Sequestra in Prostate Cancer Metastases: CT and MR Appearance

S. Kaushik, MD*†

D. A. Bluemke, MD, PhD*

E. F. McCarthy, MD‡

 

*Departments of Radiology and ‡Pathology, The Russsell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins University School of Medicine, Baltimore, Maryland

†Department of Radiology, Medical College of Virginia Hospital, Virginia Commonwealth University, Richmond, Virginia

 

KEY WORDS: sequestrum, prostate cancer metastases, magnetic resonance imaging, computed tomography

ABSTRACT

Objective: To determine the frequency and distribution of sequestra in osseous metastases from prostate cancer and define their CT and MR imaging features.

Materials and Methods: We retrospectively reviewed CT and MR imaging in 96 patients with prostate cancer osseous metastases for a 3-year period (1997–2000). Pathologic diagnosis of prostate metastases was made by core biopsy in 20 of these patients. Diagnosis of osseous metastatic disease was established by bone scans in the remaining 76 patients. The presence or absence and size of sequestra were recorded in these two groups of patients. The CT attenuation, MR signal characteristics of sequestra and surrounding marrow were noted in the two groups.

Results: Sequestra were noted in 20 of 96 patients (21%) with prostate cancer metastases or in 18% of all prostate cancer osseous metastases. Sixteen percent of lesions in the biopsy-proven group and 18% of lesions in a non-biopsy bone scan diagnosis group displayed characteristics of sequestra. The locations of 96 sequestra were as follows: thoracic spine (n = 17), lumbar spine (n = 24), proximal femur (n = 11), sacrum (n = 18), ilium (n = 22), and humerus (n = 4). The CT appearance of sequestra was that of a high-attenuation osseous lesion surrounded by a low-attenuation rim separating it from the rest of the bone marrow. Sequestra were of low signal intensity on T1-weighted and T2-weighted fast-spin echo sequence, separated from the rest of the surrounding marrow by a thin rim of low- to intermediate-signal intensity on T1-weighted and hyperintense signal intensity on T2-weighted sequences. The gadolinium-enhanced fat-saturated T1-weighted MR imaging showed enhancement of the sequestrum.

Conclusion: Sequestra can be seen with metastatic disease from prostate cancer, which should be included in the differential diagnosis of this entity in addition to other conditions, including osteomyelitis, eosinophilic granuloma, fibrosarcoma, desmoplastic fibroma, malignant fibrous histiocytoma, and primary lymphoma of bone.

INTRODUCTION

Sequestra are detached osseous fragments separated from surrounding bone by a pathologic process.1 Sequestra have been reported in patients with chronic osteomyelitis, Langerhans cell histiocytosis, primary bone tumors (fibrosarcoma, malignant fibrous histiocytoma), and lymphoma.2–5 Detection of this entity in metastatic disease has not been reported previously. Prostate cancer is a common entity and metastases from prostate cancer can be osteoblastic, osteolytic, or mixed. In the presence of sequestra, the diagnosis can be confused with the above entities. The presence of sequestra might predispose to pathologic fracture. The incidence, location and distribution, CT and MR imaging appearance of sequestra in patients with prostate cancer osseous metastatic disease is presented in our study.

MATERIAL AND METHODS

All patients with bone scan or biopsy evidence of metastatic disease from prostate cancer were identified by a retrospective review of pathology database for a 3-year period (1997–2000). Ninety-six of these patients had undergone CT and MR imaging. The diagnosis of metastases was made by bone scan (n = 76) or core biopsy (n = 20). Patients with suspected metastases from prostate cancer by bone scan only without CT or MR imaging were excluded from the study.

Imaging studies of all 96 patients were retrospectively reviewed for the presence of sequestrum appearance. CT (n = 62) and MR imaging (n = 34) were obtained in these patients. A sequestrum was defined as a focus of increased density on CT. The sequestra were expected to have a low signal intensity on T1- and T2-weighted sequences. Gadolinium enhancement of sequestra was studied.

The following findings were documented in these cases: (1) presence or absence of sequestra on CT or MR imaging, (2) number of sequestra on CT and MR imaging, (3) the location of sequestra, (4) CT density of the lesion in comparison with the surrounding bone marrow attenuation, and (5) MR signal characteristics of metastatic lesions and sequestra on T1-weighted, T2-weighted, and gadolinium-enhanced T1-weighted fat-suppressed sequence.

