Cancer Radiotherapy | Radiotherapy Treatment | Cancer Radiotherapy
 |
|
 |
 |
 |
|
Jump to
Intensity Modulated Radiation Therapy
an Introduction.
Extracts from C Nutting, DP Dearnaley, S Webb. Intensity Modulated Radiotherapy: A clinical review. Br J Radiology 2000;54:459-469 Intensity Modulated Radiotherapy (IMRT) represents a significant advance in conformal radiotherapy. In particular, this type of cancer radiotherapy treatment allows the delivery of dose distributions with concave isodose profiles such that radiosensitive normal tissues close to, or even within a concavity of a tumour may be spared from radiation injury. This article reviews the clinical application of this cancer radiotherapy treatment technique in head and neck cancer, and discusses the practical issues of treatment planning and delivery from the clinicians perspective. Introduction
The aims of radical radiotherapy are to deliver a homogeneous radiation dose to a tumour target while minimising the dose to surrounding normal tissues [1]. In this way the maximum number of clonogenic tumour cells can be eradicated with the minimum risk of normal tissue injury [2,3]. Conventional external beam radiotherapy using a small number of rectangular or simply shaped beams partly achieves these goals, but for many tumour sites this radical radiotherapy treatment leads to irradiation of unnecessarily large volumes of normal tissue [4]. Conformal radiotherapy aims to minimise the volume of normal tissue irradiated by shaping the dose distribution to tightly conform to the shape of the tumour reducing the dose to surrounding normal tissues [5]. Adequate immobilisation of the target and improved three-dimensional imaging, enable a higher degree of certainty of target localisation which permits the use of narrower margins around the target [6]. Three-dimensional conformal radiotherapy (3DCRT) entails direction of multiple beams conformed to the shape of the target from each beam’s eye view (BEV). In this type of radiotherapy treatment, three-dimensional dose distributions are calculated by a treatment-planning computer with dosimetric algorithms. Radiotherapy planning studies have confirmed that 3DCRT reduces the volume of normal-tissue within the high-dose volume compared to conventional [7-9]and is very helpful in cancer treatment. Randomised clinical trials have demonstrated a clinically significant reduction of late radiation side effects in patients with prostate cancer treated with 3DCRT compared to conventional radiotherapy [10], and at other tumour sites in non-randomised comparisons [11-15]. Dose escalation within acceptable rates of normal tissue complications is a further challenge for 3DCRT [1]. Non-randomised clinical studies in prostate cancer have suggested that dose escalation with conformal therapy improves tumour control [16,17], and several randomised clinical studies are underway. Intensity-Modulated Radiotherapy (IMRT) permits the delivery of dose distributions with concave isodose shapes. It has been suggested that 30% of tumours exhibit a concavity in the Planning Target Volume (PTV) where sparing of critical normal tissue irradiation may be achieved with the use of IMRT. What is beam intensity modulation?
Beam intensity modulation (BIM) indicates that the radiation fluence varies across the beam. This is different from a unmodulated (standard) radiation beam where the fluence is constant, producing a flat isodose profile where each isodose line at the centre of the treatment beam is perpendicular to the central axis of the beam. One of the simplest forms of BIM is produced by a wedge filter to variably attenuate the radiation beam. The radiation fluence increases across the radiation beam profile measured at a fixed distance from the source. The use of a tissue compensator, or partial transmission block produces more complex fluence profiles and are examples of BIM currently in use in most radiotherapy departments (Fig 1). Clinical applications of
IMRT
Intensity modulated radiotherapy (IMRT) is a form of conformal therapy which combines several intensity-modulated beams. In this radiotherapy treatment, the resultant isodoses are highly conformal, and uniquely can yield a concave distribution (Fig 2). IMRT therefore offers a significant advance in conformal therapy [18], by improving conformality and reducing radiation dose to radiosensitive normal tissues close to the tumour even if they lie within a concavity in the PTV [19]. In radiotherapy treatment, there are many clinical situations where radiosensitive normal tissues lie within a concavity surrounded by the PTV. The treatment of patients with tumours of the larynx, pharynx, or thyroid is a good example. The clinical target volume (CTV) often includes a midline target, and bilateral cervical lymph nodes, producing a horseshoe-shaped PTV with the spinal cord within the concavity [20]. Homogeneous irradiation of these PTV’s to radical doses (50-66 Gy) with conventional external-beam radiotherapy is difficult. Typically parallel-opposed photon portals are matched to electron