British Journal of Radiology (2005) 78, 384-393
© 2005 British Institute of Radiology
doi: 10.1259/bjr/25963871
Stereotactic radiosurgery at St. Bartholomew's hospital: third quinquennial review
A K El Hamri, MSc,
J Monk, MSc and
P N Plowman, MD, FRCR
St. Bartholomew's Hospital, London EC1, UK
Correspondence: Dr P N Plowman, Department Radiotherapy, St. Bartholomew's Hospital, London EC1, UK
 |
Introduction and "case mix"
|
|---|
This manuscript describes the stereotactic radiosurgical work of one institution over a 5 year (1999–2004) period and discusses this work in the context of the changing field of radiosurgery.
In 1949, Leksell described a system, which concentrated radiation therapy on intracranial targets within the brain [1]. He conceived image guided, multiple, cross-firing, tightly collimated and small radiation portals "focusing" a high single radiotherapy dose on an intracranial target. His first clinical work was aimed at destroying intracerebral pathways in functional disorders and he coined the name stereotactic "radiosurgery" – a nickname which has endured. The early work was hampered by inadequate equipment, as only a 200 KV X-ray apparatus was available. The concept was next furthered in 1968, again in Sweden, with the introduction of a cobalt-60 gamma unit [2]. The UK's first gamma unit became operational in Sheffield in 1985 [3]. In the current machine at St. Bartholomew's/London Radiosurgical Centre (Gamma Knife; Electa Instruments AB, Linköping, Sweden) there are 201 fixed cobalt-60 sources, each a thin rod of 1 mm diameter, the long axis of which is oriented along a radius of a hemisphere – the helmet, into which the patient's head, within a stereotactic frame, fits. The centre point (or isocentre) of this hemisphere is the point at which the stereotactic co-ordinates of the mapped intracranial target are positioned. In the last decade there have been several advances in Gamma Knife technology. The introduction of the model "C" machine has brought automation to various aspects, whilst maintaining the principal design features. The biggest single change has been the introduction of the motorized automatic positioning system (APS) for targeting each shot of the therapy. The APS has been demonstrated to significantly shorten the overall time taken to deliver a treatment and by reducing the need for human intervention, to lessen the risks of treatment errors [4]. In tandem with hardware improvements, there is continual progress in software design and functionality. The most significant has been the introduction of Gamma Knife "WizardTM". This software can be used to semi or fully automate the treatment planning process. Whereas previously, a treatment plan was formulated by a skilled operator, who placed "shots" to obtain an optimal plan, the Wizard employs a computer to automatically calculate the optimal distribution and weighting of shots to best encompass the target. At present, the unit at our centre is the "B" model, although we are capable of mimicking certain "C" model/Wizard characteristics (vide infra).
The last decade has seen a substantial advance in the quality and accuracy of three dimensional imaging technologies. The finer detail and clarity obtained improves the accuracy in delineating target lesions (e.g. the excellent imaging of the trigeminal nerve by modern MRI for radiosurgical therapy of trigeminal neuralgia – a routine part of Gamma Knife practice for drug resistant patients). With the increasing availability of PET, work by a variety of groups has started to integrate PET with conventional static, anatomical CT/MRI into the Gamma Knife treatment planning process [5]. Early results suggest that these combined methods of imaging improve target definition, particularly for infiltrating tumours whose boundaries are not so certainly defined on MR alone – we illustrated this in our last quinquennial review [6] and predicted that the technique would also prove valuable to demonstrate whether abnormalities persisting on MRI/CT, after a full cancer therapy programme, remain viable and should be targeted by radiosurgery (e.g. after chemotherapy in the case of intracranial germ cell tumours or after whole brain radiotherapy in the case of metastases) [7].
Over the last 15 years, linear accelerators (linacs) have been adapted for stereotactic delivery of radiation therapy, initially using the isocentric mounting and rotation technology but latterly both fixed fields and dynamic arcing methods have been introduced, in combination with multileaf collimator (MLC) conformational technology. In 1989, the first UK linac based radiosurgery system commenced clinical operation at St. Bartholomew's Hospital and initially employed multiple non-coplanar arcs around the stereotactically mapped target (usually as a single iso-centre) as the treatment technique. In our last quinquennial review we described our work using that technology [6]. Since that time, the routine availability of multileaf collimation in linacs (leaf widths from 3–10 mm – either as additional accessories (mMLC) or as an integral part of a linac) has allowed improved beam shaping and now most centres practising stereotactic radiation therapy employ either multiple fixed fields or dynamic arcing together with multileaf collimation to improve the conformity of the technique.
At St. Bartholomew's, the acquisition of both Gamma Knife technology and 3 mm leaf width micro-multileaf collimation (in conjunction with our stereotactic linac technology) within the last 5 years has allowed the present quinquennial review to report interesting data on many aspects of practise and be in the unique UK position to make some objective, comparative observations concerning the two major photon technologies currently utilized for stereotactic radiation therapy.
The distribution of tumour types treated in the last 5 years (circa 400 patients) has been compared with those (circa 400 patients) reported in the last quinquennial review [6] (Figure 1
). Differences are noteworthy.
View larger version (30K):
[in this window]
[in a new window]
|
Figure 1. Pie diagrams of (a) distribution of diseases treated to 1999 and (b) 1999–2004 by radiosurgery at this centre.
|
|
Despite the proportionate increase in other diagnoses, arteriovenous malformations (AVM) remain the most common disease category treated by our unit, and remain the diagnosis par excellence that established the technology as a powerful force in modern therapeutics (Figure 1
). The cavernous angioma is overall less common (but diagnosed incidentally more commonly now, during modern MR examinations for other reasons). We accept patients with recurrent symptomatic bleeds for focal radiation therapy. There have been several reports of higher complication rates following radiosurgery for cavernoma (size-for-size and dose-for-dose) than AVM and we have made some recent observations that we believe may explain a phenomenon that we have not encountered in our radiation series for ocular or other cavernous angiomas. We have utilized these observations to adapt our treatment methods. These are discussed below.
