ORIGINAL ARTICLE
Acta Orthop Traumatol Turc 2016;50(1):49–56 doi: 10.3944/AOTT.2016.14.0381
MRI delineation of the morphometric characteristics of type I split cord malformations: a retrospective analysis of 29 cases
Shandong Provincial Hospital Affiliated to Shandong University, Division of Spinal Surgery, Department of Orthopedics, Shandong, China
Objective: The purpose of this study was to elucidate the morphometric characteristics by magnetic resonance imaging (MRI) of type I split cord malformation (SCM) patients. Methods: All subjects received conventional MRI with the Achieva 3.0T system (Philips Healthcare, Andover, MA, USA), including T1WI/axial and sagittal T2WI/axial FLAIR. Transverse diameter (TD) and sagittal diameter (SD) of the split cord, TD of the convex (TDconvex) and the concave (TDconcave), cranial SD (SDcranial), and caudal SD (SDcaudal) were recorded on the sagittal image combined with the two-dimensional view. Statistical comparison was performed within and between the groups. Results: Twenty-nine type I SCM patients were included, 24 (82.8%) of whom had scoliosis. Mean TD and SD of the split cord were 0.55±0.31 cm and 7.52±4.03 cm, respectively. No statistically significant difference was observed in TD, SD, and other parameters among the 3 groups. However, mean TD of the split cord in type I SCM patients with congenital scoliosis (0.49±0.29 cm) was significantly greater than in those without congenital scoliosis (0.18±0.44 cm) (p<0.05). TDconvex was significantly smaller than TDconcave in type I SCM patients with congenital scoliosis (p<0.05). Additionally, there was no statistically significant difference between SDcranial and SDcaudal in type I SCM patients with congenital scoliosis (p>0.05); however, SDcranial was significantly smaller than SDcaudal in Group 2. Conclusion: Our study provides the first MRI characterization of the morphometric features of type I SCM, and our findings will help orthopedic surgeons in better navigating the surgical field in correc- tive surgery of congenital scoliosis of type I SCM patients. Keywords: Congenital scoliosis; measurement; morphometry; MRI; osseous septum; split cord mal- formations. Level of Evidence: Level III, Prognostic study.
Split cord malformation (SCM), which represents 3.8– It is categorized as type I if the 2 hemicords are separated 5% of all spinal cord anomalies,[1,2] is considered to be an by an osseous and cartilaginous septum and contained uncommon abnormality in which a segment of the spinal in separate dural sheaths; it is categorized as type II if cord is divided into 2 parts by a fibrous or rigid bony spur. no osseous and cartilaginous septum are present and the
Correspondence: Jian Min Sun, MD. Shandong Provincial Hospital Affiliated to Shandong University, Department of Orthopedics, Division of Spinal Surgery, 324 Jing Wu Road, Jinan, Shandong 250021, China. Tel: +86 - 15763299776 e-mail: [email protected] Available online at www.aott.org.tr Submitted: October 14, 2014 Accepted: May 28, 2015 doi: 10.3944/AOTT.2016.14.0381 ©2016 Turkish Association of Orthopaedics and Traumatology QR (Quick Response) Code 50 Acta Orthop Traumatol Turc hemicords are contained in a single dural sheath. Type I nal Surgery, Pediatric Orthopedic Surgery, and Neuro- SCM may be accompanied with congenital scoliosis. surgery of Shandong Provincial Hospital affiliated with Currently, 2 surgical options are advocated for type Shandong University between March 2006 and Sep- I SCM associated with congenital scoliosis: corrective tember 2014 were retrospectively reviewed. Type I SCM surgery of congenital scoliosis after resection of the osse- was diagnosed according to the anatomical morphology ous septum[3–7] or regardless of the osseous septum.[8–13] criteria proposed by Pang.