| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Obstetric Imaging |
1 From the Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Boston, MA 02215 (D.L., J.F.K.T., M.M., T.S.M.); and Clinical Research Program (H.A.F.) and Department of Radiology (J.A.E., C.D.R., T.Y.P., C.E.B., R.L.R.), Children's Hospital Boston, Boston, Mass. From the 2006 RSNA Annual Meeting. Received June 18, 2007; revision requested August 16; revision received August 29; accepted September 27; final version accepted October 12. Supported by National Institutes of Health, National Institute of Biomedical Imaging and Bioengineering grant 01998. Address correspondence to D.L. (e-mail: dlevine{at}caregroup.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
Materials and Methods: One hundred ninety-five women, aged 18–44 years, with 200 fetal referrals for VM, were recruited in a prospective IRB-approved, HIPAA-compliant study. Written informed consent was obtained. US scans were prospectively interpreted by three obstetric radiologists and MR examinations were read by one obstetric radiologist and three pediatric neuroradiologists. Final diagnosis was reached by consensus (198 US, 198 MR, and 196 US-MR comparisons). Gestational age, ventricular size, types of disagreements, and reasons for disagreements were recorded. Interreader agreement was assessed with 
statistics. Ventricular diameter, gestational age, and confidence scores were analyzed by using mixed-model analysis of variance, accounting for correlation within reader and fetus.
Results: There was prospective agreement on 118 (60%) of 198 US and 104 (53%) of 198 MR readings. Consensus was more likely when the final diagnosis was isolated VM (83 of 104, 80% at US; 82 of 109, 75% at MR) than when the final diagnosis included other anomalies as well (14 of 63, 22% at US; seven of 68, 10% at MR; P < .001). There was disagreement on 19 (10%) of 196 and 31 (16%) of 196 fetuses about the presence of VM at US and MR, respectively, and on 29 (15%) of 198 and 39 (20%) of 198 fetuses regarding the presence of major findings at US and MR, respectively. Reasons for discrepancies in reporting major findings included errors of observation, lack of real-time US scanning, lack of neuroradiology experience, as well as modality differences in helping depict abnormalities.
Conclusion: Of radiologists who read high-risk obstetric US and fetal MR images for VM, there is considerable variability in central nervous system diagnosis.
© RSNA, 2008
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
10 mm) or at US with a referral history of VM. US and MR examinations were performed in a manner similar to that described by Levine et al (1) and evaluated as shown in Figure 1. Each fetus was referred for VM and each fetal examination was considered as a unique study. Date of last menstrual period, fetal age according to last menstrual period, and fetal biometry (including head circumference, biparietal diameter, and gestational age according to US) were recorded. Two patients with fetuses with neural tube defects (NTDs) underwent incomplete US prior to MR and thus were not included in the US analysis. Two patients withdrew from the MR portion of the study after undergoing US, thus they were not included in the MR analysis. Therefore, the study group for consensus opinions was 198 for US comparisons, 198 for MR comparisons, and 196 for combined US-MR comparisons. One hundred ninety-five (99.5%) of 196 US-MR examination sets were performed on the same day, one (0.5%) of 196 examination sets was performed within 1 day.
|
US findings were then independently reinterpreted by the two obstetric radiologists who did not perform the US (to avoid any potential knowledge of the study prior to review), who also coded anomalies, lateral ventricle measurements, and confidence scales.
VM was diagnosed whenever one of the lateral ventricles at the level of the atrium was more than 10.0 mm in diameter. However, an obstetric radiologist could also diagnose VM subjectively, given the appearance of a dangling choroid plexus (9).
The US diagnoses of the three radiologists were compared (the reader performing US and the two from the other institution). Each US disagreement was coded as follows: (a) error of observation, (b) error of interpretation, (c) error of omission (a finding is clearly present but was not coded prospectively), (d) coding issue (two similar diagnoses with different codes, such as neuroproliferative porencephaly and encephaloclastic porencephaly), (e) disagreement regarding observation (still no agreement during consensus conference), (f) disagreement regarding interpretation (during consensus conference, the finding is seen by all but the interpretation of the finding was not agreed on), or (g) real-time scanning would have aided in diagnosis. In addition, final consensus was decided on by using a conference of three of the original readers, coded as being achieved by agreement of two or all three obstetric radiologists.