Meningiomas currently comprise a larger proportion than in our early series and are, for the majority, skull base tumours which have proved difficult to curatively resect (e.g. those involving the cavernous sinus), although parasagittal meningiomas represent another separate site group which has also grown in referral numbers. We attribute this increase in referral and treatment practice to a better awareness of the success of such therapy. In our series, atypical/aggressive meningiomas are not more heavily represented than their true incidence.
The treatment of vestibular (acoustic) neuroma (and of other intracranial neuromas) by stereotactic radiation therapy has become a standard, front-line, definitive treatment of this disease with excellent, durable, local control rates and reasonable hearing preservation results. There has been recent litigation in the UK for failing to counsel concerning such treatment by a patient who went for open resection, which was attended by complications. Not surprisingly, the representation of vestibular neuroma has risen in our latest series (Figure 1). Nowhere in stereotactic radiation therapy is the controversy between Gamma Knife and linac technology more hotly contested than in the treatment of vestibular neuroma, largely owing to the passage of the "innocent bystander", normal and radiosensitive acoustic nerve through the target volume during therapy and the desire for hearing preservation, discussed below.
Pituitary adenoma has presented an intellectual problem. On the one hand, our data and those of others have demonstrated the high efficacy of well-fractionated, moderate dose equivalent conventional radiotherapy for the long term control of pituitatry adenoma [8] and with little late morbidity (pituitary hormone endocrine deficiency being the most common late sequel). On the other hand, the adenoma is frequently a discrete and small lesion, within the fossa and well visualized on MRI; furthermore, recent radiosurgical publications suggest good tumour and endocrine control of such lesions [9]. As predicted in our last quinquennial review [6], a selective increase in the numbers of pituitary adenomas accepted onto the programme has occurred in the recent series, although conventionally fractionated radiotherapy remains the treatment recommended for the majority of patients presenting to this hospital with pituitary adenoma. Paediatric pituitary adenoma presents a particularly difficult area for us – on the one hand there is a paucity of data following radiosurgery whilst on the other hand, conventionally fractionated radiotherapy undoubtedly encompasses more hypothalamus/normal brain – even in the present conformal age – with potential attendant morbidity. This matter is further discussed below in the context of our recently reported large series of radiation treated paediatric Cushing's disease patients.
Craniopharyngioma presents a problem not dissimilar to that of vestibular neuroma in that there is a normal special sensory (radiosensitive) nerve, the optic chiasm, frequently within the target volume, and the price of harming that special sensory nerve structure is clearly devastating. Interestingly, the proportional representation of craniopharyngioma has not appreciably increased in our current series but we have reported new observations on one particular subgroup (intrasellar disease) and this is discussed below.
In our last quinquennial review [6], we noted that only 2% of our treated cases had been metastatic disease to the brain and drew attention to the fact that metastases comprised up to 33% of some Gamma Knife caseloads in other radiosurgical units. We predicted that there would be an increase in the number of metastases accepted onto our programme in the future and this has proved the case (Figure 1). The literature demonstrates the high efficacy of the technique for sterilizing individual lesions [10], but we still believe that selection of cases must be vigorous (vide infra): patients undergo high quality and thin slice MRI scans of brain to best define the extent (number) of metastatic disease and systemic staging before acceptance onto our programme – acceptance being via a multidisciplinary advisory group meeting for all patients (and all diagnoses). This is discussed below.
Other diagnoses accepted include haemangioblastoma (reviewed in our last quinquennial report [6]) and residual/recurrent discrete targets after conventional therapy (e.g. low grade glioma, central neurocytoma etc.). These are not discussed in depth here, but the point is worth making that for some other discrete and unresectable tumourous lesions, radiosurgery becomes the prime therapy choice. Similarly, the acquisition of Gamma Knife technology, with its extreme precision and ability to treat down to 4 mm collimator size portals, has led to the practice of "functional radiosurgery" (e.g. trigeminal neuralgia, the treatment of mesotemporal sclerosis causing refractory epilepsy), not previously represented in our last quinquennial review [6]; this will be the subject of future research.
|
Physics research
|
|---|
In the second half of the 1990s, and for linac based radiosurgery, there was a move away from multiple arcing beams using circular collimators to multiple static conformal beams. Conformality is produced by shaping each field either by custom made lead blocks or by multileaf collimation. A comparison by Adams et al [11] between stereotactic conformal radiation therapy treatment plans using a standard 10 mm leaf width multileaf (MLC) and custom made blocks had demonstrated that, for a similar planning target volume, there was worse conformality achieved by the MLC. However, the width of 10 mm of the MLC was regarded as the weak point of this comparison and 5 mm leaf width MLC were becoming available as standard equipment on our Varian linacs thus improving conformality. We had published a study on the comparative conformality of our early linac technology with that of Gamma Knife technology [12] and it became clear that we should repeat the study using the newer conformal static field conformal linac methodology. However, there had been recently released micro-MLC equipment (BrainLAB, Munich, Germany) with 3 mm leaf widths and it seemed clear that if we were to produce a definitive study on the respective conformalities achievable on linac versus Gamma Knife technologies then it would be sensible to first examine whether conformality was indeed significantly better using the micro-MLC and then use the perceived optimal linac technique. Added to this, there were some data suggesting that dynamic arcs might be better than fixed fields in conformal therapy and so we factored this into the comparison. Adding to the difficulties in the design of this study was the fact that the dosimetry of the Gamma Knife technology was also improving in the following regard: the newest model of the Gamma Knife had computer assisted planning and automated sequencing of fields which made it more practical to treat many more "shots" using the smallest (4 mm) collimator – thus enhancing conformality. To compare like with like we decided to mimic this facility in our comparison such that optimal technologies were compared.
