- Email: [email protected]
Serial MR imaging of intracranial metastases after radiosurgery
Magnetic ResonanceImaging, Vol. 15, No. 10, pp. 1121-l 132, 1997 6 1997 EIsevierScienceInc. All rights reserved. Printed in the USA. 0730-725X/97 $17.00 + .OO
ELSEVIER
l
PI1 SO730-725X(97)00178-1
Original Contribution SERIAL MR IMAGING OF INTRACRANIAL AFTER RADIOSURGERY
METASTASES
HANS HAWIGHORST,” MARCO ESSIG,* J~~RGENDEBUS,? MICHAEL V. KNOPP,* RITA ENGENHART-CABILIC,~ S. 0. SCH~NBERG,* GUNNAR BRIX,* IVAN ZUNA,* AND GERHARD VAN KAICK* *Department
of Radiology,
German
Cancer Research Center (dkfz), Im Neuenheimer
Feld 280, D-69120
Heidelberg,
Germany, and tDepartment of Radiation Therapy, University Hospital Heidelberg, D-69120 Heidelberg, Germany Puqme: To evahmtethe spatiotempod evolution of radkqid tismeadjacenttotheksiomby.serialmagnelkrwmance
induced changesboth in metaskmandinnommlbmiu
(MR)ilnagiqMethodsand-Tbh.ty-fiwi!’
*
‘1
metaskzs of different prima&s were treated in 25 patients by singlehigbdose mdkurgery. MR imagesacquired before radiosurgery were availablein all patienk Sixty-threefollow-up MR studieswere performed in thesepatientsh&ding T,andcontrastsnhancedT,-weightedMRimagesTbeaveragefdlow-upthnewss9~5months:(mean~standarddeviatioa [SD]). Basedon contrasknbanced T,-weigbted MR images,tumor responsewas radiologicallyda&W in the following tr&mentdidnotsbowatumorsb&kageof fourgroups:stabledkeasewasaswnediftbeavemgetumordiameterafter morethan5o%audanillcmse OfIllOll?than25%,jNUtid remi&onasashrinkageoftumorsizeofmorethan5O%,a disappearanceofcontmst-enhancing tumor asa completeremission,and an kreaseoftumordiameterofmorethan25% as tumor progress. Moreover, we analysedsignal changeson T,-weighted imagesin brain myma adjacent to the enhancing~.~~Theoveranmean~~thnewas105~7months,withal-yearacturuSalsurvivalreQ of 40%. Stabledisease,partial or complete remissionof the metastatictnmor was observedin 22 patients (gS%). Central orhomageneouslossofcontrast enhancementappearedtobeagoodprogno&signforstablediseclseorpartialrem&&n. This associationwas stati&caRy sign&ant @ < 0.05). Three patients (12%) suff& tknn tumor progressioh In eight patients (32%) witb stable diseaseor partial remission,signal changes on T,-weighted images were observed in tissue adjacenttotbecontr&enban&gksions.Apmgre&on offhehighsignalonT,-weightedimageswasseminsevenofthe eight patients between 3 and 6 months after therapy, followed by a signal regmxsion 6-18 months after irradiation. Condusion: MR imaging is a sensitiveimaging tool to evaluatetumor respome asweIl asthe presenceorabsenceofadjaamt ~~changesf~~radiosurgery..LossdhomogewousorcentralcontrastenhaneementoaGdeahancedMR imageaappeared to be a good prognosticsign for tumor response Tumor shrinkage seemsnot to be dependent on time. In addition, most casesof radiation induced changesin normal brain parenchyma observedon T,-weighted imagesseemto be sdflhnitd 01997Ekevier!kienceIn~ Keywords: Radiosurgery; MR imaging; Brain metastases.
INTRODUCTION Brain metastases are a major problem in neuro-oncology, accounting for 43% of all intracranial tumors.21 Stereotactic radiosurgery using charged particle beams, modified linear accelerators, or gamma knife units has been increasingly employed for the treatment of multiple, inoperable, or recurrent brain metastases.“‘13 Successful local tumor control has clinically been achieved in more than 80% of patients treated by this modality.” With the
rapid increase in the numbers of radiosurgical centers, it becomes more important to learn about the pattern of response and the temporal evolution of the irradiated metastases and radiation induced changes of normal brain parenchyma adjacent to the irradiated target volume. Modem imaging techniques have made in vivo studies of the reaction of the tumor as well as the adjacent normal brain tissue feasible. A number of reports on computed tomography (CT) and magnetic resonance
RECEIVED 47197; ACYCEPTED7112197. Address correspondence to: Hans Hawighorst, MD, Department of Radiology, German Cancer Research Center (dkfz), Im
Neuenheimer Feld 280, D-69120 Heidelberg, Germany. Email: [email protected] 1121
MagneticResonanceImaging0 Volume 15, Number 10,
1122
(MR) imaging of radiation induced changes following radiotherapy are now available.14.17.19 For various reasons, however, the results of these reports may not be applicable to radiosurgery. The different reactions of a pathological process to single as opposed to fractionated radiation most likely extend also to adverse reaction of the normal brain. Despite the current clinical interest and usage, no report has systematically documented the postoperative neuroimaging findings in patients with brain metastases treated by means of stereotactic radiosurgery. Therefore, the aim of this study was to evaluate the spatiotemporal evolution of radiosurgical induced changes both in metastases and in normal adjacent brain tissue adjacent to the lesions by serial MR imaging.
1997
two per patient). The target volume was defined by the contrast-enhancing tumor region enlarged by a 1 mm safety margin. An example of a 80% isocenter dose plan is depicted in Fig. 1. All patients received prophylactic steroids (20 mg dexamethason) 1-3 h before and 8 h after irradiation.
Patients From January 1988 to July 1996, 359 patients with intracranial metastases from different primary tumors were treated by stereotactic radiosurgery. However, of these only 25 patients (7%) had a complete MR followup, i.e., MR examinations prior to and at different points in time after radiosurgery. This selected population formed the basis of the present study. Patients were treated with single high-dose irradiation technique either as primary treatment (n = lo), as a treatment for a relapse (n = l), or within a combination of planned whole-brain irradiation and a radiosurgical boost therapy (n = 14). The series included 12 females and 13 males (age 52 + 11 years). All patients had a histologically confirmed primary tumor with one or more metastatic brain lesions. Eighteen patients had a solitary metastasis and 7 patients had two or three brain metastases without other widespread tumor disease. The patient data as well as the location, histology, and clinical presentation of intracranial metastases are summarized in Table 1.
MR Imaging In addition to the base-line examination, which was also used for the purpose of radiation treatment planning, all patients were followed both by neurologic and contrast-enhanced MR imaging examinations at approximately 3 monthly intervals over the first 2 years. For patients surviving the first 2 years, the follow-up timing was tailored to the individual. The mean time of MR imaging follow-up was 9 ? 5 months. MR measurements were performed on a 1.5-Tesla whole-body MR system (MAGNETOM SP 4000, Siemens AG, Erlangen, Germany). For radiofrequency transmission and detection, a circular polarized head coil with an internal diameter of 3 1 cm was used. In a first step, sagittal fast low angle shot (FLASH) images (repetition time (TR)/echo time (TE) = 30/10 ms; flip = 80’; matrix size = 128 X 128; field of view = 260; slice thickness = 6 mm; one acquisition) were acquired to get an orientation. For treatment planning and follow-up, identical contiguousslice T,-weighted spin-echo (SE) and T,-weighted SE images were measured in axial orientation (TRITE = 600/15 ms and 2800/19,93 ms, respectively, matrix size = 256 X 256, field of view = 230-270 mm, section thickness = 4 mm, one to two acquisitions). T,-weighted contrast-enhanced MR images with identical imaging parameters were obtained in the axial as well as in a coronal or sagittal plane approximately 2 min after i.v. injection of gadolinium dimeglumine (Magnevist@, Schering AG, Berlin, Germany; 0.1 mmol/kg body weight).