beams. This technique leads to dose inhomogeneity at the photon-electron matchline, and also under-doses the posterior cervical lymph nodes close to the spinal cord [21]. This shape of the PTV can be treated homogeneously using IMRT without the need for electrons (Fig 3). The dose to the spinal cord can be kept well within tolerance [20] and permits tumour dose escalation. Significant normal tissue sparing using IMRT has also been demonstrated in planning studies for tumours of the maxillary antrum, nasopharynx, lung, and prostate [22-25]. Complex dose distributions can be delivered which avoid a number of radiosensitive normal tissues close to a tumour. For example, in the treatment of nasopharyngeal cancer large parallel-opposed lateral portals are used to encompass macroscopic disease and sites of occult metastases. With this technique, both parotid glands, spinal cord and brain stem are inevitably included in the irradiated volume although these structures do not need to be included in the target volume. Complete xerostomia, and risk of myelopathy are the result. By defining concavities in the PTV, IMRT can produce a dose distribution which reduces the radiation dose to these organs (Fig 4). and promises a significant reduction in treatment morbidity. Intensity modulated radiotherapy could be used for the whole duration of a radiotherapy treatment, or simply as a boost after more conventional treatment. The appropriateness of these two treatments approaches is likely to depend on the tolerance doses of surrounding radiosensitive normal tissues. Results of IMRT from clinical
studies
At present around 5000 cancer patients have been given the Intensity modulated radiotherapy treatment worldwide. Most of these patients have been treated at a few clinical centres in the United States, and although IMRT is the focus of several research groups in the United Kingdom, only a handful of patients have been treated in radiotherapy departments to date. Head and neck cancer
At the University of Michigan, Intensity modulated radiotherapy was used to spare salivary gland tissue in patients irradiated for head and neck tumours. PTV included the primary tumour, ipsilateral cervical nodes, and contralateral cervical nodes up to and including the sub-digastric node. The contralateral parapharyngeal space and parotid gland, which were judged to be at very low risk of harbouring occult metastases, were spared, as were the submandibular salivary glands [26,27]. Patients were treated with a forward planned “step and shoot” IMRT technique (see below) using multiple non-coplanar photon beams, and low-weighted electron fields. A BEV facility was used to select beam orientations which avoided the parotid gland [28]. Unstimulated and stimulated salivary flow was measured from each parotid gland before and after radiotherapy and then at 3, 6, and 12 months. In 15 patients treated with this parotid-sparing technique, IMRT improved the minimum dose, and reduced dose inhomogeneity to the primary tumour, and lymph node regions compared to standard three-field conformal plans. IMRT reduced the radiation dose to the contralateral parotid gland to 32% compared to 93% for the standard plan. Smaller, statistically significant, reductions in the dose to the oral cavity, contralateral submandibular gland, and spinal cord were also seen but are unlikely to be clinically significant. One to three months after irradiation, the mean stimulated salivary flow from the contralateral parotid gland was 60% (SD 49%) of pre-treatment measurements [27]. Longer follow up of 11 of these patients showed that spared parotid glands, which received a mean dose of 19.9 Gy, recovered 63 % of their pre-treatment stimulated salivary flow rates at one year compared to only a 3% recovery for treated parotid glands which received 57.5 Gy [29,30] Mean stimulated salivary flow (± sd) after parotid-sparing IMRT
An analysis of 88 patients who were given treatment with parotid-sparing IMRT allowed correlation of radiotherapy dose with salivary flow measurements to produce dose-response curves for parotid gland function. A mean dose threshold was found for both stimulated (26 Gy), and unstimulated (24 Gy) saliva flow rates, such that glands receiving mean dose below or equal to the threshold showed substantial preservation of the saliva flow following radiotherapy, which may continue to improve over time. By contrast, most glands receiving mean doses above the threshold produced little saliva and had no recovery over time (A Eisbruch, personal communication). De Neve et al [31] from the University of Gent recently reported the results of treatment of 3 patients with recurrent or second primary tumours of the nasopharynx, oropharynx, and hypopharynx with IMRT. All patients had been previously treated with radical radiotherapy; tumour dose 66-70 Gy, spinal cord dose 44-45 Gy, and had inoperable disease. An Intensity modulated radiotherapy technique was used to re-treat the tumour (min target dose 48-65 Gy), with a concave dose distribution to avoid the brain stem and spinal