Thus, our first study compared the 3 mm/micro-MLC with the 5 mm leaf width collimator for linac stereotactic radiation therapy: 14 patients with various irregular shaped intracranial lesions were selected for this study. All were planned for therapy using both the micro-MLC with 3 mm leaf widths (BrainLab, Munich, Germany) and the Varian Millenium MLC (5 mm leaf widths). There was a small but significant improvement in the conformality using the 3 mm leaf-MLC. For the treating clinician, the salient point was that the mean increase in the volume of adjacent critical structure enclosed within the 50% and 70% isodoses was 5.7% and 4.9%, respectively [13]. Whilst there remains debate as to whether the size of this improved conformity mandates stereotactic radiation therapy departments to acquire the micro-MLC equipment, we thereafter accepted that we should utilize the micro-MLC technology for comparison with the latest planning methodology of Gamma Knife (Electa Instruments AB, Linköping, Sweden).
Our second comparative study therefore compared fixed field and dynamic arc linac technology using the micro (3 mm)-MLC equipment with modern Gamma Knife technology [14]. We chose acoustic neuroma as the disease for the comparison, as it has generated more controversy than perhaps any other in the discussions of optimal stereotactic radiation therapy technique. Eight patients with neuromas of various sites (intracanalicular, extracanalicular and both) and sizes were selected for study and conformity indices (total volume of prescription isodose/ volume of target covered at the prescription isodose level) and dose–volume histogram (DVH) (for target and brainstem) were computed. A clear difference between linac technology and gamma knife was found. Gamma Knife technology demonstrated a statistically significantly better conformity index (p<0.02). Additionally, when linac techniques were compared, dynamic arcing demonstrated a statistically lower conformity index (p<0.05) than fixed fields. Of course, conformity indices were worse for small targets in all techniques but the differences just described were true across the board, albeit most exaggerated for small targets.
Although the logical conclusion from the above is that the adjacent normal tissue (viz. the brainstem) should be better spared in the Gamma knife technique – and this is true, nevertheless, the highest point maximum doses received in the brainstem in our series were in patients with larger acoustic neuroma indenting/abutting the brainstem and planned for Gamma Knife. This is a feature of the higher internal dose gradient in the Gamma Knife technique, which relies on overlapping fields/shots.
The study of the minimum dose to the target (TDmin) has proved an interesting subject [15]. As a consequence of high conformality, the Gamma Knife technique delivered a lower TDmin than either linac technique. When the two linac techniques were compared, the dynamic arc still achieved a higher TDmin than the fixed field method notwithstanding the fact that it also achieved a slightly better conformity index over the fixed field technique – this was unexpected. The subject of TDmin is of potential clinical importance in that underdosing of tumourous areas may lead to treatment failure; consequently the quest for high conformity may need to be tempered when high conformity leads to a significantly lower TDmin (underdosed target/tumour areas).
The potential importance of our TDmin observations prompted the next study where the TDmin was held constant and differences in conformity were scrutinised. In short, we adjusted the Gamma Knife plans such that the TDmin on the Gamma knife plan was the same as on the linac plans and then (re-)compared the conformity index. Interestingly, the Gamma Knife plans still maintained the advantage over the linac ones. We conclude that too little attention has been made of the TDmin as a parameter that rates alongside conformity index in importance and this conclusion has applicability in other conformal radiotherapy planning outside the cranium [15].
Another physics contribution was the study of the need (or not) for distortion correction during the use of digital subtraction angiography (DSA) in the mapping of arteriovenous malformations (AVM) for Gamma Knife therapy [16]. DSA offers several advantages over conventional angiography, including increased contrast sensitivity (reducing the volume of contrast agent injected), easier image manipulation and "real time" viewing. However, the technique suffers from worse spatial resolution and is prone to geometric distortions, which may require individual correction. The degree of geometric distortion varies inversely with radial position and is significantly reduced at the centre of the image. In our stereotactic radiosurgery work it was clearly important to know if we should routinely employ such corrections, but it was also possible that this would be unnecessary as we deliberately fit the stereotactic frame such that the target to be treated is placed near the centre of the frame, where any distortion is least. In the study, patient and phantom images (with and without image distortion correction) were assessed. The errors measured in angiograms were compared with those of MR and CT images. All errors measured in the study were sub-millimetre. In the patient group, errors of distortion correction measured 0.13 mm (range 0–0.3 mm). Without distortion correction, errors averaged 0.34 mm (range 0.1–0.6 mm). By comparison, the average error for intracranial MRI was 0.3 mm (max. 0.6 mm). We concluded that distortion correction did not significantly reduce (i.e. by a factor that was clinically important) error(s) associated with the definition of the stereotactic co-ordinates in our Gamma Knife radiosurgery work for AVM [16].
A further publication related to the possible substitution of MR mapping alone rather than combined MR and conventional angiography in the mapping of AVM for stereotactic radiosurgery [17]. Current radiosurgical treatment of AVMs relies on planning protocols which integrate data from both MR and stereotactic angiography. However, angiography is invasive and carries small but not insignificant risks, whereas MR is non-invasive with multiplanar capability and can demonstrate AVM anatomical detail well. Guo et al [18] found that the delineation of the AVM nidus by MR mapping was superior to conventional stereotactic angiography particularly for medium to larger AVM. However, the authors did not believe that MR should be used alone for the mapping process because of concerns regarding geometric distortion arising from the magnetic field, inhomogeneity, MRI artefacts and the inability of MR to demonstrate the temporal sequence of vascular filling and the distribution of the feeding artery/ies. In our study, AVM planned in the conventional manner were re-planned using MR technology alone by another member of the team. In the first instance it was found that there was insignificant interindividual variation with regard to MR nidus definition on MR amongst the team – a crucial first step. There was no significant difference in the target volume between conventional dual modality planning and MR only but there was significant (random) displacement of the volume in space, that varied little between individuals. We concluded that MR mapping alone led to random displacement of the target compared with dual modality planning and such displacements could produce variation that was too large for safe radiosurgery [17].
|
Arteriovenous and cavernous malformation therapy research
|
|---|
AVM was the diagnosis for which the technique of stereotactic radiation therapy (radiosurgery) first established itself as a first line therapy and then came acceptance of the technique for other angiomas. A review of this subject was the subject of an editorial in 2002 [19] and research into the subject of AVM currently revolves around the role or otherwise of embolisation in conjunction with radiosurgery in these cases and whether there is logic, safety and efficacy in treating parts of large AVM at different radiosurgical treatment sessions over time to lessen the risks of late complications associated with treatment of large volumes.
