MR imaging findings of pilomatricomas: a radiological–pathological correlation

Introduction

Pilomatricoma is a rare benign tumor of the dermis or subcutaneous tissue that originates from pluripotential cells that normally differentiate into hair matrix cells (1). Pilomatricoma was first described by Malherbe and Chenantais in 1880 as a calcified tumor originating from the sebaceous glands, and was referred to as a “calcifying epithelioma of Malherbe” (2). However, because the most likely mechanism for developing a pilomatricoma is the inclusion of epidermal elements at abnormal locations during embryonic life that subsequently grow after birth (3), pilomatricoma has become the most commonly used term.

Pilomatricomas account for 0.12% of all skin tumors, and 2–3.5% of the reported cases involve multiple lesions (4). Pilomatricomas are usually solitary, but multiple familial pilomatricomas have been associated with autosomal dominant diseases or syndromes, such as myotonic muscular dystrophy, Gardner syndrome, and Rubinstein–Taybi syndrome (5). Pilomatricomas primarily affect children and adolescents and typically occur in the first two decades of life (5). Approximately 60% of pilomatricomas occur in patients aged under 20 years with a slight female predominance and female:male ratio of 3:2 (6,7). The treatment of choice is surgical resection and the prognosis is typically good.

In most cases, the diameter of a pilomatricomas is in the range of 0.5–3 cm (6,8). Clinically, these lesions typically present as slow-growing, asymptomatic, solitary, firm, and superficial nodules that slide freely over the underlying skin, often with a reddish-blue discoloration or ulceration of the overlying skin. Approximately half of all pilomatricomas occur in the head and neck region, and the remainder occur with decreasing frequency in the upper extremities, the trunk, and the lower extremities (4,9).

Although the CT findings of pilomatricomas have been well described (8,10–13), the magnetic resonance imaging (MRI) findings have only been reported in a few case reports (3,10,14–16). One case series of 47 pilomatricomas did assess five by MRI, but the remainder were evaluated by ultrasound and/or computed tomography (CT) (13). To the best of our knowledge, no case series has focused on the MRI findings. Accordingly, we aimed to evaluate the MRI findings of pilomatricomas and to assess how they compare with the pathological findings.

Material and Methods

Results

Imaging findings

The main MRI findings of pilomatricomas are displayed in Table 1 and Figs 1–3. The margins of all tumors were well-demarcated. Both reticular and ring-like hyperintensities were observed in seven of 12 tumors assessed by fat-suppressed T2W images, and both reticular and ring-like enhancement were observed in three of six tumors assessed by fat-suppressed gadolinium-enhanced T1W images.

OPEN IN VIEWER

Table 1.

MRI findings of pilomatricomas.

MRI findings	Value
Minimum diameter (cm)
Mean	1.7
Range	0.7–3.2
T1W image (n = 12)
Homogeneous hypointensity	12
T2W image (n = 12)
Homogeneous hypointensity	5
Reticular hyperintensity	6
Ring-like hyperintensity	0
Cystic hyperintensity	1
Fat-suppressed T2W image (n = 12)
Reticular hyperintensity	9
Ring-like hyperintensity	8
Peritumoral fat strands	5
Fat-suppressed enhanced T1W images (n = 6)
Reticular enhancement	3
Ring-like enhancement	5
No enhancement	1
Diffusion-weighted image (n = 4)
Homogeneous hypointensity	1
Heterogeneous hyperintensity	3

OPEN IN VIEWER

Fig. 1.

A 7-year-old girl with pilomatricoma in the left arm. (a) Axial T2W fast spin-echo MR image (TR/TE, 3116/102 ms) shows a well-circumscribed subcutaneous lesion (arrow) with cystic degeneration (arrow head). (b) Axial fat-suppressed gadolinium-enhanced T1W spin-echo MR image (TR/TE, 450/9 ms) also shows a well-circumscribed subcutaneous lesion (arrow) with ring-like enhancement. Obvious peritumoral fat stranding was observed (arrow head). (c) Coronal fat-suppressed T2W fast spin-echo MR image (TR/TE, 3166/105 ms) shows a well-circumscribed subcutaneous lesion (arrow) with cystic degeneration (curved arrow). Peritumoral fat stranding was observed (arrow heads), but ring-like hyperintensity was not observed.

OPEN IN VIEWER

Fig. 2.

