- Email: [email protected]
Laser stereotaxy for malignant gliomas: 2001–2004
ARTICLE IN PRESS
Medical Laser Application 21 (2006) 123–130 www.elsevier.de/mla
Laser stereotaxy for malignant gliomas: 2001–2004 Frank Ulrich, Hans-Joachim Schwarzmaier, Wernholt von Tempelhoff, Hendrik Niehoff, Frank Eickmeyer Neurochirurgische Klinik, Klinikum Krefeld, Lutherplatz 40, 47805 Krefeld, Germany Received 13 January 2006; accepted 26 January 2006
Abstract Tumours of the central nervous system represent a therapeutic challenge. Although there is a multitude of histological subgroups, the main entities are of glioma origin. The majority of these lesions have an infiltrative nature, which is, in conjunction with the vicinity of highly specialized eloquent areas, highly demanding on the surgical skills and technologies applied. These tumours will invariably recur, most often even at the very site of the apparently ‘‘complete’’ resection. This has created local therapies for direct intra cavitary treatment of the surface and for interstitial treatment of the immediately adjacent tissue. Sixteen patients suffering from recurrent Glioblastoma multiforme (GBM) were treated by Nd:YAG laser irradiation in the framework of a salvage therapy. The underlying concept is the cytoreduction by partial coagulation of the tumour. MRI follow-up examinations revealed a volume reduction of the laser-irradiated areas, while the untreated parts exhibited a more or less progression. The survival time after the diagnosis of the recurrence was about four times longer than the natural history of the disease would suggest. The conclusion is that cytoreduction by laser irradiation might be a promising option for patients suffering from infiltrative gliomas. Future work should optimize the therapeutic regimes and evaluate this treatment approach in controlled clinical trials. In this article we will summarize the technical and clinical results induced by interstitial laser application in cerebral gliomas. r 2006 Elsevier GmbH. All rights reserved. Keywords: Brain tumour; Glioma; Interstitial thermotherapy; Laser induced; MRI guided therapy; Recurrent glioblastoma multiforme; Survival
Introduction Tumours of the central nervous system represent a therapeutic challenge. Although there is a multitude of histological subgroups, the main entities are of astrocytic and oligodendrocytic origin. The majority of these lesions have an infiltrative nature, which is, in conjuncCorresponding author. Tel.: +49 2151 321320; fax: +49 2151 322033. E-mail address: [email protected] (F. Ulrich).
1615-1615/$ - see front matter r 2006 Elsevier GmbH. All rights reserved. doi:10.1016/j.mla.2006.03.006
tion with the vicinity of highly specialized eloquent areas, highly demanding of the surgical skills and technologies applied. There are various approaches to the treatment of neuroepithelial tumours. Surgery is the first treatment option for primary and recurrent lesions. However, there are cases, where either surgery predictably harbours significant morbidity or general health issues preclude long procedures. In much the same way it is certain that due to the infiltrative nature of these lesions, a surgical cure is rarely achieved. The diversity as well as the quality of the tools at our disposal has
ARTICLE IN PRESS 124
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
improved significantly over recent years. It becomes more of a challenge to comprehend and correctly employ the different therapeutic modalities. Still it is sometimes discouraging that even with such a potent armamentarium, this grave disease entity takes an unstably lethal course. Presently our best achievement is to improve or at least maintain the patient’s quality of life for as long as possible and ameliorate symptoms. In these cases, radiation or interstitial therapy, by using stereotactic principles remains an alternative option. As for interstitial therapies, radioactive seed implantation as well as interstitial laser-induced thermotherapy enable minimal interventions, which allow pathologic diagnosis combined with local treatment [1].
Interstitial laser therapy using diffusor fibre tips Laser induced interstitial thermotherapy (LITT) in glioma is performed via flexible light guides. In general, fused silica fibres of 400 or 600 mm in diameter are used. These are characterized by a damping factor of a less than 1 dB/km (1064 nm). Thus, transmission losses due to the fibre material is almost negligible with light guides of a typical length between 10 and 12 m. This length is required if LITT is monitored by magnetic resonance (MR) imaging. In this case, the laser must be positioned outside the MR-room to avoid interactions with the magnetic field. Transmission losses are predominantly due to the Fresnel reflection at the air/fibre interfaces within the fibre coupler. The major goal of interstitial thermal therapy is the heating of a given tissue volume without adverse effects such as charring or shock waves due to vaporization at the probe tip. Consequently, the peak temperature at the probe tip must not exceed 100 1C. This limits the peak power energy per given time and probe surface area. On the other hand a minimal radiant power density is required to achieve a sufficient temperature rise for tissue denaturation. The high power
density at the conventional fibre tip often resulted in an overheating of the tissue surface with consecutive vaporization and carbonization. In this situation, no longer does tissue irradiation takes place and the heating process is mediated by thermal conduction only. The technique was overcome using optical diffusors; such diffusors emit laser light over the entire length of the fibre tip, significantly reducing the power density and the penetration depth is dependent on the optical tissue properties (absorption, scattering, blood content) at the wavelength used. These devices allow to couple between 5 and 8 W into the brain tissue without adverse effects. The resulting laser lesion is also dependent on the length of the optical diffusor (see Fig. 1). At a temperature above 60 1C, a direct thermal coagulation is induced resulting in an necrosis of 2–3 cm in radial diameter. In the areas beyond the induced necrosis, temperatures are lower, however, still above 38 1C. Thus, the total thermally treated area typically extends over 3 and 4 cm in radial diameter. The final lesion sizes, however, are also dependent on the exact optical properties of the irradiated tissue as well as its blood perfusion. For on-line control, temperature monitoring was performed using the experimental software package based on the temperature-dependent phase shift of the MR signal as mentioned above. The actual temperature distribution could be displayed as colour-coded images as well as time course of the hot spot temperature (see Fig. 2). Laser irradiation was ceased when temperature monitoring revealed a steadystate temperature profile within the heated tissue [2].
Image-guided laser fibre adaption LITT procedures in the open MR Signa SP/ 2 are dependent on MR-compatible and artifact-free instruments. A guiding device (NeuroGate, Daum, Schwerin, Germany) has been used in combination with the
Fig. 1-2. (1) Laser light guide in place after repositioning and insertion of the applicator (note the artifact of the light guide’s marker (magnetite) touching the brain tumor and the artifact of the diffusing tip within the tumor). (2) The actual temperature distribution is displayed as colour-coded, near real-time MR images.
