2368
explains why the tumours are able to grow to such a tue of the enormous size of the lesion as compared with
large size before presentation either as an incidental find- the underlying nerve. The heterogeneous nature on MR
ing or with only minor symptoms.
imaging of GPSS reflects degenerative change that can
There is a wide differential diagnosis for sacrococ- be seen in long-standing schwannomas. Degenerative
cygeal and presacral masses [13]. Both CT and MR im- change, known in the histological literature as “ancient
aging are ideally suited to detect sacral pathology and schwannoma”, is typified by cyst formation, haemor-
delineate the soft tissue and bony components. While CT rhage, calcification and fibrosis [19]. These features
demonstrates the bone destruction to best effect, the were present on histology to varying degree in all four of
multiplanar capability of MR imaging and multiplanar the current cases with the exception of calcification
reconstructions readily available on state-of-the-art CT which was noted in two cases. A combination of cyst
allow for precise identification of the involved level. formation, haemorrhage and necrosis explains the pres-
This is important for surgical staging as it indicates ence of the fluid–fluid level in one case which is more
which sacral nerve may have to be sacrificed at surgery common in large lesions (Fig. 4b) [19]. Cyst formation,
[4, 10, 14]. The morphological features on cross-section- haemorrhage and necrosis are relatively common in
al imaging allow for many diagnoses to be dismissed. schwannoma but rare in neurofibroma. Similarly, the
For example, an abscess will appear predominantly cys- majority (70%) of schwannomas tend to be encapsulated,
tic and ill-defined due to extensive surrounding inflam- whereas the minority (30%) of neurofibromas are [17].
mation. An anterior meningocoele will also appear cystic This fibrous pseudocapsule can be readily identified in
but well defined. Most aneurysmal bone cysts in children GPSS as a thin low-signal rim on MR imaging (Figs. 1a,
and adolescents, giant cell tumour and chordoma in 3c). Occasionally, not seen in any of the current cases,
adults and metastases and myeloma in the elderly will be the melanotic form of schwannoma, containing pigment-
shown to arise primarily within the sacrum with second- ed schwann cells, will show raised signal intensity on
ary soft tissue extension. Giant sacral schwannoma may unenhanced T1-weighted images [20].
therefore be easily mistaken for one of these conditions,
Despite its rarity and the fact that it does not exhibit
whereas GPSS appears well defined with only minor many of the typical MR features of benign peripheral
bone involvement (Figs 1, 2, 3, 4). On occasion, the soft nerve sheath tumours, the diagnosis of GPSS can be sug-
tissue component of chordoma may greatly exceed the gested when identifying a large, well-defined, heteroge-
bony destruction and can contain mineralisation as seen neous presacral soft tissue mass with a low-signal rim,
in two cases in this series on CT (Figs. 3b, 4c). However, arising just to one side of the midline, with minor under-
the bony origin of GPSS is just lateral to the midline, lying bone erosion (Figs. 1, 2, 3, 4). Biopsy is still re-
corresponding to one of the anterior sacral foramina, un- quired as the differential diagnosis for these lesions also
like chordoma which is almost invariably a midline le- includes a malignant peripheral nerve sheath tumour
sion [13].
(MPNST) formerly known as malignant schwannoma.
In the peripheral soft tissues there are several features Indeed, features claimed to be more suggestive of a
on MR imaging which suggest the diagnosis of a neuro- MPNST than the benign counterpart include large size
genic tumour, be it schwannoma or neurofibroma; these (>5 cm) and heterogeneity [19] both of which are also
include the presence of a fusiform shape, an entering and typical as seen in this series. Indistinct margins or a lack
exiting tail representing the host nerve [15, 16], typically of a pseudocapsule are more suggestive of MPNST
eccentrically placed with a schwannoma [17], the “split- which was a cause of concern at the preoperative imag-
fat” sign [18], the “target” sign [16] and the “fasicular” ing of the case of GPSS with the satellite lesion
sign [19]. These features are not seen with GPSS by vir- (Fig. 2a).
References
1. Whittaker LD, Pemberton JJ (1938)
Tumours ventral to the sacrum. Ann
Surg 107:96–106
2. Abdelwahab IF, Herman G, Stallman
A, Wolfe D, Lewis M, Zawin J (1989)
Case report 564: giant intraosseous
schwannoma. Skeletal Radiol
18:466–469
3. Abernathey CD, Onofrio BM,
Scheithauer B, Pairolero PC, Shives
TC (1986) Surgical management of gi-
ant sacral schwannomas. J Neurosurg
65:286–295
4. Feldenzer JA, McGauley JL,
McGillicuddy JE (1989) Sacral and
presacral tumours: problems in diagno-
sis and management. Neurosurgery
25:884–891
5. Kotoura Y, Shikita J, Yamamuro T,
Kasahara K, Iwasaki R, Nakashima Y,
Yamambe H (1991) Radiation therapy
for giant intrasacral schwannoma.
Spine 16:239–242
6. Turk PS, Peters N, Libbey HJ (1992)
Diagnosis and management of giant in-
trasacral schwannoma. Cancer
70:2650–2657
7. Salvant JB, Young HF (1994) Giant in-
trasacral schwannoma: an unusual
cause of lumbosacral radiculopathy.
Surg Neurol 41:411–412
8. Ortolan EG, Sola CA, Gruenberg MF,
Vazquez FC (1996) Giant sacral
schwannoma, a case report. Spine
21:522–526
9. Dominguex J, Labato RD, Ramos A,
Rivas JJ, Gomez PA, Castro S (1997)
Giant intrasacral schwannomas: report
of 6 cases. Acta Neurochir
139:954–960