Treatment Technique Detailed descriptions of dose planning and the treatment technique using linear accelerator radiosurgery (wherein multiple gantry rotations at different table angles produce multiple non-coplanar converging arcs) have been published previously.3~5*g*15The maximum dose at the isocenter varied from lo-25 Gy (mean dose, 16 + 4 Gy). The outline of the target volume was within the 80% isodose contour in all cases. The delivered dosage depended on localization and radiosensitivity of critical structures in the vicinity of the tumor, and on the fact of whether the patients received radiosurgery as part of the initial whole or partial brain irradiation (13 patients; dose 39 ? 8 Gy). The mean field size of the collimator was 21.5 2 10 mm in diameter. Multiple irradiation isocenters were used in seven cases (average,
Data Evaluation MR image readings was separately done by two experienced radiologists (H.H., M.E.) and one radiotherapist (R.E.) using the same criteria. In all cases, the maximum tumor extension was measured on axial precontrast and on contrast-enhanced T,-weighted SE images directly from film using fine calipers. In addition, the maximum extension of parenchymal changes surrounding the metastases was measured on T,-weighted images. Based on contrast-enhanced T,-weighted MR images, the following definitions were used to radiologically classify tumor response:4 stable disease was assumed if the average tumor diameter after treatment did not show a shrinkage of more than 50% or an increase of less than 25%, partial remission as a shrinkage of tumor size of more than 50%. A disappearance of the contrast-
METHODS
AND MATERIALS
PR S
R M M N B B B Br M
22
12 20 20 15 20 20
10 25
18
Parietal lobe
Parietal lobe Cerebellum Temporal lobe Temporal lobe Parietal lobe Occipital lobe
Splenium Orbit
Thalamus
16
17 18 19 20 21 22
23 24
25
Note. B = Bronchial Contrast-enhanced;
PR PR S S S PR
Br
12
Frontal lobe
15
PR
CR
PR
PR
2.5
3.2 1.4
3.9 1.3 0.7 0.9 1.6 1.1
2.1
2.3
3.0
2.5
2.8
2.1 1.0
carcinoma; M = Melanoma; R = Renal carcinoma; Br = Breast carcinoma; *the numbers in brackets give the period (in months) at which the changes
B
M
15
PR
S
PR S
Frontal lobe
13
B M
0, 3, 6 0, 3, 6
3, 3 3, 3, 3 3,
6
6, 9 6
6, 9
data
(central loss of enhancement) (3, 9) (central loss of enhancement) (3)
complete remission (28) decreased (central loss of contrast enhancement) (6, 9) decreased (3) stable (6, 9) stable stable (3) decreased (6) decreased (central loss of contrast enhancement) (3, 6) stable (15) decreased (central loss of contrast enhancement) (6)
decreased (6, 12)
stable decreased contrast decreased contrast
increased (3, 6) increased (3, 12) increased (3) stable complete remission (3) stable (3, 6) decreased (3, 6) decreased (central loss of enhancement) (3) stable decreased (central loss of enhancement) (6) stable
*Change in size on Cd-CE image*
and MR imaging
P = Progress; S = Stable; PR = Partial remission; CR = Complete in tumor size or signal on T,-or T,-weighted images occurred.
0, 6, 12, 15, 18
0, 3, 6, 12 0, 3, 12, 15
0, 0, 0, 0, 0, 0,
0, 3, 6, 9, 12, 18, 28
0, 3, 6, 9, 12, 15
0, 3
0, 3, 6, 9
0, 3, 6
0, 3, 6 0, 3, 6
6 6, 9, 12 6 6 6 6 6
0.9 1.6
PR S
3, 3, 3, 3, 3, 3, 3,
0, 0, 0, 0, 0, 0, 0,
0.6 2.2 3.0 2.8 1.0 2.5 1.2
14
12
15 15
B Br
radiosurgical,
Time of MR (mo)
P P P S CR S PR
Br
Parietal lobe
10 11
15 12
M R B B B M B
12
Cerebellum Frontal lobe
8 9
20 15
18
20 20 15 15
Parietal lobe
Vermis Frontal lobe
1 2 3 4 5 6 7
Histology
Tumor size (cm)
Radiological response
M
Parietal lobe Cerebellum Frontal lobe Temporal lobe Cerebellum Occipital lobe Frontal lobe
Case
of brain metastases in 25 patients: clinical,
12
Location of metastases
Radiosurgery dose (80% isodose) (GY)
Table 1. Radiosurgery
remission;
mo = month;
CE =
decreased (3, 6, 9) increased (12) increased (9) decreased (6) increased (3) decreased (3) increased (3) and decreased (12) increased ( 15) increased (6, 12) and decreased (15, 18)
stable increased (3) and decreased (6, 12) increased (3, 6) and decreased (12, 18)
stable
(3)
on T2-w (MO)
increased (3, 6) increased (6) increased (3, 6) increased (3) complete remission stable (3, 6) decreased (3, 6) increased (3) and decreased (6) stable increased (3) and decreased (6) stable increased (3) and decreased (6)
*Signal
E 6 g 5 3 F
2 22 oh 9 D E P E
g
1124
MagneticResonanceImaging0 Volume 15, Number 10, 1997 changes on T,-weighted
images in brain parenchyma
adjacent the enhancing lesions. The statistical analysis was performed with use of the SAS software (SAS, Cary, NC, USA). The results of the MR image reading (tumor extension, signal changes, contrast enhancement pattern) of the three observers were separately analyzed and tested for significance. The
MR imaging data of the lesions (signal change, tumor extension, contrast enhancement pattern) measured after
radiosurgery were tested retrospectively for a difference to the values prior to radiosurgery with the paired-student t-test. A p value of less than 0.05 was accepted as
significant. RESULTS Patients As of July 1996, 11 of the 25 examined patients
(44%) had died. Fourteen patients (56%) are alive without major disability and are able to work. The overalI
mean survival time was 10.5 2 7 months, with a l-year actuarial survival rate of 40%. MR Image Reading There were no significant (p > 0.1) differences between the three observers concerning the MR image evaluation (tumor extension, signal changes, contrast enhancement pattern).
Tumor Size According to the MR imaging criteria defined above,
stable disease was observed in 15 patients (60%), partial remission in 5 patients (20%) (example in Fig. 2) and complete remission in 2 patients (8%) (example in Figs. 3,4). In the five patients with partial remission, the mean tumor size decrease was 10 mm + 3.5 mm. Three patients who had a deterioration of neurologic function
also showed an increase in the contrast-enhancing tumor region (example in Fig. 5) with a mean increase of 7.9 2 (W Fig. 1. Isodose cutyes are superimposed on a contrast-enhanced (a) axial and (b) sag&al MR image. The intracranial metastasisin a 56-year-old male was treated with stereotactic radiosurgery (20 Gy to the mmor margin at the 80% isodose
line). Note that the inner circle representsthe 80%. the following circles the 60%, 40%, and 20% isodose, respectively.The gray dots represent the target volume defined by the radiotherapist.
3.0 mm. Changes in tumor size were not significantly
associated with the initial tumor size, maximal tumor dose, additional whole brain irradiation or histology of the primary tumor (Table 1).
Tumor Enhancement In seven patients, the findings of stable disease or
tumor remission were accompanied by a loss of central tumor enhancement (10 metastases) and/or a decrease of contrast enhancement with blurred tumor margins between 3 months and 12 months after treatment (mean, 6 months) (Figs. 6, 7). Central or homogeneous loss of
enhancing tumor region was defined as a complete remission. An increase of tumor diameter of more than 25% on contrast-enhanced TI-weighted images was de-
contrast enhancement appeared to be a good prognostic sign for stable disease or tumor shrinkage and this asso-
fined as tumor progress. Moreover, we analysed signal
ciation was statisticalIy significant (p C 0.05). Radiolog-
Serial
MRI
of metastases0
HAWKHORST
1125
ET AL.
(Fig. 5). In two of mass was surgically ical examination of tumor cells without
these patients, the increased tumor excised. The result of the histologthe surgical specimen showed vital radiation-induced brain necrosis.