ARTICLE IN PRESS F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
Flashpoint-System as an instrument for planning, guiding, and performing stereotactic laser procedures in the OMR. The device allows planning of a laser fibre trajectory from the surface of the patient’s head, and puts a marker on the skin for accurate burr hole positioning. It is a tripod with an adaptor to the 3 LED-guide of the Flashpoint-System and simulates the potential 350 range of angulation of the NeuroGate. Following a skin incision, a standard burr hole is made with a vertical bone wall and suitable dimensions for fixing the NeuroGate, using a special broach and grip. The central guide element is flexible and manoeuvrable after releasing the thumb screen with a special hexagonal screw driver. During surgery, placement of the biopsy cannula or sheath could be changed at any time and was supported by the LED tracking system and MR guidance [3]. Repeated nearly real-time alignment with the trajectory plane allows monitoring of the position of the stereotactic probe in relation to the targeted structure. If the image is a fast gradient echo scan, the delay can be as little as 1 sec. However, if longer imaging times are used for improved image quality, the delay can be a few minutes. When 3D T1-weighted images show the type of the laser light guide in the targeted region, the laser therapy is performed. The Signa SP can generate full 3D volumes within some 5 min in usual resolution and any selected 2D slices within some 5 s. These nearly real-time slices together with the built-in tracking system (Flashpoint, IGT, Boulder, CO, USA) establish the real-time capability of this scanner. Real-time MR images can be selected in any plane relative to the device to give real-time device control. However, it takes the iMR several seconds to generate such a single slice. Interactive work is a little bit disturbed by this delay. Additionally for technical reasons, real-time slices have significantly reduced resolution. As long as the patient does not move, volume and patient are exactly registered with the builtin tracking system. In soft tissue, the laser intervention usually renders preoperative images inappropriate if tissue shifts occur. In these cases, intraoperative images are necessary to show the current state, to update the preoperative data, and to register preoperative images to the current situation. The update is an iterative process triggered by situational changes during the ongoing intervention [4]. The aim of magnetic resonance imaging (MRI) in monitoring LITT is target definition, trajectory selection, light guide tracking, prediction of final lesion size during therapy, and follow-up of the temporal evolution of the laser-induced lesion. The open MR-guidance in its real-time/on-line mode allows the surgeon to update his target point in its new position after shifting and
125
displacement. The on-line mode provides the option of frameless stereotaxis by flashpoint-tracking. This is helpful in the puncture of cystic lesions or in positioning the laser fibre after biopsy for LITT in tumours – so that MR-guided LITT becomes a one-time and one-place procedure. The surgeon can visualize and react to loss of cyst fluid, CSF, or necrotic regions of the neoplasms. In the T1-mode with contrast medium, small lesions can be targeted and identified easily, choosing an optimal approach over a burr hole. The intraoperative MR-imaging is a welcome improvement of the classic, static neuronavigation, adding the important feature of a dynamic on-line control of laser-induced procedure. The latest updates and developments of the MR software allow fusions of preoperative and actual intraoperative MR-data, even in 3D mode. The zonal architecture of laser-induced lesions consists of a central and peripheral zone (see Fig. 3). In the central zone, there is a generalized damage of cellular and subcellular membranes observed resulting in irreversible tissue damage with secondary effects in the periphery. The intravascular red blood cells appear empty. In low-grade gliomas, high signal intensity of T2 weighted MR images of the central zone most likely represents heat-induced methaemoglobin conversion from desoxyhaemoglobin. On the other hand, highprotein-content fluid collections may also contribute to the high signal intensity of the central zone in low-grade gliomas. The low signal intensity rim of T2-weighted images most likely represents desoxyhaemaglobin in the acute stage and haemosiderin deposition in later follow-
Fig. 3. Postoperative imaging: recurrent glioblastoma multiforme 24 h after LITT [10].
ARTICLE IN PRESS 126
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
up studies. After Gd-DTPA, an enhancement in this zone was evident in all patients independently of a blood-brain barrier disruption prior to LITT. In tumours with Gd-DTPA enhancement in the MRI prior to LITT due to a disrupted blood-brain barrier (BBB), there was no enhancement within the laser-induced lesion after LITT. The perifocal edema is not apparent immediately after LITT. It evolves 1–3 days after LITT and regresses completely within less than a month, indicating that no persistent damage occurs within this zone. In the majority of the induced lesions, the longterm development is uniform with variations in the zonal architecture in two lesions. In all patients, the decrease of lesion size followed an exponential pattern with a half-life period of a few weeks. The shrinkage of the lesion is accompanied by a corresponding reduction of the size of the neoplasm. As the laser-induced enhancing rim persists over a long period, difficulties may arise in differentiating the residual laser lesion from recurrent tumour. The sequential time course of the laser-induced enhancing rim with an ongoing reduction of size may exclude a recurrent tumour that appears as increasing volume and new onset of mass effect. It has to be considered that the central and peripheral zones together form the total lesion size [5,6].
Neuropathological findings LITT is a minimally invasive approach for the treatment of gliomas, especially in poorly accessible regions. For the treatment of glioma by LITT, the extent of lethal laser–tissue interaction and changes in tumour cells due to LITT are of particular relevance for the clinical course and the outcome of patients with brain neoplasms. The primary goal of LITT is the destruction and removal of neoplastic cells and a decrease in tumour size. However, the survival of tumour cells at the margin of the surgical lesion after treatment has been shown to determine the postoperative course. A secondary outgrowth of residual neoplastic cells may form the basis of a recurrent tumour mass. For the prevention of tumour recurrences and, therefore, the outcome of patients after surgical interventions, alterations of the cell cycle and the life span of residual tumour cells have to be considered in addition to the primary destruction of the solid tumour mass. Following laser irradiation, the laser lesion shows in experimental studies a typical architecture dependent on the interval following laser treatment. The primary lesion exhibits differences in lesion diameter in dependence of the irradiation time and procedural energy deposition to the tissue but exhibits, in general, the typical architecture of a laser lesion that consists of a central coagulation necrosis and a surrounding rim of edema adjacent to undamaged tissue. Immediately after irradiation, the central zone of
necrosis is not yet apparent but is demarcated by a rim of edema from the undamaged tissue. The necrosis becomes evident by a gradual loss of staining and early resorptive changes at their margins by day 3. A rim of granulation tissue is formed after one week and the edema further spreads to the periphery. Finally, a cystic lesion with variable remnants of unresorbed necrotic tissue and a mesenchymal and glial reaction at the emerging remains. The immunohistochemical examination of tissue obtained from these lesions shows a distinct spacio-temporal staining pattern of several antibodies reactive to glial cells and neurons in relation to the different zones of the lesion. Non-specific and diffuse immunostaining can be found in the central necrotic zone only. In several studies, no qualitative difference could be found in the general histological architecture of the LITT lesion between animals with implanted tumour cells and those without. The zones of lesion are clearly defined as a central necrosis with a rim of edema. Interestingly, however, distinct changes can be seen in the size of the lesion in areas of normal tissue in comparison to neoplastic tissue. This might be due to different histological and, therefore, optical properties of these tissues and a resulting difference in the sensitivity to laser energy. In studies on human brain tissue, differences in absorption and scattering coefficients of the tissue defined by different neuroanatomical origin and ultrastructure could be demonstrated. Additionally, the optical properties of brain tissue are altered by coagulation as a result of laser–tissue interactions. Laser-induced coagulation leads to a tissue-specific increase in absorption and scattering with a higher vulnerability of the tissue to laser light. The specific appearance of an LITT lesion in the time course is mainly defined by resorptive changes by microglial cells or macrophages and a reactive gliosis. Immunohistochemical staining for the glial fibrillary acid protein (GFAP) clearly demonstrates an astrocyte activation and the extent of gliosis. Increased numbers of GFAPpositive reactive astrocytes are present after one day in the marginal zone. The number of reactive astrocytes affected by GFAP increases after 7 days. After week 2, the marginal zone of the lesion shows a distinct threelayered structure. The central necrotic zone is surrounded by a zone of edema containing an infiltration of granulocytes, lymphocytes, and macrophages. A layer of GFAP-positive reactive astrocytes separates this zone of edema from the adjacent brain tissue that appears normal. At the cellular level, different stages of astrocyte activation are detectable. One day following LITT, clumps of GFAP-positive cytoplasmatic material coat the unclear membrane. After 3 days, GFAP staining appears in short cytoplasmatic processes. The intensity of GFAP immunoreactivity in reactive astrocytes that surround the necrotic zone increases until the end
ARTICLE IN PRESS F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
of the 2nd week. Later, sprouting capillaries and GFAPpositive reactive astrocytes extend into the zone of necrosis. After 1 month, a thin layer of GFAP-positive astrocytes surrounds a cystic defect. Following laser treatment, blood vessels generally appear engorged, containing leached erythrocytes that lay closely opposed to each other with a staining of the membranes only. The tissue structure in general, however, is well preserved. The densely packed and decolourized red blood cells seem to be a characteristic histological marker of the intravascular laser effect. Vascular effects of laser irradiation include, furthermore, a thrombotic occlusion of the vessels inside the central laser lesion. These vascular structures are resorbed by invading mesenchymal cells during the course of regeneration. Vessels outside the margin of the laser lesion become the origin of outsprouting capillaries further invading into the lesion with a functional importance for regenerative and proliferative processes. Also, these vessels clearly contain normal and undamaged erythrocytes 3 days after laser irradiation, indicating a functional blood flow and the connection to the vascular tree. Fine structural analysis of the acute and chronic changes following laser treatment consistently shows damage of cellular and subcellular membranes in the central zone surrounding the laser tip. Cell membranes and unclear membranes of nerve cells, glial cells, and endothelial cells either exhibit local defects or are broken up into fragments. Also, subcellular membranes, in particular those of mitochondria, display structural damage. The curled fragments often fuse to form small vesicles, which may be a secondary heat effect. Vascular cells such as erythrocytes similarly show membrane defects leading to a loss of haemoglobin. Only basement membranes of capillaries and blood vessels seem to resist the laser treatment. In contrast, the investigation of the edema zone adjacent to the central necrosis displays no membrane ruptures. These observations indicate that membrane alterations may play a leading role in the pathogenesis of laser-induced lesions. Intravascular erythrolysis, a common phenomenon after laser irradiation, seems to be calmed by a leakage of haemoglobin through defects in the cellular membranes of erythrocytes and may lead to vascular occlusion with secondary tissue damage. The interstitial appearance of serum proteins such as albumin, which do not cross the blood-brain barrier and typically are not detectable in the interstitium under normal conditions, can be attributed to membranous defects of the capillary endothelium. The underlying laser–tissue interaction and the resulting histological changes are well defined and have been characterized as central coagulation necrosis and peripheral edema, subsequent resorptive changes and the formation of a rim of granulation tissue. In addition, various cellular changes of degen-
127
erative and regenerative nature have been found on a functional–morphological level, both in brain and in metastasis. Such changes can be correlated with MRI appearances under both experimental and clinical conditions [3,7–9].