[1,14–16] Patients with type I Nevertheless, safe treatment of scoliosis and the osseous SCM associated with congenital scoliosis were catego- septum is an extremely challenging issue for spinal sur- rized by the position of the osseous septum and apical geons, as the extent of which the osseous septum affects vertebra into the apical vertebra group (Group 1), the the spinal cord in corrective surgery remains unknown. osseous septum above the apical vertebra group (Group To our knowledge, there is no literature on magnetic 2), and the osseous septum inferior to the apical vertebra resonance imaging (MRI) study of the morphometry of group (Group 3). SCM, even though such information would greatly ease The study protocol was approved by the local institu- the navigation by spinal surgeons of the surgical land- tional review board at the authors’ affiliated institutions. scape. In the current study, we sought to elucidate the Patient consent was not required because of the retro- MRI morphometric characteristics of the split cord and spective nature of this study. osseous septum of type I SCM patients. All subjects received conventional MRI with the Achieva 3.0T system (Philips Healthcare, Andover, Patients and methods MA, USA) using an 8-channel head coil with a mag- We retrospectively reviewed the surgical and radiologi- netic field intensity gradient of 40 mT/m and slew rate cal records of patients diagnosed with type I SCM who of 150 mT/m/ms. Conventional MRI included T1WI/ received surgical treatment at the Departments of Spi- axial and sagittal T2WI/axial FLAIR. The main param-
(a) (b) (c)
(d) (e) (f)
Fig. 1. The morphometric MRI parameter used in the current study. (a) Transverse diameter (TD) is defined as the maxi- mal distance from the right lateral border of the left hemicord to the left lateral border of the right hemicord. (b) Sagittal diameter (SD) is the straight line distance from the cephalad end to the caudal end of the split cord. (c) TD of the convex (TDconvex) and (d) the concave (TDconcave) is the distance from the lateral border of the osseous septum to the corresponding side of the split cord of the convex or concave of the spine, respectively. (e) SD from the upper border of the osseous septum to the cranial side of the split cord (SDcranial) and (f) SD from the inferior border of the osseous septum to the caudal side of the split cord (SDcaudal) are recorded on the sagittal image. [Color figures can be viewed in the online issue, which is available at www.aott.org.tr] Xia et al. Measurement, SCM, CS, MRI 51 eters were as follows: axial T1WI: TSE sequence: TR, Table 1. Demographic and baseline data of patients with type I 3056 ms; TE, 7.6 ms; TI, 860 ms; axial T2WI: TSE split cord malformation. sequence: TR, 2000 ms, TE, 200 ms, flip angle, 90°; axial Variables All patients (n=29) T2FLAIR: TR, 10002 ms, TE, 130.4 ms, TI 2400 ms, Age, years flip angle, 90°. For the above sequences, the slice thick- Mean±SD 14.4±16.01 ness was 6.0 mm, slice interval 1.0 mm, matrix 512x512, Median 13 2 FOV 240x240 mm , and mean number of excitation Range 2/12–58 (NEX) 1. Sagittal T2WI: TSE sequence: TR/TE, Female gender, n (%) 23 (79.3) 1897 s/80 ms, slice thickness 6.0 mm, slice interval 1.0 Mean height, (SD), (m) 1.21±0.29 mm, matrix 260x234, FOV 230x230x143 mm (APxR- Mean weight, (SD), kg 29.75±18.77 LxFH), and NEX 1. Mean body mass index (kg/m2) 18.67±5.12 Morphometric MRI parameters are shown in Fig- Lower extremity asymmetry, n (%) 19 (65.5) ure 1. Transverse diameter (TD) of the split cord was Rib deformity, n (%) 10 (34.5) defined as the maximal distance from the right lateral Foot deformity, n (%) 8 (27.6) border of the left hemicord to the left lateral border of Bone deformity, n (%) 24 (82.8) the right hemicord, while sagittal diameter (SD) was the Cervical scoliosis, n 1 straight line distance from the cephalad end to the cau- Thoracic scoliosis, n 19 dal end of the split cord. TD of the convex (TDconvex) Left 11 Right 8 and the concave (TDconcave) was the distance from the Lumbar scoliosis, n 3 lateral border