and the modified BED map 47, the system derives a ratio matrix 48 of values (BEDi _{j k})_mod / (BED_{i j k})_ref (step 109). The results of this ratio matrix 48 are then displayed to the clinician (step 110) to assist in the forming of a clinical judgement of the resulting improvement or deterioration in the modified treatment schedule over the reference treatment schedule.

In a preferred embodiment, a threshold value of the change represented by each value in the ratio matrix is used to determine whether the change is deemed to radiobiologically significant to any degree. For example, using a threshold value of 1 % , any values of the matrix which lie on or between the values 0.99 and 1.01 indicate that the modified treatment is deemed to be radiobiologically identical to the reference treatment for those image elements. These image elements are coded or displayed in a neutral colour such as grey or colourless, and may be displayed overlying the image 29.

Where the ratio matrix values exceed this threshold value, eg. lie outside the range 0.99 to 1.01 , the image elements are coded or displayed as shades of colour. For example, for matrix values > 1.01 these may be shades of yellow to red varying according to magnitude, and for matrix values < 0.99 these may be shades of green to blue varying according to magnitude. The colour map and neutral colour can both be overlaid on the grey-scale image identifying the various critical parts of the subject, and may also be viewed alongside or overlying the associated physical isodose distribution.

With this particular display format, a particularly good modified treatment schedule would be one in which redness is observed within the defined tumour outline and blueness in the critical normal tissues. The converse pattern would characterize a bad modified treatment schedule.

It is noted that the colours in this scheme do not represent absolute values; they represent expected changes in tumour and/or normal tissue response relative to the reference treatment schedule.

This type of display offers clinicians an immediate appreciation of the likely therapeutic differences between the reference and alternative treatment schedules without having to examine bio-effect curves and without having to attempt to interpret the complex changes in the distribution of such curves when they pass between different types of tissue.

Thus far, the physical dose distribution and generation of a physical dose map 41 has only been discussed in the context of two- or three- dimensional space within the patient. In other words, the system computes a physical dose map for each pixel of three-dimensional space of the patient, and uses these values to generate the three-dimensional biologically effective dose distribution (stored in BED map file 42).

However, in a further preferred aspect of the present system, it is recognized that the biologically effective dose distribution can be affected not only by the three-dimensional spatial distribution of the physical dose, but also its distribution in time. A modified treatment schedule may ultimately have an identical final three-dimensional physical dose distribution from the reference treatment, but a different temporal distribution, which may result in a different biological effectiveness. As an example, a total dose may be delivered as one unit of radiation thirty times in a given period, or two units of radiation fifteen times in the same or a different period.

Computation of the biological effect of this temporal distribution is possible by computing a succession of biologically effective dose distributions which are cumulative, for both the reference and modified treatment schedules, only computing the change in the biologically effective dose distribution after the final BED distributions are computed.

Alternatively, the physical dose distribution (both reference and modified) may include a fourth dimension, namely the temporal distribution as well as the spatial distribution. Thus, the specific radiobiological parameters used to compute a biologically effective dose distribution may take into account the temporal distribution of the physical dose, for generation of both the reference and modified BED maps 42, 47 when generating the ratio matrix 48 for display to the clinician.

Using the above described imaging and planning system and techniques, combined teletherapy and brachytherapy regimes become very much easier to assess and comprehend than when using prior art techniques. Experienced radiotherapists are able to judge from the colour shading of the alternative treatment schedules where improvement or detriment is likely to occur.

Where potentially undesirable hot spots exist within critical structures of the subject, these can be separately highlighted and volumetric details displayed, enabling dose-volume histogram type analysis to be carried out. With additional knowledge of the structure hierarchy of the particular tissue or organ, the clinician may make a better value judgement as to whether the hot spot is significant or not.

A figure of merit may be calculated for all of the information in the ratio matrix, or a predefined sub-component or sub-components thereof.

It will be understood that the procedures described above may be repeated for many different modified treatment schedules, either compared against the same reference treatment (the "yes" exit from step 112) or on an iterative basis. In the latter case, an improved modified treatment is found and the treatment parameters are then saved as the new reference treatment in order to improve the treatment schedule still further. Such iterative improvements may be carried out by way of manual changes to the treatment schedules by the clinician, or by automatic election of new parameters according to a preprogrammed sequence of changes within predetermined ranges of the modifiable parameters.

The system may be coupled to or formed as an integral part of a dose delivery system such as a linear accelerator. Direct control of the delivery system by the treatment planning system offers improvements in safety and reliability, since the incorrect setting of complex treatment parameters by human error is avoided. Once the improved biologically effective dose distribution has been approved by the clinician, opportunity for operator error is substantially reduced. Because the biological effects of the modified treatment are readily apparent from the colour coded display, there is less likelihood of an incorrect or inappropriate treatment schedule being administered.