Clinical results A total of 16 patients were enrolled into a representative study between May 2001 and June 2004. All patients suffered from a radiologically documented recurrence of a histologically confirmed glioblastoma grade IV WHO. All patients were non surgical candidates. In all patients, a first-line chemotherapy with temozolomide was initiated in case they did not have a respective prior treatment. A defined coagulation zone that shows a volume reduction over time as reported previously and already documented in low grade astrocytomas was exhibited. In this small group of patients, there was no inhospital lethality. No permanent neurological deficits were seen after the intervention. One patient exhibited a temporary weakness of the right arm. This is remarkable in so far as the tumours were predominantly located adjacent to functional relevant areas. Three patients developed a neutropenia and another patient exhibited a thrombocytopenia. One patient exhibited an increase of the transaminases and in one patient, a toe developed a local infection (panaritium). After LITT, the median overall survival was 6.971.7 months (95% CI 3.7–10.2). The overall survival after the diagnosis of the first relapse was 9.471.3 months (95% CI 6.8–12.0). There was, however, a substantial learning curve. During this learning period, (2001–2002), the median survival after LITT therapy was only 5.270.6 months (95% CI 4.1–6.3). Thereafter, the median survival after the first LITT increased to 11.272.0 (95% CI 7.4–15.0) months after LITT (Kaplan–Meier method, Software SPSS Version 12.0) (see Fig. 4). At the end of the study, 12 of the 16 patients were deceased. The cause of death was fatal venous thromboembolism in two cases. One patient died of a systemic mycosis, one patient from a gastrointestinal bleeding, and another due to a peritonitis after sigma perforation. The remaining deaths were due to a failure of the central regulation as typically seen in brain tumours. During the learning curve (2001–2002), the time between the diagnosis of the recurrence and the laser therapy was 2.0 months vs. 0.3 month thereafter. In addition, the KPS was lower and the tumour volume was larger in the first group of patients. It is of particular interest that five deaths were most probably more related to side effects of the concomitant chemotherapy and/or corticoid therapy, rather than to local tumour destruction [10].
ARTICLE IN PRESS 128
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
1.0
Survival function before 2003 2003 and after censored
Cumulative survival
0.8
0.6
0.4
0.2
0.0 0.00
5.00
10.00
15.00
20.00
Survival after LITT [months]
Fig. 4. Median survival: During learning period (2001–2002: 5.270.6 months) and median survival of the patients treated in 2003 and 2004: 11.272.0 months [10].
Discussion The laser-induced changes appear to be qualitatively equal, irrespective of the type of brain tumour, although differences dependent on the grade of malignancy may exist regarding the onset and size of recurrence. Primary effects of laser irradiation include a rupture of cell membranes and a fragmentation of subcellular components, resulting in irreversible tissue damage with secondary effect in the periphery. The laser–tissue interactions are followed by reactive and proliferative and BBB changes that determine in the clinical course. The BBB prevents the passage of plasma protein bound chemotherapeutica agents into the CNS. Capillaries in malignant gliomas have indeed imperfect interendothelial functions, endothelial perforations, and other ultrastructural abnormalities characteristic of neoplastic vasculature resulting in a disruption of BBB. Even though this disruption allows the diagnostically advantageous transcapillary passage of contrast material, leakage of plasma protein into gliomas and the adjacent brain does not suffice for an effective antineoplastic accumulation of chemotherapeutic agents. This is due to the fact that not all tumour capillaries are equally leaky and delivery of large molecule drugs to the tumour bed becomes thus uneven. Levin demonstrated that the BBB within the zone of infiltrations in the otherwise normal
brain adjacent to the tumour mass is unaffected. This is of particular importance, since this zone contains the leaking edge of neoplastic cells with a large fraction of cycling cells that have a high growth fraction. LITTinduced disruption of the BBB acilitates a locoregional passage of chemotherapeutic agents into the brain tissue. Of the cases of LITT, few patients are existent in whom a histological study of laser-induced changes is performed. As early as one day after LITT, surviving residual tumour cells start to grow further into the surrounding tissue increasing the size of the neoplastic lesion and outgrowth of tumour cells from the margin into the necrotic centre of the lesion can be detected. Outsprouting capillaries from the edge of the lesion are surrounded by tumour cells which soon, after several days already, form compact tumour cell masses. The diffuse infiltration of high grade gliomas into the adjacent tissue structures hinders their curative treatment. Open operations are suitable for e.g. tumour debulking in order to reduce intracranial pressure. Tumour progression, however, can usually not be prevented. Such operations impose a substantial burden on these patients and often reduce their quality of life due to postoperative neurological deficits. Local treatment offers a minimally invasive alternative cytoreduction, limiting the adverse effects associated with open surgery. Therefore, the local treatment of malignant gliomas has gained increasing interest during the past decades. Brachytherapy, stereotactic surgery, and local chemotherapy have been investigated. For some procedures, a moderately increased survival has been reported. Such results, however, have to be regarded with caution with respect to the selection bias often associated with work on this subject. Literature reports on local heating techniques in glioblastomas multiforme (GBM) are rare. This is also true for laser irradiation. Although laser irradiation of brain tumours has been described [5,6,11] most of the work refers predominantly to astrocytomas WHO grades II and III and gliosarcomas [12]. In these studies, laser irradiation was performed with comparatively low irradiation doses [5,6,11] and less sophisticated application devices [5,11]. Nevertheless, it could be shown that energy doses between 2400 and 9120 J were sufficient to induce coagulation necroses between 18 and 35 mm in diameter [13]. It could be demonstrated that this cytoreductive effect can also be achieved in recurrent Glioblastoma, furthermore, that this treatment can be repeated. Thus it is, in principle, not limited to the size of the single laser lesion but can be adapted to the geometry of the tumour and its progression over time. The most important finding, however, is the remarkable long survival of recurrent glioblastomas (GBM) associated with the laser treatment.
ARTICLE IN PRESS F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
This survival is substantially longer than the e.g. 5.4 months reported by Trent et al. [13] for a monotherapy with temozolomide. Whether the underlying mechanism is a local cytoreduction only or due to other laser effects cannot be derived from the presented data. In addition, thermal concepts may offer additional advantages in the therapy of high-grade glioma because a moderate temperature increase within brain tissue is known to modulate the BBB. Studies in normal monkey brain revealed a defined time window where the BBB remained open after local application of microwave hyperthermia. Similar findings were also reported in rabbits treated with focused ultrasound, and for laser irradiation [13]. A local disruption of the BBB after laser irradiation has also been reported for astrocytomas WHO grades II and III in patients after interstitial laser irradiation [5]. It could be shown in clinical trials that the disruption of the BBB in areas with otherwise uncompromised BBB can result in an improved clinical outcome of chemotherapy of brain tumours [14]. This is possibly due to the fact that these areas of tumour infiltration into the surrounding healthy tissue with intact BBB are the most active parts of the tumour characterized by a large fraction of cycling cells [15]. In addition, the drug concentration may be increased within the tumour tissue due to BBB modulation. If the concentration of temozolomide in the brain tumour is increased by the direct intratumoural application of temozolomide, life expectation can be increased as demonstrated in D54 human MG xenograft-challenged athymic mice [16]. The follow-up MR examinations clearly demonstrated a regression of the laser-irradiated parts of the tumour, while simultaneously the non-laser irradiated parts exhibited a significant progression. This is of particular importance because all parts of the tumour whether laser irradiated or not were prone to the same systemic chemotherapy. Consequently, we have good evidence to associate the remarkable clinical outcome with the laser irradiation. The technique of MR-guided local laser ablation has now achieved a sophisticated level. However, further exploration with well-designed and controlled clinical trials is advisable to define its final role in the treatment of gliomas. Whether the underlying mechanism is a local coagulation only or at least in part a result of the modulation of the BBB or other laser effects cannot be derived from the presented data. The answer to these questions must also be left to future investigations.
