You are on page 1of 5

Musculoskeletal Imaging Original Research

Park et al. MRI of the Lumbar Spine Musculoskeletal Imaging Original Research

Downloaded from www.ajronline.org by 202.152.202.150 on 11/10/13 from IP address 202.152.202.150. Copyright ARRS. For personal use only; all rights reserved

Incidental Findings of the Lumbar Spine at MRI During Herniated Intervertebral Disk Disease Evaluation
Hee-Jin Park1,2,3 Yong-Hwan Jeon2 Myung-Ho Rho1 Eun-Ja Lee 3 Noh-Hyuck Park 3 Sung-Il Park 4 Joon-Hee Jo 3
Park HJ, Jeon YH, Rho MH, et al.

OBJECTIVE. The objective of our study was to evaluate the frequency and types of incidental ndings of the lumbar spine during MR evaluation for herniated intervertebral disk disease. MATERIALS AND METHODS. A total of 1268 patients (male-to-female ratio, 421:847; age range, 197 years) with clinically suspected herniated intervertebral disk disease underwent MRI of the lumbar spine. Musculoskeletal radiologists evaluated the MR examinations for the presence of incidental ndings. We dened incidental nding as any abnormal nding not related to the chief complaint. Vertebral hemangioma, Tarlov cyst, brolipoma, synovial cyst, and sacral meningocele were included. Frequency distributions of the assessed imaging characteristics were calculated. For analysis of the relationship of incidental ndings with patient characteristics, the chi-square test was used. RESULTS. Overall, 107 patients (8.4%) had incidental ndings. Fibrolipoma was most common (41 cases, 3.2%), followed by Tarlov cyst (27 cases, 2.1%) and vertebral hemangioma (19 cases, 1.5%). Fibrolipoma and sacral meningocele were more common in males ( p < 0.05). There was no difference in the incidence between the sexes in the other incidental ndings ( p = 0.260.96). Four of the ve incidental ndings were signicantly more frequent in individuals younger than 50 years ( p < 0.05), whereas the incidence of vertebral hemangioma did not differ by patient age ( p = 0.32). CONCLUSION. Incidental ndings at MRI of the lumbar spine were common and associated with age and sex. Most were benign ndings. An awareness of the prevalence of the incidental ndings detected at MRI of the lumbar spine is helpful for diagnosing lesions not related to symptoms. n incidental lesion is an asymptomatic lesion found while examining a patient for an unrelated reason. PACS was introduced in many hospitals to improve reporting efciency, and its introduction has resulted in an increase in the number of reported incidental ndings and follow-up examinations [1]. Various incidental ndings can be seen on MRI of the lumbar spine during an evaluation for suspected herniated intervertebral disk disease. We focused on spinal abnormalities and excluded incidental ndings associated with other organs. The incidental ndings are vertebral hemangioma, Tarlov cyst, brolipoma, synovial cyst, and sacral meningocele. Although there have been reports of incidental ndings on spinal MRI, few studies in the literature address intraspinal incidental ndings at lumbar MRI. In this study, we assessed incidental ndings discovered in patients with suspected herni-

Keywords: cyst, herniated intervertebral disk disease, incidental ndings, MRI, perineural, spine DOI:10.2214/AJR.10.5457 Received July 31, 2010; accepted after revision November 14, 2010. Department of Radiology, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 108 Pyung-dong, Jongno-gu, Seoul 110-102, South Korea. Address correspondence to M. H. Rho (parkhiji@kangwon.ac.kr).
2 Department of Radiology, Kangwon National University School of Medicine, Gangwon-do, South Korea. 3 Department of Radiology, Myoungji Hospital, Kwandong University College of Medicine, Seoul, South Korea. 4 Department of Radiology, Shinchon Severance Hospital, Yonsei University School of Medicine, Seoul, South Korea. 1

ated intervertebral disk disease undergoing MRI of the lumbar spine and the clinical importance of those ndings. Materials and Methods Case Selection
Our study population included 421 males (33%) and 847 females (67%) with clinically suspected herniated intervertebral disk disease who underwent imaging between January 2007 and October 2009. The age distributions were as follows: 29 years old or younger, 70 patients (5.5%); 3049 years old, 328 patients (26%); 50 69 years old, 597 patients (47%); and 70 years old or older, 273 patients (22%). Most of the patients complained of chronic back pain or radiating pain with the duration of symptoms ranging from several months to several years. Patients who had a history of acute trauma were excluded from the study. This retrospective study was approved by the institutional ethics review board. There was no requirement for informed patient consent.

