Introduction

VACTERL association, characterized by the presence of three or more features including vertebral defects, anorectal malformations (ARM), tracheo-esophageal fistula/esophageal atresia, radial/renal dysplasia, and limb abnormalities, has been extensively studied since its introduction in the late 1970s. The condition now exhibits a worldwide prevalence of approximately 1 in 10,000 to 1 in 40,000 live-born infants1. The literature has not only focused on the treatment of these conditions but also on the accurate diagnosis of associated anomalies, enabling clinicians to devise precise treatment plans for patients before the onset of symptoms. Routine spinal MRI surveys to identify any spinal anomalies are recommended for all patients with ARM, due to the high prevalence of concurrent anomalies2,3,4. This recommendation has been widely adopted by clinicians across various regions and implemented in clinical practice.

Among the spinal anomalies observed in patients with VACTERL associations, tethered cords and dysraphism are the most commonly reported in comorbidity with ARM4,5; consequently, many institutions prioritize the lumbosacral (LBS) spine in their evaluations. Similarly, in our hospital, LBS MRI was routinely conducted for patients diagnosed with ARM. However, the occurrence of neurological symptoms such as limb weakness and the subsequent detection of high-level spinal lesions (in the brain, cervical, or thoracic spine) following additional MRI series prompted us to consider expanding the range of initial spinal MRI evaluations. We now routinely perform whole spine (WS) MRI for every patient diagnosed with anorectal malformation. This report presents our experiences since this change in practice and the differences observed compared to previous protocols.

Materials and methods

All patients diagnosed with ARM and treated at our institution were retrospectively reviewed from January 2011 to January 2022. The research was approved by the Institutional Review Board I of Taichung Veterans General Hospital (certificate No. CE23413A) and informed consent was exempted as per the instructions. The study was conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The classification of ARM was based on traditional classification: high, intermediate and low types6. According to clinical findings, invertograms or the cross-table lateral radiography in prone position. We ensured that at least one spinal MRI series was conducted for each patient who had not previously received one since birth.

Prior to March 2021, the scanning area was primarily limited to the LBS region, with WS MRI reserved for patients exhibiting symptoms indicative of high-level spinal lesions. However, since March 2021, our protocol has been amended to include routine WS MRI for all patients, regardless of symptomatology. We used a 1.5 T MRI scanner (MAGNETOM Aera, Siemens Healthcare, Erlangen, Germany) with 20-channel phase-array head coils and 40-channel phase-array spine coils for the imaging survey. All participants underwent conventional brain MRI focusing on the posterior fossa using axial fast-spin echo T2 weighted images (T2WI) (TR/TE 3200/115). The whole spine was covered by two sagittal acquisitions. The image parameters are as follows: field of view (FOV) = 250–280 mm, matrix size = 320 × 256 pixels for sagittal spine-echo T1WI (TR/TE = 635 ms/9.9 ms), short-TI inversion recovery (STIR; TR/TE/TI = 3000 ms/72 ms/160°), and matrix size = 448 × 358 pixels for fast-spin echo T2WI (TR/TE = 3260 ms/92 ms). Axial MR images included fast spine-echo T1WI (TR/TE = 600–700 ms/11 ms) and fast-spin echo T2WI (TR/TE = 3000–6000 ms/79 ms) with FOV 180 mm and matrix sizes of 256 × 146 pixels and 320 × 182, respectively. The imaging protocol was designed to provide comprehensive visualization of the craniospinal axis, particularly focusing on the posterior fossa, spinal cord, and vertebral bodies.

Additional MRI series were performed if patients exhibited new neurological symptoms during postoperative follow-up in the outpatient department, or for monitoring of known spinal lesions, whether treated or untreated. All MRI scans were reviewed by a radiology specialist with expertise in pediatric neuroimaging. Anomalies identified at the brain, brainstem, cervical, or thoracic spinal levels were classified as high-level for the purposes of this article.