Adjacent Parenchymal Changes on T,-Weighted
(4
(W Fig. 2. Serial contrast-enhancedT,-weighted MR images of a 57-year-old-male whose metastasisfrom a bronchial carcinoma was treated with stereotacticradiosurgery (20 Gy to the tumor margin at the 80% isodoseline). (a) PreradiosurgicalMR image depicting a contrast-enhancingmetastasisat the occipital lobe. (b) Six months after radiosurgery, partial remission occurred with decreasedcontrast enhancement and regression of tumor size.
ically determined tumor progress-which in all cases correlated with a neurological deterioration-was associated with an equally or stronger contrast enhancement compared to the tumor enhancement prior to radiosurgery
Images MR imaging performed prior to radiosurgery demonstrated in all cases an area of increased signal intensity in normal brain parenchyma adjacent to the enhancing metastases on T,-weighted MR images. In patients with radiological stable disease or tumor remission, 8 of the 25 patients (32%) demonstrated an extending parenchymal field of high signal intensity on T,-weighted images after radiosurgery. In seven of these eight cases, the parenchymal changes reverted to normal (example shown in Fig. 8) and the repeated MR examinations gave some indications about the temporal evolution of radiation induced changes and its self-limiting nature in patients with stable disease or tumor remission. The progression of the high signal areas on T,-weighted MR images was seen between 3 and 6 months after radiotherapy and followed by a regression at 6-18 months after irradiation. High signal areas with extension along the white matter tracts were observed in two cases, and a high signal region which coincided well with the 20% isodose arc in three cases. In one patient with a radiological complete remission, follow-up MR images did not depict any parenchymal changes on T,-weighted images. In each case of tumor progression, MR imaging revealed a progression of the high signal intensity on T,-weighted images coinciding with an enlarging area of tumor enhancement. The appearence of new adjacent parenchymal changes following radiosurgery did not correlate significantly with tumor size, tumor margin dose, or tumor maximum dose. However, progressively developing areas of high signal intensity were more often observed in patients with progressive tumor growth, radiosurgery of multiple metastases and/or receiving whole brain irradiation as part of the treatment protocol.
DISCUSSION In 20-40% of all patients with solitary brain metastases, surgery is not possible because the lesions are localized in or near vital structures and thus can not be resected without high risk.‘6*22 In a single treatment session, radiosurgery administers a dose of radiation presumably sufficient to achieve local tumor control. The dose can be delivered to a precisely defined volume of tissue with minimal radiation exposure to the surrounding normal structures.
1126
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
04
(a)
Fig. 3. Serial contrast-enhanced T,-weighted MR images of a 54-year-old-male patient whose metastasis from a hypemephroma was treated with stereotactic radiosurgery (22 Gy to the tumor margin at the 80% isodose line). (a) Preradiosurgical MR image depicting a large contrast-enhancing metastasis adjacent to the right ventricle. (b) Three months after radiosurgery, decreased contrast enhancement and regression of tumor size was seen. (c) Nine months after radiosurgery, further marked decrease in tumor size was seen. (d) Twenty-eight months after radiosurgery, no contrast enhancement is visible corresponding to a complete remission.
Tumor Size Sturm and coworkers*’ were the first to report the efficacy of radiosurgery in the management of brain metastases. A mean dose of 18 Gy was used, and in the 28 patients for whom follow-up were available, a major radiographic response rate of 86% was documented as determined on CT scans. Engenhart et aL6 evaluated 69
with 102 inoperable brain metastases an average of 12 months after stereotactic radiosurgery. Their assessment of tumor size found a complete remission in 20%, partial remission in 35%, stable disease in 40%, and relapse in only 5% of the patients. To our knowledge no other studies were published describing systematitally the MR imaging findings following radiosurgery in
patients
Serial MRI
of
metastases 0 HAWIGHORST
03
1127
ET AL.
03
Fig. 4. Serial contrastenhanced T,-weighted MR images of a 57-year-old-male patient whose right cerehellar metastasisfrom a bronchial carcinoma was treated with stereotacticradiosurgery (18 Gy to the tumor margin at the 80% isodoseline). Preradiosurgical MR image depicting on (a) contrast-enhanced T,-weighted images a lesion with peripheral enhancement which is homogeneous hyperintense (b) on T,-weighted images. Three months after radiosurgery, complete remission was reached. Neither (c) contrastenhancement nor (d) a hyperintense lesion on Z’,-weighted image is visible anymore. patients with brain metastases. The analysis of our MR imaging data yields a local control of brain metastases manifested as stable disease or remission in 22 of 25 patients (88%) and tumor progression in 3 of 25 patients (12%). The MR imaging control rate of 88% compares well to the radiological control rates published by other authors’S6V’3,20based on CT scans.
The radiobiological mechanism by which stereotactic radiosurgery inhibits further tumor growth is yet not fully understood. Animal studies suggest that tumor cells are irreversibly damaged after a singlefraction radiation doses as low as 20-30 Gy. Kondziolka et a1.r’ developed and studied different in vivo tumor models. The tumor grafts were treated with
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
1128
icantly larger tumor diameter at autopsy, appeared as highly cellular lesions; in contrast, the treated animals, with a smaller mean tumor diameter had large areas of hypocellularity representing cell death.
(a)
Fig. 5. Serial contrast-enhanced T,-weighted MR images of a 38-year-old female whose metastasis from a bronchial carcinoma was treated with stereotactic radiosurgery (15 Gy to the tumor margin at the 80% isodose line; in addition, the patient received 33 Gy whole brain irradiation). (a) The preradiosurgical MR image depicts a large contrast-enhancing metastasis at the frontal lobe. (b) Three months after radiosurgery, tumor progression occurred with increased contrast enhancement and progression of tumor size. The patient deteriorated clinically. The increasing mass was surgically excised. The histological examination revealed vital tumor cells without radiation necrosis.
different high doses of irradiation (15-50 Gy), compared to untreated controls and evaluation was performed at death. The untreated controls, with a signif-
(W Fig. 6. Serial contrast-enhanced Tr-weighted MR images of a 57-year-old female whose metastasis from a bronchial carcinoma was treated with stereotactic radiosurgery (18 Gy to the tumor margin at the 80% isodose line). (a) The preradiosurgical MR image depicting a large and homogeneously contrastenhancing metastasis in the left thalamus. (b) Three months after radiosurgery, tumor regression occurred with marked loss of central contrast enhancement.
Serial
MRI
of
metastases 0 HAWIGHORST
ET AL.
1129
(b)
Fig. 7. Serial contrast-enhanced MR images of a 44year-old female whose metastasisfrom a breast carcinoma was treated with stereotactic radiosurgery (18 Gy to the tumor margin at the 80% isodose line). (a) PreradiosurgicalMR image depicting a large and homogeneously contrast-enhancingmetastasisin the left frontal lobe. (b) Six months after radiosurgery, tumor regressionoccurred with marked loss of central contrast enhancement. (c) Six months after radiosurgery, further decreaseof tumor size occurred. (d) Twelve months after radiosurgery, partial tumor remission had occurred.
Tumor Enhancement In our series, postoperative MRI examinations often revealed a loss of central or homogeneous contrast enhancement which was observed 3 to 12 months after radiosurgery. Interestingly, this effect appeared to be a good prognostic sign for stable disease or tumor remis-
sion. To better define the impact of single high dose irradiation, Spiegelmann et al.” developed a cat model to study their triphasic hypothesis. According to this concept, radiosurgical masses progress through different characteristic phases: first there is acute degenerative necrosis caused by cell death and vascular injury; sec-
1130
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
(4
(cl
(d)
Fig. 8. Serial MR images of same case as in Fig. 6 show the development and regression of radiation induced changes in brain parenchyma adjacent to the irradiated metastasis.(a) PreradiosurgicalT,-weighted MR image revealsa small hyperintenserim in the left thalamus adjacent to the brain metastases.(b) The Z’,-weighted MR image obtained at 6 months after radiosurgery shows an extending parenchymal field adjacent to the metastases.(c) A further increase of the high signal field is visible at 12 months after radiosurgery. Note that the irradiated metastasesshows shrinkage.No steroidswere administered as patient was clinically free of any symptoms. (d) I’,-weighted
MR image 18 months after radiosurgery
ond, this necrotic mass is phagocytosed, and finally a prominent glial scar develops. To test this hypothesis, they irradiated 19 cats with different radiosurgical doses and evaluated sequentially with contrast-enhanced MR imaging and pathologic assessment. No central contrast enhancement was found after approximately 3 to 6
showing marked regression of the high signal lesion.
weeks. This area corresponded pathologically to necrotic debris and vascular damage. Within 12 to 29 weeks, the animals improved clinically, and the MR imaging findings revealed resorption of debris characterized by reduced lesion size and contrast enhancement. Histologitally, diminished vascular&y and gliosis predominated.
Serial
MRI of metastasesOHAw~~o~s~
Based on these observations, our findings of central loss of contrast enhancement in 10 metastases may be explained as radiation-induced vascular injury and tumor cell death. Obliteration of blood supply to the metastatic lesion may be one of the most significant mechanisms controlling tumor growth after radiosurgery. We conclude that homogeneous or central loss of contrast enhancement on T,-weighted MR images appears to be an early and good prognostic sign to predict tumor response.