of the osseous septum to the correspond- Osseous septum, n (%) ing side of the split cord of the convex or concave of the Middle to low thoracic spine 12 (41.4) spine, respectively. SD from the upper border of the os- Upper lumbar spine 17 (58.6) seous septum to the cranial side of the split cord (SD- Syringomyelia 9 (31) cranial) and SD from the inferior border of the osseous Tethered spinal cord 21 (72.4) septum to the caudal side of the split cord (SDcaudal) were recorded on the sagittal image combined with the SD: Standard deviation. two-dimensional view. For the osseous septum in the the most common symptom being asymmetric develop- vicinity of the apical vertebrae, SD included SDcranial, ment of the lower extremities (65.5%). Foot deformity SDcaudal, and SD of the osseous septum. was recorded in 8 patients (27.6%). Bony deformity was Data were expressed as mean±standard deviation recorded in 24 patients (82.8%) (Figure 3), 24 of whom and analyzed using SPSS software (version 17.0, SPSS had scoliosis (82.8%). Inc., Chicago, IL, USA). Statistical significance was set Mean TD of the split cord was 0.55±0.31 cm (range: for p values <0.05. Unpaired t-test was employed to de- 0.16–1.31 cm), and mean SD of the split cord was termine the difference of all parameters, including age, 7.52±4.03 cm (range: 0.16–1.31 cm). Mean TD of the height, weight, body mass index (BMI), TD, SD, gen- split cord in type I SCM patients with congenital sco- der, and left and right pedicles at the same vertebral level. liosis (0.49±0.29 cm) was significantly greater than that One-way analysis of variance with post hoc test was em- of patients without congenital scoliosis (0.18±0.44 cm) ployed to compare all split cord parameters among the (p<0.05). However, SD of type I SCM patients with 3 groups; least significant difference t-test was used to congenital scoliosis (7.35±4.43) showed no statistical analyze the differences among the 3 groups, and the dif- difference from that of those without congenital scoliosis ferences of intergroup parameters were determined by (6.41±4.86) (p>0.05) (Table 2). paired samples t-test Table 2. Morphometric characteristics of patients with type I split cord malformation. Results Transverse Sagittal The study flowchart is shown in Figure 2, and the de- diameter diameter mographic and baseline characteristics of 29 type I SCM patients eligible for the study are shown in Table Congenital scoliosis 1. Median age was 14.4 years (range: 2 months to 58 Yes (n=23) 0.49±0.29 7.35±4.43 years), and mean BMI was 18.67±5.12 kg/m2. Most of No (n=6) 0.18±0.44 6.41±4.86 the patients (23/29, 79.3%) were female. The majority T 4.847 0.453 of the patients (26/29, 89.7%) were symptomatic, with P <0.05 (0.000) >0.05 (0.654) 52 Acta Orthop Traumatol Turc
Relationship of OS and Incomolete data (n=10) SCM (n=46) (excluded) the apical verterbrea of CS
Group 1=OS located on the apical verterbrea
Yes Yes SCM with Type I SCM Study group complete With or without n=29 With or without CS Type I SCM with CS Group 2=OS located data, n=36 osseous septum Included n=23 above the apical verterbrea (Included) Group 3=OS located under the apical verterbrea No No
Type II SCM Study group n=7 Type I SCM with out CS excluded n=6
Fig. 2. The study flowchart. [Color figure can be viewed in the online issue, which is available at www.aott.org.tr]
(a) (b) (c) (d)
Fig. 3. Bony deformities and spinal cord anomalies accompany split cord malformation (SCM). (a) Spinal bifida occulta.(b) Block vertebrae and rib deformity. (c) Syringomyelia. (d) Tethered cord. [Color figure can be viewed in the online issue, which is available at www.aott.org.tr]
