Dachling Pang

Spinal cord injury without radiographic abnormality in children, 2 decades later.

OBJECTIVE:Much new research has emerged since1982, when the original description of spinal cord injury without radiographic abnormality (SCIWORA) as a self-contained syndrome was reported. This article reviews new and old data on SCIWORA, from the past 2 decades. METHODS:This article reviews what we have learned since 1982 about the unique biomechanical properties of the juvenile spine, the mechanisms of injuries, the profound influence of age on injury pattern and outcome, the magnetic resonance imaging (MRI) features, and management algorithms of SCIWORA. RESULTS:The increasing use of MRI in SCIWORA has yielded ample evidence of damage in virtually all nonbony supporting tissues of the juvenile vertebral column, including rupture of the anterior and posterior longitudinal ligaments, intervertebral disc disruption, muscular and interspinal ligament tears, tectorial membrane rupture, and shearing of the subepiphyseal growth zone of the vertebral endplates. These findings provide the structural basis for the postulated “occult instability” in the spine of a patient after SCIWORA. MRI also demonstrated five classes of post-SCIWORA cord findings: complete transection, major hemorrhage, minor hemorrhage, edema only, and normal. These “neural” findings are highly predictive of outcome: patients with transection and major hemorrhage had profoundly poor outcome, but 40% with minor hemorrhage improved to mild grades, whereas 75% with “edema only” attained mild grades and 25% became normal. All patients with normal cord signals made complete recovery. The large pool of clinical data from our own and other centers also lends statistical power to uphold most of our original assertions regarding incidence, causes of injury, pathophysiology, age-related changes in the malleability of the spine, vectors of deformation, and the extreme vulnerability of young children to severe cord injury, particularly high cervical cord injury. Thoracic SCIWORA has been identified as an important subset, comprising three subtypes involving high-speed direct impact, distraction from lap belts, and crush injury by slow moving vehicles. Computation of the sensitivities of MRI and somatosensory evoked potentials in detecting SCIWORA shows that both tests were normal in 12 to 15% of children with definite, persistent myelopathy; all of these children were nevertheless braced for 3 months because of their clinical syndrome. Children with transient deficits but abnormal MRI and/or somatosensory evoked potentials were also braced, but the 60% with transient deficits and normal MRI and somatosensory evoked potentials were not braced. This is a change from our original policy in 1982 of bracing all children with persistent or transient deficits, brought on by our new MRI and electrophysiology data. CONCLUSION:Injury prevention, prompt recognition, use of MRI and electrophysiological verification, and timely bracing of SCIWORA patients remain the chief measures to improve outcome.

Low-pressure hydrocephalic state and viscoelastic alterations in the brain.

Most shunt-dependent hydrocephalic patients present with predictable symptoms of headache and mental status changes when their cerebrospinal fluid shunts malfunction. Their intracranial pressure (ICP) is usually high, and they usually respond to routine shunt revision. This report describes 12 shunted patients who were admitted with the full-blown hydrocephalic syndrome but with low to low-normal ICP. All 12 patients had been maintained previously on medium-pressure shunts. Their symptoms included headache, lethargy, obtundation, and cranial neuropathies. At peak symptoms, their ventricular sizes were large (ventricular/biparietal ratio of 0.35 to 0.45) in six and massive (ventricular/biparietal ratio > 0.45) in six and their ICPs ranged from 2.2 to 6.6 mm Hg, with a mean of 4.4 +/- 1.3 mm Hg (+/- standard deviation), i.e., below or well within the pressure range of their shunts. The pressure volume index of three patients at peak symptoms ranged from 39.2 to 48.5 ml, with a mean of 43.9 +/- 4.6 ml, which represents a 190% increase from the predicted normal value. Seven patients failed to improve with multiple shunt revisions, including the use of low-pressure valves. In 11 patients, symptoms and ventriculomegaly were not reversed except with prolonged external ventricular drainage at subzero pressures (mean external ventricular drainage nadir pressure of -5.7 +/- 3.6 mm Hg, for a mean period of 22.2 days). During external ventricular drainage treatment, symptoms correlated only with ventricular size and not with ICP. All 11 were subsequently treated successfully with a new medium- or low-pressure shunt. One patient was treated successfully with prolonged shunt pumping. We postulate that: 1) the development of this low-pressure hydrocephalic state is related to alteration of the viscoelastic modulus of the brain, secondary to expulsion of extracellular water from the brain parenchyma, and to structural changes in brain tissues due to prolonged overstretching; 2) certain patients are susceptible to developing low-pressure hydrocephalic state because of an innate low brain elasticity due to bioatrophic changes; 3) low-pressure hydrocephalic state symptoms are due not to pressure changes but to brain tissue distortion and cortical ischemia secondary to severe ventricular distortion and elevated radial compressive stresses within the brain; and 4) treatment must be directed toward allowing the entry of water into the brain parenchyma and the restoration of baseline brain viscoelasticity.