Disagreements (or consensus) in each fetus were coded as having complete agreement or one or more of the following: (a) decision to call VM (difference of opinion as to whether VM was present), (b) no clinical difference owing to disagreement (such as coding for Dandy Walker variant or inferior vermian hypoplasia or not coding a small midline cyst in a fetus with agenesis of the corpus callosum [ACC]), (c) minor new finding (for example, visualization of a portion of septal leaflets in a fetus that another reader coded as having complete absence of septal leaflets), (d) major new finding (such as an examination determined as showing findings of schizencephaly by one reader but not by another reader), (e) overcall of a minor finding (at consensus, the coded anomaly was determined as absent), or (f) overcall of a major finding. Examinations with disagreement on diagnosis were those coded as having new minor or major findings or those with overcalls of minor or major findings.
Examinations with and without diagnosis disagreement were compared with respect to gestational age.
MR Interpretation and Consensus
The MR examination was interpreted by one of the same group of radiologists who performed US (readers with 2–11 years experience with fetal MR), with knowledge of the findings of the confirmatory US. The readers coded the anomalies, measured the lateral ventricles at the level of the atria, and scored confidence in diagnosis. In addition, they coded for whether MR findings affected diagnosis by using the following scoring system: 1 = no effect, 3 = some effect, and 5 = critical information added.
Three pediatric neuroradiologists (C.D.R., T.Y.P., and R.L.R., each with 10–14 years experience in pediatric neuroradiology) then independently reviewed the MR examination with knowledge of referral diagnosis but without the aid of the confirmatory US and recorded their interpretation of the MR including ventricular size, coded CNS diagnosis, and confidence rating. The confirmatory US findings were then revealed and the neuroradiologist documented any change in diagnosis. Knowledge of the effect of the confirmatory US findings at final MR diagnosis was then scored on the same five-point scale as above.
MR consensus was reached at conference with the three neuroradiologists. MR disagreements were coded as follows: (a) error of observation, (b) error of interpretation, (c) error of omission, (d) coding issue, (e) disagreement regarding observation, or (f) disagreement regarding interpretation. In addition, final consensus was coded to denote agreement of two or all three neuroradiologists. Diagnosis disagreements were coded and compared with gestational age in a similar fashion to those described for US consensus.
US-MR Comparisons
The final US and MR diagnoses were compared by one of the authors (D.L.), in consultation with the neuroradiologists (C.R., R.L.R., T.Y.P.), when any final diagnosis was unclear. Disagreements were coded as follows: (a) disagreement regarding observation, (b) disagreement regarding interpretation, (c) error of omission, (d) coding issue, (e) finding expected to be better seen at US (such as the wall of an arachnoid cyst), (f) finding expected to be better seen at MR (such as migrational abnormality), or (g) neuroradiologist experience would aid in diagnosis.
Statistical Analysis
Interreader agreement was assessed with statistics. Ventricular diameter, gestational age, and confidence scores were analyzed by using mixed-model analysis of variance, accounting for correlation within reader and fetus (Appendix).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
statistic (Table 2). Chiari malformation and spinal NTD showed excellent agreement.
|
|
|
|
|
There were 80 (40%) of 198 studies without consensus on the prospective US readings (Table 4), with 153 disagreements on specific diagnoses(Table 5). Consensus was achieved with agreement of all three obstetric radiologists for 122 (80%) of these 153 diagnoses and with agreement of two for the remaining 31 (20%) diagnoses. Twenty-seven disagreements were categorized as having no clinical implication. Of the 80 studies requiring a conference (Table 4), there was disagreement regarding the presence of VM in 19 fetuses and in 29 (15%) regarding the presence of major findings.
|
|
The level of consensus of the US readers concerning presence of VM was not affected by head circumference, biparietal diameter, or gestational age (P > .40) (Table 6). There was a significant difference among these groups when comparing ventricular diameter: those fetuses agreed on as having VM, with the largest ventricles (mean, 13.5; range, 8–44 mm); those agreed on as not having VM, with the smallest ventricles (mean, 7.5; range, 4–9 mm); and those with disagreement regarding VM, with ventricles of intermediate size (mean, 9.4; range, 7–15 mm).
