The use of quick-brain magnetic resonance imaging in the evaluation of shunt-treated hydrocephalus

Transcription

1 J Neurosurg (Pediatrics 2) 101: , 2004 The use of quick-brain magnetic resonance imaging in the evaluation of shunt-treated hydrocephalus BERMANS J. ISKANDAR, M.D., JASON M. SANSONE, B.S., JOSHUA MEDOW, M.D., AND HOWARD A. ROWLEY, M.D. Departments of Neurological Surgery and Radiology, University of Wisconsin, Madison, Wisconsin Object. Children with shunt-treated hydrocephalus are exposed to serious amounts of radiation when undergoing computerized tomography (CT) scanning. The authors report their clinical experience with single-shot fast spin echo (SSFSE) (quick-brain) magnetic resonance (MR) imaging as the modality of choice for the workup and follow up of patients in whom a shunt has been placed to treat hydrocephalus. Methods. A retrospective chart review was performed to obtain data on all cases in which a quick-brain MR image was acquired for either symptomatic workup or asymptomatic follow-up examination of shunt-treated hydrocephalus. Data regarding demographics, origin of hydrocephalus, MR imaging indications and findings, use of sedation, imaging-related complications, use of adjunctive CT scanning, details of shunt revision, and cause of shunt malfunction were collected. The authors found that SSFSE MR imaging is a sufficient, radiation-free diagnostic alternative to CT scanning that minimizes movement artifact and duration of scanning and eliminates the need for sedation. Conclusions. In light of these findings, the authors propose that quick-brain MR imaging replace CT scanning as the diagnostic modality of choice in examining and following shunt-treated patients because it offers significant advantages. KEY WORDS hydrocephalus ventriculoperitoneal shunt magnetic resonance imaging radiation exposure pediatric neurosurgery T Abbreviations used in this paper: CT = computerized tomography; MR = magnetic resonance; SSFSE = single-shot fast spin echo. HE traditional protocol for evaluating cerebrospinal fluid shunt function in patients with hydrocephalus has involved CT scanning. While its diagnostic utility is satisfactory, CT scanning possesses distinct disadvantages. Notably, the inherent radiation risk to patients is not negligible and is of greatest concern in children. 2 4, 7,8,13,21,23 Shunt-treated children with hydrocephalus are disproportionately exposed to radiation because they often require frequent CT examinations for the diagnosis of shunt malfunction. We describe our clinical experience with a radiation-free alternative to CT scanning for evaluating shunt function. Traditional MR imaging was once thought to be a viable alternative to CT scanning in following children with hydrocephalus; however, lengthy imaging times and the resultant need for sedation deterred its regular use in this population. Advancements in imaging technology have broadened clinical practice. Namely, SSFSE (quick-brain) MR imaging drastically reduces image acquisition time and thus reduces sedation requirements. Currently SSFSE is used in various clinical capacities such as establishing an intrauterine diagnosis. 1,10,16,22 In this paper, we report exclusive and successful use of SSFSE MR imaging during a 1-year period to follow all children with shunt-treated hydrocephalus. Clinical Material and Methods In April 2002, the pediatric neurosurgery service at the University of Wisconsin Hospital and Clinics changed its protocol for assessment of shunt function to include quick-brain MR imaging instead of CT scanning. After obtaining internal review board approval, we conducted a retrospective chart review of all cases in which quickbrain MR imaging was performed for either symptomatic workup or asymptomatic follow up of shunt-treated hydrocephalus (April 2002 April 2003). Data regarding demographics, origin of hydrocephalus, MR imaging indications and findings, use of sedation, imaging-related complications, use of adjunctive CT scanning, details of shunt revision, and cause of shunt malfunction were collected. The MR imaging examinations were performed using a 1.5-tesla system (GE Medical Systems, Milwaukee, WI) running standard commercially available hardware and software. The SSFSE sequence had typical parameters (TR 600 msec, TE msec, flip angle 90, Field of View cm, matrix , number of acquisitions 0.5 1, slice thickness 5 mm). Sagittal, axial, and coronal images were sequentially acquired. To obtain whole-brain 147

