Comparisons Between Heavily T2-Weighted MR and CT Myelography Studies in Two Patients with Spontaneous Intracranial Hypotension

Abstract

Our previous study demonstrated that heavily T2-weighted magnetic resonance (MR) myelography could successfully detect abnormal cerebrospinal fluid (CSF) collections in patients with spontaneous intracranial hypotension (SIH); however, its ability to demonstrate ongoing CSF leakage sites is uncertain. Currently, computed tomographic (CT) myelography is still considered the study of choice for such a purpose. In this study, we compared the results of the two imaging modalities in two patients with SIH. CSF leakages, such as C1-2 extraspinal collections, CSF along nerve root sleeves, and epidural fluid collections were noted on both studies with minor discrepancies. CSF collection along nerve root sleeves demonstrated by MR myelography was the most likely ongoing leakage site in comparison with CT myelography. Targeted epidural blood patches in both patients resulted in complete headache resolution. Noninvasive, non-contrasted and time-saving MR myelography is a promising study for patients with SIH. Further large-scale validation studies are needed.

Introduction

Spontaneous intracranial hypotension (SIH) is a disorder characterized by postural headache and accompanied with protean manifestations (1), which sometimes cause initial misdiagnosis (2). Much has been learned since the early 1990s, and evidence suggests that the most common aetiology is spinal cerebrospinal fluid (CSF) leaks (1, 3). Invasive imaging studies, such as conventional myelography, computed tomographic (CT) myelography or radioisotope cisternography, are often required to demonstrate the ongoing leakage sites (1, 3). However, sensitivities of these examinations vary among reported studies (4–7).

Heavily T2-weighted magnetic resonance (MR) can image water molecules by using a single-shot fast spin-echo pulse sequence with extremely long effective TE (8). We recently used heavily T2-weighted MR myelography to identify CSF leakage in patients with SIH and obtained high yield rates (9). Our protocol included four longitudinal planes and whole-spine transverse planes. The study was non-invasive and non-contrasted and took 13 min to finish both planes (9). Three types of abnormal CSF collections were detected in patients with SIH, including epidural fluid collection, C1-2 extraspinal collections and CSF along nerve root sleeves in the lower cervical and thoracic regions. However, because MR myelography is not a dynamic study, it is unknown whether this study could be used to demonstrate the ongoing CSF leakage sites for location-specific epidural blood patches or surgical repair. At present, CT myelography is considered the study of choice for such a purpose (1, 3).

Recently, two patients with SIH, who participated in our MR myelography studies (9), also underwent CT myelography. They both received targeted epidural blood patches later. This study compared these two myelography findings and investigated if the MR myelography could detect the ongoing leakage sites.

Case reports

Patient 1

A 30-year-old woman, who did not have a history of headache, developed a severe orthostatic headache, associated with nausea and vomiting after awakening from a nap. She described her headaches as diffuse and throbbing and occurring within seconds after sitting up or standing up and completely relieved by lying down. On the second day after headache onset, she visited a nearby hospital, where brain MRI showed diffuse pachymeningeal enhancement and lumbar puncture revealed a dry tap. With a working impression of SIH, she was admitted for intravenous hydration, but her symptoms persisted. Without a satisfactory outcome she then came to our headache clinic and was admitted to our hospital. On day 14 after headache onset, heavily T2-weighted MR myelography (9) was performed, using a 1.5-T superconducting system (Signa Excite; GE Medical System, Milwaukee, WI, USA) with a phased-array synergy-spine coil. A single-shot fast spin-echo pulse sequence with extremely long effective TE (i.e. heavily T2-weighted) was employed to acquire the MR myelography images of the entire spine. This sequence effectively demonstrated the presence of CSF and rapidly obtained images with minimal motion artefacts (10). Our protocol included four longitudinal planes [matrix size 320 × 256, TE 950 ms, TR 6000 ms, 50 mm section thickness and field of view (FOV) 280 mm] and whole-spine transverse planes (matrix size 320 × 256, TE 950 ms, TR 6000 ms, 6 mm section thickness and FOV 200 mm). The study showed CSF collections along her bilateral T1 root sleeves (Fig. 1B), extraspinal CSF collections on the dorsal side at T5 (Fig. 1D) and epidural fluid collections from C2-3 to T8-9. On day 16, CT myelography [CT scanner: Somatom Sensation 16; Siemens AG, Erlangen, Germany; CT parameters: collimation 16 × 0.75 mm, rotation time 0.75 ms, feed/rotation 3 mm, KV 120, effective mAS 330, CTDI (dose index) 50 mGy] was performed 2 h after intrathecal injection of 10 ml contrast medium [Iopamiro-300 (iodine 300 mg/ml); Bracco S.p.A., Milano, Italy] and 10 ml normal saline through lumbar puncture. The study showed leakage of contrast medium along the bilateral T1 nerve root sleeves (Fig. 1A), and over the epidural region at the C7 level. However, the extraspinal space at the T5 level was not seen (Fig. 1C). Epidural blood patches with 20 and 17 ml autologous blood over T1 and T5 levels, according to the results of MR myelography, were performed and her headache subsided immediately. Follow-up MR myelography 5 weeks later showed disappearance of the abnormal CSF collections. The patient was followed up for 2 months at our clinic without headache recurrence.

