| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Radiation Oncology |
1 From the Divisions of Radiation Oncology (J.A.K., Z.I., A.G.K., T.B., M.H.M.), Pediatric Oncology (S.G.), and Pediatric Neurosurgery (T.T.), Northwestern Memorial Hospital, 251 E Huron St, L-178, Chicago, IL 60611. From the 1999 RSNA scientific assembly. Received December 1, 1999; revision requested January 28, 2000; final revision received April 3; accepted May 11. Address correspondence to J.A.K..
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: In this retrospective analysis, stereotactic frame tolerance was assessed by recording patient discomfort or pain in the ear and nose during each treatment in 34 patients, including 21 children and 13 adults with 37 lesions treated with fractionated stereotactic radiation therapy. Radiation doses ranged from 10–60 Gy at 1.0–4.0 Gy per fraction. Repositioning accuracy was assessed by comparing portal radiographs with setup fields on computed tomographic (CT) scout images. Clinical efficacy was assessed by analyzing posttreatment CT and magnetic resonance images.
RESULTS: The stereotactic localizer was well tolerated. The mean isocenter shifts observed after studying 305 portal radiographs were x-coordinate shift of 1.0 mm ± 0.7 (SD), y-coordinate shift of 0.8 mm ± 0.8, and z-coordinate shift of 1.7 mm ± 1.0. At a median follow-up of 16 months, local control was achieved in 18 of 22 primary and in one of eight of recurrent tumors.
CONCLUSION: The Laitinen stereotactic localizer is well tolerated with accurate reproducibility during stereotactic radiation therapy. Preliminary local control rates are consistent with those in other reports.
Index terms: Brain neoplasms, 10.31, 10.32, 10.33, 10.36 • Brain neoplasms, therapeutic radiology, 10.1267, 10.1269 • Computed tomography (CT), treatment planning, 10.12112, 10.1267 • Magnetic resonance (MR), treatment planning, 10.121412, 10.1267 • Radiations, protective and therapeutic agents and devices • Stereotaxis, 10.1267
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Patient and Tumor Characteristics
Between 1995 and 1999, 34 patients with 37 benign or malignant brain and/or eye tumors were treated with SRS or SRT at Northwestern Memorial Hospital in Chicago, Ill (Table). There were 21 children and adolescents with 23 lesions and 13 adults with 14 lesions. The mean and median age of children and adolescents were both 12 years (range, 22 months to 18 years). The mean and median age of adults were both 40 years. (range, 19–77 years). The median follow up was 16 months (range, 0–44 months).
|
Immobilization and Target Localization
Since September 1995, we have used the Laitinen stereotactic frame for immobilization and stereotactic localization (Fig 1). This device has been approved by the U.S. Food and Drug Administration for clinical use. It consists of a noninvasive frame made of reinforced plastic and aluminum alloy that is mounted to the patient’s head by means of two earplugs (M, Fig 1) and a nasion support (E, Fig 1). Two cogwheel cases (F, Fig 1) serve to press the nasion support against the bridge of the nose. A threaded screw (D, Fig 1) at the nasion support serves to press the earplugs in both external auditory meatus. The lateral triangular components (G, Fig 1) of reinforced plastic are pressed tightly against the scalp by means of a connector plate over the vertex. A flexible band strapped against the occiput can be used for additional immobilization, if necessary. The lateral triangular component has four transverse bars at a distance of 25 mm from each other. The bars have a dorsoventral thickness of 2 mm. The nasion support arms (F, Fig 1) and the connector plate (I, Fig 1) are supplied with millimeter scales to facilitate exact repositioning of the adapter. A frontal pin joins the two halves of the nasion support assembly. The reference structures of the stereotactic frame are the frontal pin for the lateral x coordinate, a line on the anterior edge of the posterior bars of the lateral triangular components for the anteroposterior y coordinate, and the center of the most inferior transverse bars for the longitudinal z coordinate.
|
|
|
All patients underwent a "fitting session" in the department prior to CT or MR imaging to ascertain the appropriate sizes for the headrest, ear plugs, and settings on the frame. Individual patient data were then recorded on the stereotactic frame data sheet. The Laitinen stereotactic frame is both CT and MR compatible. All patients’ treatments were planned by using CT scans, so that the isocenter could be localized on scout views relative to radiologic landmarks. MR images were obtained when additional diagnostic information was required.