Our major vascular malformation manuscript this last 5 years [20] has dealt with the subject of cavernous angiomas of the brain and specifically the observation in several series that there is a higher incidence of late treatment complications following radiosurgery for these angiomas than AVM (equated for size, site and dose). This has not been observed in publications on radiotherapy for cavernous angiomas elsewhere in the body and certainly not in our large ocular cavernous angioma experience [21, 22]. As we say in our manuscript [20]: "speculation that a tumour/angioma may "radiosensitise" its normal tissue environs is a foreign concept in radiation therapy and presents a novel problem". Given the safety of radiation therapy for cavernous angiomas elsewhere, our interest in the problem centred on the unique presence of the haemosiderin stained fringe of normal brain tissue that always surrounds a brain cavernous angioma, well seen on modern MRI (Figure 2). The possibility that normal brain tissue surrounding a brain cavernous angioma is sensitized to radiation by virtue of "staining" by haemosiderin is a thesis that we elaborated. Under conditions of reduced pH (e.g. hypoxia, inflammation) iron is liberated from haemosiderin and can act to promote both hydroxyl free radical formation in the presence of hydrogen peroxide and lipid peroxidation in lysosomes, quite apart from any enhancement of secondary electron flux. In situations where there is impairment of protective cellular mechanisms e.g. a reduction of cellular reducing metabolites, this capability of released iron to generate reactive oxygen intermediates, leading to oxidation and peroxidation of membranes and DNA, may become unchecked and result in tissue damage. We reviewed evidence that the perilesional brain is under conditions of chronic inflammation and could well fall into the situation just described. We also called on Nelson and Stevens' study in Chinese hamster ovary cells demonstrating that the presence of iron enhanced damage in this in vitro assay [23] and the data of Hornsey demonstrating in an in vivo assay of rat spinal cord that chelation of iron reduced the late effects of radiation (by reducing the production of hydroxyl free radicals) [24] to support our thesis [20]. Lastly, we called attention to the fact that the iron impregnated tissues could influence the accuracy of MR mapping due to the magnetic susceptibility effect caused by haemosiderin [20]. (Iron containing substances cause intravoxel magnetic field inhomogeneities with line broadening on MR).
View larger version (125K):
[in this window]
[in a new window]
|
Figure 2. T1 weighted axial MR scan in the region of the upper pons, demonstrating a small brainstem cavernous angioma. The angioma is surrounded by a haemosiderin ring (black on scan), which follows chronic blood seepage into the normal surrounding brain from the angioma over time.
|
|
|
Vestibular (acoustic) neuroma research
|
|---|
In the last 5 years, the acceptance of radiation therapy as a first line curative therapy for vestibular/acoustic neuroma, and an alternative to neurosurgery, has become standard. Long term control and morbidity results following radiation therapy compete attractively with those from microsurgical units [25]. Hereafter, the controversy as to which radiation therapy method is optimal represents the largest area of disagreement in the profession at present. On the one hand, the data from the Gamma Knife centres undoubtedly contribute the majority of the world literature concerning long term, good control rates and low morbidity, arguing strongly that this method has the best pedigree. However, the very high internal dose gradient of the Gamma Knife technique, allowing double dosing of the innocent bystander cochlear nerve (particularly in intracanalicular cases), may well prejudice hearing preservation more than a technique, which has firstly a much lower internal dose gradient and secondly allows fractionation. Here enters the linac methodology into the "fray". With a low internal dose gradient (circa 10% versus circa 100% in the Gamma Knife technique) and the capability to fractionate, there is much interest in the use of linac stereotactic methods, where hearing preservation is of prime concern. Early data appear promising [26] but longer follow up is needed to be sure that this is durable and is not at the expense of loss of tumour control. The Gamma Knife offers the better validated data on tumour ablation – after all, the integral dose within the tumour will be vastly more in a technique that has such a high internal dose gradient. Our physics contributions to this debate have been reviewed above and we discussed this topical clinical controversy in a recent editorial [27].
In 2003, we presented the first UK case of a patient experiencing the return of hearing in a "dead" ear following radiosurgery (Figure 3) [28].
View larger version (94K):
[in this window]
[in a new window]
|
Figure 3. Gadolinium enhanced, T1 weighted, coronal MR scans of an acoustic neuroma (a) before and (b) after radiosurgery, demonstrating unusually dramatic shrinkage. Audiograms (c) before and (d) after demonstrate the exceptionally good return of hearing in this patient following this therapy.
|
|
Neurofibromatosis type 2 (NF-2) patients carry an oncogene in the genome and this may predispose them to oncogenesis following radiation therapy. Thus, whereas Baser et al [29] found the background rate of CNS malignancy amongst 1500 NF-2 sufferers to be 0.5%, these same authors found that of 117 neurofibromatosis patients who had received radiation therapy the incidence of late malignancy was 6% occurring within the radiation field or penumbral region – including four cases of malignant peripheral nerve sheath tumour, suggesting malignant change within the tumour itself. The very real possibility is there that the inherited loss of the NF-2 gene on the long arm of chromosome 22 leads to the loss of a critical tumour suppressor gene. As we said in our first publication on this topic: "this might well interact with radiation to lead to this higher observed incidence of late oncogenesis" [30]. We subsequently enhanced the thesis by invoking the analogy with hereditary retinoblastoma, where Knudson [31] modelled a scenario in which the first predisposition to the development of a tumour was the possession of the inherited loss of the Rb suppressor gene in the genome (the "first hit in a two hit process" in Knudson vernacular); radiation provided the second predisposition/hit by mutating the second allele. We suggested that a comparable situation applied to the NF-2 cases treated with radiation, although we called attention to the fact radiation-induced late tumours might occur either within or without the treated target (i.e. in the treated tumour itself or in the penumbral zone) and observed that in the NF-2 series data so far, it appeared more common that the second tumour occurred within the target/tumour volume [32].