cord (max spinal cord dose 21-34 Gy, max brain stem dose 67 Gy). Two patients achieved complete remission, but relapsed within one year of radiotherapy, and the other patient remained in partial remission 7 months after treatment. No patient developed myelopathy although the follow-up period was short. The same author has reported an IMRT technique for the irradiation of tumours in the neck which extend into the upper mediastinum. This technique has been used in the treatment of tumours of the thyroid, larynx and pharynx and would allow target dose escalation up to 70-80 Gy while restricting maximum spinal cord dose to 50 Gy [20]. Boyer et al [32] report the results of a planning and delivery study where three patients with head and neck tumours were planned on the Peacock inverse planning system (NOMOS Corporation, Sewickley, PA), and the plan was delivered to a humanoid phantom using nine equispaced fields by a dynamic MLC technique. For a patient with nasopharyngeal carcinoma, 96% of the primary tumour PTV reached the goal dose of 72 Gy, although part of the GTV received 90 Gy. A goal dose of 54 Gy was prescribed to the lymph node chains but 12% of this PTV was under-dosed, with a minimum dose of 26.5 Gy. Parotid and spinal cord sparing were achieved with 99% and 98% receiving <45 Gy respectively. Similarly for cancer of the larynx and ethmoid sinus, mean target doses were achieved and normal tissue structure sparing was successful, although target dose imhomogeneity was high. Goitein and Niemierko [33] have calculated that such dose inhomogeneities may lead to large reductions in the probability of tumour control. However comparison with standard radiotherapy techniques the precise location of lower dose regions and effects for example of patient movement as well as the accuracy of the planning algorithm need to be considered in determining the desired tolerance of such dose inhomogeneities. The first clinical report of 28 patients with a spectrum of head and neck tumours treated with the MIMiC tomotherapy apparatus (NOMOS Corporation, Sewickley, PA) has recently been published [34]. Ten patients were treated for tumour recurrence after previous conventional radiotherapy, and in 18 patients Intensity modulated radiotherapy was part of the primary treatment. Patients were initially immobilised using an invasive fixation device (Talon, NOMOS Corporation, Sewickley, PA) which attached to screws placed in the inner table of the skull vertex, although currently half of their patients are immobilised in a standard thermoplastic mask [59]. After CT scanning, inverse treatment planning was performed with the objective of minimising the dose to parotid glands, brain, orbits, optic nerves, and brain stem, depending on tumour site. Treatment was well tolerated with acute toxicity equivalent to conventional radical radiotherapy treatment. A high degree of parotid sparing was demonstrated in suitable patients, with less than 20% of the total parotid volume receiving greater than 20 Gy. Clinical follow up on these patients is short, and although only one of 20 patients treated definitively has recurred locally, long-term results are not yet available. Planning studies and class solutions
Planning studies offer the opportunity to assess the potential advantages of IMRT over conventional radiotherapy. By comparing the conventional treatment plan to the IMRT plan the improvements in the dose distribution can be measured and estimates can be made of the clinical benefit. For example Smitt et al [44] compared 3-D conformal planning and IMRT in patients treated for breast cancer. The IMRT plans produced more homogeneous dose distributions, and reduced the volume of lung and heart treated to high dose. Planning studies can be undertaken for each cancer site to determine those which merit further clinical study, and also suggest which end points are applicable for comparative clinical trials. For each tumour site the rival techniques can be compared and the best technique used as the basis for a “class solution” (i.e. a technique which will produce the best results when applied to a group of patients with a particular tumour type). An example of an IMRT class solution is that of De Neve et al [20] for bilateral neck node irradiation without the need to match photon and electron fields. Production of intensity-modulated beams Intensity-modulated beams (IMB) for radiotherapy can be produced in a number of ways: 1. Metal compensators A specifically manufactured metallic compensator is milled or moulded so that a variable thickness of the absorber is presented before the radiation beam. Production of compensators is simple but expensive and time consuming. They are heavy and may be difficult to position accurately in the linear accelerator head. In practice this limits the number of IMB’s that can be delivered [18]. 