Brain metastases with unusual MRI features in an infant with relapsed clear cell sarcoma of the kidney: A case report

Introduction

CCSK represents only 2–5% of all primary renal tumors in children. It carries a worse prognosis than Wilms tumor, with higher rates of bone metastases at initial diagnosis and higher rates of distant metastatic relapses. About 41% of distant metastatic relapses occur in the brain. Brain metastases have been suggested to stem from the regrowth of micrometastases that are present at an earlier stage. With the intensive chemotherapy that CCSK patients currently receive, the brain has become the most common site of distant metastatic relapse, probably due to the relative protection of the brain micrometastases from the treatment.^1,2

Even though the brain is the most common site of relapse, overall, only 7% of children with CSSK develop relapses in the brain. ¹ Because of the rarity of the condition, imaging findings of the metastatic brain lesions from CSSK have been reported in only a few cases, mostly without a detailed imaging description. ³ One case report indicated that the brain metastases from CCSK can have unusual imaging features for the metastatic brain lesions from extracranial solid tumors, such as the absence of metastasis-associated edema or contrast-enhancement, and a seemingly extra-axial location. ⁴ As MRI of the brain is being increasingly included in imaging screening and follow-up of patients with CCSK, knowledge of unusual imaging features of these metastatic lesions may aid in their early recognition.

Case report

Following the CARE guidelines, parental consent was obtained for the publication of this case. A 7-month-old female patient presented with complaints of intermittent vomiting for 1 month and abdominal enlargement for 2 days. Ultrasonography (US) and MRI of the abdomen revealed a 9 × 9 cm heterogeneous solid mass originating from the right kidney. The right renal vein and inferior vena cava showed no tumor thrombus. No abdominal lymphadenopathies were detected (Supplementary File 1). Computed tomography (CT) of the thorax and MRI of the brain showed no metastatic disease. With a provisional diagnosis of Wilms tumor, a right nephrectomy was performed, and a diagnosis of clear cell sarcoma of the kidney, a palisading variant, was made. A post-operative whole-body positron emission tomography (PET) showed no pathological fluorodeoxyglucose uptake. Chemotherapy (NWTSG-5 regime, 24-week protocol) and flank radiotherapy (1,8 Gy x 6 fractions) were applied. Interval follow-up imaging studies (MRI of the abdomen and brain, CT of the thorax) obtained at the end of the third month of therapy protocol did not show any signs of metastatic disease (Supplementary File 2).

At the end of the protocol, the patient had no complaints or symptoms. A follow-up MRI of the brain at 3T obtained 8 months after the diagnosis showed multiple well-defined rounded lesions of various sizes on the cerebral hemispheres, basal ganglia, thalamus, brain stem, and cerebellum. Most of the lesions exhibited alternating circular layers of high and low signal intensities on T2-weighted fast spin echo (SE) and T1-FLAIR images, giving them a targetoid appearance. None of the lesions showed tumor-associated vasogenic edema (Figure 1). On post-contrast, 3D T1W spoiled gradient recalled (SPGR) images most lesions showed faint enhancement of their T2 hypo- and T1 hyperintense layers. Post-contrast T1W SE images demonstrated faint but relatively diffuse enhancement of most lesions (Figure 2). Some peripheral metastatic lesions seemed to be located extra-axially at first view. A multiplanar evaluation demonstrated that the lesions were located intra-axially on the cortex-subcortical white matter interfaces and were displacing the white matter tracts inwards (Figure 3). Diffusion-weighted (DW) images showed relatively restricted diffusion of the outer T2 hyperintense signal layers. Susceptibility-weighted images did not show any apparent signal abnormalities (Figure 4). As subsequent imaging studies demonstrated synchronous metastatic bone and lung disease, and the parents declined further diagnostic or therapeutic procedures, the brain lesions were not biopsied (Supplementary File 3).OPEN IN VIEWER

Figure 1.

T2W FSE (TR:7700, TE: 102, 4 mm) (a), and T1W FLAIR (TR:2000, TE:24, TI 860, 4 mm) axial images show rounded metastatic lesions at four different levels. The lesions exhibit alternating circular layers of signal alterations on T2W and T1W images. The T2 hypointense layers are hyperintense on T1W images. Some lesions show a target-like appearance with central hypointensity and peripheral hyperintensity on T2. None of the lesions show tumor-associated vasogenic edema.

OPEN IN VIEWER

Figure 2.

T2W (a), T1W FLAIR (b), post-contrast 3D T1W SPGR (c), and post-contrast T1W SE (TR: 520, TE:10, 4 mm) (d) axial images from three different levels. T2 hypo- and T1 hyperintense components of the metastatic lesions are enhanced to a greater degree on SPGR images while T1 SE images demonstrate a more diffuse enhancement of most lesions.

OPEN IN VIEWER

Figure 3.

