Head CT scanning represents a useful diagnostic

Transcription

1 J Neurosurg Pediatrics 12: , 2013 AANS, 2013 Analysis of limited-sequence head computed tomography for children with shunted hydrocephalus: potential to reduce diagnostic radiation exposure Laboratory investigation Jonathan Pindrik, M.D., 1 Thierry A. G. M. Huisman, M.D., 2 Mahadevappa Mahesh, M.S., Ph.D., 3,4 Aylin Tekes, M.D., 2 and Edward S. Ahn, M.D. 1 1 Department of Neurosurgery, Johns Hopkins University School of Medicine; 2 Divisions of Pediatric Radiology and Pediatric Neuroradiology, and 3 Medical Imaging Physics, Russell H. Morgan Department of Radiology and Radiological Science; and 4 Division of Cardiology, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland Object. Despite its diagnostic utility, head CT scanning imparts risks of radiation exposure. Children with shunttreated hydrocephalus exhibit increased risks of radiation toxicity due to the higher vulnerability of developing, immature tissues and frequent scanning. Several methods have been used to achieve dose reduction, including modifications of CT scanner tube current and potential. This retrospective study explores the use of a newly defined limited sequence of axial head CT slices to evaluate children with shunted hydrocephalus and decrease radiation exposure from diagnostic CT scans. Methods. Consistent sequences of 7 axial slices were extracted from previously performed standard head CT scans in children with shunted hydrocephalus. Chronologically distinct limited sequences of each patient were blindly, retrospectively reviewed by 2 pediatric neuroradiologists and 1 pediatric neurosurgeon. Limited-sequence CT evaluation focused on the adequacy of portraying the ventricular system, changes in ventricular size, and visualization of the proximal catheter. Reviewers assessed all original full series head CT scans at least 4 months later for comparison. Adequacy and accuracy of the limited-sequence CT compared with the gold standard head CT was investigated using descriptive statistics. Effective dose (ED) estimates of the limited-sequence and standard head CT scans were compared using descriptive statistics and the Mann-Whitney test. Results. Two serial head CT scans from each of 50 patients (age range 0 17 years; mean age 5.5 years) were reviewed both in standard and limited-sequence forms. The limited-sequence CT adequately portrayed the ventricular system in all cases. The inaccuracy rate for assessing changes in ventricular size by majority assessment (2 of 3 reviewers evaluating inaccurately) was 3 (6%) of 50. In 1 case, the inaccurate assessment would not have altered clinical management, corresponding to a 2 (4%) of 50 clinically relevant inaccuracy rate. As compared with the gold standard complete head CT series, the limited-sequence CT exhibited high sensitivity (100%) and specificity (91%) for portraying changes in ventricular caliber. Additionally, the limited-sequence CT displayed the ventricular catheter in 91.7% of scans averaged across 3 observers. Among all scans reviewed, 97 pairs of standard head CT and complementary limited-sequence CT scans contained adequate dosing information to calculate the effective dose (ED). The ED 50 of the limited-sequence CT (0.284 msv) differed significantly from the ED 50 of the standard head CT (4.27 msv) (p < ). The limited-sequence CT reflected a median absolute reduction of 4.10 msv and a mean percent reduction of 91.8% in ED compared with standard head CT. Conclusions. Limited-sequence head CT scanning provided adequate and accurate diagnostic information in children with shunted hydrocephalus. Techniques including minimization of axial slice quantity and modification of CT scanner parameters can achieve significant dose reduction, maintaining a balance between diagnostic utility and patient safety. ( Key Words head computed tomography radiation exposure radiation toxicity limited-sequence computed tomography dose reduction shunted hydrocephalus children Abbreviations used in this paper: BEIR = Biological Effects of Ionizing Radiation; CTDIvol = volume CT dose index; DLP = doselength product; EAM = external auditory meatus; ED = effective dose; ED 50 = median ED; LNT = linear no-threshold. J Neurosurg: Pediatrics / Volume 12 / November 2013 Head CT scanning represents a useful diagnostic tool in the adult and pediatric population. Many clinical scenarios, including head trauma, intracranial tumors or hemorrhage, and craniosynostosis, 491

2 J. Pindrik et al. benefit from the potentially high resolution and imaging quality that head CT scanning offers. This imaging modality also offers significant information in the evaluation of children with shunted hydrocephalus. Portraying ventricular size, location of ventricular catheters, and other pertinent findings like subdural hygroma or hematoma, a head CT scan supplies critical information regarding shunt function. Despite its diagnostic and clinical utility, CT scanning imparts risks of radiation exposure to frequently examined children. These risks include radiation-induced malignancy, white matter injury, and subsequent delayed cognitive effects. Given the dramatic increase in CT usage over the past 3 decades and the greater susceptibility of children to radiation exposure, several research efforts have focused on radiation dose reduction strategies. Many authors have modified CT scanner parameters, including tube current and potential, to decrease CT effective dose (ED). 12,16,21 The ED reflects the total amount of radiation exposure absorbed by an individual due to an event, accounting for differing sensitivities of the organs irradiated. 