(GRC, NTZ), and Department of Neurosurgery, Faculty Of Medicine, Alexandria University, Egypt (OSA).

Address for Correspondence:
Emad N. Eskandar, M.D.
Massachusetts General Hospital
15 Parkman St. ACC # 331
Boston, MA 02114

E-mail: eeskandar@partners.org
Patient Appointments: 617.724.6590
FAX: 617.724.0339

Referrals | Stereotactic Surgery | Parkinson's Disease | Intractable Epilepsy | Movement Disorder Surgery
Guestbook | Selected Publications | Links

Intracerebral metastases are common and occur in up to 15% of cancer patients (14). The relatively well-circumscribed, non-invasive nature of brain metastases, as compared with the diffuse, infiltrating primary brain tumors, makes them appropriate candidates for focal therapy (12,29). Focal irradiation has been used successfully for both newly diagnosed and recurrent brain metastases and can administer a lethal tumor dose while minimizing exposure to normal brain (2,6,17,20,22,31).

Stereotactic radiosurgery is " the non-invasive delivery of a high dose single fraction of ionizing radiation to a limited, stereotactically-defined target volumes" (20,25). Interstitial brachytherapy provides focal intratumoral irradiation over a period of hours or days (27). Fractionated stereotactic radiotherapy combines the conformal three- dimensional dose distribution with the biologic advantage of fractionation (33,37). In all these therapeutic modalities, the objective is to deliver a high dose of radiation to a discrete volume with minimal exposure to surrounding normal brain structures.

Interstitial radiosurgery implies the delivery of ionizing radiation interstitially to the tumor (as in brachytherapy) in a high dose single fraction (as in radiosurgery) and hence the name. In this chapter, we present our experience using a battery-powered, miniature x-ray generator that can be placed stereotactically into intracranial tumors to deliver a single fraction of high dose interstitial irradiation in less than an hour. The physical characteristics of the Photon Radiosurgical System

(PRS) will be described and the preliminary clinical experience as well as possible future clinical applications will be discussed.

Conventional irradiation uses dose fractionation over time to take advantage of the differences in radiosensitivity between neoplastic and normal tissues. Dose fractionation allows repair of sublethal damage in normal tissues, which therefore receive a lesser radiobiological effect and are relatively spared. Neoplastic tissues receive a relatively greater radiobiological effect with fractionation because of reoxygenation, cell cycle redistribution and repopulation of tumor cells (20,25,33).

Stereotactic radiosurgery employs accurate target localization and steep dose gradients to deliver a single fraction of high dose irradiation. Radiosurgery is safe and efficacious if the target volume is kept as small as possible and the volume of normal tissue irradiated to high doses is minimized (20,25).

The major radiobiologic advantage of interstitial brachytherapy is due to intratumoral placement of the radioactive isotope. The dose is attenuated in tissue and a dose decline rate proportional to 1/r2 yields sharp dose fall-off, thus allowing a high dose to be delivered to the tumor while minimizing exposure to the surrounding normal brain (7). Additional advantages of conventional low dose rate interstitial irradiation (0.5-2.0 cGy/min) over days or weeks include tumor cell reoxygenation and repopulation, a lower oxygen enhancement ratio and cell cycle redistribution with resultant accumulation of tumor cells in a sensitive phase of the cell cycle during the course of irradiation (7,27,34).

The total irradiation dose required to produce a given biologic effect in tissue depends on the time period over which the tissue is irradiated (27). The dose rate is a major determinant of the biological effect of a given dose of ionizing radiation. In general, lowering the dose rate (and increasing the exposure time), reduces the biological effect of a given dose of radiation (7). Dose rates of approximately 1-2 Gy/min with the PRS, are possible and can reach up to 500 Gy/min at the tumor center (15). In addition, the irradiation dose rate varies across the target volume according to the radiation technique used. The dose rate of conventional radiotherapy varies by about 5% to 10%, whereas that of external beam radiosurgery or the PRS may vary by 50% across the target volume. In each technique, the normal tissue away from the target will receive the minimum dose rate, but a full dose rate will be delivered to any normal tissue lying within the target volume (27).

The radiobiology of the interstitial radiosurgery is probably similar to external radiosurgery; that is, both deliver a high dose in a single fraction over a period of perhaps 15 minutes to an hour. Interstitial brachytherapy using standard implanted radionuclides, generally requires days to weeks for completion of the treatment (27). The PRS combines the short treatment times of radiosurgery with the dose delivery advantages of an interstitial source.