T2W appearances of the metastatic lesions on axial (a) and corresponding sagittal (b) images. Note the inward displacement of the cerebral subcortical white matter tracts by the hemispheric lesions (upper row). The left thalamic lesion (middle row) extends into the lateral ventricle and smoothly displaces the adjacent choroid plexus. The thalamic parenchyma partially surrounds the lesion (middle row). Cerebellar lesions either smoothly displace or engulf the adjacent white matter tracts. The engulfed white matter tracts preserve their outlines (lower row). Note the absence of tumor-associated edema.

OPEN IN VIEWER

Figure 4.

Axial T2W (a) DW (b), ADC (c), and SW (d) images from three different levels. DW images show the high signal intensity of the outer T2 hyperintense circular layers of the lesions. Corresponding ADC maps demonstrate relatively decreased signals of that layers. T2 hypointense layers do not demonstrate diffusion restriction.

Discussion

The reports with T2W MR images have described the parenchymal metastatic lesions from CSSK as large, peripherally located, hyperintense tumors with mass effect. The presence or absence of peritumoral vasogenic edema was not mentioned in those reports.^3,5 The absence of peritumoral edema in a metastatic brain lesion from CCSK has been reported so far by Shajpaul et al. ⁴ A single large-sized frontal hemispheric metastatic lesion, in that case, showed no tumor-associated edema and no contrast-enhancement. Those features have been attributed to the poor vascularity and cellularity of that metastatic lesion by the authors, reflecting the histopathology of the primary tumor. Although large-sized intraparenchymal metastases of the brain mostly exhibit vasogenic edema, the presence or degree of vasogenic edema may be influenced by other factors than the lesion size, such as the primary tumor type, location of the lesion in the gray or white matter, desmoplastic effects of the chemotherapy, and the presence or absence of reactive astrogliosis around the lesion.^4,6 As the histopathological evaluation of the metastatic brain lesions in the presented case was not performed, addressing the exact cause of the absence of the peritumoral edema is not possible. A faint contrast-enhancement of the metastatic lesions, however, might be due to the lesions’ poor vascularity and the signal alteration layers with T2 hypointensity and T1-FLAIR hyperintensity within the lesions might suggest the presence of relatively hypocellular fibrotic components. The hemispheric metastatic lesions in the presented case were located at the gray-white matter interfaces; however, they exhibited an asymmetrical growth pattern into the cortex toward the cerebrospinal fluid. That growth pattern of the lesions might also have restricted the development of vasogenic edema. Non-visualization of the nearby cerebral cortex and absence of surrounding vasogenic edema gave some of the lesions a seemingly extra-axial appearance. In the case report by Shajpaul et al., ⁴ the large-sized frontal metastatic lesion was also presumed to be located extra-axially on the initial CT scan. A subsequent MRI showed the lesion’s true intra-axial location and a resection showed that the lesion was attached to the overlying dura but not originating from it.

Alternating concentric signal intensity changes in MRI of metastatic brain lesions from CCSK have not been reported before. The target sign describes central hypointensity and peripheral hyperintensity on T2W images in a lesion. Observance of that sign in most metastatic lesions in the present case might be due to the simultaneous presence of circularly arranged fibrotic and fluid-rich components, represented by T2 hypointense and T2 hyperintense signal changes, respectively. Interestingly, Shajpaul et al. ⁴ have described the presence of a metastasis-associated cyst on CT, as a rim of hypodensity surrounding the solitary metastatic lesion. The cystic appearing part of the lesion corresponded to the avascular lesion surface at resection. Although not mentioned in the report, the presented T2W image in that case also demonstrates a circular hypointense inner layer. The T2 hyperintense signal layers in the case we present showed some degree of contrast-enhancement and exhibited relatively low ADC signal, thus, were not truly cystic.

Despite the similarities between the imaging features of the presented case and Shajpaul et al.’s ⁴ case, the metastatic brain lesions from CSSK do not seem to have a common or typical imaging appearance, as Park et al. ⁷ have reported two large-sized, intensely enhancing, dura-based extra-axial metastatic lesions at relapse and Zekri et al. ⁸ have reported metastatic involvement of the brain from CSSK as diffuse leptomeningeal thickening and enhancement.

The presented imaging findings of the metastatic lesions in this report indicate that early detection of the described lesions may be difficult due to the absence of perilesional vasogenic edema, the outward growth pattern of the lesions from the gray-white matter junction, and faint enhancement of most lesions on post-contrast T1W SE images. That may be particularly true if the myelination of the subcortical white matter is not completed. Including 3D T2W and 3D gradient recalled T1W images in follow-up brain MRI protocols of the patients with CCSK may increase the early detection of such metastatic lesions.

Supplemental Material

Supplemental Material

Supplemental Material

Supplemental Material

Supplemental Material