CT scans were performed on a helical scanner (Somatom Plus 4; Siemens Medical Systems, Erlangen, Germany) with 2-mm collimation and subsequent sagittal and coronal reformats.

The MR examinations were performed on a 1.5-T MR scanner (Signa; General Electric Medical Systems, Milwaukee, WI). Sagittal and axial T1-weighted spin-echo images were obtained with a TR of 400 to 700 msec and a TE of 12 to 18 msec. T2-weighted fast spin-echo MR imaging was acquired with a TR range of 3000 to 4000 msec and a TE range of 45 to 60 msec. Gadolinium-enhanced T1-weighted imaging was acquired with fat saturation (TR: 400-700 msec, TE: 12–18 msec) in all patients. All studies were reviewed by two MSK radiologists (S.K., D.A.B.).

RESULTS

Twenty of 96 patients with CT and MR imaging of lesions underwent core biopsy. Thirty-two metastatic lesions were present in the biopsy group, five of which qualified for the appearance of sequestra. Diagnosis of osseous metastatic disease was established by bone scintigraphy in the remaining 76 patients. A total of 82 osseous lesions were discovered in these 76 patients, 15 of which demonstrated sequestra characteristics. A total of 109 lesions were noted in 96 patients, 20 of which (18%) had a sequestra appearance (Table 1).

More than one sequestrum was present in 21 of 96 patients (22%). The sequestra measured between 4 mm and 20 mm. Displacement of sequestra was not seen in any of our cases. The locations of 96 sequestra were as follows: thoracic spine (n = 17), lumbar spine (n = 24), proximal femur (n = 11), sacrum (n = 18), ilium (n = 22), and humerus (n = 4).

The CT appearance of sequestra was that of a high-attenuation osseous lesion surrounded by a low-attenuation rim separating it from the rest of the bone marrow (Fig. 1). The sequestra were of higher attenuation in comparison with the surrounding bone marrow (n = 6) on CT. The sequestra also showed CT attenuation similar to surrounding bone marrow (n = 14). Sequestra were of low signal intensity on T2-weighted fast-spin echo sequence with fat saturation separated from the rest of surrounding marrow by a thin rim of hyperintense signal intensity in patients with osteoblastic metastatic disease (Fig. 2). Figure 3 shows similar appearance of a sequestrum, however, with normal surrounding bone marrow.

T1-weighted fast-spin echo MR sequence showed sequestrum to be hypointense (Fig. 4A) with a low to intermediate signal intensity region of surrounding rim. The gadolinium-enhanced T1-weighted MR imaging acquired with fat saturation (Fig. 4B) showed enhancement of the sequestrum and surrounding abnormal marrow involved with metastatic disease. A distinct rim of unenhancing tissue was seen separating the two.

Discussion

The presence of sequestrum in association with metastatic disease has not been demonstrated previously. We do not know if sequestra among the metastatic category are only seen with prostate metastases or can be present with other metastases. The sequestra in prostate metastatic disease in our series were of high attenuation on CT, similar to sequestra seen with osteomyelitis. Sequestra associated with primary bone tumors2 might or might not have been avascular. The sequestra described with osteomyelitis were thought to be avascular,3 resulting in their dense radiographic appearance. The sequestra were of the same bone density as the surrounding bone in the cases of primary lymphoma of bone.5 The presence of enhancement of sequestra with gadolinium administration on MR imaging argues against their avascular nature in our series.

A halo sign was described with 90% of prostate metastases and 35% of metastases from other than prostate cancer in one study.6 These authors used either a diffuse hyperintense focal signal abnormality or a halo of hyperintense signal around the lesion on a T2-weighted series as a sensitive and specific marker of metastatic disease. However, direct histologic correlation of the halo sign was not provided in that study. The T1- and T2-weighted signal intensity of sequestra in our study is similar to “metastatic lesions with rim or halo” in this previous study. However, our study additionally demonstrates the appearance of sequestra on contrast-enhanced MR and on computed tomography, not known previously.

The MR features of prostate metastases to bone in our series comprised low signal intensity on T1-weighted and T2-weighted sequences. This is compatible with the osteoblastic nature of the metastatic tumor. The CT attenuation or MR signal intensity of the surrounding bone marrow depends on the presence or absence of the involvement by metastatic disease. The surrounding bone marrow showed CT attenuation or MR signal intensity similar to the sequestrum when it was infiltrated by the metastatic tumor and showed normal marrow CT attenuation and MR signal when it was free of metastatic disease

The exact mechanism of sequestrum formation in bone tumors and metastatic disease is not well known with certainty. The proposed mechanism might involve an aggressive destructive process leading to separation of sequestra from the surrounding bone marrow.