11 Diagnostic head CT scans may reveal adequate radiographic information in children with shunted hydrocephalus without the need for high resolution or imaging quality. Modified or limited imaging sequences may divulge sufficient information regarding ventricular caliber and proximal catheter position to safely impact clinical decision making. Accordingly, modified imaging techniques can offer children less radiation exposure. This retrospective study explores the use of a newly defined, limited sequence of axial head CT slices to evaluate children with shunted hydrocephalus, with the potential aim of reducing radiation exposure. Methods Composition of Limited Sequences This study represents a retrospective review of standard and limited-sequence head CT scans performed in children between the ages of 0 and 17 years who had shunted hydrocephalus. Children with a history of intracranial mass lesions or concerns for traumatic head injury were excluded from this study. These patients underwent diagnostic head CT for the evaluation of shunt function in routine follow-up, postoperative, or urgent settings. Standard head CT scans performed between 2005 and 2012 were selected for review. Scans performed during or after 2009 at our institution incorporated lower radiation dose settings (tube potential 120 kvp, constricted scanning window to avoid the corneas). Chronologically distinct serial head CT scans from different months were selected to prevent shunt failure from being suspected because of repetition of imaging within a short time interval. All serial standard head CT scans had been compared previously in documented radiology reports, ensuring the clinical relevance of comparing these serial examinations. The Johns Hopkins Medicine Institutional Review Board approved completion of this study without patient or parental informed consent. The Institutional Review Board approved this retrospective study without requiring written consent because all radiographic data had been acquired previously for clinical purposes. Requirements for the original standard head CT scan included slice thickness mm and a sufficient quantity of axial slices (at least 28) to compose a subset of slices. Consistent sequences of 7 axial slices were extracted from original head CT scans of study subjects to compose a limited-sequence complement to each standard head CT scan. Axial slices were extracted at different fractional distances (1/10, 1/5, 1/3, 2/5, 1/2, 3/5, 2/3) from the highest baseline slice traversing either external auditory meatus (EAM) to the vertex. In the occasional context where the fractional distance did not match exactly with slice location, the nearest axial slice was selected. When 2 consecutive slices lay equidistant from the fractional measurement, the higher axial slice was selected. Side-by-side comparison of the original CT scan and the scout radiograph helped to calculate distances and extract appropriate axial slices. The limited sequences of axial slices were deidentified and consolidated into a visual format convenient for review. Evaluation of Limited Sequences and Standard Head CT Scans Two pediatric neuroradiologists (T.H. and A.T.) and 1 pediatric neurosurgeon (E.S.A.) reviewed the limited sequences and standard head CT scans in blinded fashion. Chronologically distinct limited-sequence CT scans obtained in each patient were investigated for adequacy of portraying the ventricular system, changes in ventricular caliber, and visualization of the proximal shunt catheter. The latter criterion was defined as sufficient visualization to determine catheter location within or proper trajectory toward the ventricle. Reviewers completed evaluation forms regarding this information. Subsequently, the same 3 observers reviewed the original, standard head CT scans of study subjects 4 months later to decrease recall bias. The reviewers assessed the chronologically distinct standard head CT scans for changes in ventricular size and visualization of the proximal catheter. Recorded evaluations of the limited-sequence and standard head CT scans were compared for accuracy, using the standard head CT scan evaluations as the gold standard. Regarding changes in ventricular caliber, inaccuracy rates were calculated based on majority assessment among the 3 reviewers (2 of 3 reviewers assessing inaccurately). Sensitivity and specificity of the limited-sequence CT scans were calculated using the distinction between changed (positive result) and stable (negative result) ventricular size. The proportion of limited CT sequences providing adequate visualization of the ventricular system and proximal ventricular catheter was reported to describe the adequacy and clinical utility of the limited-sequence CT. Calculation of Radiation Dose Reduction Standard head CT scans contained documented information regarding radiation dose values. These parameters included tube voltage (in kvp), tube current and scan time product (in ma/sec), volume CT dose index ([CTDI- 492 J Neurosurg: Pediatrics / Volume 12 / November 2013