1131
ET AL.
which are not characterized by contrast enhancement. Furthermore, these radiation induced changes are most
often self-limited, rarely accompanied by clinical symptoms and do not require the use of steroids. In contrast, progression of high signal intensity on T,-weighted im-
ages with an enlarging area of contrast enhancing tumor is most likely caused by tumor progression, does require
the use of steroids, and shows clinical manifestations. Regression of the tissue changes was seen at 6-18
months after its appearance. As steroids may alter the Adjacent Parenchymal Changes on T,-weighted
Images In this study, parenchymal changes adjacent to contrast enhancing lesions manifested as focal high signal intensity on T,-weighted MR images and were seen in 8 of 25 patients with stable disease or partial remission. The cause of high signal regions on T,-weighted images without contrast enhancement in areas of the brain adjacent to the tumor remains controversial. The most likely cause is an edema and/or blood-brain barrier (BBB) disturbance in the adjacent brain parenchyma due to radiosurgery and not due to tumor growth. This hypothesis is supported by the absence of hypointensity on T,-weighted images, tbe temporary nature of the effect in the majority of cases, and only few, if any, newly associated neurologic symptoms despite their appearance in critical brain areas, and lacking increase of tumor size on contrast-enhanced images. The most likely explanation for the high signal aberration on T,-weighted images seems to be the transient BBB damage or edema from injured brain tissue. Demyelination is unlikely to appear outside the volumes irradiated by this imaging pattern. Thus, edema dominated the radiation induced changes in brain parenchyma between 3-6 months in the current series following radiosurgery. These interpretations are supported by the observations previously reported from animal experiments. Fike et al.’ described the occurrence of radiation edema with corresponding CT changes, Lo and colleagues12 studied the impact of high-dose, singlefraction brain irradiation in the rabbit brain using MR imaging and positron emission tomography. Within 10 months of delivering 30 Gy helium ion irradiation, BBB disturbances occurred that manifested as an early high T2 signal abnormality without contrast enhancement on MR
course of radiation induced changes, it is of particular interest to follow the serial MR images in six patients
who had a non-symptomatic adverse reaction and did not receive such treatment. In two patients, radiation induced
changes came to a stable and persistent high signal as compared to the preradiosurgical MR images. These late
appearing and persisting high signal lesions may represent demyelination and/or gliosis. Conclusion MR imaging is necessary for adequately assessing the
response of brain metastases to radiosurgery. Our analysis suggests a high initial tumor control rate of 88%
with a mean follow-up of 9 months. Loss of homogeneous or central contrast enhancement on Gd-MR ap-
pears to be a good prognostic sign for tumor response. In addition, most cases of radiation induced changes in normal brain parenchyma seem to be self limited. Further
studies with histological proof are needed to confirm the signal changes observed on MR examinations.
REFERENCES 1. Alexander, E.; Moriarty, T.M.; Davis, R.B.; Wen, P.Y.;
2. 3.
4.
images. In addition, field size increased significantly. This damage probably reflects early damage on endothe-
lial cells. In addition, higher irradiation doses may result in disruptions of the BBB. Grossmann et al.* conducted
5.
a study in cats, delivering 50 Gy with a 1.5 cm collima-
tor, and within 7 months, observed focal areas of BBB disruption,
using gadolinium-enhanced
MR imaging.
From the results of this study, we conclude that even with lower doses of 16 Gy radiosurgical induced paren-
chyma changes may be detected on T,-weighted images
6.
Fine, H.A.; Black, P.M.; Kooy, H.M.; Loeffler, J.S. Stereotactic radiosurgery for the definitive, noninvasive treatment of brain metastases.J. Nat. Cancer Ins. 87:34-40; 1995. Altschuler, E.M.; Lunsford, L.D.; Kondziolka, D., et al. Radiobiologic models for radiosurgery. Neurosurg. Clin. North Am. 3:61-77; 1992. Bauer-Kirpes, B.; Schlegel, W.; Boesecke, R.; Lorenz, W.J. Display of organs and isodosesas shaded 3-D objects for 3-D therapy planning. Int. J. Radiat. Oncol. Biol. Phys. 13:135-140; 1987. DeVita, V.T. Principles of chemotherapy. In: V.T. DeVita, S. Hellmann, S. Rosenberg, (Eds). Cancer. Principles and Practice of Oncology. Philadelphia: Lippincott Co, 1993: 276-289. Engenhart, R.; Kimmig, B.; Pastyr, 0.; Marin-Graz, M.; van Kaick, G. Entwicklung eines reproduzierbaren Fixationssystems fiir perkutane Bestrahlung im Kopf-Halsbereich. Zentralbl. Rad. 136:731-732; 1988. Engenhart, R.; Kimmig, B.N.; Hover, K.H., et al. Longterm follow-up for brain metastasestreated by percutaneous stereotactic single high-dose irradiation. Cancer 7 1: 1353-1361; 1993.
1132
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
7. Fike, J.R.; Cann, C.E.; Turowski, K., et al. Radiation responseof normal brain. Int. J. Radiat. Oncol. Biol. Phys. 14:63-70; 1988. 8. Grossmann,RI.; Hecht-Leavitt, C.M.; Evans, SM., et al. Experimental radiation injury: Combined MR imaging and spectroscopy.Radiology 169:305-309; 1988. 9. Hartmann, G.H.; Schlegel, W.; Strum, V.; Kober, B.; Pastyr, 0.; Lorenz, W.J. Cerebral radiation surgery using moving field irradiation at a linear acceleratorfacility. Int. J. Radiat. Oncol. Biol. Phys. 11:1185-I 192; 1985. 10. Kondziolka, D.; Lunsford, L.D.; Claassen,D., et al. Radiobiology of radiosurgery: Part II. The rat C6 glioma model. Neurosurgery 31:280-287; 1992. 11. Konnov, B.; Melnikov, L.; Zargrova, 0. Narrow proton beam therapy for intracranial lesions. In: International Workshop on Proton and Narrow Photon Beam Therapy, Oulu, Finland; 1989:48-55. 12. Lo, E.H.; Fabrikant, J.L. Delayed biologic reactions to stereotactic charged particle radiosurgery in the human brain. Stereotact. Funct. Neurosurg. 56:197-212; 1991. 13. Loeffler, J.S.; Kooy, H.M.; Wen, P.Y.; Fine, H.A.; Cheng, C.W.; Mannarino, E.G.; Tsai, J.S.; Alexander, E. The treatment of recurrent brain metastaseswith stereotactic radiosurgery. J. Clin. Oncol. 8:576-582; 1990. 14. Miot, E.; Hoffschir, D.; Pontvert, D.; Gaboriaud, G.; Alapetite, C.; Masse, R.; Fetissof, F.; Le Pape, A.; Akoka, S. Quantitative magnetic resonance and isotropic imaging: Early evaluation of radiation injury to the brain. Int. J. Radiation Oncology Biol. Phys. 32:121-128; 1995.
15. Pastyr, 0.; Hartmann, G.H.; Schlegel, W.; Schabbert, S.; Treuer, H.; Lorenz, W.J. Stereotactically guided convergent beam irradiation with a linear accelerator: Localization technique. Acta. Neurochir. 99:61-64; 1989. 16. Posner, J.B. Management of central nervous system metastases.Semin. Oncol. 4:81-91; 1977. 17. Rubin, P.; Gash, D.M.; Hansen, J.T.; Nelson, D.F.; Williams, J.P. Disruption of the blood-brain barrier as the primary effect of irradiation. Radiother. Oncol. 31:51-60, 1994. 18. Spiegelmann, R.; Friedmann, W.A.; Bova, F.J., et al. LINAC radiosurgery: An animal model. J. Neurosurg. 78: 638-644; 1993. 19. Steinberg, G.K.; Fabrikant, J.I.; Marks, M.P.; Levy, R.P.; Frankel, K.A.; Phillips, M.H.; Shuer, L.M.; Silverberg, G.D. Stereotacticheavy-charged particle Bragg peak radiation for intracranial arteriovenous malformations. N. Eng. J. Med. 323:96-101; 1990. 20. Stmm, V.; Kober, B.; Hover, K.H., et al. Stereotactic percutaneous single irradiation of brain metastaseswith a linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 13: 279-282; 1987. 21. Walker, A.E.; Robins, M.; Weinfeld, F.D. Epidemiology of brain tumors: The national survey of intracranial neoplasms. Neurology 35:219-226; 1985. 22. West, J.; Maor, M. Intracranial metastases:Behavioural patterns related to primary site and results of treatment by whole brain irradiation. Int. J. Radiat. Oncol. Biol. Phys. 6:11-15; 1980.