2. Multiple segments per field Each treatment field is divided into several smaller segments or sub-fields which are delivered sequentially (the “step and shoot” method). Each segment shape is defined by a multi-leaf collimator or shaped blocks. The addition of several segments produces an IMB. This type of IMRT can be delivered with technology already available in centres using an MLC to treat patients with 3DCRT, and is currently being used in Europe and the United States to treat patients with cancer of the prostate, head and neck, lung, breast, liver, brain, and other sites [20,24,27,45]. To produce the required conformality 4-9 beam directions may be required for radiotherapy treatment depending on the complexity of the PTV [20,24,27]. Each field may consist of 3-20 sub-fields which are delivered in succession, and the total number of monitor units delivered per field is often much higher than for conventional radiotherapy because of the attenuation of the beam by the MLC leaves. These factors increase treatment time from around 10 minutes for a conformal treatment delivery to at least 25 minutes for Intensity modulated radiotherapy [20]. Higher dose rates and optimisation of the sequence of delivery of segments have been used to minimise treatment times [20]. Physical problems with the use of an MLC to define segments include accuracy of MLC leaf placement, interleaf radiation leakage, the tongue and groove effect, and the accuracy of delivering small numbers of monitor units to some segments [46]. 3. Dynamic MLC (dMLC) Modulation of beam intensity by pairs of moving MLC leaves characterises this technique (also known as the “sliding window” technique). The IMB is constructed from a series of one dimensional IMB formed by the differential speed profile of the leading and trailing MLC leaves [47,48]. Each leaf pair in the MLC leaf bank moves through a series of check points determined by an interpreter which converts the required intensity distribution into speed profiles for each leaf pair [49]. Delivery times are quicker than for multiple-static segments; typical delivery time for a five-field prostate treatment is 14 minutes (personal observation). The leaf movements of the MLC during radiotherapy treatment must be accurate, as these produce the IMB. Leakage and transmission of radiation between or through MLC leaves must be taken into account in the dose calculation. Leakage occurs both between adjacent leaves, and between the ends of opposing leaf pairs. MLC’s produced by different manufacturers vary greatly in this respect. Transmission of radiation through MLC leaves is less than 1.5-2% although larger transmission of up to 2.5-3% have been measured at the interlocking leaf edge [50,51]. The phenomenon known as the “tongue and groove” effect, can be removed by “synchronisation” [52,53]. Dynamic leaf movement during the treatment delivery, combined with continuous arcing of the gantry is known as intensity modulated arc therapy [54].This technique has the advantages of quick treatment time (estimated at 5-10 minutes), and may allow the use of fewer intensity levels than dMLC [49]. 4. Tomotherapy Tomotherapy describes radiotherapy IMRT techniques which irradiate the target slice by slice. The NOMOS Corporation developed the first commercially available tomotherapy machine, the Multivane Intensity Modulating Collimator (MIMiC) which is in use in several centres in the United States [55,56,57,58,59,60,61]. This device attaches to the head of the linear accelerator which arcs about the craniocaudal axis of the patient (Fig 6). The MIMiC is a binary temporal modulator consisting of 40 tungsten leaves arranged in two banks, each irradiating a slice. Each leaf is driven by a pneumatic mechanism and defines a beam approximately 1cm square at the isocentre. Each bank of leaves treats up to a 2cm thick volume of tissue 20 cm in diameter at one time. Target volumes longer than 4cm require the couch to be indexed in the craniocaudal axis such that the next set of two slices of target volume may be treated. Accurate indexing of the couch and patient are critical to the radiotherapy treatment of large target volumes. Each slice has to be exactly matched to the previous slice and small errors of mismatch may lead to unacceptable over- or under-dose of the target. For example Carol et al [62] predicted a 41% inhomogeneity with a 2mm error in treatment set-up, although if slices were matched to 0.1 mm the inhomogeneity was only 2-3%. NOMOS Corporation recommend the use of a special indexing table called the CRANE which is said to index the patient to an accuracy of 0.1-0.2mm (NOMOS Corporation, Product Information). Measurements of delivered dose confirm homogeneity within ? 