Exclusion criteria included patients who had undergone primary surgery at another institution or those who presented solely for a second opinion. Similarly, patients whose spinal MRI was performed externally were excluded from this study. All patients were subjected to routine follow-up until the establishment of normal defecation habits, with a minimum age threshold of seven months, and neurological symptoms were assessed at each visit. Notably, cases requiring referral—those presenting symptoms—were directed to rehabilitation, neurology, or neurosurgical specialists accordingly. Patients who underwent spinal surgery at other hospitals were included if their medical records were provided by their parents or retrieved via the national health insurance inter-hospital sharing system.

Continuous variables were expressed as means with standard deviations, and categorical variables as frequencies and percentages. The T-test was utilized for continuous variable comparisons, while the Fisher’s exact test was applied to categorical variables. A p-value of less than 0.05 was deemed statistically significant. The statistical tests were performed using MedCalc Statistical Software version 22.016 (MedCalc Software bv, Ostend, Belgium; https://www.medcalc.org; 2020).

Results

From January 2011 to January 2022, a total of 85 patients with ARM were diagnosed, treated, and followed up at our institution. Of these, 8 (9.41%) were classified as high-type ARM, 9 (10.59%) as intermediate-type, and the remaining 68 (80%) as low-type. Sixteen patients were excluded from the study due to spinal MRIs being performed at other hospitals; these were predominantly diagnosed with low-type ARM, except for one with high-type and none with intermediate-type. Consequently, 69 patients who had at least one series of spinal MRI performed at our institution were included in this study (Fig. 1). Among them, 19 (27.54%) underwent WS MRI as their initial survey, while the remaining 50 (72.46%) had LBS MRI. No statistically significant differences were observed between the groups in terms of age at anoplasty (8.37 ± 5.57 months vs. 4.53 ± 1.09 months, p = 0.337), age at MRI (9.78 ± 5.70 months vs. 8.36 ± 1.39 months, p = 0.439), or combined comorbidity (n = 20, 40.0% vs. n = 7, 36.8%, p = 1.000). Within the LBS group, 36 patients (72.0%) were diagnosed with low-type ARM, 9 (18.0%) with intermediate-type, and 5 (10.0%) with high-type. In the WS group, 17 patients (89.5%) were diagnosed with low-type ARM and 2 (10.5%) with high-type. No patients in the WS group had intermediate-type ARM. No significant differences were found in ARM type between the two groups (Table 1).

Fig. 1
Fig. 1
Full size image

Patients inclusion in this study: Out of the total 85 patients, 8 patients have a high type anorectal malformation (ARM), with 7 having undergone an magnetic resonance imaging (MRI) and 1 not having an MRI. Nine patients have an intermediate type ARM, all of which have undergone an MRI. Sixty-eight patients have a low type ARM, with 53 having undergone an MRI and 15 not having an MRI. Therefore 69 patients with MRI were included in the further study.

Table 1 Patients characteristics.
Full size table

Out of the initial 69 spinal MRI series, 30 (43.48%) yielded normal results, including 27 from the LBS group and 3 from the WS group. The LBS group exhibited significantly fewer anomalies compared to the WS group, irrespective of the anomaly level (n = 27, 54.0% vs. n = 3, 15.8%; p = 0.005). The most common anomalies included occult spinal dysraphism, cord lipoma, and tethered cord. Less frequent findings were vertebral anomalies, hydromyelia, benign enlargement of the subarachnoid spaces (BESS), and caudal regression syndrome.The prevalence of occult spinal dysraphism in WS group was insignificantly higher than the LBS group (n = 20, 40.0% vs. n = 13, 68.4%; p = 0.057) (Table 2). An inter-group comparison of spinal anomalies among low-type ARM, intermediate-type ARM, and high-type ARM did not reveal statistical significance (Table 3).

Table 2 Findings of MRI.
Full size table
Table 3 Findings of spinal anomaly by type of ARM.
Full size table

Within the LBS group, 4 patients (8%) underwent subsequent WS MRI during follow-up due to clinical neurological symptoms, mostly related to neurogenic bladder, lower limb weakness, and unstable gait. In three of these cases, additional anomalies were identified at the C–T spine level, including hemivertebrae, hydromyelia, and cavernous hemangioma. Initially, all were managed with a rehabilitation program, and only one required surgical intervention for hemangioma excision.