ELSEVIER
l
PI1 SO730-725X(97)00178-1
Original Contribution SERIAL MR IMAGING OF INTRACRANIAL AFTER RADIOSURGERY
METASTASES
HANS HAWIGHORST,” MARCO ESSIG,* J~~RGENDEBUS,? MICHAEL V. KNOPP,* RITA ENGENHART-CABILIC,~ S. 0. SCH~NBERG,* GUNNAR BRIX,* IVAN ZUNA,* AND GERHARD VAN KAICK* *Department
of Radiology,
German
Cancer Research Center (dkfz), Im Neuenheimer
Feld 280, D-69120
Heidelberg,
Germany, and tDepartment of Radiation Therapy, University Hospital Heidelberg, D-69120 Heidelberg, Germany Puqme: To evahmtethe spatiotempod evolution of radkqid tismeadjacenttotheksiomby.serialmagnelkrwmance
induced changesboth in metaskmandinnommlbmiu
(MR)ilnagiqMethodsand-Tbh.ty-fiwi!’
*
‘1
metaskzs of different prima&s were treated in 25 patients by singlehigbdose mdkurgery. MR imagesacquired before radiosurgery were availablein all patienk Sixty-threefollow-up MR studieswere performed in thesepatientsh&ding T,andcontrastsnhancedT,-weightedMRimagesTbeaveragefdlow-upthnewss9~5months:(mean~standarddeviatioa [SD]). Basedon contrasknbanced T,-weigbted MR images,tumor responsewas radiologicallyda&W in the following tr&mentdidnotsbowatumorsb&kageof fourgroups:stabledkeasewasaswnediftbeavemgetumordiameterafter morethan5o%audanillcmse OfIllOll?than25%,jNUtid remi&onasashrinkageoftumorsizeofmorethan5O%,a disappearanceofcontmst-enhancing tumor asa completeremission,and an kreaseoftumordiameterofmorethan25% as tumor progress. Moreover, we analysedsignal changeson T,-weighted imagesin brain myma adjacent to the enhancing~.~~Theoveranmean~~thnewas105~7months,withal-yearacturuSalsurvivalreQ of 40%. Stabledisease,partial or complete remissionof the metastatictnmor was observedin 22 patients (gS%). Central orhomageneouslossofcontrast enhancementappearedtobeagoodprogno&signforstablediseclseorpartialrem&&n. This associationwas stati&caRy sign&ant @ < 0.05). Three patients (12%) suff& tknn tumor progressioh In eight patients (32%) witb stable diseaseor partial remission,signal changes on T,-weighted images were observed in tissue adjacenttotbecontr&enban&gksions.Apmgre&on offhehighsignalonT,-weightedimageswasseminsevenofthe eight patients between 3 and 6 months after therapy, followed by a signal regmxsion 6-18 months after irradiation. Condusion: MR imaging is a sensitiveimaging tool to evaluatetumor respome asweIl asthe presenceorabsenceofadjaamt ~~changesf~~radiosurgery..LossdhomogewousorcentralcontrastenhaneementoaGdeahancedMR imageaappeared to be a good prognosticsign for tumor response Tumor shrinkage seemsnot to be dependent on time. In addition, most casesof radiation induced changesin normal brain parenchyma observedon T,-weighted imagesseemto be sdflhnitd 01997Ekevier!kienceIn~ Keywords: Radiosurgery; MR imaging; Brain metastases.
INTRODUCTION Brain metastases are a major problem in neuro-oncology, accounting for 43% of all intracranial tumors.21 Stereotactic radiosurgery using charged particle beams, modified linear accelerators, or gamma knife units has been increasingly employed for the treatment of multiple, inoperable, or recurrent brain metastases.“‘13 Successful local tumor control has clinically been achieved in more than 80% of patients treated by this modality.” With the
rapid increase in the numbers of radiosurgical centers, it becomes more important to learn about the pattern of response and the temporal evolution of the irradiated metastases and radiation induced changes of normal brain parenchyma adjacent to the irradiated target volume. Modem imaging techniques have made in vivo studies of the reaction of the tumor as well as the adjacent normal brain tissue feasible. A number of reports on computed tomography (CT) and magnetic resonance
RECEIVED 47197; ACYCEPTED7112197. Address correspondence to: Hans Hawighorst, MD, Department of Radiology, German Cancer Research Center (dkfz), Im
Neuenheimer Feld 280, D-69120 Heidelberg, Germany. Email: [email protected] 1121
MagneticResonanceImaging0 Volume 15, Number 10,
1122
(MR) imaging of radiation induced changes following radiotherapy are now available.14.17.19 For various reasons, however, the results of these reports may not be applicable to radiosurgery. The different reactions of a pathological process to single as opposed to fractionated radiation most likely extend also to adverse reaction of the normal brain. Despite the current clinical interest and usage, no report has systematically documented the postoperative neuroimaging findings in patients with brain metastases treated by means of stereotactic radiosurgery. Therefore, the aim of this study was to evaluate the spatiotemporal evolution of radiosurgical induced changes both in metastases and in normal adjacent brain tissue adjacent to the lesions by serial MR imaging.
1997
two per patient). The target volume was defined by the contrast-enhancing tumor region enlarged by a 1 mm safety margin. An example of a 80% isocenter dose plan is depicted in Fig. 1. All patients received prophylactic steroids (20 mg dexamethason) 1-3 h before and 8 h after irradiation.
Patients From January 1988 to July 1996, 359 patients with intracranial metastases from different primary tumors were treated by stereotactic radiosurgery. However, of these only 25 patients (7%) had a complete MR followup, i.e., MR examinations prior to and at different points in time after radiosurgery. This selected population formed the basis of the present study. Patients were treated with single high-dose irradiation technique either as primary treatment (n = lo), as a treatment for a relapse (n = l), or within a combination of planned whole-brain irradiation and a radiosurgical boost therapy (n = 14). The series included 12 females and 13 males (age 52 + 11 years). All patients had a histologically confirmed primary tumor with one or more metastatic brain lesions. Eighteen patients had a solitary metastasis and 7 patients had two or three brain metastases without other widespread tumor disease. The patient data as well as the location, histology, and clinical presentation of intracranial metastases are summarized in Table 1.
MR Imaging In addition to the base-line examination, which was also used for the purpose of radiation treatment planning, all patients were followed both by neurologic and contrast-enhanced MR imaging examinations at approximately 3 monthly intervals over the first 2 years. For patients surviving the first 2 years, the follow-up timing was tailored to the individual. The mean time of MR imaging follow-up was 9 ? 5 months. MR measurements were performed on a 1.5-Tesla whole-body MR system (MAGNETOM SP 4000, Siemens AG, Erlangen, Germany). For radiofrequency transmission and detection, a circular polarized head coil with an internal diameter of 3 1 cm was used. In a first step, sagittal fast low angle shot (FLASH) images (repetition time (TR)/echo time (TE) = 30/10 ms; flip = 80’; matrix size = 128 X 128; field of view = 260; slice thickness = 6 mm; one acquisition) were acquired to get an orientation. For treatment planning and follow-up, identical contiguousslice T,-weighted spin-echo (SE) and T,-weighted SE images were measured in axial orientation (TRITE = 600/15 ms and 2800/19,93 ms, respectively, matrix size = 256 X 256, field of view = 230-270 mm, section thickness = 4 mm, one to two acquisitions). T,-weighted contrast-enhanced MR images with identical imaging parameters were obtained in the axial as well as in a coronal or sagittal plane approximately 2 min after i.v. injection of gadolinium dimeglumine (Magnevist@, Schering AG, Berlin, Germany; 0.1 mmol/kg body weight).