The Photon Radiosurgical System (Photoelectron Corporation, Waltham, MA) is a new, battery-powered miniature x-ray generator capable of delivering a prescribed therapeutic radiation dose directly to small brain lesions (15).

The device itself weighs 3.8 pounds and is designed to be compatible with current stereotactic frames (Fig. 1). It contains a step-up converter that can amplify the 9.6-V supplied by the battery to 40 KV and an internal electron gun which creates a 40 (A electron beam (approximately 0.5 mm wide). The beam is accelerated through a high-voltage field (range 15-40 KV in 5 KV increments) and then passes through a deflection chamber to control beam position and thus assure beam straightness. After traveling down the evacuated, magnetically shielded, rigid probe (3 mm in diameter and 100 mm in length), the electron beam strikes a thin gold foil target (0.5 ( m) at the probe tip producing x-ray photons whose effective energies are in the 10-20 KeV range. The gold foil is thick enough to stop the electrons but thin enough to allow the x-rays thus generated to pass through. The last 20 mm of the probe tip is constructed from beryllium, which is transparent to these low energy x-ray photons. The x-rays are emitted from the tip in a spherical symmetrical pattern resulting in a dose rate in tissue of up to 120 Gy per hour at a 10 mm radius. Two scintillation counters monitor radiation and are positioned on the stereotactic frame to detect the small number of photons that pass through the skull and compare these to the expected treatment levels. Although 99.9% of the energy created by the electron collision with the gold foil is in the form of heat and only 0.1% is generated as x-rays, hyperthermia has been excluded as a possible tumoricidal factor by limiting the power of the electron beam to maintain the temperature at the tumor margin to less than 420C. Operation of the system is directed from a low-voltage electronic control box that contains a 9.6-V rechargeable nickel-cadmium battery as the power source. A cable from the control box provides the low voltage electrical power to the unit. Battery operation and packing of all high-voltage components within a grounded housing result in a compact device with total electrical isolation for the safety of the patient (5,10,15,16,36).

The Photoelectron Device produces a radiation field similar to that of a conventional high dose rate interstitial brachytherapy source (28), but has the advantage of electronic control. The probe tip delivers point-source emission of low-energy photons directly into the lesion, with adjustable intensity and peak energies (15).

The x-ray beam behaves essentially as a point isotropic source. Ionization chamber measurements, determined that the dose rate in water for the PRS x-ray beam with a current of 40 (A and voltage of 40 KV is 15 Gy/min at a distance of 10 mm. The absolute dose is estimated to be (10% (5,16).

Because the photons produced by the PRS are low in energy, their absorption characteristics are different from standard brachytherapy sources. The "soft x-rays" of the device are attenuated rapidly within tissue and a dose decline rate of (1/r3) is obtained rather than (1/r2) as seen for standard higher energy interstitial radioactive sources. The resultant 30% dose reduction per millimeter of unit tissue creates an extremely steep dose fall off (Fig. 2). Background exposure to personnel more than 2 meters away from the probe when inserted in a patient has been measured at approximately 5-10 mrem/hour and therefore no special shielding of the patient or health care personnel is required (5,11,27).

To evaluate the dosimetry provided by the PRS with other currently available technologies, dose-volume histograms generated for the PRS, linac radiosurgery, proton beam radiation therapy were compared (16). All three treatment plans prescribed a dose to a 2 cm diameter spherical lesion, using a three-dimensional treatment planing system. The PRS plan used a single probe at the center of the lesion. The linac radiosurgery plan utilized a 6 MV linear accelerator with four arcs. The proton treatment plan used three circular beam portals. The linac and proton treatment plans were prescribed so that the 80% and 90% isodose lines, respectively, covered the treatment volume. The PRS plan prescribed 100% of the dose to the tumor volume, plus a 2 mm margin of presumed normal tissue beyond the defined tumor border (16). Comparison of dose-volume histograms from the proton beam, linac, and PRS treatment plans (Fig. 3) demonstrated relative normal tissue sparing with the PRS plan beyond that provided with the proton or linac radiosurgery plans for treatment volumes less than 50 cc. The high dose region in the PRS dose curve is due to the point-source nature of the PRS treatment with its dependence on treatment radius.

The dosimetry provided with the PRS treatment plan was similar to, and perhaps better, than that achieved with proton beam or linac three- dimensional treatment planning for small spherical intracranial lesions. However, one might not expect the dosimetric advantages of the PRS treatment plan to disappear, if the treated tumor targets were larger in size and/or more irregular in shape (16).