Zusammenfassung Laserstereotaxie maligner Gliome: 2001–2004 Tumoren des zentralen Nervensystems stellen eine therapeutische Herausforderung dar. Die Mehrzahl dieser hirneigenen Geschwu¨lste sind histologisch gese-
129
hen Gliome mit infiltrativem Wachstum in die umgebende Hirnsubstanz. Trotz mikroskopischer Resektion bildet diese Art der Tumoren in relativ kurzer Zeit Rezidive. Es wurden daher lokale Therapieformen entwickelt, die die Umgebung unscharf begrenzter Hirntumoren mit beeinflussen. 16 Patienten mit einem Glioblastomrezidiv wurden kernspingesteuert, lokal, stereotaktisch, interstitiell lasertherapeutisch behandelt. Die MRT-Verlaufsuntersuchungen zeigten eine Volumenreduktion der laserbestrahlten Areale, wa¨hrend die nicht bestrahlten Gebiete mehr oder minder eine Tumorprogression zeigten. Interessanterweise war die U¨berlebenszeit dieser Patienten 4 mal la¨nger als erwartet, woraus abzuleiten ist, dass die lokale Lasertherapie selbst bei hochmalignen, hirneigenen Tumoren in gewissen Fa¨llen eine Erfolg versprechende Alternative zu anderen Therapieformen darstellt. Mit Hilfe kontrollierter Studien sollten diese Ergebnisse daher weiter evaluiert werden. r 2006 Elsevier GmbH. All rights reserved. Schlu¨sselwo¨rter: Hirntumor, Gliom, Interstitielle Thermotherapie, Laser-induzierte MRI-gefu¨hrte Therapie, Glioblastomrezidiv, U¨berlebenszeit
Resumen Estereotaxis por la´ser de gliomas malignos: 2001–2004 Los tumores del sistema nervioso central representan un desafı´ o terape´utico. A pesar de la cantidad de subgrupos histolo´gicos, las principales entidades son de origen gliomal. La mayorı´ a de estas lesiones son de naturaleza infiltrativa, y dado que se encuentran en la proximidad de a´reas altamente especializadas del cerebro, representan una alta exigencia en lo que a habilidad quiru´rgica y a tecnologı´ as aplicadas se refiere. Estos tumores son invariablemente recurrentes, con mayor frecuencia en el mismo lugar en donde previamente se realizo´ una aparente reseccio´n ‘‘total’’. Esto ha dado lugar a terapias locales para el tratamiento directo intracavitario de la superficie, ası´ como tambie´n para el tratamiento intersticial del tejido adyacente. Diecise´is pacientes con glioblastoma multiforme recurrente fueron tratados mediante radiacio´n la´ser Nd:YAG en el marco de una terapia de rescate. El concepto de base es el de una citoreduccio´n del tumor mediante su coagulacio´n parcial. El seguimiento por RMN revelo´ una reduccio´n en el volumen de las a´reas irradiadas con la´ser, mientras que las a´reas no tratadas mostraron cierta progresio´n. El tiempo de sobrevida luego del diagno´stico de la recurrencia fue aproximadamente cuatro veces mayor que el esperado para la evolucio´n normal de la enfermedad. Como conclusio´n, la citoreduccio´n mediante radiacio´n la´ser serı´ a una opcio´n prometedora para pacientes con glioma infiltrativo.
ARTICLE IN PRESS 130
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
Nuevos trabajos servira´n para optimizar los regı´ menes terape´uticos y evaluar este tratamiento en ensayos clı´ nicos controlados. En este artı´ culo, resumiremos los resultados clı´ nicos y te´cnicos obtenidos por aplicacio´n del la´ser intersticial en gliomas cerebrales. r 2006 Elsevier GmbH. All rights reserved. Palabras claves: Tumor cerebral, Glioma, Termoterapia Intersticial, Terapia La´ser guiada por RMN, Glioblastoma multiforme recurrente, tiempo de sobrevida
[8]
[9]
[10]
References [1] Nabavi A, Blo¨mer U, Mehdorn HM. Clinical aspects of gliomas, Med. Laser Appl 2002;17:91–104. [2] Schwarzmaier H-J, Eickmeyer F, Fiedler V, Ulrich F. Basic Principles of Laser induced Interstitial Thermotherapy in Brain Tumors. Med Laser Appl 2002;17: 147–58. [3] Schulze PC, Vitzthum HE, Goldammer A, Schneider JP, Schober R. Laser-induced thermotherapy of neoplastic lesions in the brain underlying tissue alterations, MRI monitoring and clinical applicability. Acta Neurochir (Wien) 2004;146:803–12. [4] Bublat M, Grunst G, Kansy K, Schmitgen A, Meixensberger J, Kahn T, et al. Iterative multimodal computer-assisted navigation using pre- and intraoperative image data. Med Laser Appl 2000;17:123–31. [5] Kahn T, Bettag M, Ulrich F, Schwarzmaier H-J, Schober R, Fu¨rst G, et al. MRI-guided laser induced interstitial thermotherapy of cerebral neoplasms. J Comput Assist Tomogr 1994;18:519–32. [6] Schwabe B, Kahn T, Harth T, Ulrich F, Schwarzmaier HJ. Laser induced thermal lesions in the human brain; short- and long term appearance on MRI. J Comput Assist Tomogr 1997;21:818–25. [7] Schober R, Bettag M, Sabel M, Ulrich F, Hessel S. Fine structure of zonal changes in experimental Nd:YAG
[11]
[12]
[13]
[14]
[15]
[16]
laser-induced interstitial hyperthermia. Lasers Surg Med 1993;13:234–41. Schulze PC, Kahn T, Harth T, Schwarzmaier HJ, Schober R, Schulze CP. Correlation of neuropathologic findings and phase-based MRI temperature maps in experimental laser-induced interstitial thermotherapy. J Magn Reson Imag 1998;8:115–20. Schulze PC, Adams V, Busert C, Bettag M, Kahn T, Schober R. Effects of laser-induced thermotherapy (LITT) on proliferation and apoptosis of glioma cells in rat brain transplantation tumors. Lasers Surg Med 2002; 30:227–32. Eickmeyer F, Schwarzmaier H-J, von Tempelhoff W, Fiedler VU, Niehoff H, Ulrich SD, Ulrich F. MR-guided laser-induced interstitial thermotherapy of recurrent glioblastoma. Proc Intl Soc Mag Reson Med 2005;13:2140. Hata N, Morrison PR, Kettenbach J, Block PMcL, Kikinis R, Jolesz FA. Computerassisted intra-operative magnetic resonance imaging monitoring of interstitial laser therapy in the brain: a case report. J Biomed Opt 1998;3:304–10. Sabel M, Rommel F, Kondakei M, Gorol M, Willers R, Bilzer T. Locoregional opening of the rodent blood–brain barrier for paclitaxel using Nd:YAG laser-induced thermo therapy: a new concept of adjuvant glioma therapy. Lasers Surg Med 2003;33:75–80. Trent S, Kong A, Short SC, Traish D, Ashley S, Dowe A, et al. Temozolomide as second-lime chemotherapy for relapsed gliomas. J Neurooncol 2002;57:247–51. Kroll RA, Neuwelt EA. Outwitting the blood-brain barrier for therapeutic purposes: Osmotic opening and other means. Neurosurgery 1998;42:1083–110. Levin VA, Prados MD. Treatment of recurrent gliomas and metastatic brain tumors with a polydrug protocol designed to combat nitrosurea resistance. J Clin Oncol 1992;10:766–71. Heimberger AB, Archer GE, McLendon RE, Hulette C, Friedmann HS, Bigner DD, et al. Temozolomide delivered by intracerebral microinfusion is safe and efficacious against malignant glioma rats. Clin Cancer Res 2000;6:4148–53.