AJR 2011; 196:11511155 0361803X/11/19651151 American Roentgen Ray Society

AJR:196, May 2011 1151

Park et al.

Downloaded from www.ajronline.org by 202.152.202.150 on 11/10/13 from IP address 202.152.202.150. Copyright ARRS. For personal use only; all rights reserved

Fig. 127-year-old man with vertebral hemangioma. Sagittal T2-weighted image (TR/TE, 3500/120) shows vertebral hemangioma (arrow ) in second lumbar vertebral body.

A
Fig. 2 66-year-old woman with Tarlov cyst in sacrum. A and B, Sagittal ( A ) and coronal (B) T2-weighted images (TR/TE, 3500/120) reveal Tarlov cyst (arrow ) in sacrum and show dilated neural foramen in S12.

Image Analysis
MR examinations were interpreted in con sensus by two fellowship-trained academic musculo skeletal radiologists who had 12 and 10 years of experience, respectively, at the time of the study. The radiologists evaluated the MR examinations for the presence of variable incidental ndings in the spine. We dened incidental nding to include any abnormal nding not related to the chief complaint. The following pathologic conditions and anatomic variants were recorded: vertebral hemangioma, Tarlov cyst, bro lipoma, synovial cyst, and sacral meningocele. On MR images, spines were examined for marrow signal, cystic changes of the thecal sac, and fatty inltration of the lum terminale. A vertebral hemangioma was diagnosed when signal intensity was increased on T1-weighted images, signal intensity was markedly increased on T2-weighted images, and the vertebral bodies had a mottled appearance [2] (Fig. 1). Tarlov cyst was dened as dilated or ballooned areas of the sheath that cover nerve roots exiting from the sacral area of the spine (Fig. 2). Fibrolipoma of the lum terminale was diagnosed when linear fat signals were seen within the lum terminale on T1- and T2-weighted images without severe thickening or a tethered cord (Fig. 3). A synovial cyst was diagnosed when a cystic lesion abutting the adjacent facet joint showed low-to-intermediate signal intensity on T1-weighted images and had a hypointense cystic capsule, which usually is well demarcated from high-signal-intensity intrathecal uid on T2-weighted images [3] (Fig. 4). Sacral meningocele was dened as abnormal dilatation

of the meninges lying within the connes of the sacral spinal canal (Fig. 5). We collected demographic data and classied patients with regard to sex (male-to-female ratio) and age ( 29, 3049, 5069, 70 years; and younger [< 50 years] vs older [ 50 years]).

MRI Parameters
All MR examinations were performed on a 1.5T magnet (Intera, Philips Healthcare) using the same protocol with a synspine coil and fast spinecho imaging. T1- and T2-weighted images were obtained in the axial plane and T1-weighted, T2weighted, and T2 STIR images were obtained in the sagittal plane while the patient was in a supine position. For sagittal imaging, a eld of view of 32 cm, matrix of 512 256, and slice thickness of 4 mm were used; for axial imaging, a eld of view of 15 cm, matrix of 256 320, and slice thickness of 4 mm were used. The following MR sequences were performed in the sagittal plane: T1-weighted spin-echo (TR range/TE range, 500600/1217), T2 STIR (TR/TE, 2500/60), and T2-weighted (3500/120). In the axial plane, T2-weighted turbo spin-echo (TR range/TE, 30004000/60) and T1weighted turbo spin-echo (TR range/TE range, 600700/1015) sequences were performed.

Fig. 3 44-year-old man with brolipoma of lum terminale. Sagittal T1-weighted image (TR/TE, 480/12) shows brolipoma of lum terminale. Note linear fat signal within lum terminale (arrow ).