The PRS is designed for use with stereotactic frames, and initial experience has been acquired using the CRW system (Radionics Inc, Burlington, MA). After fixation of the frame to the patient, a stereotactic CT scan with contrast is obtained using 1.5-mm contiguous slices through the tumor. The antero-posterior, lateral, and vertical dimensions of the tumor are determined and target coordinates are calculated for the center of the lesion. If appropriate, a standard stereotactic biopsy is then performed through a burr hole and specimens from the target were submitted for pathological analysis.

If intra-operative frozen section analysis confirms the diagnosis of tumor, then the irradiation treatment can be instituted at the same sitting. First, the biopsy needle tract is dilated slightly to accommodate the probe tip (3-mm diameter) using a graduated series of dilators. The PRS device is then mounted on the arc of the CRW stereotactic apparatus and advanced along the biopsy tract to the target coordinates (Fig. 4). In general, lesions less than 2.0 cm in diameter are prescribed 18 Gy, and lesions greater than 3.0 cm are prescribed 15 Gy. The exact lesion dimensions obtained from the contrast-enhanced CT are combined with dose/depth and isodose curves of the PRS to determine the optimal radiation treatment plan. The probe tip position, beam voltage, and current are used to calculate the duration of treatment required to administer the prescribed dose to the periphery of the tumor with a 2-mm margin. With all components connected to the control box, the prescribed voltage and current parameters are selected, and the timer is set to the calculated treatment time. The device is activated by starting the timer and automatically terminates after completion of the selected treatment time. The PRS probe is then removed, and the incision is closed in standard fashion. The procedure is carried out under local anesthesia with intravenous sedation to minimize patient discomfort (10,11,16,27).

Interstitial radiosurgery using the PRS has been utilized at the Massachusetts General Hospital since December 1992. Case selection has been limited to those patients with either primary or secondary malignant brain tumors who were not considered suitable for conventional surgical treatment. Lesions had to be 3 cm or less in greatest dimension and supratentorial in location. All patients had to be in good general health with Karnofsky Performance Score (KPS) rating of 70% or greater and with an expected survival of more than 4 months. Brain stem or cerebellar lesions and tumors with significant cyst formation, hemorrhage, or calcification were excluded.

All patients underwent a detailed clinical evaluation, KPS rating and contrast-enhanced computarized tomography (CT) scan or magnetic resonance imaging (MRI) prior to treatment. An immediate post-treatment CT scan was obtained within 24 hours to exclude peri-operative complications. All patients were asked to return at regular intervals for follow-up clinical evaluation and imaging studies to document any changes in lesion size. Tumor response was graded independently by a neuroradiologist with direct measurement of the maximal tumor dimensions as seen on the last post-treatment MRI study and compared to pre-treatment studies. Reduction or stabilization of the tumor size was accepted as local control whereas enlargement of the tumor indicated local failure. Clinical and radiological outcomes were measured at the last follow-up visit or at the time of death. If the tumor showed signs of progression at any time following treatment, alternative methods of treatment were used as clinically necessary. Patients with metastases who had not received whole brain radiation therapy (WBRT) prior to interstitial radiosurgery were given 30 Gy in 10 fractions after treatment.

To date, 27 patients have undergone interstitial radiosurgery with the PRS. This includes 25 patients with cerebral metastases ( sixteen men, eleven women; ages 37 - 80 years, mean 56 years ) of whom twelve had solitary lesions and thirteen who had multiple (from 2 to 7) lesions. The primary source of the cerebral metastases was lung cancer in 16, malignant melanoma in 6, renal cell carcinoma in 2, and Merkle cell carcinoma in one. Lesions ranged in size from 4-30 mm (mean 16.8 mm) in greatest diameter and were located in various lobes. Treatment diameter (lesion diameter + 2 mm margin) ranged from 8-34 mm (mean 21.4 mm). Single doses of radiation between 10 and 20 Gy (mean 15.8 Gy) were administered, with an average treatment time of 19.5 minutes (range 4-75 minutes). Fourteen patients with cerebral metastases had received prior WBRT while the remainder received 30 Gy in 10 fractions within 2 weeks of PRS treatment.

Two patients were felt to have unresectable primary brain tumors. One patient was diagnosed with a primary CNS lymphoma while the other was believed to have a malignant glioma based on clinical presentation, radiological findings and intraoperative pathological analysis. Subsequent pathological evaluation suggested that the abnormal glial cells and necrosis were not due to tumor but were probably reactive astrocytes adjacent to an area of ischemic necrosis. Only necrotic tissue was irradiated and the patient has suffered no adverse effects from treatment.

