| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Musculoskeletal Imaging |
1 Department of Radiology, Mayo Clinic, 200 1st St. SW, Rochester, MN
55902.
2 Present address: Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH
44195.
3 Department of Oncology, Mayo Clinic, Rochester, MN 55902.
4 Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN 55902.
Received November 11, 2003; accepted after revision March 1, 2004.
Address correspondence to R. P. Hartman.
OBJECTIVE. Our goal was to determine the incidence and pattern of red marrow reconversion on MRI of adults receiving granulocyte-stimulating factors as part of their chemotherapy regimen for primary musculoskeletal neoplasms and correlate the changes with WBC counts.
MATERIALS AND METHODS. Twenty-five adults with soft-tissue sarcomas (n = 15) or primary bone tumors (n = 10) who underwent chemotherapy that included granulocytestimulating factors formed the study group. Two radiologists retrospectively evaluated the MRI studies by consensus before and after therapy to determine the presence of changes consistent with red marrow reconversion. Changes were categorized by appearance on T1- and T2-weighted images for location, pattern, and extent of marrow involvement. WBC counts at the time of MRI were recorded. Records were examined for evidence of marrow-infiltrating tumors or metastases.
RESULTS. Ten (40%) of the 25 patients underwent bone marrow changes consistent with red marrow reconversion. In seven (70%) of 10 patients, the changes were diffuse in visualized bones and simulated diffuse marrow-infiltrative tumor. In three (30%) of 10 patients, the changes were focal, simulating metastases. T2 signal prolongation was identified in the marrow reconversion in nine patients (90%), although all had shortened T1 signal. Four patients (40%) had elevated WBC counts at the time of the MRI after therapy, but all had shortened T1 signal. Four patients (40%) had elevated WBC counts at the time of the MRI after therapy, five (50%) had normal counts, and one (10%) had a below-normal count. No osseous metastases or marrow-infiltrating tumors were found during follow-up.
CONCLUSION. Forty percent of patients showed marrow changes mimicking tumor on MRI that were attributable to red marrow reconversion, which correlated moderately with WBC response.
Recombinant human granulocyte macrophage colony-stimulating factor (GMCSF) (Leukine, Berlex) and granulocyte colony-stimulating factor (GCSF) (Neupogen, Amgen) are granulocyte-stimulating factors (GSF) that are increasingly used as an adjunct to chemotherapy for a number of malignant tumors, including primary malignant musculoskeletal neoplasms. These adjuncts have been shown to moderate the degree of neutropenia occurring because of chemotherapy and decrease the duration of hospitalization, the amount of antibiotics used, and episodes of febrile illness [1]. This improvement allows high-dose chemotherapy to proceed and more patients to complete their prescribed chemotherapy course.
Red marrow reconversion in pediatric patients receiving GCSF and the associated MRI findings as a consequence of GCSF therapy have previously been reported in two studies describing 20 children with primary malignant musculoskeletal tumors [2, 3]. To our knowledge, the only report on adults with a primary malignant musculoskeletal tumor in the English language consists of a single case report [4]. The focus of our study was to determine the incidence of adult patients with red marrow reconversion detectable on MRI after the use of GSF as part of the treatment for musculoskeletal malignancies, to identify patterns of distribution, and to compare WBC counts with MRI findings.
Materials and Methods
Institutional review board approval was obtained to perform a chart review.
Written patient consent was obtained for all subjects included in the study. A
retrospective chart review from our institution was performed to identify
adult patients (age 
18 years) who had received GSF as an adjunct to their
chemotherapy for primary musculoskeletal tumors. A subset of 25 patients who
underwent MRI before and after GSF administration was identified and
constituted our study group. The patients ranged in age from 18 to 73 years
(mean, 40 years) and included 17 men and 8 women.
Each patient's MRI examination was performed on a 1.5-T Signa scanner (GE Healthcare) and included at least T1-weighted spin-echo and T2-weighted spin-echo or fast spin-echo (FSE) images for review. All 25 patients had T1-weighted spin-echo images obtained without fat suppression or gadolinium enhancement. In addition, four patients had gadolinium-enhanced T1-weighted spin-echo images, all with fat suppression. Thirteen had T2-weighted spin-echo images, 10 had T2-weighted FSE images, and one had both. All T2-weighted FSE and conventional spin-echo sequences were obtained with fat saturation. Two patients had T2-weighted FSE inversion recovery imaging as part of their examinations. Some variation was found with respect to the imaging parameters for the patients, but in general, the T1-weighted spin-echo images were produced with a TR range of 400–650 and a TE range of 14–15. T2-weighted conventional spin-echo images had a TR range of 2,000–3,000 and a TE of 60. T2-weighted FSE images had a TR range of 3,000–3,366 and a TE range of 75–92. The slice thickness varied from 5 to 10 mm with a 2.5- to 5-mm interslice gap.