The osseous septum was located in the upper lumbar scoliosis (p<0.05) (Table 4). Similar findings were found spine in 17 (58.6%) patients and middle-lower thoracic for patients in Groups 1 and 2; however, no statistically spine in 12 (41.4%) patients. The septum was adhered significant difference was seen between TDconvex and to the posterior junction part of the lamina, and spinal TDconcave in Group 3 (p>0.05). In addition, there was process deformity of the vertebrae became wider and transversely branched in 14 (48.3%) patients. TD of (a) (b) the osseous septum is illustrated in Figure 4. There were 12 (52.2%) patients in Group 1 whose osseous septum was located on the apical vertebrae, 6 (26.1%) patients in Group 2 whose osseous septum was above the api- cal vertebrae, and 5 (21.7%) patients in Group 3 whose osseous septum was under the apical vertebrae (Figure 5). No statistically significant difference was observed in TD, SD, and other parameters among the 3 groups Fig. 4. Transverse diameter of the osseous septum. (a) Minimal (Table 3). However, TDconvex was significantly smaller value. (b) Maximal value. [Color figure can be viewed in the than TDconcave in type I SCM patients with congenital online issue, which is available at www.aott.org.tr] Xia et al. Measurement, SCM, CS, MRI 53
(a) (b) (c)
Fig. 5. Relationship between the osseous septum and apical vertebrae of scoliosis. The osseous septum (a) directly facing, (b) above, and (c) under the apical vertebra. no statistically significant difference between SDcranial TD of the split cord in type I SCM patients with congen- and SDcaudal in type I SCM patients with congenital ital scoliosis was significantly greater than that of patients scoliosis (p>0.05); nor was statistical difference found without congenital scoliosis. In addition, TDconvex was in Groups 1 and 3 (Table 4), though SDcranial was sig- significantly smaller than TDconcave in type I SCM nificantly smaller than SDcaudal in Group 2 (p<0.05). patients with congenital scoliosis. The morphometric data of type I SCM patients from the current study will Discussion greatly facilitate the navigation by spinal surgeons of the surgical landscape in patients with type I SCM, which is Split cord malformations are often accompanied with spi- complicated by a variety of bone deformities. nal deformity and other spinal cord abnormalities,[14,16–22] and surgical correction remains the therapeutic choice in The osseous septum may impact the morphometric SCM patients. However, no literature is currently avail- parameters of the split cord. In our cohort, the osseous able on the morphometric characteristics of SCM pa- septum was located on the apical vertebrae in 52.2% of tients. Our study is the first attempt by MRI at delineat- patients, above the apical vertebrae in 26.1% of patients, ing the morphometric features of type I SCM. We found and under the apical vertebrae in 21.7% of patients. that the majority (79.3%) of type I SCM patients had When the osseous septum is located on the apical ver- congenital scoliosis and further demonstrated that mean tebrae of type I SCM patients with congenital scoliosis,
Table 3. Morphometric characteristics of patients with type I split cord malformation according to location of the osseous septum.
Group All patients Group 1 Group 2 Group 3 (n=23) (n=12) (n=6) (n=5)
TD Mean (SD) 0.49 (0.29) 3.74 (1.77) 5.16 (2.35) 4.72 (3.66) SD Mean (SD) 7.35 (4.43) 124.22 (39.78) 118.12 (17.56) 120.59 (30.88) TDconvex, mm Mean (SD) 0.91 (0.68) 0.91 (0.51) 0.66 (0.53) 1.38 (0.99) TDconcave, mm Mean (SD) 1.95 (1.03) 1.48 (0.47) 2.16 (0.94) 2.00 (1.62) SDcranial Mean (SD) 48.80 (16.88) 45.40 (18.00) 45.91 (12.47) 59.79 (23.26) SDcaudal Mean (SD) 54.87 (19.50) 61.17 (25.41) 56.00 (17.64) 44.58 (15.25)
TD: Transverse diameter; SD: Sagittal diameter. The osseous septum is located on the apical vertebra in Group 1, above the apical vertebra in Group 2, and inferior to the apical vertebra in Group 3. For all parameters, p>0.05, Group 1 vs. Group 2; p>0.05, Group 1 vs. Group 3; p>0.05, Group 1 vs. Group 3. 54 Acta Orthop Traumatol Turc
Table 4. Morphometric characteristics of patients with type I split cord malformation with congenital scoliosis.