AtlantoAxial Rotatory Fixation: Part 3—A Prospective Study of the Clinical Manifestation, Diagnosis, Management, and Outcome of Children with AlantoAxial Rotatory Fixation

OBJECTIVE:This is a prospective study of the clinical manifestations, diagnostic motion analysis, management, and outcome of children with atlantoaxial rotatory fixation (AARF). METHODS:Fifty children presenting with painful torticollis were subjected to the three-head positions diagnostic computed tomographic scanning protocol described in Part II of our AARF study. Twenty-nine children qualified as having AARF (8 Type I, 11 Type II, and 10 Type III), and six children were classified in the diagnostic gray zone (DGZ). The AARF patients were given either halter or calipers traction depending on the type and chronicity of pretreatment delay. Upon reduction, patients were immobilized with either a cervicothoracic brace or a halo. Recurrence of AARF on halo and patients whose deformity was not reducible were given posterior C1C2 fusion at the best achievable alignment. The difficulty and results of treatment were measured according to the following: duration of traction, number of reduction slippage, percent not reducible by traction, percent needing halo, percent needing fusion, total duration of treatment, total number of treatment procedures, and percent who lost normal C1C2 dynamics. Results were compared between groups stratified by AARF types, by chronicity of pretreatment delay (acute ≪ 1 mo, subacute = 1–3 mo, chronic ≥ 3 mo) and by the presence or absence of recurrence (recurrent AARF defined as having two or more slippages). DGZ patients were treated with only comfort measures for 2 weeks and then restudied. Only those children with persistent symptoms and DGZ or worse motion dynamics were given traction and bracing. RESULTS:Neither age nor etiology significantly influenced the severity of AARF. There was only a slight tendency for children younger than 5 years, and for trauma, to associate with severe C1C2 interlock. Delay of treatment up to 11 months did not result in improvement of the neck restriction or in abatement of pain. In fact, there are strong suggestions that prolonged delay could lead to worsening of the rotatory dynamics: Type I AARF are highly correlated with delays longer than 3 months and Type III with delays less than 1 month. Also, four patients who had serial motion studies during the delay period showed clear worsening in the pathological stickiness in C1C2 rotation. In addition, chronic rotatory deformity led to progressive occiput −C1 separation or laxity teleologically to compensate for a skewed visual axis. The mean occiput −C1 separation angle for chronic patients was 31.2° versus 5° for acute patients and less than 3° for normal children. The difficulty and duration of treatment, the number of reslippage after reduction, the rate of irreducibility, the need for halo and fusion, and the percentile of patients ultimately loosing normal C1C2 rotation were significantly greater with Type I patients than Type III patients, with Type II patients being intermediate. Likewise, chronic patients of all AARF types were much worse in all parameters than acute patients; subacute patients were closer to chronic patients in complexity and outcome. Severity and chronicity exerted independent effects on outcome, and the worse identifiable subgroup were the chronic Type I patients versus the best subgroup of acute Type III patients. Thirteen patients developed recurrent AARF; they had much worse prognosis in all aspects measured than nonrecurrent patients. Recurrence was adversely influenced by both the severity (type) and chronicity of AARF. Half of the DGZ patients resolved with analgesics, but two of six remained symptomatic and in DGZ dynamics, and one deteriorated to Type III AARF. Two of those three patients responded easily to traction and bracing, and one was lost to follow-up. CONCLUSION:Children with painful torticollis should be subjected to the three-position computed tomographic diagnostic protocol, not only to secure the diagnosis of AARF but also to grade the severity of the condition by virtue of the dynamic motion curve. Closed reduction with traction should be instituted immediately to avoid the serious consequences of chronic AARF. Proper typing and reckoning of the pretreatment delay are requisites for selecting treatment modalities. Recurrent dislocation and incomplete reduction should be treated with posterior C1C2 fusion in the best achievable alignment. Open reduction and halo immobilization to avoid permanent fixation can be tried with select cases.