In 14 patients, the prechemotherapy MRI and postchemotherapy MRI sequences were the same. In three patients, the examinations were the same with the exception of additional gadolinium-enhanced sequences on the prechemotherapy examinations. In one patient, gadolinium was used only on the postchemotherapy examination. In four patients, the prechemotherapy T2-weighted images were achieved with spin-echo sequences, and the postchemotherapy examination used FSE sequences. Two patients had FSE inversion recovery imaging on the prechemotherapy examination and FSE imaging on the postchemotherapy examination. One patient had FSE imaging on the prechemotherapy examination, with spin-echo T2-weighted imaging on the postchemotherapy examination. The imaged body parts included thigh (n = 12), pelvis (n = 2), shoulder (n = 2), arm (n = 6), wrist (n = 1), abdomen (n = 1), and both the shoulder and pelvis (n = 1). In 18 cases, only the body part that was the site of the primary malignancy was imaged; in seven other cases, images of the contralateral limbs were available for review. The initial MRI was performed before the commencement of chemotherapy. The second MRI to restage the primary musculoskeletal tumor followed the second course of chemotherapy with GSF. None of the patients received radiation therapy before the post-GSF MRI. The chemotherapy protocols were determined by the histologic diagnosis of the tumor. GSF, however, was administered to all patients the day after beginning chemotherapy and was continued on a daily basis until the absolute neutrophil count reached 750/mL. The dose was 5 mg/kg of body weight rounded to 300–480 mg daily. No patients received intraarterial chemotherapy.
The MRI series were evaluated by consensus of two radiologists for the presence of marrow signal changes consistent with red marrow reconversion. This change was defined as a decrease in the T1 intensity of the marrow on the post-GSF images in which healthy yellow marrow had been identified on the pre-GSF examination (Figs. 1A, 1B, 1C, and 1D). This same area was evaluated on the T2-weighted series for signs of signal changes. The T2-weighted signal intensity was considered mild if it was slightly brighter than muscle, moderate if it was significantly brighter than muscle but not as bright as fluid, and marked if it equaled or exceeded that of fluid.
|
|
|
|
The changes in patients who showed changes were further classified by location in the skeleton and extent of change (focal or diffuse) in the bone. Changes were classified as focal if the area of abnormal signal was small and ovoid or rounded with surrounding healthy marrow signal. Changes were classified as diffuse if the area of abnormal signal was occupying a significant portion of the marrow space and replacing healthy marrow signal over several centimeters in a linear or transverse orientation. Juxtaposition of red marrow reconversion and a primary bone tumor (i.e., Ewing's sarcoma) in a long bone in one patient who subsequently under-went surgery permitted MRI findings to be compared with the gross specimen.
The WBC counts of the patients were recorded before the initiation of GSF and at the time of the post-GSF MRI. The time interval from the last dose of GSF to the follow-up MRI was also recorded. In 20 of the 25 patients, the WBC counts were determined from blood drawn on the day of the follow-up MRI. For other patients, the WBC counts were determined from blood drawn at different times: for one patient it was 10 days before, for one it was 1 day before, for one it was 1 day after, for one it was 4 days after, and for one it was 7 days after the respective follow-up MRI. Nuclear medicine bone scans after the initiation of GSF were available for review in six patients. One of these patients had an FDG PET scan as well.
After imaging evaluation, a chart review was performed to ensure that none of the patients had developed metastases or marrow-infiltrative tumors. The follow-up period ranged from 1 to 65 months (mean, 20.4 months).
Results
The types of primary malignant musculoskeletal tumors in the 25 patients were as follows: Among the soft-tissue sarcomas were three leiomyosarcomas (two thighs and one arm) (12%), two malignant fibrous histiocytomas (one thigh and one arm) (8%), two pleomorphic sarcomas (thigh) (8%), two fibrosarcomas (thigh) (8%), two synovial sarcomas (one shoulder and one arm) (8%), two malignant nerve sheath tumors (one pelvis and one arm) (8%), one embryonal sarcoma (abdomen) (4%), and one sarcoma not otherwise specified (thigh) (4%).
The primary bone tumors were the following: eight Ewing's sarcomas (two femur, two humerus, one shoulder, one pelvis, and one wrist; one patient had shoulder and pelvic tumors) (32%), and two osteosarcomas (femur) (8%). The anatomic sites imaged corresponded to the location of the primary tumors and included 12 thighs, four forearms, three arms, three shoulders, two pelves, and one abdomen.
Findings in the marrow on the post-GSF MRI scans show that 10 patients (40%) of this group had red marrow reconversion. None of the patients were found to have skeletal metastases, and none of the patients with soft-tissue sarcomas had infiltrative tumors at the time of surgery or follow-up imaging and chart review. The patient who only had 1 month of follow-up died from sepsis.