Group All patients Group 1 Group 2 Group 3 (n=23) (n=12) (n=6) (n=5)
TDconvex, mm Mean (SD) 0.91 (0.68) 0.68 (0.53) 0.99 (0.51) 1.38 (0.99) TDconcave, mm Mean (SD) 1.95 (1.03) 2.16 (0.94) 1.48 (0.47) 2.00 (1.62) t 6.658 7.865 5.517 1.827 p <0.05 (0.000) <0.05 (0.000) <0.05 (0.003) >0.05 (0.142) SDcranial Mean (SD) 48.80 (16.88) 45.91 (12.47) 45.40 (18.00) 59.79 (23.26) SDcaudal Mean (SD) 54.87 (19.50) 56.00 (17.64) 61.17 (25.41) 44.58 (15.25) t 1.167 1.330 2.604 1.527 p >0.05 (0.256) >0.05 (0.210) <0.05 (0.048) >0.05 (0.201) the semicord of the convex side is restricted by the bony or radiological studies for further delineation. The cur- spur, while the semicord of the concave side maintains rent study found no statistical difference in SD of type distance from the bony spur because of the principle of I SCM patients with and without congenital scoliosis, minimal travel path. Consistently, we found that TD- and location of the osseous septum also exerted no ap- convex was not significantly different from TDconcave parent impact on SD. The findings indicate that the os- in type I SCM patients whose osseous septum was lo- seous septum has no significant influence on SD of the cated on the apical vertebrae. On the other hand, in type split cord, regardless of congenital scoliosis. These above I SCM patients with congenital scoliosis, TDconvex findings together strongly indicate that the presence of was significantly smaller than TDconcave. These find- the osseous septum has more meaningful impact on the ings suggest that the presence of congenital scoliosis may width than the length of the split cord. have greater influence on TD of the split cord than the location of the osseous septum. However, this phenome- Previous studies[23–27] showed that the caudal end of non appears to be limited to patients whose osseous sep- the split cord and the osseous septum were at the same tum is located on or above the apical vertebra, indicating level, and the cephalad end of the split cord was much a complex interplay of congenital scoliosis and location higher than the end of the osseous septum. These find- of the osseous septum, which requires additional clinical ings suggest that growth and development of the spine
(a) (b)
Fig. 6. (a, b) The split cord may not be located in the vicinity of the apical vertebra of patients with congenital scoliosis. Xia et al. Measurement, SCM, CS, MRI 55 was faster than growth and development of the spinal 5. Kumar R, Bansal KK, Chhabra DK. Split cord malfor- cord, which could be restrained and immobilized by the mation (scm) in paediatric patients: outcome of 19 cases. presence of the osseous septum. However, as shown in Neurol India 2001;49:128–33. our study, the osseous septum is not always situated in 6. Sinha S, Agarwal D, Mahapatra AK. Split cord malfor- the caudal end of the split cord (Figure 6). Furthermore, mations: an experience of 203 cases. Childs Nerv Syst we found no statistically significant difference between 2006;22:3–7. CrossRef SDcranial and SDcaudal. However, if the osseous sep- 7. Ayvaz M, Akalan N, Yazici M, Alanay A, Acaroglu RE. Is tum was located above the apical vertebrae, SDcaudal it necessary to operate all split cord malformations before was significantly greater than SDcranial. We speculated corrective surgery for patients with congenital spinal defor- that the osseous septum may not be present at birth but mities? Spine (Phila Pa 1976) 2009;34:2413–8. CrossRef may later on be present on the bony spur sprouted from 8. Akay KM, Izci Y, Baysefer A, Timurkaynak E. Split cord the junction of bilateral lamina. Gupta et al.[28] reported malformation in adults. Neurosurg Rev 2004;27:99–105. a 3-year-old female patient who developed recurrent fi- 9. Zuccaro G. Split spinal cord malformation. Childs Nerv bro-osseous spur after primary excision. Both her spine Syst 2003;19:104–5. and spinal cord presented the bowstring shape, and the 10. Tubbs RS, Salter EG, Oakes WJ. Split spinal cord malfor- spinal cord was shifted to the opposite directions of sco- mation. Clin Anat 2007;20:15–8. CrossRef liosis. Consistently, we found significant difference be- 11. Peart O. Split cord malformation. Radiol Technol tween TDconvex and TDconcave. 