Thoracoscopic Techniques for the Treatment of Scoliosis: Early Results in Procedure Development

OBJECTIVE The goal of this study was to determine the effectiveness of an endoscopic option for anterior approaches to the thoracolumbar spine for scoliosis treatment. Fifty patients with 24 to 45 months of follow-up data were retrospectively studied. Techniques for endoscopic treatment of spinal disorders have been under development since 1993. The benefits of thoracoscopic surgery for the treatment of spinal deformities have been documented as improved observation of the spine, enhanced access to the extremes of the curve, decreased operative times and blood losses, shorter hospital stays and recuperative periods, and decreased overall costs. After more than 150 endoscopic procedures for the treatment of these spinal deformities had been performed, the next step was to develop a thoracoscopic technique for instrumentation, correction, and fusion for the treatment of primary thoracic scoliosis. Our goal has been to develop a safe, reproducible, and effective endoscopic technique for the treatment of scoliosis that can provide equal or better outcomes, compared with formal open surgical techniques. METHODS Between October 1996 and October 1998, 50 patients with a diagnosis of primary thoracic scoliosis were selected to undergo thoracoscopic instrumentation, correction, and fusion. Postoperatively, patients were assessed with respect to restoration of spinal alignment, axial derotation, pain management, and incidence of complications. RESULTS Endoscopic instrumentation was successfully performed for all patients. Curve correction averaged 50.2%, improving to 68.6% in the last 10 cases. Patients with hypokyphosis averaged 20.7 degrees of correction. The preoperative axial rotation (as measured with a scoliometer) averaged 16 degrees, which was corrected to 5 degrees. Postoperative pain was less, and patients could discontinue the use of all pain medications by 1 to 3 weeks, compared with patients who underwent formal open procedures, who required pain medication for 6 to 12 weeks. The hospital stays averaged 2.9 days. Our initial complication rate was high, which could be attributed to the development of a new technique. Keys to successful fusions include total discectomy, complete endplate removal, and the use of autogenous bone graft. CONCLUSION Although these techniques are still in early development, the initial results for our thoracoscopic techniques are promising. With experience, surgical times are decreasing and fusion and curve correction rates are improving. With further evolution, patients should realize shortened hospitalizations, decreased rehabilitation times, and decreased levels of postoperative pain. This is a technically demanding procedure that requires demonstrated skills in endoscopic discectomy and fusion.

Long-term outcome of total and near-total resection of spinal cord lipomas and radical reconstruction of the neural placode: part I-surgical technique.