On the day of planning CT, the CT adapter was attached to the scanner couch. The patient was prepared to receive contrast material and was laid on the preselected headrest. The frame was then repositioned on the patient by using the settings from the fitting session. Adjustments were made to head tilt by using polystyrene foam slabs of varying thickness under the headrest to make the posterior ear bars as horizontal as possible to minimize the CT gantry tilt angle. The frame and patient were then squared up with the line lasers of the scanner and were rigidly attached to the couch by means of a multijoint system on the CT adapter insert. A lateral scout view was acquired, from which the gantry tilt angle (<3°) was determined such that the CT sections would be parallel to the transverse bars. This procedure took about 5 minutes.
After administration of contrast material (iopromide, Ultravist; Berlex Laboratories, Wayne, NJ), CT scanning was performed. For CT scans, the patients underwent imaging with a Quick CT scanner (GE Medical Systems) by using the routine head scanning technique (120kVp, 170 mAs, 2 seconds per section, 25-cm field of view). The thinnest section was 3 mm. The data (512 x 512 matrix) were transferred to the treatment-planning computer by means of nine-track magnetic tape. Hard copies ("one on one") of anteroposterior and lateral scout views were obtained to be used for verification of isocenter location.
For MR imaging, the three-dimensional volume was positioned on localizer images to include all aspects of the stereotactic frame. High-spatial-resolution T1-weighted anatomic MR images were acquired (Vision; Siemens Medical Systems, Iselin, NJ) by using a three-dimensional fast low-angle shot sequence with the following imaging parameters: repetition time msec/echo time msec, 15/5.6; flip angle, 20°; field of view, 256 mm; matrix, 256 x 256; section thickness, 1.0 mm. All anatomic sections were obtained in the transverse plane and were transferred to the planning workstation via the hospital network.
Treatment Planning
CT scans and/or MR images obtained with the Laitinen stereotactic frame were utilized for SRT planning by two of the authors (A.G.K., T.B.). Treatment planning was performed with a three-dimensional treatment-planning computer system (Plato; Nucletron, Columbia, Md). Noncoplanar static or rotational beams were used, and the prescribed isodose lines ranged from 90%–100%. All patients were treated with one isocenter for a given lesion. The isocenter was localized on the lateral and anteroposterior scout images, and a localization field was drawn on each. The y and z coordinates were drawn on the target plates, and the x coordinate was referenced to the patient’s midline. The field size varied from 2.5 to 6.0 cm. The number of radiation fields varied from three to eight, with a mean of five. All treatments were delivered by using 6- or 10-MV x rays (model SL 15 linear accelerator; Elekta Instruments, Norcross, Ga).
Treatment Setup and Verification
For each treatment, a couch adapter was attached to the treatment table (Fig 1a). The Laitinen stereotactic frame was mounted on the patient in a fashion similar to the setup in the CT scanner or MR imager. Adjustments were made to level the posterior ear bars relative to the side-localizing lasers before rigidly affixing the frame to the couch adapter by means of a multijoint mechanism. To facilitate target setup, white plastic plates were attached to each of the triangle components of the adapter. The y and z coordinates of the isocenter were marked on the target plates.
The projection of the isocenter on each target plate was then lined up with the side-localization lasers by adjusting the treatment couch. Couch rotation about its midpoint (not about the isocenter) facilitated squaring up of the frame (and patient) relative to the axis of gantry rotation. The mismatch between the target points on the lateral plates and the crosses defined by the side-localization lasers was corrected by using tabletop rotation to correct for half of the error and table in-out motion to correct for the other half of the error. Within two iterations, the patient was on target in the y-z plane. The lateral x coordinate was set relative to the midline of the patient on a linear millimeter scale that was placed perpendicularly on the frontal pin of the adapter. By moving the couch laterally, the frontal laser was brought to coincide with the lateral x coordinate of the target. Thus, the target center was positioned at the isocenter of the linear accelerator.
Orthogonal localization radiographs were obtained before the first treatment and then twice weekly. During each treatment, either the radiation physicist (A.G.K.) or the dosimetrist (T.B.) was present to assist with treatment setup. During treatment, the patient was continuously monitored on a television screen, and no patient required an interruption of treatment because of head motion or discomfort involving the stereotactic frame.