5% of predicted dose [60]. Mackie et al [63] proposed the construction of a more complex tomotherapy treatment machine which would avoid indexing by arcing the gantry while the patient was translated slowly through the machine defining a spiral treatment geometry. The prototype is currently being assembled at the University of Wisconsin [64]. Other methods of producing IMB’s include an attenuating bar which moves across the open radiation field [65], and a scanning pencil beams [66]. Perhaps the ultimate IMRT tool might be a linac mounted on a robotic arm capable of pointing in arbitrary directions. Such a system does exist [67], and recently a detailed inverse planning study has shown that complex 3D dose distributions can be sculptured for concave targets in the vacinity of radiosensitive structures [68] Inverse planning Immobilisation IMRT may produce high dose gradients close to critical radiosensitive normal-tissue structures, and with some techniques precise matching of segments is essential. Accurate and reproducible patient positioning with the use of appropriate immobilisation devices is therefore required. For cancer of the head and neck or brain, the use of customised thermoplastic shell [58], stereotactic frame or invasive system is used to immobilise the patient to within 1-2mm [60]. For the radiotherapy treatment of prostate cancer, several centres recommend the use of a pelvic immobilisation device [71]. Treatment Verification All radiation therapy requires accurate verification of treatment delivery. This can be broadly divided into quality assurance of planning and delivery techniques within a radiotherapy department, and the verification of the radiotherapy treatment on a per-patient basis. For IMRT the delivery of each fraction of a course of radiotherapy requires careful verification. When the patient is in the treatment position it is possible to check the treatment isocentre position by taking orthogonal portal images either with film or more satisfactorily using an electronic portal imaging device (EPID) [75-78]. The process of treatment verification depends on the treatment delivery method. If a step-and-shoot segmented treatment technique is being used, then portal images can be obtained of each segment, and these segment shapes can be checked against the MLC shapes from the treatment planning system [78]. Often however the segments themselves are too small for accurate identification of bony anatomical landmarks by which the segment placement can be checked, and larger orthogonal images will be required to check the isocentre position. The accuracy of the radiotherapy treatment plans can be verified using phantoms with measurement of the dose distribution by film, TLD, or polyacrylamide gel [32,58,79,80]. For dynamic sliding window delivery techniques, there is no accepted verification procedure. The treatment radiographer can watch the movement of MLC leaves on a computer display in the treatment area, and some groups have suggested the leaf movements could be followed by a camera within the treatment head [81] or by correlating portal snapshot images with predicted leaf positions [82]. Manufacturers claim leaf position tolerances of <2mm during dynamic delivery, very low rates of treatment error, and provide overrides to switch the beam off if errors exceed certain preset levels. Dosimetry studies in humanoid and water phantoms support the claims of high levels of accuracy of delivered dose, although some groups propose dose measurements at points within the patient (in vivo dosimetry). In reality the few centres (eg MSKCC) which have implemented dMLC delivery use weekly isocentre checks on each patient, and rely on the machine QA and overrides for accuracy of of radiotherapy treatment delivery. Conclusions It seems likely that the benefits of Intensity modulated radiotherapy treatment will be greatest for patients with cancer targets which are concave, and where normal-tissue avoidance is clinically important. On this basis it has been suggested that around 30% of current radiotherapy patients could significantly benefit if they were treated with an IMRT technique. The delivery of IMRT is complex, time intensive in planning and treatment, and for some patients may offer little or no advantage over conventional more simple techniques. In a health service of limited resources it is important to examine the potential benefits of any new technique, to estimate how many patients are likely to benefit, and to investigate the size of potential the benefits [83]. Radiotherapy planning studies offer the best opportunity to do this. By comparing the current radiotherapy technique to an IMRT technique for a particular cancer site, the potential improvements which might be expected, if IMRT was introduced, can be estimated. The advantage of planning studies is that they are relatively quick to perform, and data from previously treated patients can be used as the “control group”. Several rival radiotherapy treatment techniques can be compared to one another, and objective end points (such as radiation dose to a particular organ) can be used. The result of a planning study requires careful interpretation, but may help to identify tumour sites suitable for further clinical investigation. A planning study does not however provide clinical data, and is no substitute for suitable clinical trials, which are needed to confirm the safety and efficacy of new radiotherapy techniques as well as to compare clinically important end-points (tumour control, side effects, quality of life, health economics) between rival treatment approaches. References |
|