Figure 1

(A) shows contrast medium leakage along bilateral T1 nerve root sleeves (arrows) in the axial view of 2-h computed tomographic (CT) myelography. Compatible findings are shown in magnetic resonance (MR) myelography (B, arrowheads). The epidural cerebrospinal fluid (CSF) collections are separated from the thecal sac by a lower signal dura (D, arrowheads). However, the extraspinal CSF collections on the dorsal side of the T5 level (D, arrows) in MR myelography are not shown in the CT myelography at the same level (C).

Patient 2

A 29-year-old woman, with a history of episodic migraine without aura, suffered from acute onset of severe orthostatic headaches accompanied by nausea, vomiting and bilateral tinnitus. The headache occurred within 10 min in an upright position, and could be relived by lying down within 10 min. Brain MRI on day 11 after headache onset showed diffuse pachymeningeal enhancement. MR myelography on day 40 showed C1-2 extraspinal fluid collections (Figs 2C and 3B), CSF along bilateral C5-6, C6-7 and C7-T1 nerve root sleeves (Fig. 2I) as well as whole-spine epidural fluid collections. On the next day, CT myelography was performed. In order to evaluate the dynamic change of CSF leakage, we conducted CT myelography twice, i.e. immediately (around 15 min) and 2 h after intrathecal injection of contrast medium [10 ml Iopamiro-300 (iodine 300 mg/ml); Bracco S.p.A., Milano, Italy; and 10 ml normal saline]. The immediate CT myelography showed leakage of contrast medium in the C1-2 extra-spinal region (Figs 2A and 3A), along bilateral C2-3, C3-4, C4-5, C5-6, C6-7 and C7-T1 root sleeves (Fig. 2D,G) and epidural collections from C1-2 to T2-3. The 2-h CT myelography disclosed contrast medium leakage in the C1-2 extraspinal region (Fig. 2B), along bilateral C4-5, C5-6, C6-7 and C7-T1 nerve root sleeves (Fig. 2H) and epidural collections from C1-2 to T1-2. Nerve root sleeves were not shown at C2-3 (Fig. 2E) or C3-4 levels. The patient received epidural blood patch with 12 ml autologous blood at the C7-T1 level, which relieved her headache on the same day. Three weeks later, a follow-up MR myelography showed only slight epidural fluid collections from C2-3 to T1 on the ventral side of the spinal cord. The CSF collections at the C1-2 extraspinal region and along multiple nerve root sleeves disappeared. The patient was followed up for 2 months at our clinic without headache recurrence.

Figure 2

(A,D,G) Axial views at C1-2, C2-3 and C7-T1 levels in the immediate computed tomographic (CT) myelography. (B,E,H) Two-hour CT myelography and (C,F,I) transverse views of magnetic resonance (MR) myelography at the same levels. (A,B) Contrast medium leakage in the C1-2 extraspinal region in both CT myelography studies, and (C) compatible findings in the MR myelography (arrow). (D) Contrast medium leakage at the C2-3 level in the immediate CT myelography, but not in 2-h CT myelography (E) or MR myelography (F). (F) Epidural fluid collection, which was separated from the thecal sac by the dura (arrowheads). Both G and H display contrast medium leakage along the bilateral C7 nerve root sleeves in the immediate and 2-h CT myelography studies. Compatible cerebrospinal fluid collections are shown in the MR myelography (I, arrow).