Repositioning Accuracy
The accuracy of the setup was assessed by comparing the portal radiographs obtained during SRT with the setup fields drawn on the CT scout images. The portal radiographs were analyzed in consensus by three of the authors (J.A.K., Z.I., and A.G.K.) to determine the reproducibility of the setup. A transparency was placed over the CT or MR scout image. The central axis and three points of reference at fixed bone landmarks were marked on this image. A line was drawn from the central axis to each point and beyond. Extension of the lines beyond the points allowed coregistration of landmarks independent of magnification. The transparency was superimposed on each portal image. The bone landmarks were then coregistered between the scout and portal images. On the basis of this coregistration, the shift in the central axis was measured directly by using a graduated ruler in the x, y, and z dimensions. This shift measurement had a resolution of 1 mm. Owing to magnification on the portal radiograph (approximately x1.4), this translated to approximately 0.7 mm on the patient.
Tolerance
Patient tolerance to the stereotactic frame was assessed by recording patient discomfort or pain in the ear and nose during each SRT treatment. No specific patient discomfort scales were utilized.
Follow-up
All children were followed up at the multidisciplinary clinic of Northwestern and Children’s Memorial Hospital. Complete neurologic evaluation and MR and/or CT images were obtained at approximately 3-month intervals to assess the response to treatment.
Efficacy
The efficacy of treatment was assessed by comparing manual measurement of tumor dimensions on pre- and posttreatment CT or MR images. Local tumor control was considered to have been achieved if tumor size remained stable or decreased at last follow-up.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Accuracy of Stereotactic Frame Repositioning
A total of 305 portal radiographs, including 160 lateral and 145 anteroposterior portal radiographs, were obtained during the study. The shifts in isocenter in the three coordinates were (a) x-coordinate shift, 0–3.5 mm (mean, 1.0 mm ± 0.7 [SD]); (b) y-coordinate shift, 0–3.0 mm (mean, 0.8 mm ± 0.8); and (c) z-coordinate shift, 0–3.5 mm (mean, 1.7 mm ± 1.0). The SD represents the precision with which patient setup could be performed daily, as measured by the observer (J.A.K., Z.I., A.G.K.).
Local Control and Sequelae
At last follow-up, local control was achieved in five of five patients with craniopharyngioma, in one of four patients with recurrent medulloblastoma, in two of three patients with low-grade glioma, in neither of two patients with high-grade glioma, in both patients with mixed germ cell tumor, in all three patients with choroidal hemangioma, in the patient with choroidal melanoma, in none of four patients with metastases, in both patients with meningioma, in the patient with jugular glomus, in the patient with esthesioneuroblastoma, and in the patient with orbital rhabdomyosarcoma; local control was not achieved in the patient with retinoblastoma. The local control rate was 82% (18 of 22) for primary tumors and 12% (one of eight) for recurrent tumors. These local control rates are preliminary, given the short follow-up. Follow-up in four children was not sufficient to allow comment on local control.
Seventeen patients were followed up for more than 2 years, and at last follow-up there was no evidence of radiation necrosis or radiation-induced damage to cranial nerves in any patient. The incidence of radiation-induced hypopituitarism could not be assessed in the six children with craniopharyngioma because all six had panhypopituitarism after radical surgery. At the time of this writing, none of the children had developed radiation-induced cataract.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
|
For SRS, the stereotactic frame is fixed to the skull of the patient to maximize treatment accuracy and minimize errors due to patient motion. The Leksell frame (Elekta Instruments) and the Brown-Roberts-Wells frame (Radionics, Burlington, Mass) are used for SRS with a gamma knife and linear accelerator, respectively (2,3). In a study at the University of Pittsburgh (23), errors in stereotactic coordinate settings occurred in 24 of 200 (12%) isocenters that were studied. This constituted 2.2% (n = 26) of a total of 1,200 individual coordinate settings. Errors in the x-coordinate setting were identified in only one of 400 cases, and this was a 0.25-mm error, resulting in an isocenter displacement of 0.125 mm. A total of eight of 400 y-coordinate and 17 of 400 z-coordinate setting errors greater than 0.25 mm were identified and corrected. There was a significantly higher incidence of y- and z-coordinate errors, as compared with the incidence of x-coordinate errors.