2 B. J. Iskandar, et al. coverage, this required approximately 20 slices per plane and resulted in an imaging duration of approximately 15 seconds per plane. The mean time for total in-magnet imaging was less than 4 minutes, which included positioning, preimaging, slice prescription, and combined imaging duration for all three planes. Results In 72 patients a total of 131 quick-brain MR images were obtained (mean 1.8 images per patient) for either symptomatic workup or asymptomatic follow-up assessment of hydrocephalus after placement of a shunt. There were 40 male (56%) and 32 female (44%) patients who ranged in age from birth to 62 years, although the vast majority of patients were children (median age 3.46 years). The origin of hydrocephalus included spina bifida in 27.8% of cases, congenital/idiopathic in 27.8%, intraventricular hemorrhage of prematurity in 19.4%, intracranial cyst in 9.7%, aqueductal stenosis in 6.9%, tumor in 4.2%, trauma in 2.8%, and meningitis in 1.4%. Symptomatic Workup Forty-two patients initially presented to the emergency department or neurosurgery clinic with symptoms of shunt malfunction, and a total of 76 quick-brain MR images were acquired. The distribution of presenting symptoms and signs is shown in Table 1 and is consistent with those described in previous reports. 6,17 Of the 76 images obtained, 27 revealed interval ventricular enlargement, 44 demonstrated slit or small ventricles, and five revealed unchanged large ventricles. Multicystic hydrocephalus was known to be present in five patients, and all twelve images obtained in these patients demonstrated excellent multiplanar visualization of the ventricular loculations. Follow-Up Assessment in Asymptomatic Patients Forty-eight patients presented for yearly follow-up assessment in the neurosurgery clinic. All were asymptomatic, and a total of 55 SSFSE MR images were acquired. All four images obtained in cases involving loculated hydrocephalus successfully revealed the ventricular compartments. In three asymptomatic patients (5.8%) SSFSE imaging demonstrated incidental interval enlargement of the ventricular system, and all underwent successful shunt revision after intraoperative confirmation of the device s malfunction. One of these presumed asymptomatic patients experienced significant postoperative improvement in memory and general energy level, indicating that there was, indeed, subtle preoperative symptomatology. Duration of the Imaging Session The quick-brain MR imaging studies required a mean of 3.4 minutes to complete. The patients were usually on the imaging table for 5 minutes or less. Intraimaging Sedation No patient required sedation for the completion of a quick-brain MR imaging session. Movement Artifact Although several children moved significantly during acquisition of the SSFSE sequence, no image demonstrated evidence of movement artifact sufficient to obscure ventricular anatomy. An example is shown in Fig. 1. In a few cases, gross movement artifact present on a single image plane or movement between imaging studies resulted in either movement-degraded images or suboptimal prescription. In these cases, our technologists repeated the pertinent study to achieve satisfactory results.we found that obtaining three-plane images of satisfactory diagnostic quality required a mean of 3.3 images per patient. This indicates that approximately 10% of the sequences had to be repeated. Because each sequence required only 15 seconds to acquire, however, the duration of the imaging session was not significantly prolonged, nor was sedation required. Shunt Revision TABLE 1 Presenting signs and symptoms in 72 patients with shunt-treated hydrocephalus Sign/Symptom Twenty-six patients underwent 49 shunt revisions after presenting for symptomatic workup. Ventricular enlargement was observed in 24 cases (49%), small or slit ventricles in 23 (47%), and unchanged large ventricles were observed in two cases (4%) on quick-brain MR imaging examination. Three patients underwent a shunt revision after presenting without symptoms to follow-up examination. In all of these patients abnormal findings were demonstrated on SSFSE MR imaging prior to revision. The number of shunt procedures reported in this study represents a fraction of the number of shunt revisions performed at our institution in the past year. This occurred because early in the year CT scanning was still used frequently for the evaluation of shunt malfunction; in addition, several patients presented after undergoing imaging examination at other facilities and consequently additional studies were not required. Computerized Tomography Examination % of Cases headache 54.8 lethargy 30.9 irritability 19.0 vomiting 14.3 bradycardia 11.9 behavioral problems 7.1 large head circumference 7.1 nausea 7.1 dysphagia 4.8 leg weakness 4.8 stridor 2.4 back pain 2.4 bulging fontanelle 2.4 CSF leak 2.4 decreased appetite 2.4 diplopia 2.4 seizures 2.4 * CSF = cerebrospinal fluid. In seven (5.3%) of the 131 studies, quick-brain MR 148