Treatment Technique Detailed descriptions of dose planning and the treatment technique using linear accelerator radiosurgery (wherein multiple gantry rotations at different table angles produce multiple non-coplanar converging arcs) have been published previously.3~5*g*15The maximum dose at the isocenter varied from lo-25 Gy (mean dose, 16 + 4 Gy). The outline of the target volume was within the 80% isodose contour in all cases. The delivered dosage depended on localization and radiosensitivity of critical structures in the vicinity of the tumor, and on the fact of whether the patients received radiosurgery as part of the initial whole or partial brain irradiation (13 patients; dose 39 ? 8 Gy). The mean field size of the collimator was 21.5 2 10 mm in diameter. Multiple irradiation isocenters were used in seven cases (average,
Data Evaluation MR image readings was separately done by two experienced radiologists (H.H., M.E.) and one radiotherapist (R.E.) using the same criteria. In all cases, the maximum tumor extension was measured on axial precontrast and on contrast-enhanced T,-weighted SE images directly from film using fine calipers. In addition, the maximum extension of parenchymal changes surrounding the metastases was measured on T,-weighted images. Based on contrast-enhanced T,-weighted MR images, the following definitions were used to radiologically classify tumor response:4 stable disease was assumed if the average tumor diameter after treatment did not show a shrinkage of more than 50% or an increase of less than 25%, partial remission as a shrinkage of tumor size of more than 50%. A disappearance of the contrast-
METHODS
AND MATERIALS
PR S
R M M N B B B Br M
22
12 20 20 15 20 20
10 25
18
Parietal lobe
Parietal lobe Cerebellum Temporal lobe Temporal lobe Parietal lobe Occipital lobe
Splenium Orbit
Thalamus
16
17 18 19 20 21 22
23 24
25
Note. B = Bronchial Contrast-enhanced;
PR PR S S S PR
Br
12
Frontal lobe
15
PR
CR
PR
PR
2.5
3.2 1.4
3.9 1.3 0.7 0.9 1.6 1.1
2.1
2.3
3.0
2.5
2.8
2.1 1.0
carcinoma; M = Melanoma; R = Renal carcinoma; Br = Breast carcinoma; *the numbers in brackets give the period (in months) at which the changes
B
M
15
PR
S
PR S
Frontal lobe
13
B M
0, 3, 6 0, 3, 6
3, 3 3, 3, 3 3,
6
6, 9 6
6, 9
data
(central loss of enhancement) (3, 9) (central loss of enhancement) (3)
complete remission (28) decreased (central loss of contrast enhancement) (6, 9) decreased (3) stable (6, 9) stable stable (3) decreased (6) decreased (central loss of contrast enhancement) (3, 6) stable (15) decreased (central loss of contrast enhancement) (6)
decreased (6, 12)
stable decreased contrast decreased contrast
increased (3, 6) increased (3, 12) increased (3) stable complete remission (3) stable (3, 6) decreased (3, 6) decreased (central loss of enhancement) (3) stable decreased (central loss of enhancement) (6) stable
*Change in size on Cd-CE image*
and MR imaging
P = Progress; S = Stable; PR = Partial remission; CR = Complete in tumor size or signal on T,-or T,-weighted images occurred.
0, 6, 12, 15, 18
0, 3, 6, 12 0, 3, 12, 15
0, 0, 0, 0, 0, 0,
0, 3, 6, 9, 12, 18, 28
0, 3, 6, 9, 12, 15
0, 3
0, 3, 6, 9
0, 3, 6
0, 3, 6 0, 3, 6
6 6, 9, 12 6 6 6 6 6
0.9 1.6
PR S
3, 3, 3, 3, 3, 3, 3,
0, 0, 0, 0, 0, 0, 0,
0.6 2.2 3.0 2.8 1.0 2.5 1.2
14
12
15 15
B Br
radiosurgical,
Time of MR (mo)
P P P S CR S PR
Br
Parietal lobe
10 11
15 12
M R B B B M B
12
Cerebellum Frontal lobe
8 9
20 15
18
20 20 15 15
Parietal lobe
Vermis Frontal lobe
1 2 3 4 5 6 7
Histology
Tumor size (cm)
Radiological response
M
Parietal lobe Cerebellum Frontal lobe Temporal lobe Cerebellum Occipital lobe Frontal lobe
Case
of brain metastases in 25 patients: clinical,
12
Location of metastases
Radiosurgery dose (80% isodose) (GY)
Table 1. Radiosurgery
remission;
mo = month;
CE =
decreased (3, 6, 9) increased (12) increased (9) decreased (6) increased (3) decreased (3) increased (3) and decreased (12) increased ( 15) increased (6, 12) and decreased (15, 18)
stable increased (3) and decreased (6, 12) increased (3, 6) and decreased (12, 18)
stable
(3)
on T2-w (MO)
increased (3, 6) increased (6) increased (3, 6) increased (3) complete remission stable (3, 6) decreased (3, 6) increased (3) and decreased (6) stable increased (3) and decreased (6) stable increased (3) and decreased (6)
*Signal
E 6 g 5 3 F
2 22 oh 9 D E P E
g
1124
MagneticResonanceImaging0 Volume 15, Number 10, 1997 changes on T,-weighted
images in brain parenchyma
adjacent the enhancing lesions. The statistical analysis was performed with use of the SAS software (SAS, Cary, NC, USA). The results of the MR image reading (tumor extension, signal changes, contrast enhancement pattern) of the three observers were separately analyzed and tested for significance. The
MR imaging data of the lesions (signal change, tumor extension, contrast enhancement pattern) measured after
radiosurgery were tested retrospectively for a difference to the values prior to radiosurgery with the paired-student t-test. A p value of less than 0.05 was accepted as
significant. RESULTS Patients As of July 1996, 11 of the 25 examined patients
(44%) had died. Fourteen patients (56%) are alive without major disability and are able to work. The overalI
mean survival time was 10.5 2 7 months, with a l-year actuarial survival rate of 40%. MR Image Reading There were no significant (p > 0.1) differences between the three observers concerning the MR image evaluation (tumor extension, signal changes, contrast enhancement pattern).
Tumor Size According to the MR imaging criteria defined above,
stable disease was observed in 15 patients (60%), partial remission in 5 patients (20%) (example in Fig. 2) and complete remission in 2 patients (8%) (example in Figs. 3,4). In the five patients with partial remission, the mean tumor size decrease was 10 mm + 3.5 mm. Three patients who had a deterioration of neurologic function
also showed an increase in the contrast-enhancing tumor region (example in Fig. 5) with a mean increase of 7.9 2 (W Fig. 1. Isodose cutyes are superimposed on a contrast-enhanced (a) axial and (b) sag&al MR image. The intracranial metastasisin a 56-year-old male was treated with stereotactic radiosurgery (20 Gy to the mmor margin at the 80% isodose
line). Note that the inner circle representsthe 80%. the following circles the 60%, 40%, and 20% isodose, respectively.The gray dots represent the target volume defined by the radiotherapist.
3.0 mm. Changes in tumor size were not significantly
associated with the initial tumor size, maximal tumor dose, additional whole brain irradiation or histology of the primary tumor (Table 1).
Tumor Enhancement In seven patients, the findings of stable disease or
tumor remission were accompanied by a loss of central tumor enhancement (10 metastases) and/or a decrease of contrast enhancement with blurred tumor margins between 3 months and 12 months after treatment (mean, 6 months) (Figs. 6, 7). Central or homogeneous loss of
enhancing tumor region was defined as a complete remission. An increase of tumor diameter of more than 25% on contrast-enhanced TI-weighted images was de-
contrast enhancement appeared to be a good prognostic sign for stable disease or tumor shrinkage and this asso-
fined as tumor progress. Moreover, we analysed signal
ciation was statisticalIy significant (p C 0.05). Radiolog-
Serial
MRI
of metastases0
HAWKHORST
1125
ET AL.
(Fig. 5). In two of mass was surgically ical examination of tumor cells without
these patients, the increased tumor excised. The result of the histologthe surgical specimen showed vital radiation-induced brain necrosis.
Adjacent Parenchymal Changes on T,-Weighted
(4
(W Fig. 2. Serial contrast-enhancedT,-weighted MR images of a 57-year-old-male whose metastasisfrom a bronchial carcinoma was treated with stereotacticradiosurgery (20 Gy to the tumor margin at the 80% isodoseline). (a) PreradiosurgicalMR image depicting a contrast-enhancingmetastasisat the occipital lobe. (b) Six months after radiosurgery, partial remission occurred with decreasedcontrast enhancement and regression of tumor size.
ically determined tumor progress-which in all cases correlated with a neurological deterioration-was associated with an equally or stronger contrast enhancement compared to the tumor enhancement prior to radiosurgery
Images MR imaging performed prior to radiosurgery demonstrated in all cases an area of increased signal intensity in normal brain parenchyma adjacent to the enhancing metastases on T,-weighted MR images. In patients with radiological stable disease or tumor remission, 8 of the 25 patients (32%) demonstrated an extending parenchymal field of high signal intensity on T,-weighted images after radiosurgery. In seven of these eight cases, the parenchymal changes reverted to normal (example shown in Fig. 8) and the repeated MR examinations gave some indications about the temporal evolution of radiation induced changes and its self-limiting nature in patients with stable disease or tumor remission. The progression of the high signal areas on T,-weighted MR images was seen between 3 and 6 months after radiotherapy and followed by a regression at 6-18 months after irradiation. High signal areas with extension along the white matter tracts were observed in two cases, and a high signal region which coincided well with the 20% isodose arc in three cases. In one patient with a radiological complete remission, follow-up MR images did not depict any parenchymal changes on T,-weighted images. In each case of tumor progression, MR imaging revealed a progression of the high signal intensity on T,-weighted images coinciding with an enlarging area of tumor enhancement. The appearence of new adjacent parenchymal changes following radiosurgery did not correlate significantly with tumor size, tumor margin dose, or tumor maximum dose. However, progressively developing areas of high signal intensity were more often observed in patients with progressive tumor growth, radiosurgery of multiple metastases and/or receiving whole brain irradiation as part of the treatment protocol.