Statistical Analysis
Frequency distributions of the assessed imaging characteristics were calculated. For analysis of the relationship of the incidental ndings with patient characteristics, the chi-square test was used. The association of the incidence of each incidental

nding was evaluated with regard to patient sex and age (< 50 years, 50 years). A p value < 0.05 was considered to indicate statistical signicance.

Results A total of 1268 patients underwent MRI of the lumbar spine as part of an evaluation for clinically suspected herniated intervertebral

1152

AJR:196, May 2011

MRI of the Lumbar Spine

Downloaded from www.ajronline.org by 202.152.202.150 on 11/10/13 from IP address 202.152.202.150. Copyright ARRS. For personal use only; all rights reserved

A
Fig. 422-year-old man with synovial cyst. A and B, Axial ( A ) and sagittal (B) T2-weighted images (TR/TE, 4400/120 and 3500/120, respectively) show synovial cyst (arrow ) adjacent to right facet joint.

Fig. 5 42-year-old man with sacral meningocele. Axial T2-weighted image (TR/TE, 4400/120) reveals sacral meningocele (arrow ) as intradural uid-lled cystic lesion in central portion of spinal canal with signal intensity identical to CSF.

TABLE 1: Incidence of Incidental Findings of the Lumbar Spine at MRI Categorized by Sex
Sex No. (%) of patients Men Women p Total, no. (%) of patients 6 (1.4) 13 (1.5) 0.86 19 (1.5) 11 (2.6) 16 (1.9) 0.96 27 (2.1) 21 (5.0) 20 (2.4) 0.01 41 (3.2) 5 (1.2) 5 (0.6) 0.26 10 (0.8) 8 (1.9) 2 (0.2) 0.0018 10 (0.8) 1268 421 847 Vertebral Hemangioma Tarlov Cyst Fibrolipoma Synovial Cyst Sacral Meningocele Total

disk disease. Overall, 107 patients (8.4%) had incidental ndings. Fibrolipoma was most common (41 cases, 3.2%), followed by Tarlov cyst (27 cases, 2.1%) and vertebral hemangioma (19 cases, 1.5%) (Table 1). The incidence of synovial cyst and sacral meningocele was 0.8% and 0.8%, respectively. Of these 107 patients, 103 patients (96.3%) had one nding, three patients (2.8%) had two ndings, and one patient (0.9%) had three ndings. Detected brolipomas extended over multiple spinal segments and were more conspicuous on axial images than on sagittal images. The cornus medullaris terminated at the normal level (L12) in all patients. The locations of the visible Tarlov cysts were not related to the symptoms of the patients. The incidences of the incidental ndings of the lumbar spine are shown by patient age in Table 2. Fibrolipoma and sacral meningocele were observed more commonly in males ( p < 0.05). There were no differences between the sexes in the incidences of the other incidental ndings ( p = 0.260.96). Most incidental ndings except vertebral hemangioma ( p = 0.32) were signicantly more frequent in the younger (< 50 years) age group (p < 0.05). Other rare incidental ndings were simple bone cyst (one patient) and sacral dysra-

phism (two patients). Table 2 identies patient age characteristics associated with the likelihood of the incidental ndings. Discussion The discovery of incidental lesions is and always has been a part of practicing medicine. An incidental lesion is an asymptomatic lesion found while examining a patient for an unrelated reason. The impact of nding incidental lesions on patient health outcome is not certain [4], but it is worth remembering that an incidental nding may be more signicant than the suspected disease that prompted imaging [5]. The common incidental ndings in the lumbar spine are vertebral hemangioma, perineural cyst, brolipoma, synovial cyst, and sacral meningocele. Most are benign lesions but sometimes hidden malignancy can be found. Vertebral hemangiomas are benign vascular tumors that have been shown in 11% of spines at autopsy; despite their common occurrence and usual benign course, they occasionally produce spinal cord compression [2]. Less than 1% of vertebral hemangiomas produce symptoms owing to collapse (pathologic fracture) or cord compression [6]. Most