3 Limited-sequence head CT for shunted hydrocephalus vol]; in mgy), and dose-length product ([DLP]; in mgy cm). 11 The CTDIvol represents a fundamental dose indicator reflecting the radiation dose exposure for a given CT scanner and set of parameters (tube current and voltage). 11 Documentation of CTDIvol for most CT studies allows calculation of the DLP: DLP (mgy cm) = CTDIvol scan length [Eq. 1] The DLP represents the total amount of radiation dose delivered during a specific CT scan, accounting for scan length. 11,18 Similarly, the ED quantifies the total amount of radiation absorbed by the body and accounts for each organ s sensitivity to radiation toxicity. 3,6,11,18 Calculation of ED (in msv) for each head CT scan follows from the product of DLP and the reported conversion factor k (in msv/[mgy cm]) 11 : ED (msv) = k DLP [Eq. 2] The conversion factor k varies depending on age and body region irradiated. This study focused on head radiation exposure in children ages 0 17 years, resulting in the following k values: for age 0 11 months; for age 1 4 years; for age 5 9 years; and for age years. 5,11 Calculation of ED for the limited-sequence CT required estimation of a new DLP for each limited sequence. The limited-sequence scan length (in cm) represents the product of the number of axial slices (7) and the documented slice thickness for each CT examination (ranging from 0.5 to 5 mm). The dose parameter CTDIvol remained constant between the standard and limited-sequence head CT scans. Subsequently, the ED for each limited sequence could be calculated using DLP, k, and Eq. 2. The calculated ED for each standard head CT and limited-sequence CT scan were compared using descriptive statistics. For each CT type, the mean, median, range, and SD of ED were calculated among all scans performed with adequate dosing information. The ED 50 values between both types of CT scans were compared using the Mann-Whitney test. Furthermore, the absolute and percent reductions in ED were calculated for each complementary pair of standard head CT scan and limited-sequence CT. The mean, median, and SD of these values were calculated and compared. These calculations helped investigate the dose reduction afforded by the limited-sequence CT compared with the standard head CT scan. Additionally, the number of head CT scans obtained in each patient per year (interpreted at the authors institution) during was recorded to further appreciate the volume and burden of CT scanning in this population. Results Two serial limited sequences and standard head CT scans were reviewed for each of 50 patients (age range 0 17 years; mean age 5.5 years) (Figs. 1 and 2). In total, 100 limited sequences and 100 standard head CT scans were reviewed by all 3 observers. The limited-sequence CT adequately portrayed the ventricular system in all 50 patients, based on unanimous evaluations from all 3 reviewers. The observers did not respond that they were J Neurosurg: Pediatrics / Volume 12 / November 2013 unable to determine ventricular caliber in any limited sequence of axial slices. Regarding changes in ventricular size, the limited-sequence CT reflected a 6% (3 of 50 patients) inaccuracy rate by majority assessment (2 of 3 reviewers assessing incorrectly) (Table 1). In 1 case, 2 of 3 reviewers reported a decrease in ventricular size based on the limited-sequence CT, but stable ventricular size based on the standard head CT scan. Because this determination probably would not alter clinical management, the clinically relevant inaccuracy rate (by majority assessment) was 4% (2 of 50). The mean failure rate of visualizing the proximal catheter in 100 total limited CT sequences was 8.3% among all reviewers, corresponding to an average 91.7% success rate of ventricular catheter visualization. Compared with the gold standard complete head CT series, the limited-sequence CT sequences exhibited a sensitivity of 100% (18 of 18) and a specificity of 91% (29 of 32) for demonstrating changes in ventricular size (Table 2). There were no cases of false negatives, defined as perceived changes in ventricular caliber based on the gold standard head CT but assumed stability of ventricular size based on the limited-sequence CT. Cases reflecting stable ventricular size based on assessment of the gold standard head CT but changed (increased or decreased) ventricular size based on evaluation of the limited-sequence CT represented false positives. There were 3 cases of false-positive results. Among 100 standard head CT scans and 100 complementary limited sequences, radiation dose levels could be calculated for 97 scans within each group. Comparison of ED between standard head CT and limited-sequence CT showed significant differences and dramatic reductions in dose levels (Table 3). Out of 97 standard head CT scans with dose calculations, 34 scans (35%) exhibited EDs above 7 msv and 9 scans (9.3%) reflected doses above 40 msv. The median ED for the standard head CT scans (4.27 msv) differed significantly from the median ED for the limited-sequence CT (0.284 msv) (Mann-Whitney test, p < ). The limited-sequence CT provided a median absolute reduction of 4.10 msv and a mean percent reduction of 91.8% (SD 8.71%) in ED compared with standard head CT. Scout views (topograms) accompanying each CT scan (standard and limited sequence) incorporated a tube potential range of kvp, but did not contain complete dosing information (CTDIvol and DLP). Review of CT scanning records for all study subjects at our institution revealed 21 patients (42%) with at least 6 head CT scans in 1 year during (Table 4). The average number of total head CT scans performed in each patient during was 13.4 (range 1 94). Thirteen patients (26%) underwent at least 18 head CT scans total since 2008 (Fig. 3). Discussion Volume of CT Usage The dramatic increase in CT usage over the past 2 3 decades reflects its clinical and diagnostic utility. 3,18 Population statistics in Norway reveal that 10% 15% of the 4 493