|
|
MR Diagnoses and Agreement
Of 198 MR studies, the four readers reached consensus prospectively in 104 (53%). These final diagnoses included isolated VM (n = 82), normal (n = 14), holoprosencephaly (n = 1), and VM and another diagnosis (n = 7; four with ACC, two with spinal NTD, and one with CNS neoplasm). Consensus was significantly more likely when VM was isolated (82 of 109, 75%) than when other anomalies were present in association with VM (seven of 68, 10%; P < .001)
Excluding two fetuses with holoprosencephaly, there was disagreement about the presence of VM at MR in 31 (16%) of 196 fetuses. Ventricular diameter was largest in fetuses with preconference agreement on VM (mean, 14.4 mm) and smallest in fetuses with preconference agreement that VM was absent (mean, 7.4 mm; P < .001) (Table 3). In studies with disagreement, readers favoring VM recorded larger ventricular diameter for a given fetus (10.9 vs 9.0 mm, P < .001). Variation among the readers for a given fetus and reader classification (obstetric radiologist or neuroradiologist) was 0.4 mm, representing less than 1% of total variance. The intraclass correlation was 93%.
There were 94 (47%) of 198 studies without consensus on the prospective MR readings (Table 4), with 223 disagreements on particular diagnoses (Table 5). The three neuroradiologists reached unanimous consensus in conference on 218 (98%) of those 223 disagreements and resolved the remaining five (2%) with a two-to-three consensus. Of the 77 diagnoses where the disagreement included something other than the decision to call VM, 32 concerned coding issues of no clinical consequence. There were 39 (20%) fetuses with disagreement regarding the presence of major findings (Table 4).
Coding issues (82 of 223, 37%) and errors of observation (86 of 223, 39%) were most common types of disagreement (Table 5). Among the MR readers, head circumference did not vary according to the level of consensus concerning presence of VM (P = .24) but biparietal diameter was larger (mean, 10 mm) and gestational age was older (mean, 3.3 weeks) in fetuses where the ultimate MR consensus included VM (Table 6).
MR readers' confidence in the presence, character, or type of abnormality did not differ between examinations with agreement and disagreement (Fig 3), but their confidence in additional findings was significantly higher in examinations where they agreed (P < .03).
In five (3%) of 198 studies where neuroradiologists interpreted the images before and after use of US findings, the knowledge from the US findings changed the MR interpretation, including one fetus with intraventricular hemorrhage and porencephaly (seen on only three MR images in the study); and one fetus each with small subdural hematoma, megacisterna magna, cerebellar hypoplasia, and callosal dysgenesis. In one study, this was scored as having no effect on reading, in two studies as having some effect, and in three studies as having a critical effect on reading.
There was no relationship between gestational age and diagnosis disagreement at MR.
Comparing US and MR Readings
Of 147 US-MR pairs of examinations, the obstetric radiologist who performed and interpreted US and MR images was the same, including 99 (99%) of 100 from one institution and 46 (48%) of 96 from the second. There were five fetuses with diagnosis disagreement, with the final diagnosis being visualized on US and not MR, including two fetuses with cysts associated with ACC and one fetus each with ACC, porencephaly, and cerebellar hypoplasia with heterotopias (visualized at MR, as coded by neuroradiologists). There were 14 pregnancies with diagnosis disagreement with final diagnosis seen at MR and not at US when read by the same reader, including four fetuses with porencephaly; three fetuses with polymicrogyria; two fetuses with schizencephaly; and one fetus each with subependymal hemorrhage, spinal meningocele, encephalocele, congenital infarction, and septooptic dysplasia.
When comparing ventricular measurements from all seven readers, three for US and four for MR (Table 3), the recorded diameter was largest in fetuses for which US and MR consensuses included VM (mean, 13.5 mm) and smallest in fetuses where US and MR consensuses excluded VM (mean, 7.6 mm; P < .001). In studies where US and MR consensuses did not match, those readers favoring VM recorded larger ventricular diameter for a given fetus by a margin of over 2 mm in the mean diameter (10.8 vs 8.5 mm, P < .001). Variation among the readers for a given fetus and reader class was 0.4 mm, representing less than 1% of total variance. The intraclass correlation was 92%. MR measurements of ventricular diameter were greater than those from US by 0.6 mm on average (standard error, 0.3 mm; P = .05) for a given fetus, all other factors being equal (reader class and degree of agreement).