DISCUSSION In 20-40% of all patients with solitary brain metastases, surgery is not possible because the lesions are localized in or near vital structures and thus can not be resected without high risk.‘6*22 In a single treatment session, radiosurgery administers a dose of radiation presumably sufficient to achieve local tumor control. The dose can be delivered to a precisely defined volume of tissue with minimal radiation exposure to the surrounding normal structures.
1126
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
04
(a)
Fig. 3. Serial contrast-enhanced T,-weighted MR images of a 54-year-old-male patient whose metastasis from a hypemephroma was treated with stereotactic radiosurgery (22 Gy to the tumor margin at the 80% isodose line). (a) Preradiosurgical MR image depicting a large contrast-enhancing metastasis adjacent to the right ventricle. (b) Three months after radiosurgery, decreased contrast enhancement and regression of tumor size was seen. (c) Nine months after radiosurgery, further marked decrease in tumor size was seen. (d) Twenty-eight months after radiosurgery, no contrast enhancement is visible corresponding to a complete remission.
Tumor Size Sturm and coworkers*’ were the first to report the efficacy of radiosurgery in the management of brain metastases. A mean dose of 18 Gy was used, and in the 28 patients for whom follow-up were available, a major radiographic response rate of 86% was documented as determined on CT scans. Engenhart et aL6 evaluated 69
with 102 inoperable brain metastases an average of 12 months after stereotactic radiosurgery. Their assessment of tumor size found a complete remission in 20%, partial remission in 35%, stable disease in 40%, and relapse in only 5% of the patients. To our knowledge no other studies were published describing systematitally the MR imaging findings following radiosurgery in
patients
Serial MRI
of
metastases 0 HAWIGHORST
03
1127
ET AL.
03
Fig. 4. Serial contrastenhanced T,-weighted MR images of a 57-year-old-male patient whose right cerehellar metastasisfrom a bronchial carcinoma was treated with stereotacticradiosurgery (18 Gy to the tumor margin at the 80% isodoseline). Preradiosurgical MR image depicting on (a) contrast-enhanced T,-weighted images a lesion with peripheral enhancement which is homogeneous hyperintense (b) on T,-weighted images. Three months after radiosurgery, complete remission was reached. Neither (c) contrastenhancement nor (d) a hyperintense lesion on Z’,-weighted image is visible anymore. patients with brain metastases. The analysis of our MR imaging data yields a local control of brain metastases manifested as stable disease or remission in 22 of 25 patients (88%) and tumor progression in 3 of 25 patients (12%). The MR imaging control rate of 88% compares well to the radiological control rates published by other authors’S6V’3,20based on CT scans.
The radiobiological mechanism by which stereotactic radiosurgery inhibits further tumor growth is yet not fully understood. Animal studies suggest that tumor cells are irreversibly damaged after a singlefraction radiation doses as low as 20-30 Gy. Kondziolka et a1.r’ developed and studied different in vivo tumor models. The tumor grafts were treated with
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
1128
icantly larger tumor diameter at autopsy, appeared as highly cellular lesions; in contrast, the treated animals, with a smaller mean tumor diameter had large areas of hypocellularity representing cell death.
(a)
Fig. 5. Serial contrast-enhanced T,-weighted MR images of a 38-year-old female whose metastasis from a bronchial carcinoma was treated with stereotactic radiosurgery (15 Gy to the tumor margin at the 80% isodose line; in addition, the patient received 33 Gy whole brain irradiation). (a) The preradiosurgical MR image depicts a large contrast-enhancing metastasis at the frontal lobe. (b) Three months after radiosurgery, tumor progression occurred with increased contrast enhancement and progression of tumor size. The patient deteriorated clinically. The increasing mass was surgically excised. The histological examination revealed vital tumor cells without radiation necrosis.
different high doses of irradiation (15-50 Gy), compared to untreated controls and evaluation was performed at death. The untreated controls, with a signif-
(W Fig. 6. Serial contrast-enhanced Tr-weighted MR images of a 57-year-old female whose metastasis from a bronchial carcinoma was treated with stereotactic radiosurgery (18 Gy to the tumor margin at the 80% isodose line). (a) The preradiosurgical MR image depicting a large and homogeneously contrastenhancing metastasis in the left thalamus. (b) Three months after radiosurgery, tumor regression occurred with marked loss of central contrast enhancement.
Serial
MRI
of
metastases 0 HAWIGHORST
ET AL.
1129
(b)
Fig. 7. Serial contrast-enhanced MR images of a 44year-old female whose metastasisfrom a breast carcinoma was treated with stereotactic radiosurgery (18 Gy to the tumor margin at the 80% isodose line). (a) PreradiosurgicalMR image depicting a large and homogeneously contrast-enhancingmetastasisin the left frontal lobe. (b) Six months after radiosurgery, tumor regressionoccurred with marked loss of central contrast enhancement. (c) Six months after radiosurgery, further decreaseof tumor size occurred. (d) Twelve months after radiosurgery, partial tumor remission had occurred.
Tumor Enhancement In our series, postoperative MRI examinations often revealed a loss of central or homogeneous contrast enhancement which was observed 3 to 12 months after radiosurgery. Interestingly, this effect appeared to be a good prognostic sign for stable disease or tumor remis-
sion. To better define the impact of single high dose irradiation, Spiegelmann et al.” developed a cat model to study their triphasic hypothesis. According to this concept, radiosurgical masses progress through different characteristic phases: first there is acute degenerative necrosis caused by cell death and vascular injury; sec-
1130
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
(4
(cl
(d)
Fig. 8. Serial MR images of same case as in Fig. 6 show the development and regression of radiation induced changes in brain parenchyma adjacent to the irradiated metastasis.(a) PreradiosurgicalT,-weighted MR image revealsa small hyperintenserim in the left thalamus adjacent to the brain metastases.(b) The Z’,-weighted MR image obtained at 6 months after radiosurgery shows an extending parenchymal field adjacent to the metastases.(c) A further increase of the high signal field is visible at 12 months after radiosurgery. Note that the irradiated metastasesshows shrinkage.No steroidswere administered as patient was clinically free of any symptoms. (d) I’,-weighted
MR image 18 months after radiosurgery
ond, this necrotic mass is phagocytosed, and finally a prominent glial scar develops. To test this hypothesis, they irradiated 19 cats with different radiosurgical doses and evaluated sequentially with contrast-enhanced MR imaging and pathologic assessment. No central contrast enhancement was found after approximately 3 to 6
showing marked regression of the high signal lesion.
weeks. This area corresponded pathologically to necrotic debris and vascular damage. Within 12 to 29 weeks, the animals improved clinically, and the MR imaging findings revealed resorption of debris characterized by reduced lesion size and contrast enhancement. Histologitally, diminished vascular&y and gliosis predominated.
Serial
MRI of metastasesOHAw~~o~s~
Based on these observations, our findings of central loss of contrast enhancement in 10 metastases may be explained as radiation-induced vascular injury and tumor cell death. Obliteration of blood supply to the metastatic lesion may be one of the most significant mechanisms controlling tumor growth after radiosurgery. We conclude that homogeneous or central loss of contrast enhancement on T,-weighted MR images appears to be an early and good prognostic sign to predict tumor response.