Medical Laser Application 21 (2006) 123–130 www.elsevier.de/mla
Laser stereotaxy for malignant gliomas: 2001–2004 Frank Ulrich, Hans-Joachim Schwarzmaier, Wernholt von Tempelhoff, Hendrik Niehoff, Frank Eickmeyer Neurochirurgische Klinik, Klinikum Krefeld, Lutherplatz 40, 47805 Krefeld, Germany Received 13 January 2006; accepted 26 January 2006
Abstract Tumours of the central nervous system represent a therapeutic challenge. Although there is a multitude of histological subgroups, the main entities are of glioma origin. The majority of these lesions have an infiltrative nature, which is, in conjunction with the vicinity of highly specialized eloquent areas, highly demanding on the surgical skills and technologies applied. These tumours will invariably recur, most often even at the very site of the apparently ‘‘complete’’ resection. This has created local therapies for direct intra cavitary treatment of the surface and for interstitial treatment of the immediately adjacent tissue. Sixteen patients suffering from recurrent Glioblastoma multiforme (GBM) were treated by Nd:YAG laser irradiation in the framework of a salvage therapy. The underlying concept is the cytoreduction by partial coagulation of the tumour. MRI follow-up examinations revealed a volume reduction of the laser-irradiated areas, while the untreated parts exhibited a more or less progression. The survival time after the diagnosis of the recurrence was about four times longer than the natural history of the disease would suggest. The conclusion is that cytoreduction by laser irradiation might be a promising option for patients suffering from infiltrative gliomas. Future work should optimize the therapeutic regimes and evaluate this treatment approach in controlled clinical trials. In this article we will summarize the technical and clinical results induced by interstitial laser application in cerebral gliomas. r 2006 Elsevier GmbH. All rights reserved. Keywords: Brain tumour; Glioma; Interstitial thermotherapy; Laser induced; MRI guided therapy; Recurrent glioblastoma multiforme; Survival
Introduction Tumours of the central nervous system represent a therapeutic challenge. Although there is a multitude of histological subgroups, the main entities are of astrocytic and oligodendrocytic origin. The majority of these lesions have an infiltrative nature, which is, in conjuncCorresponding author. Tel.: +49 2151 321320; fax: +49 2151 322033. E-mail address: [email protected] (F. Ulrich).
1615-1615/$ - see front matter r 2006 Elsevier GmbH. All rights reserved. doi:10.1016/j.mla.2006.03.006
tion with the vicinity of highly specialized eloquent areas, highly demanding of the surgical skills and technologies applied. There are various approaches to the treatment of neuroepithelial tumours. Surgery is the first treatment option for primary and recurrent lesions. However, there are cases, where either surgery predictably harbours significant morbidity or general health issues preclude long procedures. In much the same way it is certain that due to the infiltrative nature of these lesions, a surgical cure is rarely achieved. The diversity as well as the quality of the tools at our disposal has
ARTICLE IN PRESS 124
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
improved significantly over recent years. It becomes more of a challenge to comprehend and correctly employ the different therapeutic modalities. Still it is sometimes discouraging that even with such a potent armamentarium, this grave disease entity takes an unstably lethal course. Presently our best achievement is to improve or at least maintain the patient’s quality of life for as long as possible and ameliorate symptoms. In these cases, radiation or interstitial therapy, by using stereotactic principles remains an alternative option. As for interstitial therapies, radioactive seed implantation as well as interstitial laser-induced thermotherapy enable minimal interventions, which allow pathologic diagnosis combined with local treatment [1].
Interstitial laser therapy using diffusor fibre tips Laser induced interstitial thermotherapy (LITT) in glioma is performed via flexible light guides. In general, fused silica fibres of 400 or 600 mm in diameter are used. These are characterized by a damping factor of a less than 1 dB/km (1064 nm). Thus, transmission losses due to the fibre material is almost negligible with light guides of a typical length between 10 and 12 m. This length is required if LITT is monitored by magnetic resonance (MR) imaging. In this case, the laser must be positioned outside the MR-room to avoid interactions with the magnetic field. Transmission losses are predominantly due to the Fresnel reflection at the air/fibre interfaces within the fibre coupler. The major goal of interstitial thermal therapy is the heating of a given tissue volume without adverse effects such as charring or shock waves due to vaporization at the probe tip. Consequently, the peak temperature at the probe tip must not exceed 100 1C. This limits the peak power energy per given time and probe surface area. On the other hand a minimal radiant power density is required to achieve a sufficient temperature rise for tissue denaturation. The high power
density at the conventional fibre tip often resulted in an overheating of the tissue surface with consecutive vaporization and carbonization. In this situation, no longer does tissue irradiation takes place and the heating process is mediated by thermal conduction only. The technique was overcome using optical diffusors; such diffusors emit laser light over the entire length of the fibre tip, significantly reducing the power density and the penetration depth is dependent on the optical tissue properties (absorption, scattering, blood content) at the wavelength used. These devices allow to couple between 5 and 8 W into the brain tissue without adverse effects. The resulting laser lesion is also dependent on the length of the optical diffusor (see Fig. 1). At a temperature above 60 1C, a direct thermal coagulation is induced resulting in an necrosis of 2–3 cm in radial diameter. In the areas beyond the induced necrosis, temperatures are lower, however, still above 38 1C. Thus, the total thermally treated area typically extends over 3 and 4 cm in radial diameter. The final lesion sizes, however, are also dependent on the exact optical properties of the irradiated tissue as well as its blood perfusion. For on-line control, temperature monitoring was performed using the experimental software package based on the temperature-dependent phase shift of the MR signal as mentioned above. The actual temperature distribution could be displayed as colour-coded images as well as time course of the hot spot temperature (see Fig. 2). Laser irradiation was ceased when temperature monitoring revealed a steadystate temperature profile within the heated tissue [2].
Image-guided laser fibre adaption LITT procedures in the open MR Signa SP/ 2 are dependent on MR-compatible and artifact-free instruments. A guiding device (NeuroGate, Daum, Schwerin, Germany) has been used in combination with the
Fig. 1-2. (1) Laser light guide in place after repositioning and insertion of the applicator (note the artifact of the light guide’s marker (magnetite) touching the brain tumor and the artifact of the diffusing tip within the tumor). (2) The actual temperature distribution is displayed as colour-coded, near real-time MR images.
ARTICLE IN PRESS F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
Flashpoint-System as an instrument for planning, guiding, and performing stereotactic laser procedures in the OMR. The device allows planning of a laser fibre trajectory from the surface of the patient’s head, and puts a marker on the skin for accurate burr hole positioning. It is a tripod with an adaptor to the 3 LED-guide of the Flashpoint-System and simulates the potential 350 range of angulation of the NeuroGate. Following a skin incision, a standard burr hole is made with a vertical bone wall and suitable dimensions for fixing the NeuroGate, using a special broach and grip. The central guide element is flexible and manoeuvrable after releasing the thumb screen with a special hexagonal screw driver. During surgery, placement of the biopsy cannula or sheath could be changed at any time and was supported by the LED tracking system and MR guidance [3]. Repeated nearly real-time alignment with the trajectory plane allows monitoring of the position of the stereotactic probe in relation to the targeted structure. If the image is a fast gradient echo scan, the delay can be as little as 1 sec. However, if longer imaging times are used for improved image quality, the delay can be a few minutes. When 3D T1-weighted images show the type of the laser light guide in the targeted region, the laser therapy is performed. The Signa SP can generate full 3D volumes within some 5 min in usual resolution and any selected 2D slices within some 5 s. These nearly real-time slices together with the built-in tracking system (Flashpoint, IGT, Boulder, CO, USA) establish the real-time capability of this scanner. Real-time MR images can be selected in any plane relative to the device to give real-time device control. However, it takes the iMR several seconds to generate such a single slice. Interactive work is a little bit disturbed by this delay. Additionally for technical reasons, real-time slices have significantly reduced resolution. As long as the patient does not move, volume and patient are exactly registered with the builtin tracking system. In soft tissue, the laser intervention usually renders preoperative images inappropriate if tissue shifts occur. In these cases, intraoperative images are necessary to show the current state, to update the preoperative data, and to register preoperative images to the current situation. The update is an iterative process triggered by situational changes during the ongoing intervention [4]. The aim of magnetic resonance imaging (MRI) in monitoring LITT is target definition, trajectory selection, light guide tracking, prediction of final lesion size during therapy, and follow-up of the temporal evolution of the laser-induced lesion. The open MR-guidance in its real-time/on-line mode allows the surgeon to update his target point in its new position after shifting and
125
displacement. The on-line mode provides the option of frameless stereotaxis by flashpoint-tracking. This is helpful in the puncture of cystic lesions or in positioning the laser fibre after biopsy for LITT in tumours – so that MR-guided LITT becomes a one-time and one-place procedure. The surgeon can visualize and react to loss of cyst fluid, CSF, or necrotic regions of the neoplasms. In the T1-mode with contrast medium, small lesions can be targeted and identified easily, choosing an optimal approach over a burr hole. The intraoperative MR-imaging is a welcome improvement of the classic, static neuronavigation, adding the important feature of a dynamic on-line control of laser-induced procedure. The latest updates and developments of the MR software allow fusions of preoperative and actual intraoperative MR-data, even in 3D mode. The zonal architecture of laser-induced lesions consists of a central and peripheral zone (see Fig. 3). In the central zone, there is a generalized damage of cellular and subcellular membranes observed resulting in irreversible tissue damage with secondary effects in the periphery. The intravascular red blood cells appear empty. In low-grade gliomas, high signal intensity of T2 weighted MR images of the central zone most likely represents heat-induced methaemoglobin conversion from desoxyhaemoglobin. On the other hand, highprotein-content fluid collections may also contribute to the high signal intensity of the central zone in low-grade gliomas. The low signal intensity rim of T2-weighted images most likely represents desoxyhaemaglobin in the acute stage and haemosiderin deposition in later follow-
Fig. 3. Postoperative imaging: recurrent glioblastoma multiforme 24 h after LITT [10].