3 Quick-brain MR imaging for shunt-treated hydrocephalus imaging was deemed clinically insufficient, and adjunctive CT scanning was also performed. These studies were completed for one of four reasons: 1) the need to demonstrate the ventricular catheter clearly (two images); 2) the need for intraventricular contrast imaging to delineate communication of multicystic loculations (two images); 3) the suspicion of postoperative intraventricular hemorrhage (two images); and 4) the suspicion of postoperative pneumocephalus (one image). Discussion The introduction of SSFSE MR imaging into our protocol for the evaluation of shunt-treated hydrocephalus allowed us to reduce the radiation- and sedation-related risks to our patient population with no detectable compromise of the CT scan s traditional diagnostic utility. 20 In fact, the multiplanar capacity and higher resolution of MR imaging were advantageous in several cases involving multiloculated hydrocephalus. Additionally, this MR imaging sequence was found to be less prone to motion artifact than CT scanning. Quick-brain MR imaging is used at our institution as the standard diagnostic study for symptomatic patients who present for a workup of shunt malfunction as well as for routine evaluation of shunt function on a yearly basis in the outpatient clinic. FIG. 1. A: Snapshot images of a child moving significantly during acquisition of the SSFSE MR images. B: Sagittal, coronal, and axial images demonstrating absence of movement artifact. Radiation Exposure The radiation risk inherent in a CT scanning examination is of considerable concern, particularly in children. The effective radiation dose during a head CT examination is fourfold greater in a newborn than an adult. 9 Adverse consequences of repeated radiation exposure include cataract development, genetic effects, and, most notably, carcinogenesis. 2 4,7,8,13,21,23 In a recent study by Brenner, 3 the author estimated the lifetime risk of cancer-related mortality attributable to CT examinations as a function of age and organ system. The author reports that the lifetime cancer mortality risk attributable to a single head CT in a 1-year-old patient was 0.07%. This is only slightly greater than what would be expected due to natural background radiation. Although these statistics, however, are of minimal concern to the average patient in whom a single CT scan is obtained, they become especially important in children with hydrocephalus who require several CT examinations during their lifetime. In this population, the radiation doses become compound, and radiation-induced neoplasia becomes a substantial health risk. 7 In our study, 11 patients (15.3%) presented on multiple occasions for symptomatic workup due to shunt failure. In these cases, 42 quick-brain MR images were acquired (mean 3.8 examinations per patient). Brenner s calculations 3 would imply that if CT scanning were used instead of MR imaging, the mortality rate due to radiation-induced neoplasia would have increased by 0.27% in the past year. Furthermore, our most problematic patient presented for workup after shunt failure on 10 separate occasions within this 1-year period, each time undergoing quick-brain MR imaging. A similar calculation establishes that MR imaging instead of CT imaging eliminated what would have been a 0.7% increased lifetime risk of cancer-related death. Furthermore, if one were to obtain a yearly scan for rou- 149

4 B. J. Iskandar, et al. TABLE 2 Advantages and disadvantages of quick-brain MR imaging in assessment of shunt function* advantages speed no radiation exposure no movement artifact no need for sedation multiplanar images excellent visualization of intraventricular & multicystic loculations disadvantages could not study communication btwn ventricular & cystic loculations poor shunt catheter visualization poor visualization of ICH & air only T 2 -weighted images w/ limited tissue contrast * ICH = intracranial hemorrhage. Unable to document such a communication because the intracontrast material is not available presently for this modality. tine follow up of shunt function, as is common practice at many neurosurgical centers, the risk of cancer increases markedly. Extrapolation of the data provided by Brenner indicates that a patient receiving an annual scan through the first decade of life will experience an increased cancer mortality risk of approximately 0.7%. Movement Artifact and Sedation A significant advantage of quick-brain MR imaging is its ability to acquire requisite data for image reconstruction with a single excitation pulse. Image slices are obtained in less than 1 second, essentially eliminating movement artifact and the need for sedation. This too eliminates a source of potential risk for pediatric patients. 19 Despite the presence of excellent pediatric sedation programs at most large medical centers, the incidence of adverse events associated with sedation in children was found to be as high as 4%. 