OBJECTIVEPartial resection of complex spinal cord lipomas is associated with a high rate of symptomatic recurrence caused by retethering, presumably promoted by a tight content-container relationship between the spinal cord and the dural sac, and incomplete detachment of the terminal neural placode from residual lipoma. Since 1991, we have performed more than 250 total/near-total resections of complex lipomas with radical reconstruction of the neural placodes. Sixteen years of follow-up have proven the long-term benefits of this technique. Part I of this series introduces our technique of total resection and reports the immediate surgical results. Part II will analyze the long-term outcomes of both total and partial resection and identify the factors affecting outcome. METHODSFrom 1991 to 2006, 238 patients (age range, 2 months–72 years) with dorsal, transitional, and chaotic lipomas underwent total or near-total lipoma resection and radical placode reconstruction. Eighty-four percent of the patients were children younger than 18 years and 16% were adults. The technique consisted of wide bony exposure, complete unhinging of the lateral adhesions of the lipoma-placode assembly from the inner dura, untethering of the terminal conus, radical resection of the fat off the neural plate along a white fibrous plane at the cord-lipoma interface, meticulous pia-to-pia neurulation of the supple neural placode with microsutures, and expansile duraplasty with a bovine pericardial graft. Elaborate electrophysiological monitoring was used. RESULTSThree postoperative observations concern us. The first is that of the 238 patients, 138 (58%) had no residual fat on postoperative magnetic resonance imaging; 81 patients (36%) had less than 20 mm3 of residual fat, the majority of which were small bits enclosed by neurulation; and 19 patients (8%), mainly of the chaotic lipoma group, had more than 20 mm3 of fat. There are no significant differences in the amount of residual fat among lipoma types, but redo lipomas are more likely than virgin (previously unoperated on) lipomas to have residual fat by a factor of 2 (P = 0.0214). The second concern is that the state of the reconstructed placode is objectively measured by the cord-sac ratio, obtained by dividing the sagittal diameter of the reconstructed neural tube by the sagittal diameter of the thecal sac. A total of 162 patients (68%) had cord-sac ratios less than 30% (low), 61 (25.6%) had ratios between 30% and 50% (medium), and only 15 (6.3%) had high ratios of more than 50%. Seventy-four percent of patients with virgin lipomas had low cord-sac ratios compared with 56.3% in the redo lipoma patients. The overall distribution of cord-sac ratio is significantly different between redo and virgin lipomas (P = 0.00376) but not among lipoma types. Finally, the incidence of combined neurological and urological complications was 4.2%. The combined cerebrospinal fluid leak and wound infection/dehiscence incidence was 2.5%. Both sets of surgical morbidity compared favorably with the published rates reported for partial resection. CONCLUSIONTotal/near-total resection of spinal cord lipomas and complete reconstruction of the neural placode can be achieved with low surgical morbidity and a high yield of agreeable postoperative cord-sac relationship. Some large rambling transitional lipomas and most chaotic lipomas are the most difficult lesions to resect and tend to have less favorable results on postresection magnetic resonance imaging.

ATLANTO-OCCIPITAL DISLOCATIONPART 1—NORMAL OCCIPITAL CONDYLE-C1 INTERVAL IN 89 CHILDREN

OBJECTIVEAlthough atlanto-occipital dislocation has long been recognized as an extremely unstable and often lethal injury, no single radiodiagnostic criterion published to date has achieved failure-proof status. This is because most existing diagnostic tests exploit bony landmarks remote from the injured condyle–C1 (OC1) joint so that patient positioning could inadvertently line up these landmarks and conceal actual disruption of the joint. Many of the landmarks used are wide apart and/or noncoplanar; their measurements are subject to errors related to x-ray angle, target–film distance, and superimposed bony outlines. We propose using the actual occipital condyle–C1 interval (CCI) obtained from high-resolution reconstructed computed tomographic scans as the indicator for OC1 joint disruption. We hypothesize that the normal CCI is very small and has great left-right symmetry and that atlanto-occipital dislocation is always manifested by an abnormal widening of the CCI and/or by left-right joint asymmetry irrespective of the shifting of other remote bony landmarks. Part I of this study establishes standard normal values for CCI in children. METHODSagittal and coronal reformatted images were obtained from thin axial computed tomographic scans performed on 89 children, 18 for nontraumatic complaints and 71 as part of a minor head trauma protocol but later exonerated for cervical injury. The interval between condyle and C1 was measured at four equidistant points on the joint surface on the sagittal and coronal images of all 178 joints (left and right) in the group. Sagittal and coronal CCIs are the means of four sagittal and four coronal measurements, respectively. The combined or true CCI for an individual joint was taken as the mean of both the sagittal and coronal (total of eight) measurements. RESULTSThe mean combined CCI of all 178 joints was 1.28 mm ± 0.26 (standard deviation [SD]). None of the 178 CCIs exceeded 1.95 mm, and none of the individual joint interval measurements exceeded 2.5 mm. Left-right symmetry was tested by computing the mean left-right difference in CCI from all 89 subjects equal to 0.047 mm ± 0.002 (SD); or only 3% of the mean combined CCI. When the mean of all right CCIs (1.333 mm ± 0.31 SD) is contrasted with the mean of all left CCIs (1.327 mm ± 0.30 SD), the difference is 0.006 mm, or 1.09% of the mean total CCI (P = 0.792). Left-right symmetry is also apparent in conformational anatomy in both sagittal and coronal images. Linear regression analysis between CCI and age shows no statistical difference in CCI between age groups from 0.5 to 18 years. Linear regression performed separately on the right and left CCIs suggests that left-right symmetry is also stable through this age range. CONCLUSIONThe normal OC1 joint in children 0 to 18 years is tightly held together by ligaments with a mean CCI of 1.28 mm in the 89 subjects tested. There is great left–right joint symmetry in both CCI and conformational anatomy. CCI and left-right symmetry do not appear to change significantly with age. It is reasonable to set a maximum CCI as a discriminator between normal and disrupted OC1 joints to indicate atlanto-occipital dislocation.