The Joint Center for Radiation Therapy (3) has reported a maximum mechanical error for linear accelerator radiosurgery of 0.90 mm ± 0.30, with a mean of 0.30 mm ± 0.30. Thus, for high-dose single-fraction radiosurgery, invasive stereotactic frames provide accurate patient positioning and precise dose delivery.
For SRT, noninvasive relocatable frames are utilized to provide stereotactic treatment planning capability and accurate daily repositioning during treatment. A number of relocatable frames have been used for SRT, including the Gill-Thomas-Cosman (GTC) frame (Radionics), the Boston Children’s frame (4), and the Laitinen frame (5,6) used in the present study. The GTC frame has an aluminum alloy base ring compatible with the Brown-Roberts-Wells stereotactic system. Mechanical reference for accurate repositioning is provided by an individually constructed oral appliance and an occipital support. The Boston Children’s frame uses ear canal supports because the mouthpiece cannot be reliably used in children. The frame derives its fixation from the glabellar spectacle frame and occipital head supports. Unlike the Laitinen stereotactic frame, the GTC treatment protocol is not reliant on portal radiographs for verification of the accuracy of helmet placement during treatment. Instead, a depth conformation helmet is utilized to measure the depth from 25 fixed positions to the cranial surface in the Brown-Roberts-Wells frame coordinate system. The depth measurements from the depth conformation helmet in 20 patients using the GTC frame indicates a mean measurement error of 0.71 mm ± 0.06 (range, 0.31–1.22 mm), mean lateral movement of 0.35 mm ± 0.06 (range, 0.07–0.79 mm), mean superior movement of 0.52 mm ± 0.09 (range, 0–1.77 mm), and mean occipital movement of 0.34 mm ± 0.09 (range, 0–1.30 mm). The maximum range in individual measurements was 3.4 mm ± 0.6 (4).
As described earlier, the Laitinen stereotactic frame consists of a metal frame that is mounted to the patient’s head by means of two earplugs and a nasion support. This frame has shown a high degree of reproducibility, with an error of no more than 1–2 mm in all coordinates in phantom experiments, volunteers, and patients (5,7,24–26). In a series of 10 patients in whom repositioning accuracy was analyzed for 50 setups, the mean distance errors in the three coordinates were 1.82 mm ± 0.7 for x coordinates, 1.5 mm ± 0.5 for y coordinates, and 1.9 mm ± 0.9 for z coordinates (24). An important advantage of the Laitinen stereotactic frame over the GTC frame is its compatibility with MR imagers and lack of need for CT and MR image fusion software. Another advantage is the ability to treat low-lying targets such as tumors below the base of the skull.
One of the advantages of this study was that, to our knowledge, it constituted the largest experience in the United States with the Laitinen stereotactic frame for SRT. The previously reported (24) repositioning accuracy of the stereotactic frame was confirmed by measuring isocenter shifts on over 300 portal radiographs. This frame was well tolerated by the patients, as was demonstrated by the lack of need for anesthesia or sedation in any of the children, including the youngest (22-month-old) child. The local control rates for the different tumors in our series were similar to those in other published reports (13,16–18,20–22). A limitation of this analysis was the lack of long-term follow-up for assessment of efficacy and complications of stereotactic frame–based SRT. Our results demonstrated that the Laitinen stereotactic frame is well tolerated, with accurate reproducibility of patient fixation during SRT in children and adults with brain and eye tumors.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, J.A.K., S.G., M.H.M., T.T.; study concepts and design, J.A.K., M.H.M., A.G.K.; definition of intellectual content, J.A.K., M.H.M.; literature research, J.A.K., Z.I., M.H.M.; clinical studies, J.A.K., M.H.M., T.T., S.G., A.G.K., T.B.; data acquisition, J.A.K., Z.I., A.G.K.; data analysis, J.A.K., Z.I., A.G.K., M.H.M., S.G., T.T.; statistical analysis, A.G.K.; manuscript preparation, all authors; manuscript editing, M.H.M., S.G., T.T., A.G.K.; manuscript review, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