Of the 196 US-MR cases, there were 83 (42%) diagnoses with complete agreement by the seven readers. These diagnoses included normal (n = 11), isolated VM (n = 68), VM with NTD (n = 1), and VM with ACC (n = 3).
The consensus US opinion and the consensus MR opinion were in agreement on all diagnoses in 134 (68%) of 196 fetuses. The 62 (32%) fetuses with disagreements included 30 where the only disagreements were either to diagnose VM or were differences in coding judged to have no clinical effect (Table 4). The remaining 32 (16%) with diagnosis disagreement included 21 major new findings, six minor changes in diagnosis, four overcalls of major findings, and two overcalls of a minor finding. The 21 major new findings included fetuses with meningocele, porencephaly, hemorrhage, schizencephaly, and other migrational abnormalities (Table 1). The overcalls of major findings included a fetus at 22 weeks with an MR-aided diagnosis of ACC not agreed on at consensus, a fetus at 21 weeks with VM and ACC and a question of encephaloclastic porencephaly that was not agreed on at consensus, and a fetus at 35 weeks with VM and porencephaly with a question of Dandy Walker variant at US where the Dandy Walker variant was not agreed on at consensus.
Head size and gestational age were not associated with the US-MR consensus (Table 6).
The four radiologists performing MR varied considerably in their ratings of the effect of MR on diagnostic certainty beyond what US alone would provide (Table 7). Two obstetric radiologists had lower confidence scores at US than did the other two; those with the lowest confidence score at US rated the highest effect of MR. Adjusting for these individual tendencies, the effect of MR was significantly greater in examinations where the performing reader disagreed with the neuroradiologists on the diagnosis of VM by 0.4–0.8 points on a scale of 1–5, than in examinations where they agreed or judged their difference to be clinically unimportant.
|
There was no relationship between gestational age and diagnosis disagreement on US-MR comparisons.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
In addition, measurements obtained with MR were larger than those obtained with US, which can also affect the rate of diagnosis of VM. It is important to acknowledge that these measurement differences occur, since patients often have more than one imaging study in the work-up of VM. While the degree of VM may change over time, it is also possible that some differences in diagnosis result from individual variability in measurements and modality. It may be that fetuses diagnosed with mild VM in whom reviewers cannot agree if VM is present will have better outcomes than fetuses with consensus about the diagnosis of VM.
US and MR are both used to characterize fetal CNS abnormalities; each modality has specific strengths and weaknesses in helping depict abnormalities. For example, US may show the wall of an arachnoid cyst to better advantage than does MR, whereas MR frequently shows cortical migrational abnormalities to better advantage than does US. While these generalities are well recognized, the variability in individual interpretation of studies has not, to our knowledge, been assessed.
Diagnostic image interpretation is dependent on detection and characterization of findings. US-aided diagnosis is known to depend on the skill of the person obtaining the images to correctly display the anatomy and the skill of the reader to appropriately diagnose the abnormality. When assessing US, we are limited by the images that have been obtained, with real-time scanning at times aiding in diagnosis. The Routine Antenatal Diagnostic Imaging with US study demonstrated the importance of user experience in fetal anomaly detection, since highly experienced ultrasonographers detected 35% of anomalies compared with only 13% by less-experienced US readers (13). In our study, we demonstrated that even individuals with expertise in obstetric US, when evaluating the same images, can arrive at different fetal diagnoses.
When assessing CNS anomalies, another factor to consider is the subspecialty training of the individual interpreting the findings of the examination. The pediatric neuroradiologists in our study diagnosed more abnormalities when interpreting fetal MR than did the obstetric radiologists. It is likely that some of the perceived benefit of MR in previously published studies has come from the effect of having neuroradiology input with the final diagnosis.