Long-term outcome of total and near-total resection of spinal cord lipomas and radical reconstruction of the neural placode, part II: outcome analysis and preoperative profiling.

OBJECTIVETo show the long-term benefits of total and near-total resection of complex spinal cord lipomas and reconstruction of the neural placode. METHODSWe analyzed 238 patients with dorsal, transitional, and chaotic lipomas who had total resection as described in part I for overall progression-free survival probability (PFS, Kaplan-Meier analysis) over 16 years. We also analyzed subgroup proportional recurrence hazard (Cox analysis) of 6 outcome predictors of sex, lipoma type, age, preoperative symptoms, previous surgery, and postoperative cord-sac ratio. These results were compared with an age-matched, lesion-matched series of 116 patients followed for 11 years after partial lipoma resection and with the Parisian series of nonsurgical treatment. RESULTSThe immediate effects of surgery were similar between total and partial resection: both achieved greater than 95% symptom stabilization or improvement rate. The neuro-urologic complication rates for the groups were also similar, 4.2% and 5.2% for total and partial resection, respectively. The combined cerebrospinal fluid leakage and wound complication rate of total resection was much lower at 2.5% than the 6.9% for partial resection, but both were better than published rates. The overall PFS for total resection was 82.8% at 16 years, comparing much more favorably with 34.6% for partial resection at 10.5 years (P < .0001). Culling only the asymptomatic patients with virgin (previously unoperated) lipomas to match the patient profile of the Parisian series, the PFS for prophylactic total resection for this subgroup increased to 98.4% at 16 years, versus 67% at 9 years for no surgery and 43.3% at 10.5 years for our own partial resection series, with a remarkable statistical difference between total and partial resection (P = .00001). Subgroup analyses showed that sex and lipoma type did not affect outcome. For the other predictor variables, while univariate analyses showed that young age, absence of symptom, and virgin lipomas correlated with better statistical PFS than older age, symptoms, and redo lipomas, these effects vanished with multivariate analyses. Cord-sac ratio stood alone as the only influential outcome predictor in multivariate analysis, with a 96.6% PFS for a low ratio of <30% and an 80.6% progression-free probability for a high ratio of >50%, and a 3-fold increase in recurrence hazard for high ratios (P = .0009). This suggested that all the individual effects of the other predictor variables could be reduced to whether a low cord-sac ratio could be achieved with total lipoma resection and placode reconstruction. Cord-sac ratio was the obvious factor that differentiated the outcomes between total and partial resection, the latter associated with a >90% chance of having a high cord-sac ratio. CONCLUSIONTotal and near-total resection of lipomas and complete reconstruction of the neural placode produced a much better long-term progression-free probability than partial resection and nonsurgical treatment. The perioperative complications for total resection were low and compared favorably with published results. A low postoperative cord-sac ratio and well-executed placode neurulation were strongly correlated with good outcome. The ideal preoperative patient profile with early disease stabilization and the best recurrence-free probability is an asymptomatic child less than 2 years without previous lipoma surgery. There are strong indications that partial resection in many cases produces worse scarring on the neural placode and worse prognosis than no surgery.