14 These events include hypoxemia, failed sedation, and paradoxical reactions. In a study conducted by Malviya, et al., % of sedated pediatric patients experienced hypoxemia, 3.6% experienced medication-related adverse events, and in 7% of patients sedation failed altogether. Even death has been reported to occur as a direct result of sedation. In a study in which the Food and Drug Administration, the US Pharmacopoeia, and pediatric specialists compiled data concerning adverse events related to pediatric sedations, 51 deaths and nine permanent neurological injuries were reported regardless of drug class and route of administration. 5 Catheter Visualization Although shunt catheters are often seen on quick-brain MR images, such barium-impregnated catheters are much better visualized on CT scans. The limitation of SSFSE MR imaging in this capacity may be resolved in the future by improving its technology and/or producing catheters with ferromagnetic properties. Nonetheless, there are only rare situations in which the shunt catheter needs to be very clearly depicted for adequate diagnosis. Need for Additional CT Studies We propose that for the majority of shunt evaluations, quick-brain MR imaging adequately and sufficiently provides the data essential for diagnosis (Table 2). There are three clinical situations, however, in which CT scanning is an efficacious alternative or adjunctive evaluative measure: 1) when better-quality imaging of the ventricular catheter is needed (for instance, patients who have not undergone imaging for several years may require more accurate demonstration of the ventricular catheter to confirm its integrity and the location of its tip in the ventricle); 2) when it is important to evaluate the communication pattern of multicystic loculations by administering intraventricular contrast; and 3) when patients suffer postoperative neurological deterioration requiring further imaging to rule out hemorrhage or pneumocephalus. Indeed, CT examination continues to be a valuable adjunct to MR imaging in some circumstances. We propose, however, that the relegation of CT scanning to a secondary, case-sensitive role poses a minimal disadvantage to patient health. Limitations of Quick-Brain MR Imaging This MR imaging protocol was designed for the evaluation of moving or uncooperative patients and those with shunt-treated hydrocephalus. It should not be considered a full diagnostic MR imaging protocol in the traditional sense. Quick-brain MR imaging uses only heavily T 2 - weighted sequences, which have relatively flat image contrast. Although this image contrast is optimal for evaluation of ventricular anatomy, it is poorly suited for detailed evaluation of myelination, gray white distinction, and characterization of lesions such as migration abnormalities or tumors. It also has very little T 2 * susceptibility features, making it relatively insensitive to blood products, air, calcification, and implanted devices. Institution of the protocol therefore requires that the referring clinical services and prescribing radiologists understand these limitations. Conclusions The establishment of a clinical protocol in which quickbrain MR imaging is used to evaluate shunt-treated hydrocephalus proved to be advantageous. Quick-brain MR imaging allowed clinically adequate demonstration of the neuroanatomy necessary for proper evaluation of shunt function (namely, ventricular anatomy). In addition, the risk to patients is profoundly reduced: radiation exposure is eliminated, thus decreasing the long-term risks for radiation-induced neoplasia. Furthermore, movement artifact is reduced, eliminating the potential complications associated with deep sedation or anesthesia. Instituting this change in clinical protocol at our institution was facilitated by our close collaboration with the radiology department, with which an agreement was made to obtain emergency quick-brain MR images on a sameday basis between scheduled MR images. The flexible same-day turnaround schedule for SSFSE MR imaging examinations allowed an expeditious diagnostic evaluation of our patients without significantly delaying conventional radiology scheduling or MR imaging. In addition, billing arrangements were made with the hospital to lower the charges for these limited MR imaging studies to approximate more closely the charges for a noncontrast head CT scan. 150