4 J. Pindrik et al. Fig. 1. Case 8. Example of limited-sequence head CT. This compressed view displays the limited sequence of 7 axial head CT slices for this patient at 2 chronologically distinct times. The second sequence (lower row) demonstrates a decrease in ventricular caliber compared with the prior sequence (upper row; standard CT performed 11 months earlier). Both sequences provided adequate visualization of the ventricular catheter, as determined by all reviewers. million yearly radiographic examinations or image-guided procedures occur in children younger than 15 years. 8 Many of these diagnostic and procedure-related studies involve CT scanning. Consequently, radiation scientists estimate that CT scanning accounts for 40% 60% of the total Norwegian population dose exposure. Statistics in the US show similar trends, with more than 62 million CT scans performed per year, accounting for nearly 50% of the collective radiation dose for medical purposes. 3,10 More than 4 million (6% 11%) of these scans occur in children; pediatric diagnosis represents one of the top two categories exhibiting the greatest increase in CT usage since Radiation scientists confirm that low dose levels from medical diagnostic imaging represent the largest manmade source of radiation exposure in the human population. 15 Children with shunted hydrocephalus represent a subpopulation that is highly susceptible to frequent imaging. The high complication and failure rates of ventricular shunting, reaching 40% 50% over the first 2 postoperative years and 4% 5% per year thereafter, partially justifies the need for repeat imaging studies. 7,9,17 Evaluation of potential shunt malfunction usually requires some form of rapid radiographic examination to gauge ventricular size and evaluate proximal catheter location. Children with profound developmental delay or neurological debility may offer limited physical examinations, placing greater emphasis on imaging data. Furthermore, common presentations of pediatric patients to the emergency department (headache, emesis, fever, viral syndromes, and decreased oral intake) often overlap with presenting symptoms and signs of shunt malfunction or infection. As reflected by the dramatic increase in CT usage since 1980, the medical community has appeared to adopt a lower threshold for diagnostic imaging as well. 3,19 Given the need for routine follow-up scans and urgent Fig. 2. Case 23. Example of limited-sequence head CT. This compressed view displays the limited sequence of 7 axial head CT slices for this patient at 2 chronologically distinct times. The second sequence (lower row) demonstrates an increase in ventricular caliber compared with the prior sequence (upper row; standard CT performed 3 months earlier). Although providing less visualization of the catheter than in the patient in Case 8, both limited sequences display the proximal tip within the lateral ventricles, as determined by all reviewers. 494 J Neurosurg: Pediatrics / Volume 12 / November 2013

5 Limited-sequence head CT for shunted hydrocephalus TABLE 1: Inaccurate assessment of ventricular size using limited-sequence head CT* Case No. Interval Btwn Studies (mos) Interpretation of Limited-Sequence CT obs 1: decrease obs 2: stable obs 3: decrease obs 1: stable obs 2: increase obs 3: increase obs 1: stable obs 2: increase obs 3: increase * obs = observer. Interpretation of Standard Head CT obs 1: stable obs 2: stable obs 3: stable obs 1: stable obs 2: stable obs 3: stable obs 1: stable obs 2: stable obs 3: stable Comments inaccurate assessment would probably not alter clinical management dysmorphic ventricular system per original radiology report original radiology report: degree of ventricular dilatation not appreciably changed from previous study, considering differences in technique radiographic studies to evaluate potential shunt malfunction, children with shunted hydrocephalus exhibit high frequencies of head CT scanning. Results from this study reflect the high volume of diagnostic CT imaging in this subpopulation. Study subjects underwent an average of 13 head CT examinations over 5 years, and more than 40% underwent at least 6 head CT scans in 1 year. A large subgroup (25%) of the study population underwent at least 18 head CT scans total during the 5-year study period (Fig. 3). These statistics reflect the tremendous volume of head CT scanning in children with shunted hydrocephalus and underlie a serious health concern. 14 Risks of Diagnostic Radiation Exposure Although it represents an efficient and widely applicable imaging modality for diagnostic and treatment purposes, CT scanning exposes patients to low levels of ionizing radiation. Since 1980, the US per capita dose of ionizing radiation exposure from medical imaging has increased by a factor of nearly 6. 6,15 These radiation doses, although low in magnitude, result in nonnegligible risks of carcinogenesis and later cognitive deficits when exposure occurs at a younger age. 6,8,15,19 Based on the Biological Effects of Ionizing Radiation (BEIR) VII Phase 2 report, several studies have demonstrated increased lifetime attributable risk of developing solid malignancies or leukemia and cancer-related death from diagnostic radiation exposure. 1,2,4,6,10,14,18,19 Smith-Bindman et al. 18 reported a median lifetime attributable risk of cancer development in 0.23 per 1000 patients undergoing a single CT scan. TABLE 2: Sensitivity and specificity of limited-sequence head CT* Limited-Sequence Head CT Gold Standard Head CT Change Stable change 18 (true positive) 3 (false positive) stable 0 (false negative) 29 (true negative) sensitivity 100% (18/18) specificity 91% (29/32) * Change indicates change in ventricular caliber (increase or decrease in size); stable indicates no change in ventricular caliber. J Neurosurg: Pediatrics / Volume 12 / November 2013 This represents a low but nonnegligible risk considering the volume of CT scans performed globally. Epidemiologists estimate nearly 0.4% 2% of cancer diagnoses in developed nations may be related to CT radiation exposure, corresponding to at least 700 cases per year in the United Kingdom. 