ARTICLE IN PRESS 126
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
up studies. After Gd-DTPA, an enhancement in this zone was evident in all patients independently of a blood-brain barrier disruption prior to LITT. In tumours with Gd-DTPA enhancement in the MRI prior to LITT due to a disrupted blood-brain barrier (BBB), there was no enhancement within the laser-induced lesion after LITT. The perifocal edema is not apparent immediately after LITT. It evolves 1–3 days after LITT and regresses completely within less than a month, indicating that no persistent damage occurs within this zone. In the majority of the induced lesions, the longterm development is uniform with variations in the zonal architecture in two lesions. In all patients, the decrease of lesion size followed an exponential pattern with a half-life period of a few weeks. The shrinkage of the lesion is accompanied by a corresponding reduction of the size of the neoplasm. As the laser-induced enhancing rim persists over a long period, difficulties may arise in differentiating the residual laser lesion from recurrent tumour. The sequential time course of the laser-induced enhancing rim with an ongoing reduction of size may exclude a recurrent tumour that appears as increasing volume and new onset of mass effect. It has to be considered that the central and peripheral zones together form the total lesion size [5,6].
Neuropathological findings LITT is a minimally invasive approach for the treatment of gliomas, especially in poorly accessible regions. For the treatment of glioma by LITT, the extent of lethal laser–tissue interaction and changes in tumour cells due to LITT are of particular relevance for the clinical course and the outcome of patients with brain neoplasms. The primary goal of LITT is the destruction and removal of neoplastic cells and a decrease in tumour size. However, the survival of tumour cells at the margin of the surgical lesion after treatment has been shown to determine the postoperative course. A secondary outgrowth of residual neoplastic cells may form the basis of a recurrent tumour mass. For the prevention of tumour recurrences and, therefore, the outcome of patients after surgical interventions, alterations of the cell cycle and the life span of residual tumour cells have to be considered in addition to the primary destruction of the solid tumour mass. Following laser irradiation, the laser lesion shows in experimental studies a typical architecture dependent on the interval following laser treatment. The primary lesion exhibits differences in lesion diameter in dependence of the irradiation time and procedural energy deposition to the tissue but exhibits, in general, the typical architecture of a laser lesion that consists of a central coagulation necrosis and a surrounding rim of edema adjacent to undamaged tissue. Immediately after irradiation, the central zone of
necrosis is not yet apparent but is demarcated by a rim of edema from the undamaged tissue. The necrosis becomes evident by a gradual loss of staining and early resorptive changes at their margins by day 3. A rim of granulation tissue is formed after one week and the edema further spreads to the periphery. Finally, a cystic lesion with variable remnants of unresorbed necrotic tissue and a mesenchymal and glial reaction at the emerging remains. The immunohistochemical examination of tissue obtained from these lesions shows a distinct spacio-temporal staining pattern of several antibodies reactive to glial cells and neurons in relation to the different zones of the lesion. Non-specific and diffuse immunostaining can be found in the central necrotic zone only. In several studies, no qualitative difference could be found in the general histological architecture of the LITT lesion between animals with implanted tumour cells and those without. The zones of lesion are clearly defined as a central necrosis with a rim of edema. Interestingly, however, distinct changes can be seen in the size of the lesion in areas of normal tissue in comparison to neoplastic tissue. This might be due to different histological and, therefore, optical properties of these tissues and a resulting difference in the sensitivity to laser energy. In studies on human brain tissue, differences in absorption and scattering coefficients of the tissue defined by different neuroanatomical origin and ultrastructure could be demonstrated. Additionally, the optical properties of brain tissue are altered by coagulation as a result of laser–tissue interactions. Laser-induced coagulation leads to a tissue-specific increase in absorption and scattering with a higher vulnerability of the tissue to laser light. The specific appearance of an LITT lesion in the time course is mainly defined by resorptive changes by microglial cells or macrophages and a reactive gliosis. Immunohistochemical staining for the glial fibrillary acid protein (GFAP) clearly demonstrates an astrocyte activation and the extent of gliosis. Increased numbers of GFAPpositive reactive astrocytes are present after one day in the marginal zone. The number of reactive astrocytes affected by GFAP increases after 7 days. After week 2, the marginal zone of the lesion shows a distinct threelayered structure. The central necrotic zone is surrounded by a zone of edema containing an infiltration of granulocytes, lymphocytes, and macrophages. A layer of GFAP-positive reactive astrocytes separates this zone of edema from the adjacent brain tissue that appears normal. At the cellular level, different stages of astrocyte activation are detectable. One day following LITT, clumps of GFAP-positive cytoplasmatic material coat the unclear membrane. After 3 days, GFAP staining appears in short cytoplasmatic processes. The intensity of GFAP immunoreactivity in reactive astrocytes that surround the necrotic zone increases until the end
ARTICLE IN PRESS F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
of the 2nd week. Later, sprouting capillaries and GFAPpositive reactive astrocytes extend into the zone of necrosis. After 1 month, a thin layer of GFAP-positive astrocytes surrounds a cystic defect. Following laser treatment, blood vessels generally appear engorged, containing leached erythrocytes that lay closely opposed to each other with a staining of the membranes only. The tissue structure in general, however, is well preserved. The densely packed and decolourized red blood cells seem to be a characteristic histological marker of the intravascular laser effect. Vascular effects of laser irradiation include, furthermore, a thrombotic occlusion of the vessels inside the central laser lesion. These vascular structures are resorbed by invading mesenchymal cells during the course of regeneration. Vessels outside the margin of the laser lesion become the origin of outsprouting capillaries further invading into the lesion with a functional importance for regenerative and proliferative processes. Also, these vessels clearly contain normal and undamaged erythrocytes 3 days after laser irradiation, indicating a functional blood flow and the connection to the vascular tree. Fine structural analysis of the acute and chronic changes following laser treatment consistently shows damage of cellular and subcellular membranes in the central zone surrounding the laser tip. Cell membranes and unclear membranes of nerve cells, glial cells, and endothelial cells either exhibit local defects or are broken up into fragments. Also, subcellular membranes, in particular those of mitochondria, display structural damage. The curled fragments often fuse to form small vesicles, which may be a secondary heat effect. Vascular cells such as erythrocytes similarly show membrane defects leading to a loss of haemoglobin. Only basement membranes of capillaries and blood vessels seem to resist the laser treatment. In contrast, the investigation of the edema zone adjacent to the central necrosis displays no membrane ruptures. These observations indicate that membrane alterations may play a leading role in the pathogenesis of laser-induced lesions. Intravascular erythrolysis, a common phenomenon after laser irradiation, seems to be calmed by a leakage of haemoglobin through defects in the cellular membranes of erythrocytes and may lead to vascular occlusion with secondary tissue damage. The interstitial appearance of serum proteins such as albumin, which do not cross the blood-brain barrier and typically are not detectable in the interstitium under normal conditions, can be attributed to membranous defects of the capillary endothelium. The underlying laser–tissue interaction and the resulting histological changes are well defined and have been characterized as central coagulation necrosis and peripheral edema, subsequent resorptive changes and the formation of a rim of granulation tissue. In addition, various cellular changes of degen-
127
erative and regenerative nature have been found on a functional–morphological level, both in brain and in metastasis. Such changes can be correlated with MRI appearances under both experimental and clinical conditions [3,7–9].