The system enables the clinician to see prospective biological consequences, overlaid on a specific patient image, of a change to the radiotherapy treatment schedule, thus greatly reducing the risks of harm to the patient by incorrect interpretation of isodose curves in conjunction with radiobiological parameters. The system thereby also offers improved prospects of identifying an optimum treatment thereby reducing the chance of subsequent treatments being required to supplement less effective treatments and/or reducing the likelihood that the patient will develop longer-term complications resulting from excessive injury of normal structures. This offers clear advantages by reducing the burden on health care providers, or by allowing increasing patient throughput on expensive radiotherapy treatment systems. The system may also be used as a teaching tool, demonstrating (among other things) the difficulties associated with achieving true radiobiological equivalence of any two treatments. The effect of changing crucial treatment parameters can be displayed in a clinically relevant way.

The system may be further developed to take into account effects of organ movement and shrinkage etc during the treatment.

The system is completely flexible and allows for inclusion of new methods of radiobiological parameterization, eg. by means of genetic characterization, simply by updating the clinical database 20.

Claims

1. A method of enabling the optimisation of a radiotherapy treatment schedule by imaging the changes in biological effectiveness of a modified physical, dose distribution compared with a reference physical dose distribution, comprising the steps of:

(a) providing image data representing a subject;

(b) identifying component parts of the subject, within the image, having differing radiotherapy responses and ascribing a specific radiobiological parameter to each component part;

(c) defining a reference radiotherapy treatment schedule;

(d) computing a reference physical dose distribution within the subject;

(e) using said specific radiobiological parameters to compute a reference biologically effective dose distribution within the subject;

(f) defining a modified radiotherapy treatment schedule;

(g) computing a modified physical dose distribution within the subject;

(h) using said specific radiobiological parameters to compute a consequent change to the biologically effective dose distribution within the subject and imaging said change distribution.

2. A method according to claim 1 in which step (h) comprises the steps of: using said specific radiobiological parameters to compute a modified biologically effective dose distribution, and computing the difference between said reference biologically effective dose distribution and said modified biologically effective dose distribution.

3. A method according to claim 1 or claim 2 in which said physical dose distributions and said biologically effective dose distributions comprise a plurality of image elements each having at least one quantitative value ascribed thereto.

4. A method according to claim 3 in which said physical dose distributions include temporal distribution information.

5. A method according to any preceding claim further including the step of displaying the consequent changes to the biologically effective dose distribution on a two- or three-dimensional image of the subject.

6. A method according to claim 5 in which the degrees of change to the biologically effective dose distribution are displayed using a colour scale.

7. A method according to any preceding claim in which the specific radiobiological parameters are derived from a look-up table in a clinical database.

8. A method according to any preceding claim in which the specific radiobiological parameters are derived from the linear quadratic model to obtain quantitative cell survival characteristics.

9. A method according to claim 5 in which the image data for the subject is displayed as a grey-scale image, and the degrees of change in the biologically effective dose distribution are displayed as a colour image overlaid thereon where the degrees of change exceed a predetermined value.

10. A method according to claim 9 in which the predetermined value is greater than or equal to 1 % .

11. A method according to claim 9 or claim 10 in which positive changes are shown in a first range of colours and negative changes are shown in a second range of colours.

12. A method according to claim 11 in which the different colours within said first and second ranges of colours represent different degrees of positive and negative changes.

13. A method according to claim 1 in which the image data of the subject represents output from an imaging device such as a CT scanner or MRI imaging system.

14. A method according to claim 1 in which the image data of the subject represents a computer generated simulation of a subject.

15. Apparatus for imaging the changes in biological effectiveness of a modified physical dose distribution compared with a reference physical dose distribution, comprising:

(a) means for providing image data representing a subject;

(b) means for identifying component parts of the subject, within the image, having differing radiotherapy responses and ascribing a specific radiobiological parameter to each component part;

(c) means for inputting a reference radiotherapy treatment schedule;

(d) means for computing a reference physical dose distribution within the subject; (e) means for using said specific radiobiological parameters to compute a reference biologically effective dose distribution within the subject;

(f) means for inputting a modified radiotherapy treatment schedule;

(g) means for computing a modified physical dose distribution within the subject;

(h) means for using said specific radiobiological parameters to compute a consequent change to the biologically effective dose distribution within the subject and imaging said change distribution.

16. Apparatus according to claim 15 further including a dose delivery apparatus and control means for transferring at least some of said modified radiotherapy treatment schedule parameters to said dose delivery apparatus.

17. A method of displaying, in a coded map form, changes in clinical effectiveness of a radiotherapy treatment schedule as a function of spatial position within a subject, resulting from changes in the treatment parameters.

18. Apparatus for displaying, in a coded map form, changes in clinical effectiveness of a radiotherapy treatment schedule as a function of spatial position within a subject, resulting from changes in the treatment parameters.