5 Quick-brain MR imaging for shunt-treated hydrocephalus The relative safety and diagnostic utility of quick-brain MR imaging provides greater justification for developing outpatient screening protocols in the shunt-treated hydrocephalic population. Instituting such screening protocols (that is, yearly imaging sessions) could reduce both shuntrelated complications and mortality rates and has been suggested as a means of improving care for this patient population. 12,15 After developing a screening program that included an annual CT scan, Liptak, et al., 15 found that 8% of shunt malfunctions were detected on routine CT scanning in otherwise asymptomatic patients. Furthermore, it has been reported that approximately two thirds of patients who die of shunt malfunction never develop symptoms prior to their death. 12 Although this is a controversial issue, it is reasonable to assume that serial brain images could detect some of the asymptomatic shunt malfunctions, potentially preventing some deaths. In our study, three asymptomatic patients were found to require shunt revision based on quick-brain MR imaging evidence alone. Finally, despite their utility, neither quick-brain MR imaging nor CT scanning is entirely reliable for evaluating shunt function. It is well recognized that ventricle size does not uniformly dilate in response to shunt failure, which is evident in patients who suffer from the slit-ventricle syndrome or those with decreased ventricular compliance due to chronic shunting. 11 In fact in 50% of our patients with intraoperatively proven shunt malfunction, the ventricles were small. Vigilant clinical observation remains a critical partner of any imaging study. Disclaimer No author has a financial interest in the technology discussed in this paper. References 1. Arcement CM, Meza MP, Arumanla S, et al: MRCP in the evaluation of pancreaticobiliary disease in children. Pediatr Radiol 31:92 97, Boice JD Jr, Land CE, Shore RE, et al: Risk of breast cancer following low-dose radiation exposure. Radiology 131: , Brenner DJ: Estimating cancer risks from pediatric CT: going from the qualitative to the quantitative. Pediatric Radiology 32: , Bross ID, Ball M, Falen S: A dosage response curve for the one rad range: adult risks from diagnostic radiation. Am J Public Health 69: , Coté CJ, Karl HW, Notterman DA, et al: Adverse sedation events in pediatrics: analysis of medications used for sedation. Pediatrics 106: , Garton HJ, Kestle Jr, Drake JM: Predicting shunt failure on the basis of clinical symptoms and signs in children. J Neurosurg 94: , Gaskill SJ, Marlin AE: Radiation exposure in the myelomeningocele population. Pediatr Neurosurg 28:63 66, Gustafsson M, Mortensson W: Radiation exposure and estimate of late effects of chest roentgen examination in children. Acta Radiol Diagn (Stockh) 24: , Huda W, Atherton JV, Ware DE, et al: An approach for the estimation of effective radiation dose at CT in pediatric patients. Radiology 203: , Ikeda K, Hokuto I, Mori K, et al: Intrauterine MRI with singleshot fast-spin echo imaging showed different signal intensities in hypoplastic lungs. J Perinat Med 28: , Iskandar BJ, McLaughlin C, Mapstone TB, et al: Pitfalls in the diagnosis of ventricular shunt dysfunction: radiology reports and ventricular size. Pediatrics 101: , Iskandar BJ, Tubbs S, Mapstone TB, et al: Death in shunted hydrocephalic children in the 1990s. Pediatr Neurosurg 28: , Jablon S, Bailar JC III: The contribution of ionizing radiation to cancer mortality in the United States. Prev Med 9: , Karian VE, Burrows PE, Zurakowski D, et al: The development of a pediatric radiology sedation program. Pediatr Radiol 32: , Liptak GS, Bolander HM, Langworthy K: Screening for ventricular shunt function in children with hydrocephalus secondary to meningomyelocele. Pediatr Neurosurg 34: , Liu X, Ashtari M, Leonidas JC, et al: Magnetic resonance imaging of the fetus in congenital intrathoracic disorders: preliminary observations. Pediatr Radiol 31: , Madikians A, Conway EE Jr: Cerebrospinal fluid shunt problems in pediatric patients. Pediatr Ann 26: , Malviya S, Voepel-Lewis T, Eldevik OP, et al: Sedation and general anaesthesia in children undergoing MRI and CT: adverse events and outcomes. Br J Anaesth 84: , Mitchell AA, Louik C, Lacouture P, et al: Risks to children from computed tomographic scan premedication. JAMA 247: , Mittal TK, Halpin SFS, Bourne MW, et al: A prospective comparison of brain contrast characteristics and lesion detection using single-shot fast spin-echo and fast spin-echo. Neuroradiology 41: , Russell JG: Diagnostic radiography in children. Arch Dis Child 63: , Takehara Y: Fast MR imaging for evaluating the pancreaticobiliary system. Eur J Radiol 29: , Webster EW: Garland Lecture. On the question of cancer induction by small X-ray doses. AJR Am J Roentgenol 137: , 1981 Manuscript received September 9, Accepted in final form June 17, Address reprint requests to: Bermans J. Iskandar, M.D., Department of Neurological Surgery, University of Wisconsin Hospital and Clinics, 600 Highland Avenue, K4/832, Madison, Wisconsin iskandar@neurosurg.wisc.edu. 151