Meanwhile, in the last 5 years, the group in Sheffield have produced good data demonstrating that acoustic neuromas which are secondary to NF-2, respond less well to radiosurgery than sporadic cases with the volume of the treated tumour being very influential in the likelihood of control [33]. Having said this, the Sheffield experience is that overall 50% of NF-2 cases will be controlled by radiosurgery and 20% will have definitely failed by 8 years, whilst 30% will have some concerns overall durable control [34].
In a disease that is bilateral and therefore threatens total hearing loss there will remain a place for radiosurgery but the last 5 years' literature has brought increasing knowledge upon which to base decisions.
|
Pituitary adenoma research
|
|---|
In our last quinquennial review [6], we outlined our intellectual problem with introducing stereotactic radiosurgery as primary definitive radiation therapy for pituitary adenoma, when conventionally fractionated radiotherapy produced excellent results and had been used in this department over decades. However, we were building up a series of patients who had failed such conventional radiation therapy. Specifically, we published the results of radiosurgery using the linac on 21 patients all of whom had undergone conventional radiotherapy to 45–50 Gy; 18 of these patients had also been operated at least once. Clearly this was a difficult population of patients in that the optic chiasm, just rostral to the target, had received the full dose of the previous radiotherapy and this prejudiced the dose of radiosurgery that could be safely be prescribed.
Swords et al [35] reported the follow-up of these 21 patients. 13 patients had somatotroph adenoma, four corticotroph adenoma, three non-functional adenoma and one prolactinoma. The radiosurgical dose varied between 8 Gy and 15 Gy (modal prescription dose 10 Gy), depending on the optic chiasmal dose; the median follow up was 33 months with a range of 3–72 months. Radiosurgery by this linac technique proved safe and effective: amongst our somatotroph ademoas, we observed a more rapid reduction in growth hormone (GH) and insulin-like growth factor (IGF-1) levels than we are accustomed after conventional radiotherapy (Figure 4), with normalization of GH and IGF-1 levels in 58%. Although Landolt et al [9] had made a previous observation regarding faster falls in GH after radiosurgery versus conventional radiotherapy, our observation was in previously irradiated patients and was therefore new. Mean GH levels fell from 21.6 mU l–1 to 7.9 mU l–1 (Figure 4) and the mean IGF-1 level from 624 ng ml–1 to 384 ng ml–1. In this stereotactic radiosurgical series of previously irradiated patients, there was clear cut evidence of high efficacy of stereotactic radiosurgery. The dose to the optic chiasm was limited to 3 Gy in this series and no visual complications have been observed to date. We believe that the dosimetrically superior Gamma Knife technique (with regard to sparing dose to the rostral chiasm) will take on our work in this area and we will report on this subject again, including primarily treated patients in our next review.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 4. Mean serum growth hormone levels over time (each data point being the mean value of a five point day curve) in a population of previously irradiated (conventionally fractionated to a minimum dose of 45 Gy) patients with acromegaly, who had failed this therapy (in terms of growth hormone normalization or subsequent rise) and who were then treated by linac radiosurgery (data from Swords et al [35]).
|
|
There is a continuing controversy as to whether stereotactic radiosurgery should supplant conventional radiotherapy in the primary therapy of pituitary adenoma. For example, it could well be argued that the dose of ionizing radiation received by adjacent structures (e.g. hypothalamus) would be considerably less if patients were routinely treated by radiosurgical methods. On the other hand, not infrequently pituitary adenoma is a "trans-fossa" disease and there is a necessity to fully irradiate the whole fossa and adjacent cavernous sinus up to the carotid arteries and rostrally above the diaphragma, coupled to the fact that early radiosurgical results were suboptimal [36]. These features, which remain relevant in the post-operative setting, may well argue in favour of conventional radiotherapy, which after all has an excellent pedigree of durable success/control in this disease reviewed in 1999 [36], although we are now looking at selected use of primary radiosurgery for small discrete adenomas. However, it should be born in mind that these latter cases are usually exactly the cases that do well with modern surgery.
We have recently reported on a rare form of pituitary adenoma: viz. paediatric Cushing's disease [37]. Here is a benign adenoma, usually small and in a vulnerable patient population with regard to late morbidity of therapy, but one in which first time cure is vital as subsequent attempts are not attended by good long term control rates. The early radiosurgical literature in this disease in adults was not persuasive [38]. Following transphenoidal surgery, we studied seven paediatric Cushing's disease children who were not cured by the surgery. All received conventionally fractionated radiotherapy and with a mean follow up of 6.9 years, all remained in endocrinological remission and probably cured [37]. The late morbidity included GH deficiency but surprisingly normal gonadotropin secretion, although one male demonstrated early puberty; adrenocorticotrophic hormone (ACTH), thyroid stimulating hormone (TSH) and interestingly prolactin axes remained normal. With these good conventional radiotherapy results we have elected to continue with conventionally fractionated radiotherapy in this patient group for the present – two children so treated in the first quarter of 2004.