are benign, but rarely aggressive vertebral he mangiomas are noted. In this case, intense localized spinal pain, myelopathy, or radiculopathy from osseous expansion may be seen. Histopathologically, hemangiomas consist of thin-walled, blood-lled vessels and sinuses lined by endothelium and interspersed among the longitudinally oriented trabeculae of bones. The dilated vascular channels are set in a substratum of fatty marrow [7]. They continue to show a distinctive appearance of increased signal, at least in the osseous portions of the tumor on both T1- and T2-weighted images. Increased signal intensity on T1-weighted images, marked degree of increased signal on T2-weighted images, and mottled appearance of the vertebral bodies appear to make a constellation of ndings specic for vertebral hemangioma [2]. The incidence of vertebral hemangioma in our study group was 1.5%, and there were no differences between the sexes and age groups in terms of incidence. Barzin and Maleki [8] reported that the incidence of vertebral he mangioma ranged from 10% to 27% on autopsy and that vertebral hemangioma was more common in older age groups and had a slight female preponderance (3:2). It is not clear

AJR:196, May 2011 1153

Park et al. TABLE 2: Incidence of Incidental Findings of the Lumbar Spine at MRI Categorized by Age
Age (y) No. (%) of patients Downloaded from www.ajronline.org by 202.152.202.150 on 11/10/13 from IP address 202.152.202.150. Copyright ARRS. For personal use only; all rights reserved 29 3049 5069 70 < 50 50 p Total no. (%) of patients 0 (0) 4 (1.2) 8 (1.3) 7 (2.6) 4 (1.0) 15 (1.7) 0.32 19 (1.5) 1 (1.4) 15 (4.6) 9 (1.5) 2 (0.7) 16 (4.0) 11 (1.3) 0.0026 27 (2.1) 8 (11.4) 14 (4.3) 13 (2.2) 6 (2.2) 22 (5.5) 19 (2.2) 0.0012 41 (3.2) 2 (2.9) 6 (1.8) 1 (0.2) 1 (0.4) 8 (2.0) 2 (0.2) 0.0004 10 (0.8) 0 (0) 9 (2.7) 1 (0.2) 0 (0) 9 (2.3) 1 (0.1) < 0.0001 10 (0.8) 1268 70 328 597 273 398 870 Vertebral Hemangioma Tarlov Cyst Fibrolipoma Synovial Cyst Sacral Meningocele Total

why our results do not match theirs. The differences may be due to the fact that only the lumbar spine was included in our MR study; because the thoracic and cervical spines were excluded from our study, a smaller incidence may have resulted. Tarlov cysts, also known as perineural cysts, are CSF-lled sacs located in the spinal canal of S1S4 region of vertebrae and can be distinguished from other meningeal cysts by nerve berlled walls. Tarlov cysts are dened as cysts formed within the nerve root sheath at the dorsal root ganglion. The space or cysts created by the dilated sheaths are directly connected to the subarachnoid area of the spinal column, the area through which CSF ows. Most people with these cysts have no symptoms. However, when conditions cause these perineural cysts to ll with CSF and expand in size, they can begin to compress important neighboring nerve bers, resulting in a variety of symptoms including pain, weakness, and abnormal sensation [9]. In our study, the detected Tarlov cysts had no relation with the symptoms of the patients. The incidence of the lesion in our study was 2.1%, and Tarlov cysts were predominantly in the younger population ( p < 0.05) but incidence did not differ between the sexes. In previous studies, investigators reported a 15% incidence and slight female predominance [9, 10]. Fibrolipoma of the lum terminale is also called fatty lum terminale and usually is not combined with the tethered cord and cornus position. On MRI, linear fat signal within the lum terminale is seen on T1-weighted images and the size of the lum appears normal. If the lum is larger than 2 mm, then intraspinal lipoma can be suspected. Postmortem studies have reported a 46% incidence of occult brolipomas of the lum terminale in what were thought to be otherwise normal spinal