1,3 Children exhibit higher susceptibility to radiation toxicity than adults, resulting in greater risks of carcinogenesis. 2,8,10,12,15,18,19,21 Several factors account for the increased exposure risk in children. Smaller body volumes and cross-sectional areas of exposure result in greater densities of dose absorption. 19,21 Smaller body habitus and less subcutaneous soft tissue also provide less shielding of potentially harmful ionizing radiation. 2 Growing, immature tissues in children exhibit a higher sensitivity to nucleic acid damage by irradiation. 8,21 Furthermore, children possess a longer latency period to develop radiationinduced malignancy, given their young age at exposure and expected survival. 2,19,21 The factors described above result in higher EDs from diagnostic CT and greater risks of carcinogenesis in children compared with adults. 15 Radiation doses from standard head CT in adults can increase by a factor of in young children and neonates. 1,10,19 Holmedal et al. 8 reported single head CT mean EDs of msv in children and 4 8 msv in infants. These figures resemble the range of msv ED per head CT in children who underwent imaging at other institutions nationally and abroad. 10,13 Given the necessity for higher imaging quality, multiple series, or image reconstruction, certain CT scans may deliver EDs significantly higher than the reported means. Furthermore, repetition of scanning results in higher cumulative doses and greater risks of radiation toxicity. 19 For instance, a retrospective study with mean follow-up of more than 7 years reported cumulative EDs ranging between 2.3 msv and 63.8 msv in a group of 67 children with treated hydrocephalus. 8 Smyth et al. 19 reported 2 cases of pediatric patients in whom shunts were placed during early infancy, with numerous revisions and head CT scans (at least 14 23) during childhood and adolescence. One patient developed Hodgkin lymphoma in the neck at the age of 17 years, and the other died from a left parietooccipital gliosarcoma at the age of 19 years. Although the authors could not confirm radiation-induced 495

6 J. Pindrik et al. TABLE 3: Comparison of ED between standard head and limited-sequence CT Variable ED Mean Median Range SD standard head CT 15.3 msv 4.27 msv msv 18.6 msv limited-sequence CT 1.10 msv msv msv 1.96 msv absolute reduction in ED* 14.2 msv 4.10 msv msv 17.7 msv % reduction in ED 91.8% 96.7% 49.2% 98.3% 8.71% * Absolute reduction in ED = ED standard head CT ED limited-sequence CT. Percent reduction in ED = [(absolute reduction in ED)/ED standard head CT ] 100. malignancy, they strongly suspected a causative role for diagnostic radiation exposure. Several articles have reported increased risks of carcinogenesis and cancer-related death related to CT scanning at early ages. Brenner et al. 2 reported a 0.07% radiation-induced cancer mortality rate from single head CT examinations in 1-year-old infants. Using similar reported frequencies, Smyth et al. 19 estimated a cumulative 1% lifetime risk of developing radiation-induced malignancy from 15 head CT scans performed during infancy. In a retrospective cohort study, Pearce et al. 14 demonstrated a positive association between organ-specific CT radiation absorption and the development of certain malignancies. Cumulative red bone marrow exposure exceeding 30 mgy, corresponding to 5 10 head CT scans in patients younger than 15 years of age, correlated to increased relative risk (3.18) of developing leukemia. Two to 3 head CT scans in the same age group could result in mgy of cumulative brain exposure, associated with an elevated relative risk (2.82) of developing brain tumors. 14 Although the absolute risk remains small, frequent scanning and inherent susceptibility factors place children with shunted hydrocephalus at elevated, nonnegligible risk of carcinogenesis from diagnostic radiation exposure. Fig. 3. Bar graph showing frequency of head CT scanning in study subjects, This graph shows the number of study subjects (y axis) within each specified group of head CT numbers (x axis) at our institution from 2008 to The percentages of subjects within each group are indicated above the respective bars. Controversy Regarding Diagnostic Radiation Exposure The risks of diagnostic radiation exposure have been heavily debated. Some authors contest application of the linear no-threshold (LNT) model to low radiation levels, arguing that doses experienced in CT scanning fall well below atomic bomb or occupational exposures. 10,13,14,20 Critics also note that cancer risks estimated from the BEIR VII Phase 2 report reflect population rather than individual risks. 10 Pauwels and Bourguignon 13 argued that the paucity of epidemiological data for cancer induction due to CT EDs less than 50 msv questions the presumed toxicity of diagnostic radiation. Reported low dose absorption levels in survivors of the Hiroshima and Nagasaki atomic bombs ranged between 10 and 100 msv. 3,18 Nearly two-thirds of atomic bomb survivors encountered radiation doses at or below 100 msv. 15 Undermining the argument of Pauwels and Bourguignon, subgroups of atomic bomb survivors and nuclear industry workers exposed to low levels of radiation (average msv) have exhibited increased risk of carcinogenesis and cancer-related mortality. 3 Accordingly, occupational regulations in the nuclear industry restrict workers to an average exposure of 20 msv per year (maximum of 50 msv in 1 year). 