Atlantoaxial rotatory fixation: part 2--new diagnostic paradigm and a new classification based on motion analysis using computed tomographic imaging.

OBJECTIVE:This is Part II of a study on atlantoaxial rotatory fixation (AARF) that aims to introduce a new diagnostic paradigm and a new classification of this condition based on motion analysis of C1C2 rotation using computed tomographic (CT) imaging. This phase of the study is possible because Part I succeeded in defining physiological C1C2 axial rotation with CT data from 21 normal children, displayed in a highly concordant composite motion curve, which is used as the normal template for the present study. AARF is defined as flagrant departure from normal motion dynamics as delineated by abnormal motion curves. The new classification is predicated on the graded amount of pathological stickiness in the restricted rotation. METHODS:Forty children age 1.5 to 14 years with painful “cock-robin” necks resulting from minor trauma or otolaryngological procedures were subjected to 3 CT examinations: 1) in the presenting (P) position; 2) with the nose pointing up (P0 position); and 3) with the head forcefully turned to the opposite side as much as the patient could tolerate (P_ position). The angles made by C1 and C2 and the separation angle C1C2° (C1 minus C2 degrees) were obtained as described in Part I. The test motion curve was generated by plotting C1 against C1C2 angles, and all motion curves were analyzed in the context of the normal template. RESULTS:Five distinct groups with highly characteristic motion curves could be identified. Group 1 (n = 5) patients showed essentially unaltered (“locked”) C1C2 coupled configurations regardless of corrective counterrotation, with curves that are horizontal lines in the upper two quadrants of the template. Group 2 (n = 7) patients had reduction of the C1C2 separation angle with forced correction, but C1 could not be made to cross C2. Their curves slope downward from right to left in the upper quadrants but never traverse the x axis. Group 3 (n = 9) patients showed C1C2 crossover, but only when the head was cranked far to the opposite side. Their motion curves traverse the x axis left of C1 = −20 degrees. Groups 1, 2, and 3 motion dynamics are respectively classified as Types I, II, and III AARF in descending degree of pathological stickiness, which is in essence a resistance against closure of the C1C2 angle to counterrotation. Group 4 (n = 14) patients had normal dynamics, and Group 5 (n = 5) patients showed motion curve features between normal and Type III AARF, designated as belonging to the diagnostic gray zone, an uncertain group that may or may not revert to normal dynamics with only comfort measures. CONCLUSION:AARF can be reliably diagnosed with a simple and practical CT protocol and construction of a three-point motion curve superimposed on a reusable normal template. The type of AARF, reflective of the severity of pathological stickiness of rotation, can be identified readily by the shape of the motion curve. This system of classification is useful in selecting the best regimen of management.

Atlantoaxial rotatory fixation.