Clinical results A total of 16 patients were enrolled into a representative study between May 2001 and June 2004. All patients suffered from a radiologically documented recurrence of a histologically confirmed glioblastoma grade IV WHO. All patients were non surgical candidates. In all patients, a first-line chemotherapy with temozolomide was initiated in case they did not have a respective prior treatment. A defined coagulation zone that shows a volume reduction over time as reported previously and already documented in low grade astrocytomas was exhibited. In this small group of patients, there was no inhospital lethality. No permanent neurological deficits were seen after the intervention. One patient exhibited a temporary weakness of the right arm. This is remarkable in so far as the tumours were predominantly located adjacent to functional relevant areas. Three patients developed a neutropenia and another patient exhibited a thrombocytopenia. One patient exhibited an increase of the transaminases and in one patient, a toe developed a local infection (panaritium). After LITT, the median overall survival was 6.971.7 months (95% CI 3.7–10.2). The overall survival after the diagnosis of the first relapse was 9.471.3 months (95% CI 6.8–12.0). There was, however, a substantial learning curve. During this learning period, (2001–2002), the median survival after LITT therapy was only 5.270.6 months (95% CI 4.1–6.3). Thereafter, the median survival after the first LITT increased to 11.272.0 (95% CI 7.4–15.0) months after LITT (Kaplan–Meier method, Software SPSS Version 12.0) (see Fig. 4). At the end of the study, 12 of the 16 patients were deceased. The cause of death was fatal venous thromboembolism in two cases. One patient died of a systemic mycosis, one patient from a gastrointestinal bleeding, and another due to a peritonitis after sigma perforation. The remaining deaths were due to a failure of the central regulation as typically seen in brain tumours. During the learning curve (2001–2002), the time between the diagnosis of the recurrence and the laser therapy was 2.0 months vs. 0.3 month thereafter. In addition, the KPS was lower and the tumour volume was larger in the first group of patients. It is of particular interest that five deaths were most probably more related to side effects of the concomitant chemotherapy and/or corticoid therapy, rather than to local tumour destruction [10].
ARTICLE IN PRESS 128
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
1.0
Survival function before 2003 2003 and after censored
Cumulative survival
0.8
0.6
0.4
0.2
0.0 0.00
5.00
10.00
15.00
20.00
Survival after LITT [months]
Fig. 4. Median survival: During learning period (2001–2002: 5.270.6 months) and median survival of the patients treated in 2003 and 2004: 11.272.0 months [10].
Discussion The laser-induced changes appear to be qualitatively equal, irrespective of the type of brain tumour, although differences dependent on the grade of malignancy may exist regarding the onset and size of recurrence. Primary effects of laser irradiation include a rupture of cell membranes and a fragmentation of subcellular components, resulting in irreversible tissue damage with secondary effect in the periphery. The laser–tissue interactions are followed by reactive and proliferative and BBB changes that determine in the clinical course. The BBB prevents the passage of plasma protein bound chemotherapeutica agents into the CNS. Capillaries in malignant gliomas have indeed imperfect interendothelial functions, endothelial perforations, and other ultrastructural abnormalities characteristic of neoplastic vasculature resulting in a disruption of BBB. Even though this disruption allows the diagnostically advantageous transcapillary passage of contrast material, leakage of plasma protein into gliomas and the adjacent brain does not suffice for an effective antineoplastic accumulation of chemotherapeutic agents. This is due to the fact that not all tumour capillaries are equally leaky and delivery of large molecule drugs to the tumour bed becomes thus uneven. Levin demonstrated that the BBB within the zone of infiltrations in the otherwise normal
brain adjacent to the tumour mass is unaffected. This is of particular importance, since this zone contains the leaking edge of neoplastic cells with a large fraction of cycling cells that have a high growth fraction. LITTinduced disruption of the BBB acilitates a locoregional passage of chemotherapeutic agents into the brain tissue. Of the cases of LITT, few patients are existent in whom a histological study of laser-induced changes is performed. As early as one day after LITT, surviving residual tumour cells start to grow further into the surrounding tissue increasing the size of the neoplastic lesion and outgrowth of tumour cells from the margin into the necrotic centre of the lesion can be detected. Outsprouting capillaries from the edge of the lesion are surrounded by tumour cells which soon, after several days already, form compact tumour cell masses. The diffuse infiltration of high grade gliomas into the adjacent tissue structures hinders their curative treatment. Open operations are suitable for e.g. tumour debulking in order to reduce intracranial pressure. Tumour progression, however, can usually not be prevented. Such operations impose a substantial burden on these patients and often reduce their quality of life due to postoperative neurological deficits. Local treatment offers a minimally invasive alternative cytoreduction, limiting the adverse effects associated with open surgery. Therefore, the local treatment of malignant gliomas has gained increasing interest during the past decades. Brachytherapy, stereotactic surgery, and local chemotherapy have been investigated. For some procedures, a moderately increased survival has been reported. Such results, however, have to be regarded with caution with respect to the selection bias often associated with work on this subject. Literature reports on local heating techniques in glioblastomas multiforme (GBM) are rare. This is also true for laser irradiation. Although laser irradiation of brain tumours has been described [5,6,11] most of the work refers predominantly to astrocytomas WHO grades II and III and gliosarcomas [12]. In these studies, laser irradiation was performed with comparatively low irradiation doses [5,6,11] and less sophisticated application devices [5,11]. Nevertheless, it could be shown that energy doses between 2400 and 9120 J were sufficient to induce coagulation necroses between 18 and 35 mm in diameter [13]. It could be demonstrated that this cytoreductive effect can also be achieved in recurrent Glioblastoma, furthermore, that this treatment can be repeated. Thus it is, in principle, not limited to the size of the single laser lesion but can be adapted to the geometry of the tumour and its progression over time. The most important finding, however, is the remarkable long survival of recurrent glioblastomas (GBM) associated with the laser treatment.
ARTICLE IN PRESS F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
This survival is substantially longer than the e.g. 5.4 months reported by Trent et al. [13] for a monotherapy with temozolomide. Whether the underlying mechanism is a local cytoreduction only or due to other laser effects cannot be derived from the presented data. In addition, thermal concepts may offer additional advantages in the therapy of high-grade glioma because a moderate temperature increase within brain tissue is known to modulate the BBB. Studies in normal monkey brain revealed a defined time window where the BBB remained open after local application of microwave hyperthermia. Similar findings were also reported in rabbits treated with focused ultrasound, and for laser irradiation [13]. A local disruption of the BBB after laser irradiation has also been reported for astrocytomas WHO grades II and III in patients after interstitial laser irradiation [5]. It could be shown in clinical trials that the disruption of the BBB in areas with otherwise uncompromised BBB can result in an improved clinical outcome of chemotherapy of brain tumours [14]. This is possibly due to the fact that these areas of tumour infiltration into the surrounding healthy tissue with intact BBB are the most active parts of the tumour characterized by a large fraction of cycling cells [15]. In addition, the drug concentration may be increased within the tumour tissue due to BBB modulation. If the concentration of temozolomide in the brain tumour is increased by the direct intratumoural application of temozolomide, life expectation can be increased as demonstrated in D54 human MG xenograft-challenged athymic mice [16]. The follow-up MR examinations clearly demonstrated a regression of the laser-irradiated parts of the tumour, while simultaneously the non-laser irradiated parts exhibited a significant progression. This is of particular importance because all parts of the tumour whether laser irradiated or not were prone to the same systemic chemotherapy. Consequently, we have good evidence to associate the remarkable clinical outcome with the laser irradiation. The technique of MR-guided local laser ablation has now achieved a sophisticated level. However, further exploration with well-designed and controlled clinical trials is advisable to define its final role in the treatment of gliomas. Whether the underlying mechanism is a local coagulation only or at least in part a result of the modulation of the BBB or other laser effects cannot be derived from the presented data. The answer to these questions must also be left to future investigations.