Figure 3

Both the sagittal reconstruction of the computed tomographic myelography (A) and the sagittal magnetic resonance myelogram (B) disclose the leakage of contrast medium at the C1-2 extraspinal region (arrows). Of note, (A) shows anterior epidural fluid collection in the middle thoracic region (arrowhead), separated from thecal sac by a faint lower signal dura.

Discussion

This pilot study made direct comparisons between CT and heavily T2-weighted MR myelography studies in two patients with SIH. The headache characteristics in our two patients fulfilled all the diagnostic criteria of International Classification of Headache Disorders-2 code 7.2.3, ‘Headache attributed to spontaneous low CSF pressure’ (11). The intervals between these two studies in both patients were 1 and 2 days. Abnormal CSF leakages demonstrated by the CT myelography including CSF along the nerve root sleeves, C1-2 extraspinal region collections and epidural fluid collections were all documented in the MR myelography. However, some discrepancies were also noted. In patient 1, the findings of the two studies were quite compatible except that the extraspinal CSF collection at the T5 level was not shown on 2-h CT myelography. The exact reason is unknown. Nearby bone tissue may have been obscuring; however, different CSF dynamics could not be completely excluded (12). In patient 2, the immediate CT myelography showed more levels of leakage along nerve root sleeves than MR myelography (C2 to T1 vs. C5 to T1). We postulate that intrathecal injection of 20 ml contrast medium and saline might raise intrathecal pressure during the immediate CT myelography study; therefore, not only ‘large’ but also ‘small’ CSF leakage sites were shown. The CSF pressure then gradually declined, and only large leakage sites were shown in the 2-h CT myelography. Nerve root sleeve enhancement demonstrated by the CT myelography has been deemed as the ongoing leakage sites for patients with SIH (13). Due to the compatible findings, we believe the CSF collections along the nerve root sleeves seen by MR myelography can represent ongoing leakage sites. In addition, epidural blood patches specifically targeted at these sites based on our MR myelography resulted in headache resolution in both our patients on the same day.

Regarding the epidural fluid collections, CT myelography showed fewer levels of contrast medium collection (T7 in patient 1, C1-2 to T2-3 in patient 2) than the MR myelography (C2-3 to T8-9 in patient 1 and whole spine in patient 2). Because the epidural collection of contrast medium in CT myelography is leaked directly from the intrathecal space after intrathecal injection, the collection sites may represent the leakage sites (13), whereas MR myelography shows the overall distribution of CSF collections in the epidural space.

C1-2 extraspinal fluid collection is considered to be a false localization sign of ongoing CSF leakage sites because it is hypothesized to result from a sequence of events of CSF leakage, i.e. CSF leaking into the epidural space, then extending within the spinal canal, and eventually escaping from the epidural space into the soft tissue at the C1–2 levels (14). However, our study has failed to demonstrate these sequential events after intrathecal injection of contrast medium in patient 2. In fact, C1-2 extraspinal fluid collection was found in both the immediate and 2-h CT myelography and MR myelography. The results suggest that patient 2 might have a high-flow CSF leakage.

It is not easy to make comparisons with previous studies of MR myelography in patients with SIH, because most of them were single case reports and different methodologies were utilized. Extraspinal fluid collection (n = 1) (15), sacral bilateral meningeal diverticula (n = 1) (16), cystic perineural dilation (n = 1) and cystic dilation of nerve root sheath (n = 1) (17) were reported as ‘positive findings’ in previous MR myelography. ‘Irregular root sleeves’ shown in longitudinal MR myelography were reported in five cases, but their clinical significance was not known because no direct comparisons with CT myelography were made (4, 17, 18). Recently, MR myelography with intrathecal injection of gadolinium has also been used to study patients with SIH (19–21). All three showed contrast leakage at single (19) or multiple spinal levels (20, 21). This technique has the potential to evaluate patients with SIH; however, it is invasive and the safety of intrathecal injection of gadolinium still needs further investigation. In Taiwan, intrathecal injection of gadolinium has not yet been approved.

Conclusions

In spite of some minor discrepancies, the results of heavily T2-weighted MR and CT myelography studies were quite compatible in two patients with SIH. CSF along the nerve root sleeves in MR myelography may represent ongoing leakage sites for further targeted treatment. Non-invasive, non-contrasted and time-saving MR myelography is a promising study for patients with SIH. More data are needed to validate this examination.