2004;76:168. 12. Yu B, Wang YP, Qiu GX, Zhang JG, Li JY, Shen JX, et al. Our study is limited by the retrospective nature Corrective surgery of congenital scoliosis with type II split of the study, and the data was obtained from a small spinal cord malformation. Chin Med Sci J 2006;21:48– number of cases. In addition, the outcomes of some pa- 52. rameters were relative values. In conclusion, our study 13. Qureshi MA, Asad A, Pasha IF, Malik AS, Arlet V. Staged provides the first MRI characterization of the morpho- corrective surgery for complex congenital scoliosis and split metric features of type I SCM, and our findings would cord malformation. Eur Spine J 2009;18:1249–54. CrossRef help orthopedic surgeons in better navigating the surgi- 14. Lao L, Zhong G, Li X, Liu Z. Split spinal cord malforma- cal field in corrective surgery of congenital scoliosis of tion: report of 5 cases in a single Chinese center and review type I SCM patients. of the literature. Pediatr Neurosurg 2013;49:69–74. CrossRef 15. Tsai TJ, Michaud LJ. Split cord malformation. Am J Phys Acknowledgements Med Rehabil 2013;92:839. CrossRef The authors would like to thank the staff at the Imaging 16. Erşahin Y. Split cord malformation types I and II: a person- Research Institute of Shandong Province and the Imag- al series of 131 patients. Childs Nerv Syst 2013;29:1515– ing Department of Shandong Provincial Hospital Affili- 26. CrossRef ated to Shandong University for their technical support. 17. Jindal A, Mahapatra AK. Split cord malformations--a clin- ical study of 48 cases. Indian Pediatr 2000;37:603–7. Conflics of Interest: No conflicts declared. 18. Tortori-Donati P, Rossi A, Biancheri R, Cama A. Mag- netic resonance imaging of spinal dysraphism. Top Magn References Reson Imaging 2001;12:375–409. CrossRef 1. Pang D. Split cord malformation: Part II: Clinical syn- 19. Schijman E. Split spinal cord malformations: report of drome. Neurosurgery 1992;31:481–500. CrossRef 22 cases and review of the literature. Childs Nerv Syst 2. Zani DD, De Zani D, Morandi N, Biggi M, Belloli AG, 2003;19(2):96–103. Riccaboni P, et al. Imaging diagnosis-split cord malforma- 20. Kumar R, Singh SN, Bansal KK, Singh V. Comparative tion. Vet Radiol Ultrasound 2010;51:57–60. CrossRef study of complex spina bifida and split cord malformation. 3. Alzhrani GA, Al-Jehani HM, Melançon D. Multi-level Indian J Pediatr 2005;72:109–15. CrossRef Split Cord Malformation: Do We Need a New Classifica- 21. Mahapatra AK, Gupta DK. Split cord malformations: tion? J Clin Imaging Sci 2014;4:32. CrossRef a clinical study of 254 patients and a proposal for a new 4. Liljenqvist UR. Expert’s comment concerning Grand clinical-imaging classification. J Neurosurg 2005;103(6 Rounds case entitled “vertebral column resection for com- Suppl):531–6. CrossRef plex congenital kyphoscoliosis and type I split cord mal- 22. Ozturk E, Sonmez G, Mutlu H, Sildiroglu HO, Velioglu formation” (Hua Hui, Zhen-Xing Zhang, Tuan-Min Yang, M, Basekim CC, et al. Split-cord malformation and accom- Bao-Rong He, Ding-Jun Hao). Eur Spine J 2014;23:1164– panying anomalies. J Neuroradiol 2008;35:150–6. CrossRef 5. CrossRef 23. Proctor MR, Scott RM. Long-term outcome for patients 56 Acta Orthop Traumatol Turc
with split cord malformation. Neurosurg Focus 2001;10:5. 300 cases at AIIMS 1990-2006. J Pediatr Neurosci 24. Shokouhi G, Tubbs RS, Shoja MM. Type I split cord mal- 2011;6(Suppl 1):41–5. CrossRef formation with an usual bony morphology. Folia Morphol 27. Borkar SA, Mahapatra AK. Split cord malformations: (Warsz) 2007;66:78–9. A two years experience at AIIMS. Asian J Neurosurg 25. Liu W, Zheng D, Cui S, Zhang C, Liu Y, Jia Y, et al. Char- 2012;7:56–60. CrossRef acteristics of osseous septum of split cord malformation in 28. Gupta DK, Mahapatra AK. Proposal for a new clinico- patients presenting with scoliosis: a retrospective study of radiological classification of type I split-cord malforma- 48 cases. Pediatr Neurosurg 2009;45:350–3. CrossRef tions: a prospective study of 25 cases. Pediatr Neurosurg 26. Mahapatra AK. Split cord malformation - A study of 2006;42:341–6. CrossRef