Another factor to consider in US-aided diagnosis of fetal CNS anomalies is experience with visualization of the specific anomalies. We have found improved recognition of such diagnoses as ACC and heterotopias at US in our study compared with prior studies (14,15), likely owing to improved understanding of how these anomalies appear in utero.
Prior studies that compared US with MR in aiding the diagnosis of fetal brain abnormalities stated the importance of having high-quality US images available for comparison to assess the incremental benefit of MR (1). However, in clinical practice this may not be as important a factor. In our study of a highly enriched population of fetal CNS abnormalities, knowledge of the confirmatory US findings was only judged to have affected diagnosis in 2.5% of examinations. This is a reassuring finding for clinical use of fetal MR because in many centers, the imaging expertise with obstetric US and fetal MR do not coexist in the same radiology department.
We found 29 (14.8%) of 196 US and 39 (19.9%) of 196 MR examinations had disagreements about major findings being present. Our disagreements and types of errors are similar to those originally categorized by Smith (16) in 1967 and updated by Renfrew et al (17) These studies show false-positive errors of overreading and misinterpretation, true-positive readings with misclassification, and false-negative errors. We found false-negative diagnoses in 33% and 39% of US and MR disagreements, respectively.
An example of a false-negative finding that can be attributed to satisfaction of search was in a fetus with VM with a small meningocele without Chiari malformation. It is likely that a more extensive US search of the spine and cranium would have depicted this small lesion. An example of false-negative findings at MR is the difficulty in navigation through large data sets where error can occur, as in a fetus with porencephaly in the frontal horn that was noted by one neuroradiologist only after confirmatory US results were given, indicating the region of abnormality.
Other reasons for false-negative findings were lack of real-time scanning for US diagnoses in 3% of US disagreements and lack of neuroradiology training for US-MR comparisons in 15% of MR disagreements.
After consensus, we found that MR added additional diagnoses beyond those found with US. In particular, fetuses with migrational abnormalities and porencephaly were diagnosed at MR but not at US. These findings are similar to findings in other studies that have assessed the effect of MR on prenatal diagnosis of CNS anomalies (1,18–23). This reinforces the added value of MR in fetuses referred for VM. Another factor that affects the perceived benefit of MR is the confidence in diagnosis of the radiologist. In our study, two of the obstetric radiologists had lower confidence scores than did the other two. The individual with the lowest confidence score for US rated the highest effect for MR.
A limitation of our study was the lack of a true reference standard. We used the opinion of the US-MR consensus conference as our final diagnosis. Because our study assessed the reproducibility of reporting of individual findings rather than the sensitivity of detecting a final outcome at birth, we did not feel that postnatal outcome was needed. This reference standard of imaging concordance has been used by others (24–26).
A second limitation was that the study situation may have led the radiologists to identify more subtle abnormalities than would otherwise be noted. However, in a clinical setting of a high-risk population of fetuses referred for VM, the threshold level for suggesting abnormality is likely similar. The alternative is also possible, in that the effects of the study situation might have led second-opinion readers to take less care than usual, knowing that their impressions would not be used for immediate patient care.
A third limitation was that by coding abnormalities, there was no "gray zone." Radiologists had to decide if an anomaly was present, with no option to indicate uncertainty in each finding. This was necessary to compare the readings of the individuals but probably led to an overestimate of discrepancies.
A fourth limitation was that the results are only applicable to fetuses with VM. Our patient population does not reflect that which would be found in a screening population or among those referred for MR for any other indication. In a screening population, we would expect much higher concordance in findings, since a large percentage of the fetuses would be normal.
