Spinal angiolipomas (SAL) are rare benign tumors, accounting for about 0.14–1.2% of all spinal axis tumors and 2–3% of spinal epidural tumors [1]. They most commonly occur in mid-thoracic region, purely lumbar location is rarely reported [2, 3]. Histologically, SAL is comprised of fat tissue and abnormal vascular architecture. Depending on the predominance of the tissue components, imaging findings of the lesion varies. MRI has become the imaging modality of choice for detecting and characterizing SALs.

Here we report 3 cases of epidual spinal angiolipoma, one lumbar and two thoracic SALs, and we evaluated them by employing multiple MRI techniques including diffusion weight imaging (DWI) and dynamic contrast-enhanced (DCE) MRI.

First reported in 1890 by Berenbruch [4] spinal epidural angiolipomas are rare benign tumors composed of mature adipocytes and abnormal vasculature in varied proportions. There were 177 patients with spinal epidural angiolipoma identified up to June 2015, with an average age of 46 ± 16 years old and 59% of patients were females [5]. The tumors are predominantly located in thoracic spine, pure lumbar angiolipomas are rather rare with less than 20 cases ever reported [2, 3]. About 90% of lesion arises in the posterior epidural spaces, extending over several vertebral bodies [6].

The clinical symptoms of SALs are consistent with other benign, space-occupying spinal lesions that are related to spinal cord/nerve root compression, or infiltration of surrounding tissue. The most common initial symptoms include lower limb numbness or paraesthesia, leg weakness and back pain. It can evolve over a span of months to years, but occasionally sudden onset or worsening of neurological symptoms occurred when there is a rapid increase in tumor size due to intratumoral thrombosis, haemorrhage [7, 8].

Etiology and pathogenesis of spinal angiolipoma remain largely unknown. It was proposed that an insufficiency of blood supply in thoracic spine contributes to the thoracic predominance of SAL [9]. In addition, it was reported that coexistence of vertebral hemangiomas is common in SAL patients, 23.8% cases (5/21) in one study [10] and one case (#2) in our report, indicating an abnormal vessel formation and growth mechanism. In fact, some authors consider SAL in the middle of a disease spectrum with lipoma at one end and hemangioma in the other end [16]. Several predisposing factors were also suggested to be associated with SAL, such as weight gaining or hormonal changes in pregnancy [11] and elevated body mass index [10], however, systematic studies on the etiology of SALS are still lacking.

The first pathological report describing anatomopathological composition of SAL was done by Howard et al. [12]. Under gross examination, the lesion of SAL is grey reddish in color, soft texture, with none or incomplete encapsulation, easy to separate from peripheral tissue. The tumor grows in epidural space along the longitudinal axis, normally in spindle shape. It can spread along the intervertebral foramen and presented as dumbbell shaped lesion. Traditionally, SALs are categorized into two types: noninfiltrating and infiltrating SAL. The noninfiltrating type is more common, usually well encapsulated and demarcated, while the infiltrating types are rare, partially or entirely unencapsulated, ill defined, and commonly infiltrate the surrounding tissues especially the bone [13], therefore should be considered and treated differently [14]. However, in recent years, some researchers indicated most patients achieved good recovery and there is no difference in the outcomes of the two types [15]. All 3 cases in this report can be categorized into the none-infiltrating type as they showed clear demarcation and no invasion to the bone.

Histological, SAL is a lesion with two distinct components: mature adipose tissue and proliferating abnormal blood vessels including capillary, sinusoidal, venous or arterial vascular elements [13]. The ratio of fat to vessels is variable [16], and atypia, pleomorphism, and mitotic figures of both adipose and angiomatous component were never observed [17]. Immunohistochemical assays were performed in 2 cases, and we observed a positive stain for CD31, S-100, SMA, but low proliferation rate was found with Ki-67, which is in agreement with others’ findings [18, 19].

Currently, MRI is the best imaging modality of choice in the diagnosis of SAL. The common features of SALs in MRI are presented as a spindle shaped lesion with both ends pen-like, usually located in the posterior epidural space extending longitudinally along the spine. The normal dura displayed as a slit-like low signal on T2-weighted imaging between the tumor and the spinal cord, and the subarachnoid space was narrowed and the spinal cord displaced. Occasionally the lesion may extend into the adjacent intervertebral forame with foraminal widening. The MR signal of SAL is comprised of two components: adipose tissue and vascular architecture. The fatty content was hyperintense on both T1- and T2-weighted images, hypointense on fat-suppressed images. The vascular component was hypointense on T1-weighted and hyperintense on T2-weighted, and showed intense enhancement in post-contrast images. Depending on its composite ratio of fat tissue versus vascular component, the lesion displayed as inhomogenous in pre-contrast images. Vascular flow voids sign is rarely seen on the MR images, as there is usually lack of well-developed arterioles inside the tumor.

In this case report, the MRI images of case 2 and case 3 demonstrated features of vascular component dominating SAL, as hypointense in T1, hypertintense in T2 and remarkable enhancement post-contrast. This type of SAL can be differentiated from other vascular originated tumors like hemangioma in that the fat signals from SAL are part of the tumor itself while in hemangioma the surrounding fat tissue clearly delineated from the tumor. Vascular void sign can also used to distinguish from arteriovenous malformation which is rarely seen in SALs. However, in this report one strip of vascular void signal was shown in case 2.

Case 1 can be easily confused with lipoma as it showed hyperintense in T2, reduced signal in fat suppressed T2, and no contrast enhancement. In this particular case, contrary to previous report by Hu et al. [6] fat suppression is not very helpful to reveal any vascular component. SAL was finally confirmed by histology and immunochemistry.

Dynamic contrast-enhanced MRI (DCE-MRI) was performed to assess the tumor vascular characteristics. The three SAL cases showed distinct shape in the time-intensity curve. No enhancement was observed in case one, while there were initial onsets of wash-in phase observed in both case two and three, followed by a wave-like washout phase in case two, and slow decaying washout phase in case three.

Diffusion-weighted imaging (DWI) is now widely appreciated as an indispensable tool in the examination of the central nervous system (CNS), especially for detection of acute ischaemic stroke and for characterization and differentiation of CNS tumors. In this study, all three cases of SAL showed low signal on DWI map, but high signal on ADC diagram, with no obvious diffusion restriction. A large scale of study would be helpful in order to assess the utility of DWI in the diagnosis of SAL.

SALs are treated exclusively by surgical removal of the lesion, gross total resection for noninfiltrating SALs and sub-total excision for infiltrating ones. The outcomes of surgery are good and recurrence is unusual in SALs.

In summary, we reported three cases of SAL with the emphasis on their MRI features. MRI findings indicate that one case is fat tissue predominant and the other two cases are vascular component predominant. By employing multiple MRI technology, better characterization of SAL can be obtained in regard to the composition and infiltration of the tumor. In spite of the ever-growing acuity of MRI imaging, diagnosis of SAL should be made in combination with clinical, radiological and pathological examination findings.