OBJECTIVEAtlantoaxial rotatory fixation (AARF) remains a recondite entity. Our normative study using CT motion analysis shows that there is a high degree of concordance for rotational behavior of C1 and C2 in children 0 to 18 years. C1 always crosses C2 at or near 0 degree. The predictable relationship between C1 and C2 is depicted by 3 distinct regions on the motion curve: when C1 rotates from 0 to 23 degrees, it moves alone while C2 remains stationary at 0 (the single-motion phase). When C1 rotates from 24 to 65 degrees, C1 and C2 move together (the double-motion phase), but C1 always moves faster as C2 is being pulled by yoking ligaments. From 65 degrees onward, C1 and C2 move in unison (the unison-motion phase) with a fixed, maximal separation angle of approximately 43 degrees, the head rotation being carried exclusively by the subaxial segments. Because of this high concordance among patients and a relatively narrow variance from the mean, the physiological composite motion curve can be used as a normal template for the diagnosis and classification of AARF. METHODSUsing a 3-position CT protocol to obtain the diagnostic motion curve, we identified 3 distinct types of AARF. Type I AARF patients show essentially unaltered (“locked”) C1–C2 coupling regardless of corrective counterrotation, with curves that are horizontal lines in the upper 2 quadrants of the template. Type II AARF patients show reduction of the C1–C2 separation angle with forced correction, but C1 cannot be made to cross C2. Their curves slope downward from the right to left upper quadrants but never traverse the x axis. Type III AARF patients show C1–C2 crossover but only when the head is cranked far to the opposite side. Their motion curves traverse the x axis far left of 0 degree ( C1 < −20). Thus, type I, II, and III AARF are in descending degrees of pathological stickiness. A fourth group of patients showing motion curve features between normal and type III AARF are designated as belonging to a diagnostic gray zone (DGZ). The AARF patients are further classified as acute if treatment is started less than 1 month from the onset of symptoms, as subacute if the delay in treatment is 1 to 3 months, and chronic if treatment delay exceeds 3 months. The treatment protocol for AARF consists of reduction using either halter or caliper traction and then immobilization with brace or halo, depending on the AARF type and chronicity. Recurrent slippage and irreducibility are treated with C1–C2 fusion. RESULTSThe treatment course and outcome of AARF are analyzed according to the AARF type and chronicity. The difficulty and duration of treatment, the number of recurrent slippage, the rate of irreducibility, the need for halo and fusion, and the percentage ultimately losing normal C1–C2 rotation are significantly greater in type I patients than type III patients, with type II patients somewhere in between. Likewise, all parameters are much worse in patients with any type of chronic AARF than acute AARF. The worse subgroup is chronic type I versus the best subgroup of acute type III. Recurrent AARF patients do much worse than nonrecurrent AARF patients. Recurrence is, in turn, adversely influenced by both the severity (type) and chronicity of AARF. The symptoms of most DGZ patients will resolve with analgesics, but a few remain symptomatic or deteriorate to true AARF requiring the full treatment. CONCLUSIONThus, children with painful torticollis should undergo the 3-position CT protocol not only to confirm the diagnosis of AARF but also to grade its severity. Closed reduction with traction should be instituted immediately to avoid the serious consequences of chronicity. Proper typing and reckoning of the pretreatment delay are requisites for selecting treatment modalities. Recurrent dislocation and incomplete reduction should be treated with posterior C1–C2 fusion in the best achievable alignment.

A magnetic resonance template for normal cerebellar development in the human fetus

OBJECTIVE Although ultrasound is the primary imaging modality for prenatal anatomic evaluation, some central nervous system malformations may be better defined with high-resolution magnetic resonance imaging (MRI). MRI allows us to visualize the features of brain development that were previously only seen histologically by embryologists and anatomists. Although there are several reports of the postnatal development of the cerebellum as revealed on magnetic resonance (MR) images, systematic MR studies of cerebellar development during the fetal period are lacking. Our objective was to use high-resolution MRI to provide a template of cerebellar development during the late first and early second trimesters, a period when the diagnosis of congenital malformations is most medicoethically relevant. The MR findings were then correlated with histological data. METHODS Twenty-six normal formalin-fixed fetal specimens with a gestational age of 9 to 24 weeks were examined with high-resolution MRI using a conventional clinical magnet and pulse sequences. The MR findings were correlated with the whole-mount histological specimens catalogued in a well-known fetal atlas. RESULTS Resolution of the morphological features of cerebellar development in fetuses greater than 10 weeks gestational age was possible. Development of the rhombic lips, vermis, fourth ventricular roof, foramen of Magendie, and the cerebellar fissures was documented. Development of the cerebellum as revealed on MR images lagged behind the known stages of development by as much as 5 weeks. Features of cerebellar histogenesis were beyond the resolution of MRI. However, differences in signal intensity between gray and white matter of the developing cerebellum were detected and are postulated to represent differences in cellularity and water content of the constituent tissues. CONCLUSION Direct correlation of MR images of fetuses during the late first and early second trimesters with anatomic atlases could result in a mistaken diagnosis of delayed or abnormal development of the posterior fossa contents because of a time lag in the detection of structures on MR images. An MR template of normal cerebellar development would be useful to avoid confusion of normal development with abnormal development and to identify the expected developmental features when provided the estimated gestational age of a fetus.