A 4-year-old girl with pilomatricoma in the left posterior neck. (a) Sagittal fat-suppressed T2W fast spin-echo MR image (TR/TE, 4063/90 ms) shows a well-circumscribed subcutaneous lesion (arrow) with reticular and ring-like hyperintensities. Peritumoral fat stranding was not observed. (b) Sagittal fat-suppressed gadolinium-enhanced T1W spin-echo MR image (TR/TE, 575/10 ms) also shows a well-circumscribed subcutaneous lesion (arrow) with reticular and ring-like enhancement. (c) ADC map shows low ADC value (1.19 × 10⁻³mm²/s) (arrow). (d) Histopathological specimen (HE stain, low-power magnification) reveals epithelial cell islands componsed of basaloid cells (asterisks) and shadow cells (star). Abundant edematous and fibrous stroma with inflammatory cell infiltration and vascular proliferation (arrows) is observed between epithelial cell islands. Tumor capsule with collagen fiber and inflammatory cell infiltration is observed (arrow heads).

OPEN IN VIEWER

Fig. 3.

A 14-year-old girl with pilomatricoma in the left posterior neck. (a) Sagittal fat-suppressed T2W fast spin-echo MR image (TR/TE, 3050/90 ms) shows a well-circumscribed subcutaneous lesion (arrow) with reticular and ring-like hyperintensities. Mild peritumoral fat stranding was observed (arrow head). (b) Sagittal fat-suppressed gadolinium-enhanced T1W spin-echo MR image (TR/TE, 770/9 ms) also shows a well-circumscribed subcutaneous lesion (arrow) with reticular and ring-like enhancement. (c) Histopathological specimen (HE stain, low-power magnification) reveals epithelial cell islands componsed of nest of shadow cells (asterisks). Reticular edematous and fibrous stroma with inflammatory cell infiltration and vascular proliferation (arrows) is observed between epithelial cell islands. Thick tumor capsule with collagen fiber, inflammatory cell infiltration, and vascular proliferation is observed (arrow heads). Prominent lymphocytic infiltration of peritumoral fat tissue is also observed (curved arrow).

No significant difference was found between the maximum diameter of the tumors with (17.3 ± 7.1 mm) and without (14.0 ± 1.7 mm) reticular hyperintensity on fat-suppressed T2W images (P = 0.45) (P = 0.24 for Levene’s tests). However, the maximum diameter of tumors was significantly larger in those with (19.4 ± 6.3 mm) than without (12.4 ± 1.2 mm) ring-like hyperintensity on fat-suppressed T2W images (P < 0.05) (P = 0.30 for Levene’s tests).

On DW images, one tumor showed homogeneous hypointensity and three showed heterogeneous hyperintensity. Four pilomatricomas showed low ADC values (range, 0.93–1.23 × 10⁻³mm²/s; mean, 1.13 ± 0.17 × 10⁻³mm²/s) (Fig. 2).

Discussion

The radiographic finding of a solitary, sharply demarcated subcutaneous tumor with extensive sand-like or dense focal calcification is probably pathognomonic of a pilomatricoma (17). The characteristic CT findings are reported as a well-defined soft-tissue density mass with various amounts of calcification and various degrees of enhancement (10–12). However, although CT images revealed the presence of macroscopic calcifications, they could not observe microscopic calcifications (12). Our histopathological examinations revealed intratumoral calcification in 11 of the 12 tumors. Although all 12 tumors were accompanied by focal or diffuse hypointensity on T2W images, T2 hypointensity cannot only be caused by calcification but can also be caused by fibrosis or epithelial cell islands, making it difficult to assess the presence of calcification on MR images. The degree of enhancement has also been said to depend on the epithelial cell components, fibrous tissue stroma, and vascular tissue (12). Finally, given that the unusual cystic degeneration of pilomatricoma appearing as a fluid density without enhancement on CT has only been reported in two cases (13), we also presented the first MRI report of pilomatricoma with cystic degeneration (Fig. 1).

Pilomatricoma has been described as showing a homogeneous intermediate signal intensity on T1W images and an inhomogeneous intermediate signal intensity on T2W images (3,10,14,15). The hyperintense reticulations have been reported as a characteristic MR finding of pilomatricomas on T2W images (13,15,16) and gadolinium-enhanced T1W images (13,16). At first, the internal reticulations on T2W images were reported to correspond to the basaloid cells (15). Later, internal reticulations and septations that were equally well seen on T2W and fat-suppressed gadolinium-enhanced T1W images were reported to correspond to the surrounding edematous stroma rather than the basaloid cells because basaloid cells were seen as sheets of avascular epithelial cells (13,16). On our histological examination, edematous and fibrous stroma with various degrees of inflammatory cell infiltration and vascular proliferation was observed in all nine tumors that showed evidence of reticular hyperintensity on fat-suppressed T2W images. Therefore, we considered the reticulation on MR images to be edematous and fibrous stroma with inflammatory cell infiltration and vascular proliferation.

Because fat-suppressed sequences augment tissue contrast by improving the dynamic range, they can emphasize the internal structures of soft tissue tumors. In addition, fat suppression can be helpful for reducing the severity of artifacts (18). Therefore, fat-suppressed T2W/STIR imaging should be performed for all soft tissue tumors. As with other soft tissue tumors, internal reticulations were observed more frequently and more clearly on fat-suppressed T2W images than on T2W images in our series.