1131
ET AL.
which are not characterized by contrast enhancement. Furthermore, these radiation induced changes are most
often self-limited, rarely accompanied by clinical symptoms and do not require the use of steroids. In contrast, progression of high signal intensity on T,-weighted im-
ages with an enlarging area of contrast enhancing tumor is most likely caused by tumor progression, does require
the use of steroids, and shows clinical manifestations. Regression of the tissue changes was seen at 6-18
months after its appearance. As steroids may alter the Adjacent Parenchymal Changes on T,-weighted
Images In this study, parenchymal changes adjacent to contrast enhancing lesions manifested as focal high signal intensity on T,-weighted MR images and were seen in 8 of 25 patients with stable disease or partial remission. The cause of high signal regions on T,-weighted images without contrast enhancement in areas of the brain adjacent to the tumor remains controversial. The most likely cause is an edema and/or blood-brain barrier (BBB) disturbance in the adjacent brain parenchyma due to radiosurgery and not due to tumor growth. This hypothesis is supported by the absence of hypointensity on T,-weighted images, tbe temporary nature of the effect in the majority of cases, and only few, if any, newly associated neurologic symptoms despite their appearance in critical brain areas, and lacking increase of tumor size on contrast-enhanced images. The most likely explanation for the high signal aberration on T,-weighted images seems to be the transient BBB damage or edema from injured brain tissue. Demyelination is unlikely to appear outside the volumes irradiated by this imaging pattern. Thus, edema dominated the radiation induced changes in brain parenchyma between 3-6 months in the current series following radiosurgery. These interpretations are supported by the observations previously reported from animal experiments. Fike et al.’ described the occurrence of radiation edema with corresponding CT changes, Lo and colleagues12 studied the impact of high-dose, singlefraction brain irradiation in the rabbit brain using MR imaging and positron emission tomography. Within 10 months of delivering 30 Gy helium ion irradiation, BBB disturbances occurred that manifested as an early high T2 signal abnormality without contrast enhancement on MR
course of radiation induced changes, it is of particular interest to follow the serial MR images in six patients
who had a non-symptomatic adverse reaction and did not receive such treatment. In two patients, radiation induced
changes came to a stable and persistent high signal as compared to the preradiosurgical MR images. These late
appearing and persisting high signal lesions may represent demyelination and/or gliosis. Conclusion MR imaging is necessary for adequately assessing the
response of brain metastases to radiosurgery. Our analysis suggests a high initial tumor control rate of 88%
with a mean follow-up of 9 months. Loss of homogeneous or central contrast enhancement on Gd-MR ap-
pears to be a good prognostic sign for tumor response. In addition, most cases of radiation induced changes in normal brain parenchyma seem to be self limited. Further
studies with histological proof are needed to confirm the signal changes observed on MR examinations.
REFERENCES 1. Alexander, E.; Moriarty, T.M.; Davis, R.B.; Wen, P.Y.;
2. 3.
4.
images. In addition, field size increased significantly. This damage probably reflects early damage on endothe-
lial cells. In addition, higher irradiation doses may result in disruptions of the BBB. Grossmann et al.* conducted
5.
a study in cats, delivering 50 Gy with a 1.5 cm collima-
tor, and within 7 months, observed focal areas of BBB disruption,
using gadolinium-enhanced
MR imaging.
From the results of this study, we conclude that even with lower doses of 16 Gy radiosurgical induced paren-
chyma changes may be detected on T,-weighted images
6.
Fine, H.A.; Black, P.M.; Kooy, H.M.; Loeffler, J.S. Stereotactic radiosurgery for the definitive, noninvasive treatment of brain metastases.J. Nat. Cancer Ins. 87:34-40; 1995. Altschuler, E.M.; Lunsford, L.D.; Kondziolka, D., et al. Radiobiologic models for radiosurgery. Neurosurg. Clin. North Am. 3:61-77; 1992. Bauer-Kirpes, B.; Schlegel, W.; Boesecke, R.; Lorenz, W.J. Display of organs and isodosesas shaded 3-D objects for 3-D therapy planning. Int. J. Radiat. Oncol. Biol. Phys. 13:135-140; 1987. DeVita, V.T. Principles of chemotherapy. In: V.T. DeVita, S. Hellmann, S. Rosenberg, (Eds). Cancer. Principles and Practice of Oncology. Philadelphia: Lippincott Co, 1993: 276-289. Engenhart, R.; Kimmig, B.; Pastyr, 0.; Marin-Graz, M.; van Kaick, G. Entwicklung eines reproduzierbaren Fixationssystems fiir perkutane Bestrahlung im Kopf-Halsbereich. Zentralbl. Rad. 136:731-732; 1988. Engenhart, R.; Kimmig, B.N.; Hover, K.H., et al. Longterm follow-up for brain metastasestreated by percutaneous stereotactic single high-dose irradiation. Cancer 7 1: 1353-1361; 1993.
1132
Magnetic Resonance Imaging 0 Volume 15, Number 10, 1997
7. Fike, J.R.; Cann, C.E.; Turowski, K., et al. Radiation responseof normal brain. Int. J. Radiat. Oncol. Biol. Phys. 14:63-70; 1988. 8. Grossmann,RI.; Hecht-Leavitt, C.M.; Evans, SM., et al. Experimental radiation injury: Combined MR imaging and spectroscopy.Radiology 169:305-309; 1988. 9. Hartmann, G.H.; Schlegel, W.; Strum, V.; Kober, B.; Pastyr, 0.; Lorenz, W.J. Cerebral radiation surgery using moving field irradiation at a linear acceleratorfacility. Int. J. Radiat. Oncol. Biol. Phys. 11:1185-I 192; 1985. 10. Kondziolka, D.; Lunsford, L.D.; Claassen,D., et al. Radiobiology of radiosurgery: Part II. The rat C6 glioma model. Neurosurgery 31:280-287; 1992. 11. Konnov, B.; Melnikov, L.; Zargrova, 0. Narrow proton beam therapy for intracranial lesions. In: International Workshop on Proton and Narrow Photon Beam Therapy, Oulu, Finland; 1989:48-55. 12. Lo, E.H.; Fabrikant, J.L. Delayed biologic reactions to stereotactic charged particle radiosurgery in the human brain. Stereotact. Funct. Neurosurg. 56:197-212; 1991. 13. Loeffler, J.S.; Kooy, H.M.; Wen, P.Y.; Fine, H.A.; Cheng, C.W.; Mannarino, E.G.; Tsai, J.S.; Alexander, E. The treatment of recurrent brain metastaseswith stereotactic radiosurgery. J. Clin. Oncol. 8:576-582; 1990. 14. Miot, E.; Hoffschir, D.; Pontvert, D.; Gaboriaud, G.; Alapetite, C.; Masse, R.; Fetissof, F.; Le Pape, A.; Akoka, S. Quantitative magnetic resonance and isotropic imaging: Early evaluation of radiation injury to the brain. Int. J. Radiation Oncology Biol. Phys. 32:121-128; 1995.
15. Pastyr, 0.; Hartmann, G.H.; Schlegel, W.; Schabbert, S.; Treuer, H.; Lorenz, W.J. Stereotactically guided convergent beam irradiation with a linear accelerator: Localization technique. Acta. Neurochir. 99:61-64; 1989. 16. Posner, J.B. Management of central nervous system metastases.Semin. Oncol. 4:81-91; 1977. 17. Rubin, P.; Gash, D.M.; Hansen, J.T.; Nelson, D.F.; Williams, J.P. Disruption of the blood-brain barrier as the primary effect of irradiation. Radiother. Oncol. 31:51-60, 1994. 18. Spiegelmann, R.; Friedmann, W.A.; Bova, F.J., et al. LINAC radiosurgery: An animal model. J. Neurosurg. 78: 638-644; 1993. 19. Steinberg, G.K.; Fabrikant, J.I.; Marks, M.P.; Levy, R.P.; Frankel, K.A.; Phillips, M.H.; Shuer, L.M.; Silverberg, G.D. Stereotacticheavy-charged particle Bragg peak radiation for intracranial arteriovenous malformations. N. Eng. J. Med. 323:96-101; 1990. 20. Stmm, V.; Kober, B.; Hover, K.H., et al. Stereotactic percutaneous single irradiation of brain metastaseswith a linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 13: 279-282; 1987. 21. Walker, A.E.; Robins, M.; Weinfeld, F.D. Epidemiology of brain tumors: The national survey of intracranial neoplasms. Neurology 35:219-226; 1985. 22. West, J.; Maor, M. Intracranial metastases:Behavioural patterns related to primary site and results of treatment by whole brain irradiation. Int. J. Radiat. Oncol. Biol. Phys. 6:11-15; 1980.