In conclusion, among radiologists who read high-risk obstetric US and fetal MR for VM, there is considerable variability in CNS diagnoses. A reduction in variability will require more consistent criteria for diagnostic interpretation. We assessed some of the reasons for missing and misdiagnosing some of the anomalies. Knowledge of the appearance of anomalies seen in association with VM should help to improve the accuracy of prenatal diagnosis.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
statistic, a dimensionless index with = 0–0.20 indicating poor agreement, = 0.21–0.40 indicating fair agreement, = 0.41–0.60 indicating moderate agreement, = 0.61–0.80 indicating good agreement, and = 0.81–1.00 indicating excellent agreement (27,28). Statistics were separately calculated for the most common diagnoses for the following groups: the group of four obstetric radiologists (three of whom read a given scan); the group of three pediatric neuroradiologists (all of whom read each MR); the group of four readers who participated in MR diagnoses (one obstetric radiologist and three neuroradiologists for each image); and the entire group of seven US and MR readers and the consensus US and MR readings. The association between head circumference, biparietal diameter, and gestational age and diagnostic agreement concerning VM was evaluated by dividing the cases into four classes according to the degree of agreement achieved in subsequent stages of the study: (a) preconference consensus of VM, (b) preconference consensus on absence of VM, (c) initial disagreement resolved as a final diagnosis of VM, and (d) initial disagreement resolved as absence of VM. This classification was assigned separately for each stage of the study (US, MR, and US-MR). Comparison across the classes was made by using analysis of variance. To account for within-reader correlation, the analysis of variance included a random effect identifying the obstetric radiologist who performed the measurement.
Ventricular diameter was measured by all readers, permitting us to examine the association between individual reader measurements of ventricular size with respect to fetal head size and/or gestational age and the subsequent degree of disagreement concerning presence of VM. We divided the cases into three classes: (a) all readers at preconference who agreed on presence of VM, (b) all readers at preconference who agreed on absence of VM, and (c) readers who disagreed regarding VM. Measurements for the third class were further subdivided as those from individual readers who scored VM as either present or absent.
The entire set of measurements (three per fetus for US, four per fetus for MR, and seven per fetus for US-MR) was subjected to analysis of variance, with the class of agreement as independent variable. Analyses that included MR measurements were adjusted for a dichotomous indicator of whether the reader was an obstetric radiologist or a neuroradiologist. Analysis of the combined MR-US measurements was additionally adjusted for modality, enabling us to compare US with MR, controlling for all other factors. To account for variability among fetuses and systematic variability among individual readers, the analysis of variance included random effects. Intraclass correlation, defined as the fraction of variance attributable to true variability in the population (as opposed to measurement artifact), was calculated as interfetus variance divided by the sum of interfetus, interreader, and residual variance.
In two fetuses that were coded as holoprosencephaly by all readers, lateral ventricles were not measured by some readers, and VM was not coded by some readers. These cases were excluded from the above analyses of ventricular size.
The effect of MR on visualization of anomalies was assessed by comparing scores from readers who read US and MR images, and by comparing the consensus diagnoses reached at US and MR. Confidence scores were compared for fetuses with and without diagnosis disagreement. Analysis of variance was employed for the analyses of confidence and effect, with reader effects assessed from fitted parameters of the model.
The association between gestational age and diagnosis disagreement was performed with the Student t test.
A P value of less than .05 was considered to indicate a significant difference. All computations were performed with software (SAS, version 9.1; SAS Institute, Cary, NC).
|
ADVANCES IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
|
IMPLICATION FOR PATIENT CARE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
|
FOOTNOTES |
|---|
Abbreviations: ACC = agenesis of the corpus callosum • CNS = central nervous system • NTD = neural tube defect • VM = ventriculomegaly
Author contributions: Guarantors of integrity of entire study, D.L., C.D.R.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, D.L.; clinical studies, D.L., J.F.K.T., J.A.E., C.D.R., T.Y.P., C.E.B., T.S.M., R.L.R.; statistical analysis, D.L., H.A.F.; and manuscript editing, D.L., H.A.F., J.A.E., T.Y.P., T.S.M., R.L.R.
Authors stated no financial relationship to disclose.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXADVANCES IN KNOWLEDGEIMPLICATION FOR PATIENT CAREReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  K Melchiorre, M Liberati, C Celentano, S Domizio, C Puglielli, S Buoni, M Strambi, and R Zannolli Neurological outcome following isolated 10-12 mm fetal ventriculomegaly Arch. Dis. Child. Fetal Neonatal Ed., July 1, 2009; 94(4): F311 - F312. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
= Two fetal US with NTD findings were incomplete at MR and not included in US consensus. 
= MR consensus performed by three pediatric neuroradiologists used to assess need for consensus diagnosis by obstetric radiologist who read MR, diagnoses of neuroradiologists.