On ultrasound, the hypoechoic rims of pilomatricomas have been reported to correspond to the connective tissue capsule (19). Lim et al. (13) reported that a connective tissue capsule was histologically confirmed in 37 (79%) of 47 tumors, and that rim enhancement on contrast-enhanced T1W images was observed in five (100%) of five tumors. Our pathological examination also included the rim of the tumor capsule with collagen fibers, inflammatory cell infiltration, and vascular proliferation. While the fibrous capsules of Schwannomas and pleomorphic adenomas usually present as hypointense rims on T2W images (20,21), the capsules of polimatricomas presented as hyperintense rims on fat-suppressed T2W images. We attributed this to the prominent inflammatory cell infiltration and vascular proliferation seen at this site in most of our histopathological specimens. In our series, although ring-like hyperintensity was not observed on T2W images, it was observed in eight of the 12 tumors on fat-suppressed T2W images. Because ring-like hyperintensity was hidden by surrounding hyperintense fat tissue on non-fat-suppressed sequences, fat-suppressed sequences were useful for its detection. However, one of six tumors showed no enhancement on fat-suppressed contrast-enhanced T1W images, because the intratumoral stroma and the tumor capsule were histologically obscure. In addition, our results showed that the tumors with ring-like hyperintensity on fat-suppressed T2W images were significantly larger than those without that finding.

Higher signal bands radiating from the lesion center to higher signal periphery on T2W images were reported as another characteristic of pilomatricoma on MR finding (15). However, it seems that this MR finding corresponds to the combination of a hyperintense reticulation and a hyperintense rim. In our series, both reticular and ring-like hyperintensities were observed in seven of the 12 tumors on fat-suppressed T2W images, and both reticular and ring-like enhancement were observed in three of six tumors on fat-suppressed gadolinium-enhanced T1W images.

Lim et al. (13) have reported that chronic inflammation with a foreign body reaction of the peritumoral subcutaneous fat tissue was histologically identified in 30 (64%) of 47 tumors, and that peritumoral changes were present in 47% of lesions on ultrasound, 32% on CT, and 60% on MRI. Our histological examination revealed various degrees of inflammatory cell infiltration in the peritumoral fat of all five tumors with evidence of peritumoral fat stranding on fat-suppressed T2W images. Therefore, we considered peritumoral fat stranding on MR images to signify inflammatory cell infiltration in peritumoral fat. Foreign body giant cells in the peritumoral fat were also observed in some cases.

To our knowledge, the DW imaging findings of pilomatricomas have not been reported prior to this study. In our series, three of four pilomatricomas showed heterogeneous hyperintensity on DW images and had low ADC values. We speculate that inflammatory cell infiltration caused this hyperintensity on DW images and a small amount of tumor matrix caused restriction of water protons.

The differential diagnosis for pilomatricoma includes many other histological diagnoses, notably Schwannomas and epidermal cysts. In particular, the imaging findings of pilomatricomas on fat-suppressed T2W images are similar to those of Schwannoma. This is because the “fascicular” sign, which indicates multiple, small, ring-like hypointense structures accompanied by peripheral hyperintense areas, resembles the hyperintense reticulation of pilomatricomas on T2W images (21). However, unlike pilomatricomas, Schwannomas have a hypointense rim that corresponds to the fibrous capsule on T2W images and they are not accompanied by peritumoral fat stranding. Furthermore, because of their superficial location, pilomatricomas do not show the “split-fat” sign characteristic of Schwannomas. Although epidermal cysts are arguably the most important clinical differential diagnosis (22), they are easily differentiated from pilomatricomas because they always show pure cystic appearances on MRI. Other important differential diagnoses include epidermal inclusion cyst, dermoid cyst, foreign body reaction, lipoma, calcinosis cutis, osteoma cutis, ossifying hematoma, cutaneous tuberculosis, squamous cell carcinoma, cutaneous T-cell lymphoma, dermatofibrosarcoma protuberans, malignant melanoma, Merkel cell carcinoma, and metastatic carcinoma of the skin.

The present study has several limitations. First, because pilomatricomas are rare soft tissue tumors, our sample was small. Second, our study had a retrospective design, which meant that we could not obtain contrast-enhanced MR images in all patients. Rare lesions, however, make it difficult to conduct prospective studies. Third, the interobserver variation in assessing the findings in the study was not determined, with all discrepancies being resolved by consensus.

In conclusion, the reticular appearance, ring-like appearance, and presence of peritumoral fat stranding were common MRI features of pilomatricomas. The fat-suppressed sequences allowed easy recognition of these characteristic findings. In addition, tumors with ring-like hyperintensity on fat-suppressed T2W images were significantly larger than those without.