cords [11]. Ross [12] described equal sex distribution and no predominant age group. Our cases showed an incidence of 3.2% and were more predominant in males ( p = 0.01) and in the younger population ( p = 0.0012). A synovial cyst is a common juxtaarticular lesion that occurs most frequently in the extremities but also can occur in the lumbar spine. Synovial cysts usually show a communication with the adjacent facet joint and have a lining of synovial cells at histologic analysis [13]. Synovial cysts are thought to be the result of degenerative changes of the ligamentum avum that occur in conjunction with hypermobility associated with degenerative disk disease, degenerative spondylolisthesis, and subluxation of the facet joints [14]. Lumbar facet synovial cysts have been reported as a cause of low back pain, radiculopathy, and neurogenic claudication. These cysts can hemorrhage and cause acute symptoms [15]. Synovial cysts appear as lesions with lowto-intermediate signal intensity on T1-weighted MR images. On T2-weighted MR images, the cyst capsule appears as a hypointense line that usually is well demarcated from the high-signal-intensity intrathecal CSF. The signal intensity may be heterogeneous owing to the presence of hemorrhage, calcication, and vacuum phenomenon. Enhancement of the synovial cyst wall and, occasionally, enhancement of the adjacent facet joint have been shown after IV administration of gadolinium-based contrast material. The relationship of the cyst with the adjacent facet joint is best depicted on transverse images [13]. Liu et al. [16] reported that synovial cysts showed a female predominance and a mean age of 58 years, but somewhat different results were seen in our study. The incidence of synovial cysts in our study was 0.8% with similar sex distribution but showed prepon-

derance in the younger (< 50 years) population ( p = 0.0004). Sacral meningoceles are usually a form of dysraphism and are thought to arise during development. Sacral meningocele is dened as a protrusion of the spinal meninges through a defect in the vertebral column or foramina and usually is associated with a congenitally dysraphic vertebra, with the spinal cord remaining entirely conned to the vertebral canal. These lesions are frequently identied in the posterior location over the thoracic and sacral areas at birth and constitute about 10% of all patients with spina bida [17]. Usually the meningocele cavity communicates with the subarachnoid cul-de-sac by a narrow channel. The meningocele cavity may show evidence of loculation by septa, or separate isolated cavities can exist. Sacral nerves are found to be incorporated in the wall of the cavity or lie just outside it, compressed against the bony wall [18]. Sacral meningocele should be differentiated from sacral perineural cyst; the latter, as its name suggests, completely surrounds an individual nerve [19]. The primary MRI nding of sacral meningocele is an intradural, extramedullary spaceoccupying lesion with T1- and T2-weighted signal intensities identical to CSF. Heterogeneous signal intensity on T2-weighted MRI, depending on the ow effect in the cyst uid, may be the only evidence to indicate the presence of an abnormal uid collection [20]. Dahlgren et al. [17] suggested that more than three quarters of cases are found in female patients suggest that the result may reect the fact that women of reproductive age are more likely to undergo pelvic examination than men. Lee et al. [21] reported that men and women seemed to be equally affected, usually in the third through fth decades of