In the WS group, 4 out of 19 patients were found to have high-level lesions, which could have been overlooked with lumbosacral spinal MRI alone. Anomalies at the higher level included hydromyelia at the C-spine level, prominent CSF space in the bilateral frontoparietal region, syringomyelia extending from the T-spine level, and anterior displacement of C2 on C3. These patients underwent MRI as part of a perioperative survey and presented no significant motor dysfunction or muscle weakness. All were enrolled in a rehabilitation program for their imaging findings and defecation training, performing well without the need for spinal surgery during follow-up.

We present two cases with high-level spinal lesions. The first was a boy diagnosed with low-type ARM with an anocutaneous fistula, who underwent posterior sagittal anorectoplasty at three months. Preoperative L–S spine MRI revealed a lipoma in the spinal canal at the S1–S3 level, hemivertebrae at T10 and T11, and spina bifida at the lower sacrum. These findings resulted in a neurogenic bladder and left leg weakness with an unstable gait, requiring an extensive rehabilitation program. No spinal surgery was performed. After three years, due to claudication following long-distance walking, he underwent another MRI series. This whole spine MRI showed a tethered cord with lipoma extending from L4 to the sacral region, hemivertebrae at T10 and T11, and spina bifida consistent with the previous image. Additional findings included hydromyelia or syrinx extending from C5 to the conus medullaris and scoliosis of the T–L spine (Fig. 2). Despite these findings, no surgery was required, and he continued with the rehabilitation program. The second patient, diagnosed with low-type ARM with an anoperineal fistula, underwent surgery at two months of age. Her WS MRI at 11 months showed normal findings in the lumbosacral region but hydromyelia extending over the C2–C7 spinal cord (Fig. 3). She exhibited no neurological symptoms associated with the hydromyelia during outpatient follow-up, and no spinal surgery was performed.

Fig. 2
Fig. 2
Full size image

Magnetic resonance imaging (MRI) of a boy diagnosed with low-type anorectal malformation: (a,b) Regional spine MRI at the age of 3 months revealed a lipoma in the spinal canal at the S1–S3 level (arrow) and spina bifida at the lower sacrum. (c,d) Whole spine MRI 3 years later revealed additional findings of hydromyelia or syrinx extending from C5 to the conus medullaris and scoliosis of the T–L spine.

Fig. 3
Fig. 3
Full size image

Magnetic resonance imaging (MRI) of a girl having low-type anorectal malformation with an anoperineal fistula. Whole spine MRI at 11 months showed (a) normal findings in the lumbosacral region but (b,c) hydromyelia extending over the C2–C7 spinal cord (arrows).

Discussion

Since the introduction of VACTERL associations in the late twentieth century, spinal MRI has been recommended as a routine assessment for all patients diagnosed with ARM to determine the presence of spinal cord anomalies2. Although the most commonly reported associated anomalies in the literature occur within the LBS region, such as tethered cords and dysraphism5,7, this does not preclude the occurrence of higher-level lesions. Thoracic, cervical spine, and even brain anomalies have been repeatedly observed in some studies and case reports8,9,10. The precise prevalence of anomalies among the brain, cervical spine, thoracic spine, lumbar spine, or sacral spine has not been reported, nor has the recommended range of the spine to be scanned been thoroughly studied to date. Many institutions, including ours, have focused predominantly on the lumbosacral area and conducted spinal MRI only at this level. The prevalence of high-level anomalies in this patient group is unknown, and a lack of recognition of such conditions could result in delayed treatment.

In this article, we retrospectively reviewed all cases of ARM treated at our institution from January 2011 to January 2022, where MRI was conducted as part of the perioperative assessment. Of the 69 patients included in this study, 7 were identified with high-level lesions, including hydromyelia at the C-spine level, prominent CSF space in the bilateral frontoparietal region, syringomyelia extending from the T-spine level, anterior displacement of C2 on C3, hemivertebrae of the T-spine, and cavernous hemangioma at the T-spine level. All 7 patients presented with additional anomalies within the lumbosacral area, with none exhibiting exclusively high-level anomalies. It is noteworthy that 46 patients from the LBS group never underwent imaging that included their brain, brainstem, cervical, or thoracic spine, and it is uncertain if additional high-level anomalies would have been detected with whole spine MRI. High-level anomalies were infrequent, and no correlation was found with the type of anorectal malformation or other characteristics that could identify risk factors.