Zusammenfassung Laserstereotaxie maligner Gliome: 2001–2004 Tumoren des zentralen Nervensystems stellen eine therapeutische Herausforderung dar. Die Mehrzahl dieser hirneigenen Geschwu¨lste sind histologisch gese-
129
hen Gliome mit infiltrativem Wachstum in die umgebende Hirnsubstanz. Trotz mikroskopischer Resektion bildet diese Art der Tumoren in relativ kurzer Zeit Rezidive. Es wurden daher lokale Therapieformen entwickelt, die die Umgebung unscharf begrenzter Hirntumoren mit beeinflussen. 16 Patienten mit einem Glioblastomrezidiv wurden kernspingesteuert, lokal, stereotaktisch, interstitiell lasertherapeutisch behandelt. Die MRT-Verlaufsuntersuchungen zeigten eine Volumenreduktion der laserbestrahlten Areale, wa¨hrend die nicht bestrahlten Gebiete mehr oder minder eine Tumorprogression zeigten. Interessanterweise war die U¨berlebenszeit dieser Patienten 4 mal la¨nger als erwartet, woraus abzuleiten ist, dass die lokale Lasertherapie selbst bei hochmalignen, hirneigenen Tumoren in gewissen Fa¨llen eine Erfolg versprechende Alternative zu anderen Therapieformen darstellt. Mit Hilfe kontrollierter Studien sollten diese Ergebnisse daher weiter evaluiert werden. r 2006 Elsevier GmbH. All rights reserved. Schlu¨sselwo¨rter: Hirntumor, Gliom, Interstitielle Thermotherapie, Laser-induzierte MRI-gefu¨hrte Therapie, Glioblastomrezidiv, U¨berlebenszeit
Resumen Estereotaxis por la´ser de gliomas malignos: 2001–2004 Los tumores del sistema nervioso central representan un desafı´ o terape´utico. A pesar de la cantidad de subgrupos histolo´gicos, las principales entidades son de origen gliomal. La mayorı´ a de estas lesiones son de naturaleza infiltrativa, y dado que se encuentran en la proximidad de a´reas altamente especializadas del cerebro, representan una alta exigencia en lo que a habilidad quiru´rgica y a tecnologı´ as aplicadas se refiere. Estos tumores son invariablemente recurrentes, con mayor frecuencia en el mismo lugar en donde previamente se realizo´ una aparente reseccio´n ‘‘total’’. Esto ha dado lugar a terapias locales para el tratamiento directo intracavitario de la superficie, ası´ como tambie´n para el tratamiento intersticial del tejido adyacente. Diecise´is pacientes con glioblastoma multiforme recurrente fueron tratados mediante radiacio´n la´ser Nd:YAG en el marco de una terapia de rescate. El concepto de base es el de una citoreduccio´n del tumor mediante su coagulacio´n parcial. El seguimiento por RMN revelo´ una reduccio´n en el volumen de las a´reas irradiadas con la´ser, mientras que las a´reas no tratadas mostraron cierta progresio´n. El tiempo de sobrevida luego del diagno´stico de la recurrencia fue aproximadamente cuatro veces mayor que el esperado para la evolucio´n normal de la enfermedad. Como conclusio´n, la citoreduccio´n mediante radiacio´n la´ser serı´ a una opcio´n prometedora para pacientes con glioma infiltrativo.
ARTICLE IN PRESS 130
F. Ulrich et al. / Medical Laser Application 21 (2006) 123–130
Nuevos trabajos servira´n para optimizar los regı´ menes terape´uticos y evaluar este tratamiento en ensayos clı´ nicos controlados. En este artı´ culo, resumiremos los resultados clı´ nicos y te´cnicos obtenidos por aplicacio´n del la´ser intersticial en gliomas cerebrales. r 2006 Elsevier GmbH. All rights reserved. Palabras claves: Tumor cerebral, Glioma, Termoterapia Intersticial, Terapia La´ser guiada por RMN, Glioblastoma multiforme recurrente, tiempo de sobrevida
[8]
[9]
[10]
References [1] Nabavi A, Blo¨mer U, Mehdorn HM. Clinical aspects of gliomas, Med. Laser Appl 2002;17:91–104. [2] Schwarzmaier H-J, Eickmeyer F, Fiedler V, Ulrich F. Basic Principles of Laser induced Interstitial Thermotherapy in Brain Tumors. Med Laser Appl 2002;17: 147–58. [3] Schulze PC, Vitzthum HE, Goldammer A, Schneider JP, Schober R. Laser-induced thermotherapy of neoplastic lesions in the brain underlying tissue alterations, MRI monitoring and clinical applicability. Acta Neurochir (Wien) 2004;146:803–12. [4] Bublat M, Grunst G, Kansy K, Schmitgen A, Meixensberger J, Kahn T, et al. Iterative multimodal computer-assisted navigation using pre- and intraoperative image data. Med Laser Appl 2000;17:123–31. [5] Kahn T, Bettag M, Ulrich F, Schwarzmaier H-J, Schober R, Fu¨rst G, et al. MRI-guided laser induced interstitial thermotherapy of cerebral neoplasms. J Comput Assist Tomogr 1994;18:519–32. [6] Schwabe B, Kahn T, Harth T, Ulrich F, Schwarzmaier HJ. Laser induced thermal lesions in the human brain; short- and long term appearance on MRI. J Comput Assist Tomogr 1997;21:818–25. [7] Schober R, Bettag M, Sabel M, Ulrich F, Hessel S. Fine structure of zonal changes in experimental Nd:YAG
[11]
[12]
[13]
[14]
[15]
[16]
laser-induced interstitial hyperthermia. Lasers Surg Med 1993;13:234–41. Schulze PC, Kahn T, Harth T, Schwarzmaier HJ, Schober R, Schulze CP. Correlation of neuropathologic findings and phase-based MRI temperature maps in experimental laser-induced interstitial thermotherapy. J Magn Reson Imag 1998;8:115–20. Schulze PC, Adams V, Busert C, Bettag M, Kahn T, Schober R. Effects of laser-induced thermotherapy (LITT) on proliferation and apoptosis of glioma cells in rat brain transplantation tumors. Lasers Surg Med 2002; 30:227–32. Eickmeyer F, Schwarzmaier H-J, von Tempelhoff W, Fiedler VU, Niehoff H, Ulrich SD, Ulrich F. MR-guided laser-induced interstitial thermotherapy of recurrent glioblastoma. Proc Intl Soc Mag Reson Med 2005;13:2140. Hata N, Morrison PR, Kettenbach J, Block PMcL, Kikinis R, Jolesz FA. Computerassisted intra-operative magnetic resonance imaging monitoring of interstitial laser therapy in the brain: a case report. J Biomed Opt 1998;3:304–10. Sabel M, Rommel F, Kondakei M, Gorol M, Willers R, Bilzer T. Locoregional opening of the rodent blood–brain barrier for paclitaxel using Nd:YAG laser-induced thermo therapy: a new concept of adjuvant glioma therapy. Lasers Surg Med 2003;33:75–80. Trent S, Kong A, Short SC, Traish D, Ashley S, Dowe A, et al. Temozolomide as second-lime chemotherapy for relapsed gliomas. J Neurooncol 2002;57:247–51. Kroll RA, Neuwelt EA. Outwitting the blood-brain barrier for therapeutic purposes: Osmotic opening and other means. Neurosurgery 1998;42:1083–110. Levin VA, Prados MD. Treatment of recurrent gliomas and metastatic brain tumors with a polydrug protocol designed to combat nitrosurea resistance. J Clin Oncol 1992;10:766–71. Heimberger AB, Archer GE, McLendon RE, Hulette C, Friedmann HS, Bigner DD, et al. Temozolomide delivered by intracerebral microinfusion is safe and efficacious against malignant glioma rats. Clin Cancer Res 2000;6:4148–53.