1154

AJR:196, May 2011

MRI of the Lumbar Spine life. However, we found more cases in male patients ( p = 0.0018) and in patients in the third and fourth decades (90%). Our study of incidental ndings at lumbar MRI of patients with suspected herniated intervertebral disk disease showed incidental ndings in approximately 8% and revealed differences between the sexes and age groups with regard to the likelihood of most of the incidental ndings. Clinicians and radiologists treating similar patient populations should expect to encounter relevant incidental ndings fairly frequently. The limitation of our study is that it was not based on surgical conrmation. Each positive nding was diagnosed through MRI only. Also, we did not review the patients medical charts for documentation of referral or follow-up of incidental ndings, so we could not evaluate patient outcome; however, because most of the ndings were benign, a particular treatment or close follow-up might not have been needed. In conclusion, incidental ndings of the lumbar spinal MRI were common and some were associated with age and sex. Most were benign pathology. An awareness of the prevalence of incidental ndings is helpful for the detection and diagnosis of a lesion not related to symptoms. References
1. Wagner SC, Morrison WB, Carrino JA, Schweitzer ME, Nothnagel H. Picture archiving and communication system: effect on reporting of incidental ndings. Radiology 2002; 225:500505 2. Ross JS, Masaryk TJ, Modic MT, Carter JR, Mapstone T, Dengel FH. Vertebral hemangiomas: MR imaging. Radiology 1987; 165:165169 3. Jackson DE, Atlas SW, Mani JR, Norman D. Intraspinal synovial cysts: MR imaging. Radiology 1989; 170:527530 4. Westbrook JI, Braighwaite J, Mclntosh JH. The outcomes for patients with incidental lesions: serendipitous or iatrogenic? AJR 1998; 171:11931196 5. Kamath S, Jain N, Goyal N, Mansour R, Mukherjee K. Incidental ndings on MRI of the spine. Clin Radiol 2009; 64:353361 6. Doppman JL, Oldeld EH, Heiss JD. Sympto matic vertebral hemangiomas: treatment by means of direct intralesional injection of ethanol. Radiology 2000; 214:341348 7. Murray RO, Jacobson HG. Radiology of skeletal disorders, 2nd ed. New York, NY: Churchill Livingstone, 1977:518 8. Barzin M, Maleki I. Incidence of vertebral he mangioma on spinal magnetic resonance imaging in northern Iran. Pak J Biol Sci 2009; 12:542544 9. Voyadzis JM, Bhargava P, Henderson FC. Tarlov cysts: a study of 10 cases with review of the literature. J Neurosurg 2001; 95:2532 10. Nabors MW, Pait TG, Byrd EB, et al. Updated assessment and current classication of spinal meningeal cysts. J Neurosurg 1988; 68:366377 11. Brown E, Matthes JC, Bazan C 3rd, Jinkins JR. Prevalence of incidental intraspinal lipoma of the lumbosacral spine as determined by MRI. Spine (Phila Pa 1976) 1994; 19:833836 12. Ross JS, ed. Diagnostic imaging, spine. Salt Lake City, UT: Amirsys, 2004 13. Bureau NJ, Kaplan PA, Dussault RG. Lumbar facet joint synovial cyst: percutaneous treatment with steroid injections and distensionclinical and imaging follow-up in 12 patients. Radiology 2001; 221:179185 14. Abdullah AF, Chamber TW, Daut DP. Lumbar nerve root compression by synovial cysts of the ligament avum. J Neurosurg 1984; 60:617620 15. Howling SJ, Kessel D. Case report: acute radiculopathy due to a haemorrhagic lumbar synovial cyst. Clin Radiol 1997; 52:7374 16. Liu SS, Williams KD, Drayer BP, Spetzler RF, Sonntag VK. Synovial cysts of the lumbosacral spine: diagnosis by MR imaging. AJR 1990; 154: 163166 17. Dahlgren RM, Baron EM, Vaccaro AR. Pathophysiology, diagnosis, and treatment of spinal meningoceles and arachnoid cysts. (dissertation) Philadelphia, PA: Department of Orthopaedic Surgery, Thomas Jefferson University, 2007 18. Young IS, Bruwer AJ. The occult intrasacral meningocele. Am J Roentgenol Radium Ther Nucl Med 1969; 105:390399 19. Abbott KH, Retter RH, Leinbach WH. Role of perineurial sacral cysts in sciatic and sacrococcygeal syndromes: review of literature and report of nine cases. J Neurosurg 1957; 14:521 20. Shimizu H, Tominaga T, Takahashi A. Cine mag netic resonance imaging of spinal intradural arachnoid cysts. Neurosurgery 1997; 41:95100 21. Lee HJ, Cho DY. Symptomatic spinal intradural arachnoid cysts in the pediatric age group: description of three new cases and review of the literature. Pediatr Neurosurg 2001; 35:181187

Downloaded from www.ajronline.org by 202.152.202.150 on 11/10/13 from IP address 202.152.202.150. Copyright ARRS. For personal use only; all rights reserved

AJR:196, May 2011 1155

You might also like