All patients tolerated the procedure well and most patients were discharged home the day after treatment. Follow up has ranged from 10 days to more than 30 months (mean 7 months). Two patients experienced isolated focal motor seizures within 24-48 hours of surgery and required anti-convulsants. Two patients developed brain edema that resolved on corticosteroids and diuretics. One patient developed a subdural hematoma, three months after treatment but it was on the side opposite the treated lesion. There were no deaths, infections, or serious neurological morbidity.

Local control of the lesion (defined as stabilization or reduction of tumor size) was obtained in 19 (73%) of the 26 patients with tumors. Tumor progression was observed at 3 months post-PRS treatment in 5 patients, at 6 months in one patient, and at 10 months in one patient. Surgical resection was performed in two patients with enlarging tumors at 4 and 10 months post-PRS treatment and pathological analysis demonstrated central necrosis along with a thin rim of viable tumor around the periphery of the lesion. Two patients had post-mortem examination. One patient with malignant melanoma who died twelve months after treatment, documented absence of tumor cells in the treated lesion. The second patient with adenocarcinoma of lung died twenty months after treatment and demonstrated no tumor cells in the treated frontal lesion, but rare viable tumor cells were detected in the treated temporal lesion.

At last follow-up, 19 patients with cerebral metastases have died; 13 from systemic disease (between 12 days and 31 months after treatment), 5 from distant central nervous system (CNS) metastases (between 5.5 and 12 months after treatment), and one patient 10 days after percutaneous gastrostomy for recurrent aspiration pneumonia. Local control of the treated tumors was documented in all 5 patients who died from progressive CNS metastases and in 9 patients of those who died from systemic disease, while local failure was evident in the others. Six patients with cerebral metastases are alive and well. Median survival was 13 months.

The goal of any radiosurgical procedure is to accurately deliver a large single fraction of radiation to a small target volume while minimizing exposure to surrounding tissue (23). With external radiosurgical systems, the desired dose/volume distribution is obtained when the focused beams intersect and summate at the target point after traversing cranial tissue from varied angles (11). Conventional interstitial brachytherapy obtains its desired dose/volume distribution by placing a radiation source directly in the tumor. Dose rates for conventional implanted radionuclides, are in the range of 0.5 to 2.0 cGy/min and the treatment takes many hours to several days (28).The dose rate during radiosurgery, either with a linear accelerator (linac), or gamma knife is typically several hundred cGy/min with treatment times averaging 15 to 60 minutes (27).

The PRS has operational characteristics that combine the short treatment times of external beam radiosurgery with the radiobiological advantages of interstitial brachytherapy (10). The major radiobiological advantage of interstitial brachytherapy is related to intratumoral placement of the radioactive isotope. The emitted dose is attenuated in tissue with increasing distance(1/r2) from the source which results in maximal delivery of radiation to the tumor with relative sparing of the surrounding normal brain (7). Of two commonly used isotopes, I125 generates low-energy photons (28-35 KeV) but has a very low dose rate typically requiring treatment times of several days. Radioisotopes with very low dose rates may require permanent implantation (10,11). Ir192 emits high-energy photons (300-610 KeV) and higher dose rates that can shorten treatment times but has less sharply-defined dosimetry. These conventional radioactive isotopes have been effective in the treatment of certain brain tumors but do decay in activity during storage and require special facilities and procedures for their safe handling and disposal. Protective shielding of health personnel must also be considered. Compared with interstitial brachytherapy using standard implanted radionuclides, the PRS treatment has significant advantages because of the adjustable dose rate and steeper dose gradient it provides and hence less dose to normal tissue outside the target. Also, PRS treatment eliminates the need to handle radioactive sources (5).

Unlike external radiosurgical systems, which require very large and expensive facilities, the portable PRS provides a unique cost effective opportunity for providing interstitial radiation therapy immediately following histologic confirmation of malignancy in patients undergoing biopsy of intracranial lesions (16). Moreover, the dosimetry provided by the PRS is similar to, and perhaps better, than that achieved with proton beam or linac three-dimensional treatment planning for small spherical intracranial lesions (16).

An obvious difference between interstitial approaches ( brachytherapy and PRS) and other non-invasive radiosurgical techniques (gamma knife, linac ,or proton beam therapy) is that with the former, a catheter(s) or probe tip must be placed into the tumor. This imparts a small but definite risk of hemorrhage, injury to the surrounding brain, or infection. Given the need for tissue diagnosis prior to many forms of radiation treatment, however, the added risk of the treatment procedure can be minimized by inserting the probe at the time of the stereotactic biopsy. The PRS should not impart any more risk than conventional interstitial procedures and since the procedure is performed over a very short time period, the risk of infection is less (11,27).