|
Craniopharyngioma research
|
|---|
In our last quinquennial review we overviewed the role of radiation therapy in craniopharyngioma and illustrated good response of this disease to focal single dose radiation therapy. However, we debated the merits of focal single fraction radiation therapy versus fractionated therapy in a situation where there is a radiosensitive special sensory nerve (viz. the optic chiasm) running adjacent or more frequently within the target volume. For example, the case illustrated in our publication demonstrated complete response of the mass lesion after radiosurgery but the patient subsequently lost vision in the ipsilateral eye due to the radiation (Figure 4) [6]. (The situation is highly comparable with that with the radiovulnerable cochlear nerve running within the radiosurgical target volume of an intracanalicular vestibular neuroma). We concluded then that there was a real argument in favour of conventionally fractionated radiotherapy for this disease under these circumstances and that view point was highlighted when the Karolinska workers reported in 2002 on 21 consecutive patients (11 children) with craniopharyngioma who had been treated by Gamma Knife radiosurgery, with a mean of 16.8 years' follow up [39]. 13 patients received a marginal dose of 6 Gy or less to respect chiasmal single dose tolerance (which is not perceived to be a radical radiosurgical dose and certainly equates to a very much lower dose equivalent than could have been achieved by standard/conventional fractionated radiotherapy); 11 of the 13 patients had relapsed. Not only was the control rate in this series poor but there was visual deterioration in 8 of 21 patients and there was uncertainty as to the aetiology of this visual loss. The authors concluded, as have many others, that there must be a space between the chiasm and the craniopharyngioma for safe and effective radiosurgery to be indicated as a therapeutic modality; although, as in many radiosurgery papers, conventionally fractionated radiation therapy data are not reviewed as a highly effective alternative in their manuscript, highlighting the sorry divide between neurosurgeons practising (usually Gamma Knife) radiation therapy divorced from multidisciplinary teams with good radiotherapy facilities.
In 2003, we published a series of three patients who had been referred to us with recurrent craniopharyngioma entirely within the pituitary fossa and we had been safely able to deliver radical radiosurgical dosage and who had achieved complete responses [40]. We concluded that radiosurgery for intrasellar craniopharyngioma was indeed a very good treatment option.
In 1999, we published our intracystic Y-90 instillation results for cystic craniopharyngioma [41], and the only other point to highlight in this review is the combined use of intracystic Y-90 and stereotactic radiosurgery for semicystic, semisolid craniopharyngiomas for which we have only limited data at present.
|
The management of metastases
|
|---|
The optimal management of brain metastases has continued to be a controversial topic in the last 5 years, with the benefits of whole brain radiotherapy continually being questioned for some patient subgroups [42]. Undoubtedly, patient performance status/extent of disease predicts success or failure [43]. Notwithstanding the literature demonstrating increased control of brain disease by whole brain radiotherapy after neurosurgical resection of an apparently isolated brain metastasis, the delegates at a recent multidisciplinary conference reported in 2001 [44] demonstrated that all clinicians involved in the care of these patients were dissatisfied with the achievements of whole brain radiotherapy. In 1999, Kondziolka et al [10], reported a small trial of patients with two to four metastases, none greater than 2.5 cm in diameter and with a KPS of >70. These patients either received whole brain radiotherapy (WBRT) alone or WBRT with a radiosurgical boost to the established metastases. The rate of failure at 1 year was 100% after WBRT alone as compared with 8% after WBRT plus radiosurgical boost. Our "straw poll" [44] of the delegates of a multidisciplinary symposium in London reflected the views of an editorial of the time [45] viz. that stereotactic radiosurgery should be further explored for better prognosis patients. Since that time, an RTOG study [46] has concluded that patients with 1–3 cerebral metastases would be best treated (as judged by functional autonomy at 6 months after therapy or improved survival in the single brain metastatic cases) by WBRT and radiosurgical boost. The foregoing goes some way to explain the higher incidence of brain metastases patients in our recent cohort of treated patients (Figure 1); a trend we expect to continue.
|
Research into reactions to therapy
|
|---|
With increasing numbers of patients treated, we have observed more early and late reactions. It has been well established that in the first 24+ hours after radiosurgery for a AVM, particularly when located in the subcortex, there is an increased tendency to fit, owing to a presumed acute vasogenic reaction (comparable with primary erythema in skin). What we have once observed is a patient (who had suffered no fit for a long time) actually having such a fit during the treatment, causing cessation of therapy and immediate scanning demonstrated a vasogenic, oedematous response that had not been present on a comparable scan the previous week – representing a hyperacute reaction that we had not previously encountered [47]. We have also reported facial spasm complicating radiosurgery for acoustic neuroma and drawn attention to this unusual "stimulatory" reaction to radiation therapy [48]. Neither of these reactions fit comfortably into the classification that we published in 1999 [49], both being stimulatory in nature, and demonstrate the interesting new spectrum of radiation reactions that are being observed in this developing field of radiation therapy.
We have published one study regarding the possibility of modifying the post therapy radiation reaction. In our early linac based series of AVM treatments, using a single isocentre and rotational technology, we took on a quantity of large AVM cases, accepting a higher risk than is associated with treatment of smaller AVM. Following experimental animal evidence that essential fatty acids could modify that development of post-radiation transverse myelitis in pigs [50, 51], we introduced "post hoc" administration of gamma linolenic acid after radiosurgery for larger AVM. In the study that we reported in 2001 [52], any form of complication after radiosurgery for large (more than 10 cm3) AVM fell from 20% to nil with the introduction of gamma linolenic administration. However, the obliteration rate of these AVM fell from 41% to 5%, suggesting to us that not only was the post hoc administration of this lipid altering the radiation response of the normal nervous system but also that it was altering it in the AVM. Now, AVM are an unusual target in radiation medicine in that we are here relying, to cure the AVM, on a late radiation reaction that is shared by normal tissue. Although our study did not achieve the desired objective of improving the therapeutic ratio for this target, we argued that the situation is very different in the treatment of neoplasms; here the kill of cancer cells occurs near instantaneously after radiation as distinct from the late radiation reaction that governs the late morbidity of the nervous system. Thus, that gamma linolenic acid works post hoc might well allow an improved therapeutic ratio in the therapy of brain tumours and further research is required [51].
|
Conclusions
|
|---|
Radiosurgery is a fast expanding field with the technique replacing some conventional neurosurgical procedures and allowing effective therapy where previously there was none. This manuscript is the third quinquennial review of work in the subject at St. Bartholomew's Hospital, London.
Received for publication September 14, 2004.
Revision received November 19, 2004.