The prevalence of concurrent high-level spinal anomalies with ARM or within a VACTERL association remains undetermined, as there is a lack of literature specifically addressing this issue. Among our 19 patients who underwent whole spine MRI as an initial assessment, 6 were imaged before this became routine practice, prompted by clinical suspicion of a high-level anomaly based on neurological symptoms. Of the remaining 13 patients, who received whole spine MRI as standard practice regardless of symptoms, 3 (23.08%) were found to have high-level anomalies. This unexpectedly high prevalence, although potentially misleading due to the small sample size, highlights the possibility of overlooking significant anomalies when MRI is limited to the LBS region. In WS MRI, the entire range from the posterior cranial fossa to the sacrum can be examined with the highest image quality maintained throughout the spinal tract. In comparison with LBS MRI, it includes other anatomical regions without affecting the quality of the imaging series. Importantly, there is no statistically significant difference between WS MRI and LBS MRI groups on the incidence rates of tethered cord and occult spinal dysraphism.

Not all spinal anomalies necessitate intensive treatment; many are managed adequately with conservative measures such as rehabilitation programs. Some remain asymptomatic and may not significantly impact the patient if undetected. Nonetheless, Totonelli et al. reported a 12% surgical rate in patients with anorectal malformation and concurrent spinal anomalies, a figure too substantial to disregard11. Their study included spinal anomalies at all levels, without isolating those above the thoracic spine. In our cohort, one patient with a high-level anomaly underwent surgical intervention to address a cavernous hemangioma that was compressing the spinal cord and causing symptoms. For the remaining anomalies within the LBS area, occult spinal dysraphism (n = 33, 47.83%), cord lipoma (n = 13, 18.84%), and tethered cord (n = 9, 13.04%) were among the most commonly found, aligning with previous literature2,7,12,13,14. A female patient with anterior anus in the current series was referred for WS MRI at 1.5 years old at the time of the scan, and the MRI showed tethered cord with the conus medullaris at the L4 level, a syrinx extending from T8 to L3 and spinal dysraphism with an intradural lipoma. This was because the patient developed improved neurological signs by physiotherapists and therefore surgical intervention was prevented. In particular, the T8 lesion would not have been identified if the MRI of the lumbosacral area has been performed only. Though risks of sedation may also rise, it is better that all the abnormalities are identified in a single sedation session rather than having to do multiple procedures on the patient. To have base-line images for long term follow-up is also essential for those patients.

The increasing trend of abnormal findings detected by whole spine MRI may be attributed to retrospective bias and the heightened awareness of such anomalies among the team. Conducting MRI surveys in pediatric patients presents challenges, often necessitating sedation, which carries its own risks. At our institution, sedation for up to 50 min is required to complete an MRI series in young infants. Podberesky et al. described a similar protocol, with 30 min for imaging and nearly 60 min for total preparation15. Clinical doctors may be hesitant to order a second MRI series due to the incompleteness of the first. However, substituting LBS MRI with WS MRI as the initial survey does not significantly extend the time required, hence not prolonging sedation. In Taiwan, the national health insurance system covers all patients, making the cost of WS MRI comparable to LBS MRI. This cost efficiency may not translate to other regions, and clinicians should weigh the benefit of routine WS MRI against the risk of missing high-level anomalies.

Limitations

This retrospective study has inherent limitations, including uncontrolled baseline characteristics across different patient groups. While all patients underwent anoplasty and follow-up at our institution, some received treatment for other VACTERL associations elsewhere, potentially leading to incomplete medical records.

Conclusions

This study substantiates the value of comprehensive spinal imaging in the management of ARM, emphasizing the need for a multidisciplinary approach to diagnose and address the full spectrum of spinal anomalies in these patients. Future studies with larger cohorts and prospective designs are warranted to further validate these findings and refine imaging protocols for optimal patient care.