6 Lower radiation levels investigated in epidemiological studies do not differ dramatically from reported standard head CT EDs of 2 8 msv, especially considering variations in dose across different CT parameters, CT scanners, and institutions. 1,6,10,18 Increased susceptibility factors and frequent scanning in children with shunted hydrocephalus often result in cumulative doses above the threshold for consideration in radiation toxicity models. 8,13 Finally, compelling evidence has not been presented to suggest the existence of a threshold dose or to refute the LNT model and its applicability to diagnostic radiation exposure. 2,15 The BEIR VII report upholds the LNT model as the most practical and effective method of estimating carcinogenic risk from low-level radiation exposure. 15 Although controversy exists about the perceived effects of radiation toxicity from diagnostic imaging and application of the LNT model to CT, most authors recognize the potential hazards of increased scanning and support initiatives to decrease radiation exposure in children. 10,12,13 Methods to Decrease Radiation Exposure Several potential strategies exist to decrease the amount of ionizing radiation exposure related to diagnos- 496 J Neurosurg: Pediatrics / Volume 12 / November 2013

7 Limited-sequence head CT for shunted hydrocephalus tic CT. The medical community can decrease the overall volume of CT scanning by performing fewer studies and using other elements of clinical information to assist diagnosis. However, this represents a less useful approach in children with shunted hydrocephalus given the necessity of frequent scanning. Other imaging modalities could replace CT scanning to evaluate ventricular size and predict shunt function. In neonates with an open anterior fontanel, head ultrasonography typically offers sufficient diagnostic information. Fast-sequence MRI may supplant CT scanning in certain contexts. 13 However, the limited availability of MRI technicians and units during urgent evaluation or across medical centers may hinder ubiquitous implementation of this approach. Furthermore, fast MRI sequences may require more time to complete than low-dose CT protocols and provide suboptimal visualization of the proximal ventricular catheter. 19 Modification of CT scanning techniques to achieve dose reduction represents another compelling strategy embraced by multiple research groups nationally and abroad. Important CT parameters affecting radiation dose include tube current, scanning time, tube potential, anatomical range of study, scan pitch, and number of axial slices. Modification of one or more imaging factors may reduce dose absorption while increasing image noise or diminishing overall image quality. 3,11 13,19,21 However, the resulting scan may still provide adequate, clinically useful information for the appropriate subgroup of patients. 2,12 Adjusting tube current or potential represent practical methods of achieving dose reduction, because these factors primarily influence beam energy. 10,13 Achieving dose reductions of 35% 60%, modified low-dose head CT protocols offer balance between diagnostic utility and patient safety. 13 At Polish Mother s Memorial Hospital, Rybka et al. 16 conducted a trial of a low-radiation-dose head CT protocol with decreased tube voltage and current, showing 60% 70% ED reduction in scans that maintained diagnostic utility for children with hydrocephalus. Udayasan kar et al. 21 conducted a similar study at Emory University Hospital, implementing a modified head CT protocol with lower tube current (80 ma/sec as opposed to 220 ma/sec) in postoperative follow-up scanning performed in children with shunted hydrocephalus. Despite producing images of lower quality and higher noise levels, the low-dose protocol provided sufficient diagnostic information while achieving a mean dose reduction of 63%. Other authors have reported similar results, with modified head CT protocols exhibiting lower imaging quality but acceptable diagnostic utility and significant dose reductions. 12 Head CT scans obtained during or after 2009 in hydrocephalic children treated at the authors institution have incorporated lower radiation dose settings (tube potential 120 kvp, constricted scanning window). Limited-Sequence CT Reducing the number of axial slices in head CT scans represents another useful dose-reduction strategy reflected in this study. Despite providing less radiographic data (and therefore exhibiting the potential to reduce the ED), the limited-sequence head CT did not sacrifice diagnostic information. Although portraying the ventricular system J Neurosurg: Pediatrics / Volume 12 / November 2013 in all cases and providing adequate visualization of the proximal catheter in 91.7% of scans, the limited-sequence CT exhibited a low clinically relevant inaccuracy rate (4%). All cases of inaccurate assessment involved CT scans showing diminutive changes in ventricular size, based on prior radiology reports. High sensitivity (100%) and lack of false-negative results indicate that the limitedsequence head CT did not fail to portray any changes in ventricular size (Table 2). The lower specificity (91%) and the presence of 3 false-positive results reflect the tendency of the limited-sequence head CT to over-call certain cases. However, this tendency to over-call seems less detrimental than under-recognition of increased ventricular size reflecting shunt failure. The results of this study suggest that neurosurgical clinicians and radiologists may use a limited amount of information available from head CT scans obtained to evaluate children with shunted hydrocephalus. Elimination of the unused radiographic information could lower