Accepted for publication December 6, 2004.
|
References
|
|---|
- Leksell L. A stereotaxic apparatus for intracerebral surgery. Acta Chir Scand 1949;99:231.
- Leksell L. Cerebral radiosurgery I. Gamma thalamotomy in two cases of intractable pain. Acta Chir Scand 1968;134:585–95.[Medline]
- Walton L, Bomford CK, Ramsden D. The Sheffield stereotactic radiosurgery unit: physical characteristics and principles of operation. Br J Radiol 1987;60:897–906.[Abstract]
- Gerhard A, Horstmann MD, Shogens H, Albertus TCJ, Van Eck MD, Kreiner HJ, et al. First clinical experience with automatic position system and Leksell Gamma Knife Model C. J Neurosurg 2000;93(Suppl. 3):193–7.
- Levivier M, Wikier D, Goldman S, David P, Metens T, Massager N, et al. Integration of the metabolic data of positron emission tomography in the dosimetry planning of radiosurgery with the Gamma Knife: early experience with brain tumours. J Neurosurg 2000;93(Suppl. 3):233–8.
- Sims E, Doughty D, Macaulay E, Royle N, Wraith C, Darlison R, et al. Stereotactically delivered cranial radiation therapy: a ten year experience of linac based radiosurgery in the UK. Clin Oncol 1999;11:303–20.[CrossRef]
- St. George EJ, Plowman PN. The role of Gamma Knife and PET in neuro-oncology. CME Cancer Med 2002;1:3–8.
- Plowman PN. Pituitary adenoma radiotherapy – when, who and how? Clin Endocrinol 1999;51:265–71.[CrossRef][Medline]
- Landolt AM, Haller D, Lomax N, Scheib S, Schubiger O, Siegfried J, et al. Stereotactic radiosurgery for recurrent surgically treated acromegaly: comparison with fractionated radiotherapy. J Neurosurg 1998;88:1002–8.[Medline]
- Kondziolka D, Patel A, Lunsford D, et al. Stereotactic radiosurgery plus whole brain radiotherapy versus whole brain radiotherapy alone for patients with multiple brain metastases. Int J Radiat Oncol Biol Phys 1999;45:427–34.[Medline]
- Adams EJ, Cosgrove VP, Shepherd SF, et al. Comparison of a multileaf collimator with conformal blocks for the delivery of stereotactically guided conformal radiotherapy. Radiother Oncol 1999;51:205–9.[CrossRef][Medline]
- Plowman PN, Doughty D. Stereotactic radiosurgery X. Clinical isodosimetry of gamma knife versus linear accelerator X-knife for pituitary and acoustic tumours. Clin Oncol 1999;11:321–9.[CrossRef]
- Monk JE, Perks JR, Doughty D, Plowman PN. Comparison of a micro-multileaf collimator with a 5mm leaf width collimator for intracranial stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2003;57:1443–9.[CrossRef][Medline]
- Perks JL, St. George EJ, El Hamri K, Blackburn P, Plowman PN. Stereotactic radiosurgery XVI. Isodosimetric comparison of Gamma Knife versus linear accelerator micro-leaf MLC stereotactic radiotherapy methods for the therapy of acoustic neuroma – potential clinical importance. Int J Radiat Oncol Biol Phys 2003;57:1450–9.[CrossRef][Medline]
- El Hamri AK, Blackburn P, Perks JP, Plowman PN. Relationship of conformity index (CI) and minimum target dose (min TD) for two stereotactic treatment delivery systems. Proc Eur Gamma Knife Soc (Lugo, Italy. Dec 2003) Abstr.
- Perks J, St. George EJ, Doughty D, Plowamn PN. Is distortion correction necessary for digital subtraction angiography in the Gamma Knife treatment of intracranial arteriovenous malformations. Stereotact Funct Neurosurg 2001;76:94–105.[CrossRef][Medline]
- St George EJ, Butler P, Plowman PN. Can magnetic resonance imaging alone accurately define the arteriovenous nidus for gamma knife radiosurgery? J Neurosurg 2002;97(Suppl. 5):464–70.
- Guo WY, Nordell B, Karlsson B, et al. Target delineation in radiosurgery for cerebral arteriovenous malformations. Acta radiol 1993;34:457–63.[Medline]
- Plowman PN. Radiation therapy for vascular malformations – a changing scene. Clin Oncol 2002;14:91–6.[CrossRef]
- St George J, Perks J, Plowman PN. Stereotactic radiosurgery XV. The role of the haemosiderin ring in the development of adverse reactions following radiosurgery for intracranial cavernous malformations - a hypothesis. Br J Neurosurg 2002;16:385–91.[CrossRef][Medline]
- Madreperla SA, Hungerford JL, Plowman PN. Choroidal haemangiomas: visual and anatomic results of treatment by photocoagulation or radiotherapy. Ophthalmology 1997;104:1773–9.[Medline]
- Plowman PN, Hungerford JL. Radiotherapy for ocular angioma. Br J Ophthalmol 1997;81:258–9.[Free Full Text]
- Nelson JM, Stevens RG. Ferritin ion increases killing of Chinese hamster ovary cells by x-irradiation. Cell Prolif 1992;25:579–85.[Medline]
- Hornsey S, Myers R, Jenkinson T. The reduction of radiation damage to the spinal cord by post-irradiation administration of vaso-active drugs. Int J Radiat Oncol Biol Phys 1990;18:1437–42.[Medline]
- Kondziolka D, Lunsford D, McLaughlin MR, et al. Long term outcomes after radiosurgery for acoustic neuroma. N Engl J Med 1998;339:1426–33.[Abstract/Free Full Text]
- Fuss M, Debus J, Lohr F, et al. Conventionally fractionated stereotactic radiotherapy (FSRT) for acoustic neuromas. Int J Radiat Oncol Biol Phys 2000;48:1381–7.[CrossRef][Medline]
- Plowman PN. Post-radiation sensorineuronal hearing loss. Int J Radiat Oncol Biol Phys 2002;52:589–91.[CrossRef][Medline]
- Rowlands RG, Plowman PN, Millard R, Wareing M. Recovery of hearing in a dead ear following stereotactic radiosurgery for vestibular schwannoma. Proc Fourth International conference on vestibular schwannoma and other CPA lesions. (Cambridge UK, July 2003). Abstr.