Potential inaccuracies involving target selection and volume determination exist with any radiosurgical system, but the extremely sharp dose distribution curves of this device require targeting accuracy. Opening of the cranial vault with shift of intracranial structures may induce additional targeting error. Placement of the probe itself could displace the tumor volume from its imaged position, and may increase the inaccuracy related to dose distribution. The same problem of mechanical perturbation of tissues exist with interstitial brachytherapy using temporary I125 implants. The ability, however, to check the actual position of the implanted sources with post-operative radiographs, and to modify the source positions by placing or adjusting the catheters positions, may be one potential advantage of the I125 catheter implant system (10,27).

The spherical and non-infiltrative nature of cerebral metastases makes them ideal candidates for interstitial brachytherapy and radiosurgical treatment (20,27). Whole brain radiation therapy (WBRT) in patients with cerebral metastases extend median survival time to between 3 and 6 months depending on the type of primary cancer, extent of systemic disease, and a variety of other factors, but overall prognosis is poor (8). Patients with solitary cerebral metastases who undergo surgical excision plus WBRT experience improved overall survival (10-12 months) and better local control rates as compared to WBRT alone (32,35,38). Resection of multiple metastases may also lead to improved survival in those patients with controlled systemic disease, but resective surgery necessitates hospitalization and operative risk which could be minimized by less invasive techniques. Palliation of symptoms and improvement in long-term survival has already been demonstrated in patients with both new and recurrent solitary brain metastases using I125 implant (31). If the PRS device is similarly effective, it could represent an attractive alternative to conventional surgery in this population of patients (10,11).

Treatment that provides local control of single, multiple, or recurrent brain metastases imparts clinical benefit in terms of quality of life and survival (3,24,32). Patchell et al. demonstrated that the combination of surgery and radiotherapy for solitary brain metastases improved local control from 48% to 80% and increased median survival from 15 to 40 weeks (32). Alexander et al. demonstrated local control rates of 89% and median survival of 9.4 months in 248 radiosurgically treated patients with brain metastases (3). The results of radiosurgery are encouraging and suggest that radiosurgery can be an effective alternative to surgery in patients with single, multiple, or recurrent cerebral metastases (1,3,9,16,17,18).

Interstitial radiosurgery using the photoelectron device offers a safe alternative for the treatment of brain metastases. Results of treatment using the PRS technique, demonstrate a 73% local control rate and median survival of 13 months, which are within the range previously reported by others and appear to be comparable to those of other radiosurgical modalities.

When tissue diagnosis is required, the ability to perform stereotactic biopsy and interstitial radiosurgery at the same procedure could be advantageous and cost-effective. However, the PRS is unlikely to replace conventional radiosurgery devices such as gamma knife, linac, or proton beam especially with regard to treatment of patients with arteriovenous malformations or acoustic neurinomas (11).

Interstitial radiosurgery imparts less dose to non-target normal tissue when compared to other forms of radiosurgery. This reduction in dose at larger volumes may represent a clinical advantage for selected cases and may allow for treatment of metastases too large to be treated with external radiosurgery. It is not yet clear what role the photoelectron device will have in the treatment of primary brain tumors. Malignant and even low-grade gliomas are generally larger than 3-4 cm at the time of diagnosis and tend to infiltrate locally along white matter tracts. However, the device may be of use in small, deep gliomas or recurrences, and treatment plans incorporating multiple isocenters are feasible (10,11).

Perhaps its most useful role will be for intraoperative irradiation during open neurosurgical procedures in a fashion similar to that employed by other surgical specialities. The probe tip can be easily shielded to allow for unidirectional irradiation of resection cavities, tumor margins, dural attachments, or skull base lesions. Conversely, sensitive neural structures could be shielded by thin radio-opaque metals to prevent radiation injury during treatment. It may have even broader application in other surgical subspecialities including breast cancer surgery, urology, gynecology, orthopedics, and otolaryngology (10,11).

The PRS is a new device for the interstitial radiosurgical treatment of cerebral metastases and possibly other brain tumors. It offers a convenient and cost efficient method for providing interstitial radiation therapy in patients undergoing diagnostic biopsy of intracranial lesions.

When treating small, circumscribed, nearly spherical intracranial lesions, the PRS treatment appears to offer superior dosimetric advantages over linac, gamma knife, or proton beam therapy in sparing the surrounding normal brain tissue. Given its operational characteristics and dosimetry profile, it compares favorably to current external radiosurgical devices and has certain logistical advantages over standard interstitial brachytherapy.