diagnostic radiation exposure without compromising clinical decision making. Although readily implemented in retrospective fashion, this method of acquiring axial head CT slices would require validation prospectively. We have initiated a prospective trial comparing the limited-sequence CT protocol to standard head CT scans in the evaluation of hydrocephalic pediatric patients. Of note, both the limited-sequence CT and standard head CT incorporate lower radiation dose settings (tube potential 120 kvp, tube current approximating 150 ma/sec). The novel imaging sequence, called the limited-sequence shunt CT, uses landmarks of the EAM and auditory canal to assign an appropriate scanning window (incorporating 7 axial slices starting above the EAM) based on the scout view. This algorithm has yielded reproducible head CT limited sequences displaying the ventricles and proximal catheter in an ongoing series of children with shunted hydrocephalus (Fig. 4). The limited-sequence CT exhibited the potential to achieve median absolute reduction of 4.10 msv and mean percent reduction of 91.8% in the ED. Incomplete dosing information precluded calculation of the total ED for the CT scan and scout view (topogram) combined. However, incorporation of the scout view dose would not alter calculations of absolute reduction because topogram dose estimates for the standard and limited-sequence CT scans would match and cancel out. Furthermore, the diminutive increase in the total ED from the scout view (approximately msv) would have little impact on overall percent dose reduction. Exclusion of scout view dose estimates from CT ED calculations models prior studies, given the absent or minimal impact on radiation reduction assumptions. 16,21 Concerning dose estimates for the standard head CT scan, outliers occurred in high-resolution scans spanning large areas of study with many axial slices. These scans contained elevated DLP values and were obtained in children 1 4 years of age with a greater conversion factor k (0.067). In this study we paid close attention to age at exposure, which greatly impacts dose estimates, and resulted in outlier EDs greater than norms reported in prior studies. Given its sensitivity on outliers, the mean 497

8 J. Pindrik et al. Fig. 4. Limited-sequence shunt CT, example from prospective study. This compressed view displays an example of the limited-sequence shunt CT implemented in a current prospective study at our institution. The limited sequence of 7 axial slices starts above the EAM and terminates below the vertex, displaying the ventricular system and proximal ventricular catheter. ED (standard CT: 15.3 msv; limited-sequence CT: 1.10 msv) exceeded the ED 50 (standard CT: 4.27 msv; limitedsequence CT: msv) for both CT types. Because it provided a skewed representation of typical diagnostic CT scans, the mean ED was abandoned in favor of the ED 50 for comparison statistics. The ED 50 (0.284 msv) of the limited-sequence CT closely resembles the mean EDs (0.20 msv, 0.58 msv) of other modified CT protocols. 16,21 Of note, scout views add an another msv to the total ED based on dosing information from our institution. Prior studies report mean CT ED reductions as high as 60% 70%. 16,21 In this study we report mean ED and ED 50 reductions of 91.8% and 96.7%, respectively. Most prior studies restricted imaging to routine, postoperative follow-up scenarios, whereas this study retrospectively analyzed cases involving routine and urgent presentations with concern for shunt malfunction. Furthermore, this cohort of patients included those with documented complex ventricular systems and many patients who had undergone several shunt revisions in the past. Modified CT protocols must contain adequate and accurate radiographic information to assist and safely guide clinical decision making. A limited sequence of axial slices that captures the ventricular system and proximal catheter may reveal sufficient diagnostic information to support or refute shunt malfunction. The limited-sequence CT used in this study accurately reflects ventricular size and ventricular catheter position. The adoption of this and other strategies may spawn optimal techniques to decrease diagnostic radiation exposure in children with shunted hydrocephalus. Furthermore, the standardization of CT protocols for specific patient populations can help reduce the ED 50 per individual scan and limit the wide variation in dose across different scanners and sites. 18,19 The limited-sequence head CT may prove especially useful for repeat, routine, or follow-up studies in children with shunted hydrocephalus, who are likely to undergo frequent scanning (Fig. 3; Table 4). Study Limitations Limitations of this study include its retrospective design and moderate sample size of 50 patients. A prospective study that is being conducted at our institution will provide for limited-sequence CT acquisition in real time. The potential for recall bias could not be completely eliminated, despite a 4-month separation time between review of the limited-sequence and standard head CT scans. However, this method of review seems preferable to comparing observers assessments against the original radiologist report based on the standard head CT. The latter method would have introduced interobserver error. Finally, the limited-sequence CT applies narrowly to a subset of patients, specifically children with shunted hydrocephalus in postoperative, follow-up, and urgent clinical settings. The limited-sequence head CT protocol would not be applicable in the following scenarios: 1) initial screening or diagnostic evaluations in children with suspected hydrocephalus; 2) visualization of lesions or abnormalities outside the ventricular system (for example, intraparenchymal or subarachnoid hemorrhage); 3) trauma; or 4) evaluation of a known or suspected intracranial mass or lesion. Conclusions The limited-sequence head CT provided adequate and accurate radiographic information regarding the ventricular system and proximal ventricular catheter in children with shunted hydrocephalus. Additionally, the limited-sequence CT reflected significant ED reductions of more than 90% 498 J Neurosurg: Pediatrics / Volume 12 / November 2013