- Baser ME, Evans DGR, Jackler RK, Sujansky E, Rubinstein A. Malignant nervous system tumours, radiotherapy and neurofibromatosis. Br J Cancer 2000;82:998–1002.
- Plowman PN, Evans DGR. Stereotactic radiosurgery XI. Acoustic neuroma therapy and radiation oncogenesis. Br J Neurosurg 2000;14:93–5.[CrossRef][Medline]
- Knudson AG. Mutation and cancer: a statistical study of retinoblastoma. Proc Natl Acad Sci 1971;68:820–3.[Abstract/Free Full Text]
- Blackburn TPD, Perks JR, Plowman PN. Is there a predisposition to radiation oncogenesis in NF-2 patients and is the situation comparable to hereditary retinoblastoma. Proc Europ Gamma Knife Soc Mtg. Nov 2002 (London), Abstr.
- Rowe JG, Radatz M, Walton L, Kemeny AA. Stereotactic radiosurgery for Type 2 neurofibromatosis acoustic neuromas: patient selection and tumour size. Stereotact Funct Neurosurg 2002;79:107–16.[CrossRef][Medline]
- Rowe JG, Radatz MWR, Walton L, Soanes T, Rodgers J, Kemeny AA. Clinical experience with Gamma knife stereotactic radiosurgery in the management of vestibular schwannomas secondary to type 2 neurofibromatosis. J Neurol Neurosurg Psych 2003;74:1288–93.[Abstract/Free Full Text]
- Swords FM, Allan CA, Plowman PN, Evanson J, Chew SL, Grossman AB, et al. Steretactic radiosurgery XVI: a treatment for previously irradiated pituitary adenomas. J Clin Endocrinol Metab 2003;88:5334–40.[Abstract/Free Full Text]
- Plowman PN. Pituitary adenoma radiotherapy. Clin Endocrinol 1999;51:265–7.
- Storr HL, Plowman PN, Carroll PV, Krassas GE, Afshar F, Besser GM, et al. Clinical and endocrine responses to pituitary radiotherapy in paediatric Cushing's disease: an effective second-line treatment. J Clin Endocrinol Metab 2003;88:34–7.[Abstract/Free Full Text]
- Degerblad M, Rahn T, Bergstrand G, Thoren M. Long term results of stereotactic radiosurgery to the pituitary gland in Cushing's disease. Acta Endocrinologica 1986;112:310–4.[Abstract/Free Full Text]
- Ulfarrson E, Roberts M, Rahn T, Linquist M, Thoren M, Lindquist C. Gamma Knife surgery for craniopharyngiomas: long term results of the first patients. Proc 11th Annual Meeting of the Leksell Gamma Knife Society, Prague, May 2002.
- Jackson ASN, St. George EJU, Hayward RJ, Plowman PN. Stereotactic radiosurgery XVII. Recurrent intrasellar craniopharyngioma. Br J Neurosurg 2003;17:138–43.[CrossRef][Medline]
- Blackburn TPD, Doughty D, Plowman PN. Stereotactic intracavity therapy of recurrent cystic craniopharyngioma by instillation of 90-Yttrium. Br J Neurosurg 1999;13:359–65.[CrossRef][Medline]
- Tsao N, Sultanem K, Chiu D, Copps F, Dixon P, Easton D, et al. Supportive care management of brain metastases: what is known and what we need to know. Clin Oncol 2003;15:429–34.[CrossRef]
- Gaspar L, Scott C, Rotman M, et al. Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastasis trials. Int J Radiat Oncol Biol Phys 1997;37:745–51.[CrossRef][Medline]
- Marcou Y, Lindquist C, Adams C, Retsas S, Plowman PN. What is optimal therapy of brain metastases? Clin Oncol 2001;13:105–11.
- Shaw EG. Radiotherapeutic management of multiple brain metastases: 3000 in 10 whole brain is no longer a no brainer. Int J Radiat Oncol Biol Phys 1999;45:253–4.[Medline]
- Andrews DW, Scott CB, Sperduto PW, et al. Whole brain radiation therapy with or without stereotactic radiosurgery boost for patients with one to three brain metastases: phase 111 results of RTOG 9508 randomised trial. Lancet 2004;363:1665–72.[CrossRef][Medline]
- Blackburn TPD, El Hanmri K, Boxhall YL, Barett S, Plowman PN, Bullock P. Acute cerebral oedema during radiosurgery – natural history or complication? Proc Eur Gamma Knife meeting (London 2002 – Abstr.).
- Blackburn TPD, Jeanson De Damoiseau KM, Sabin HI, Plowman PN. Facial spasm complicating stereotactic radiosurgery/therapy for acoustic schwannoma. World Gamma Knife Users Meeting (Vienna, 2004) Abstr.
- Plowman PN. Stereotactic radiosurgery VIII. The classification of post-radiosurgery reactions. Br J Neurosurg 1999;13:256–64.[CrossRef][Medline]
- Hopewell JW, Van den Aardweg GJM, Morris GM. Unsaturated lipids as modulators of radiation damage in normal tissues. In: Horobin DF, editor. New approaches to cancer treatment: unsaturated lipids and photodynamic therapy. London: Churchill Communication Europe, 1994;88–106.
- El Agamawi AY, Hopewell JW, Plowman PN, Rezvani M, Wilding D. Modulation of normal tissue response to radiation. Br J Radiol 1996;69:374–5.[Medline]
- Sims EC, Plowman PN. Stereotactic radiosurgery XII. Large AVM and the failure of the radiation response modifier gamma linolenic acid to improve the therapeutic ratio. Br J Neurosurg 2001;15:28–34.[CrossRef][Medline]
Top
Introduction and "case mix"
Physics research