The presence of sequestra is the hallmark of our study because this entity has not been described previously with prostate cancer metastases. We acknowledge the small sample size and retrospective nature of our series as limitations, which are the result of the fact that a large number of patients with prostate cancer metastatic disease on bone scan but without CT or MR imaging studies were not included in the study. The exclusion of these patients from the study would alter the prevalence of sequestra in prostate cancer metastases; however, it will not refute its presence. Thus, our study in fact reflects the prevalence of a sequestrum sign in patients with metastases from prostate cancer who undergo cross-sectional (CT, MR) imaging. Many patients had biopsies performed from sites other than the ones showing a CT or MR imaging appearance of sequestra (eg, from a fracture site to exclude pathologic fracture or from a more accessible location for a percutaneous image-guided approach to the lesion). However, the fact that these patients had pathologically proven diagnoses of metastatic disease from prostate cancer and a typical appearance of sequestra on CT and MR imaging studies curtails the effects of these limitations.

In conclusion, osseous metastatic disease from prostate cancer can result in a sequestrum appearance on CT and MR imaging and should be included in the differential diagnosis of sequestra in addition to osteomyelitis and primary bone tumors, lymphoma.

REFERENCES

1. Blakiston’s Gould Medical Dictionary, 4th ed. New York: McGraw-Hill; 1979.

2. Dahlin DC, Unni KK: Malignant lymphoma of bone (reticulum cell sarcoma). In: Bone Tumors, 4th ed. Springfield, IL: Thomas; 1986:208–226.

3. Resnick D, Niwayama G: Osteomyelitis, septic arthritis, and soft tissue infection: the mechanisms and situations. In: Diagnosis of Bone and Joint Disorders, 2nd ed. Philadelphia: Saunders; 1988:2525–2754.

4. Edeiken-Monroe B, Edeiken J, Kim EE: Radiologic concepts of lymphoma of bone. Radiol Clin North Am 28:841–86, 1990.

5. Mulligan ME, Kransdorf MJ: Sequestra in primary lymphoma of bone: prevalence and radiologic features. Am J Roentgenol 160:1245–1248, 1993.

6. Schweitzer ME, Levine C, Mitchell DG, Gomella L: Bull’s eyes and halo’s useful MRI discriminators of osseous metastases. Radiology 188:249–252, 1993.

 

Table 1. Prostate Cancer Metastases on CT/MRI

 

                  Biopsy Proven     Bone Scan—Based

Patients (n)         20                           76

Lesions (n)         32                           82

Sequestra (n)      5                            15

Sequestra (%)    16                           18

 

Figure 1. Sagittal reformatted image from thoracic spine CT shows an osteoblastic metastatic lesion (black dot) seen separated from the rest of the vertebral body marrow by a low attenuation rim (arrowheads). This is a characteristic CT appearance of sequestrum seen in a 58-year-old man with prostate cancer.

 

Figure 2. Sagittal T2-weighted (TR: 2400, TE: 90) MR image of thoracic spine. The sequestrum is of hypointense signal intensity similar to the remaining vertebral body, however, separated by a hyperintense signal rim (arrowheads). There is expansion of the vertebral body resulting in cord compression.

 

Figure 3. T2-weighted (TR: 2600, TE: 90) axial MR image of the left femur in a 64-year-old man with prostate cancer and abnormal corresponding bone scan (not shown). There is a low signal intensity lesion (sequestrum) surrounded by a hyperintense signal intensity rim (arrowheads).

 

Figure 4A

 

Figure 4B

 

Figure 4. (A) T1-weighted (TR: 500, TE: 14) axial MR image without fat saturation shows a hypointense left femoral lesion (thick arrows) at the level of the greater trochanter. The surrounding rim of abnormal signal is not distinguishable from low signal marrow (thin arrows) at the posterior aspect of the lesion. (B) Gadolinium-enhanced T1-weighted (TR: 600, TE: 18) axial MR image with fat saturation corresponding to Figure 6A shows enhancement of metastatic focus and posteriorly located marrow, also involved with the metastatic disease. Note the distinct hypointense signal unenhancing rim (arrowheads) is better appreciated on this sequence.