9 Limited-sequence head CT for shunted hydrocephalus TABLE 4: Patients with at least 6 head CT scans in 1 year at authors institution, Case No. Age (yrs) at 1st Study CT No. of Head CT Scans/Yr Total compared with standard head CT, without sacrificing diagnostic information. Modification of CT scanning parameters (tube current or potential) and implementation of the limited-sequence CT may be extended to several pediatric institutions nationally and abroad. The limited-sequence head CT may prove especially useful for repeat, routine, or follow-up studies in children with hydrocephalus, who are likely to undergo frequent scanning. Limited-sequence CT and other techniques may help decrease diagnostic radiation exposure in children with shunted hydrocephalus, achieving an appropriate balance between diagnostic utility and patient safety. Disclosure The first author (Dr. Pindrik) receives research salary funding for this project from the Hartwell Foundation Biomedical Research Fellowship ( ). These authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Ahn, Pindrik, Huisman, Mahesh. Acquisition of data: Pindrik, Huisman, Tekes. Analysis and interpretation of data: all authors. Drafting the article: Pindrik, Huisman. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Ahn. Statistical analysis: Pindrik. Administrative/technical/material support: Mahesh. Study supervision: Ahn. References J Neurosurg: Pediatrics / Volume 12 / November Berrington de González A, Darby S: Risk of cancer from diagnostic X-rays: estimates for the UK and 14 other countries. Lancet 363: , Brenner D, Elliston C, Hall E, Berdon W: Estimated risks of radiation-induced fatal cancer from pediatric CT. AJR Am J Roentgenol 176: , Brenner DJ, Hall EJ: Computed tomography an increasing source of radiation exposure. N Engl J Med 357: , Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII Phase 2. Washington, DC: National Academies Press, Diagnostic Imaging Council CT Committee: AAPM Report No. 96: The Measurement, Reporting, and Management of Radiation Dose in CT. College Park, MD: American Association of Physicists in Medicine, 2008 ( org/pubs/reports/rpt_96.pdf) [Accessed August 9, 2013] 6. Fazel R, Krumholz HM, Wang Y, Ross JS, Chen J, Ting HH, et al: Exposure to low-dose ionizing radiation from medical imaging procedures. N Engl J Med 361: , Gupta N, Park J, Solomon C, Kranz DA, Wrensch M, Wu YW: Long-term outcomes in patients with treated childhood hydrocephalus. J Neurosurg 106 (5 Suppl): , Holmedal LJ, Friberg EG, Børretzen I, Olerud H, Laegreid L, Rosendahl K: Radiation doses to children with shunt-treated hydrocephalus. Pediatr Radiol 37: , Kestle JR: Pediatric hydrocephalus: current management. Neurol Clin 21: , vii, King MA, Kanal KM, Relyea-Chew A, Bittles M, Vavilala MS, Hollingworth W: Radiation exposure from pediatric head CT: a bi-institutional study. Pediatr Radiol 39: , Mahesh M: MDCT Physics: The Basics: Technology, Image Quality and Radiation Dose. Philadelphia: Lippincott Williams & Wilkins, Mullins ME, Lev MH, Bove P, O Reilly CE, Saini S, Rhea JT, 499

10 J. Pindrik et al. et al: Comparison of image quality between conventional and low-dose nonenhanced head CT. AJNR Am J Neuroradiol 25: , Pauwels EK, Bourguignon M: Cancer induction caused by radiation due to computed tomography: a critical note. Acta Radiol 52: , Pearce MS, Salotti JA, Little MP, McHugh K, Lee C, Kim KP, et al: Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 380: , Royal HD: Effects of low level radiation what s new? Semin Nucl Med 38: , Rybka K, Staniszewska AM, Biegański T: Low-dose protocol for head CT in monitoring hydrocephalus in children. Med Sci Monit 13 (Suppl 1): , Sandberg DI: Endoscopic management of hydrocephalus in pediatric patients: a review of indications, techniques, and outcomes. J Child Neurol 23: , Smith-Bindman R, Lipson J, Marcus R, Kim KP, Mahesh M, Gould R, et al: Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 169: , Smyth MD, Narayan P, Tubbs RS, Leonard JR, Park TS, Loukas M, et al: Cumulative diagnostic radiation exposure in children with ventriculoperitoneal shunts: a review. Childs Nerv Syst 24: , Tubiana M: Computed tomography and radiation exposure. N Engl J Med 358:850, 2008 (Letter) 21. Udayasankar UK, Braithwaite K, Arvaniti M, Tudorascu D, Small WC, Little S, et al: Low-dose nonenhanced head CT protocol for follow-up evaluation of children with ventriculoperitoneal shunt: reduction of radiation and effect on image quality. AJNR Am J Neuroradiol 29: , 2008 Manuscript submitted February 18, Accepted August 8, Portions of this work were presented in oral format at the 41st Annual Meeting of the AANS/CNS Section on Pediatric Neurological Surgery in St. Louis, Missouri, in November Please include this information when citing this paper: published online September 20, 2013; DOI: / PEDS1322. Address correspondence to: Edward S. Ahn, M.D., Division of Pediatric Neurosurgery, Johns Hopkins University School of Medicine, 600 North Wolfe Street, Phipps 560A, Baltimore, MD eahn4@jhmi.edu. 500 J Neurosurg: Pediatrics / Volume 12 / November 2013