The distribution of soft-tissue sarcomas present in this subset of 10 patients were two leiomyosarcomas (thighs) (20%), two malignant fibrous histiocytomas (one thigh and one arm) (20%), two pleomorphic sarcomas not otherwise subclassified (thighs) (20%), and one malignant nerve sheath tumor (arm) (10%). The primary bone tumors in this group were the following: three Ewing's sarcomas (pelvis, femur, and one patient with shoulder and pelvic tumors) (30%). The patients ranged in age from 21 to 73 years (mean, 44 years) and included seven men and three women.
The changes in the marrow occurred in the proximal segments of long bones or in the pelves in all cases. No T2 signal change was evident in one patient who had T1 changes (Figs. 2A, 2B, 2C, 2D, 2E, and 2F). The T2 signal intensity of the marrow changes in these 10 patients were mild in eight (80%) (Figs. 3A, 3B, 3C, and 3D), moderate in one (10%) (Figs. 4A, 4B, 4C, 4D, and 4E), and marked in none.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The changes in the bone were graded as diffuse in seven cases (70%) (Figs. 2A, 2B, 2C, 2D, 2E, and 2F) and focal in three cases (30%) (Figs. 3A, 3B, 3C, and 3D). The diffuse changes were seen in three Ewing's sarcomas, two leiomyosarcomas, one pleomorphic sarcoma, and one malignant fibrous histiocytoma. The focal changes were seen in one malignant nerve sheath tumor, one pleomorphic sarcoma, and one malignant fibrous histiocytoma. Nine (90%) of the 10 areas of red marrow reconversion exhibited increased intensity on the T2-weighted images. In the three patients with a primary bone malignancy, the marrow reconversion changes were contiguous with the bone tumor but showed different signal characteristics from the malignancy (Figs. 4A, 4B, 4C, 4D, and 4E). No abnormal enhancement in the bone marrow was seen on the gadolinium-enhanced images except enhancement attributable to the primary bone tumor.
Correlation with the WBC count at the time of the post-GSF MRI showed an elevation in the WBC count in four (40%) of the 10 patients with red marrow reconversion. Three of these patients had markedly elevated counts. Five (50%) of the remaining patients had counts in the normal range, with only one (10%) having a WBC count below normal. Only three (20%) of the 15 patients without red marrow reconversion had elevated WBC counts at the time of the post-GSF MRI. Eleven (73%) of 15 had counts in the normal range, and one (7%) had a below-normal count.
The time interval after the last GSF dose to the follow-up MRI in the group with red marrow reconversion was 0–42 days (mean, 9 days). In the group without marrow reconversion, the range was 0–280 days (mean, 40 days).
One patient with a primary malignant bone tumor underwent resection after chemotherapy. In this case, the primary bone tumor showed significant necrosis, and no tumor was found in the nonfatty marrow contiguous areas, which were accurately interpreted as sites of red marrow reconversion (Figs. 5A, 5B, 5C, 5D, 5E, 5F, 5G, 5H, and 5I).
|
|
|
|
|
|
|
|
|
One patient had a PET scan in addition to the follow-up MRI that showed no metabolic activity at the site of the primary bone tumor, but exhibited diffuse uptake throughout the rest of the visualized skeleton consistent with red marrow reconversion (Fig. 4E). Findings from patients with nuclear medicine bone scans were all interpreted as negative for bone metastases.
Discussion
GSF, in the form of GMCSF and GCSF, is becoming a widely used adjuvant to chemotherapy for the treatment of many musculoskeletal malignancies because it decreases the incidence of neutropenic events and infections and reduces the number and length of hospitalizations and duration of antibiotic therapy [1].
GSF primarily affects cells of myeloid lineage in the marrow, with its primary role of regulating the proliferation and differentiation of hematopoietic progenitor cells. In addition, GSF may enhance the function of mature blood cells [5]. Administering GSF to patients undergoing chemotherapy stimulates the marrow to produce more WBC and usually prevents neutropenia.
Red marrow reconversion in pediatric patients receiving GCSF has been reported [6]. Altehoefer et al. [7] have reported on the MRI characteristics seen in healthy blood stem cell donors receiving GCSF. However, only one case report [4] describes the effects in the adult population receiving GCSF with chemotherapy for a primary musculoskeletal malignancy.
As the healthy human skeleton matures, a red–yellow marrow conversion begins in childhood and is usually completed by age 25 [8]. In general, red–yellow marrow conversion proceeds from distal to proximal areas in the limbs. In adults the largest areas of red marrow remain in the vertebrae, pelvis, ribs, and sternum, with visible red marrow in the proximal shafts of the femora and humeri [9]. In patients experiencing red marrow reconversion, the sites in which red marrow first appears are those areas that last converted to yellow marrow, and this process then proceeds in reverse physiologic order [9].
Red marrow exhibits T1 and T2 relaxation times that are different from those of yellow marrow. In practice, these areas of red marrow are darker than those of healthy yellow marrow on T1-weighted scans and may have a slightly increased signal intensity on T2-weighted scans [8]. These findings may mimic changes caused by bone metastases and infiltrative diseases of the marrow. Differentiating these without knowledge of the effect of GSF can be difficult.
In our study, 40% of the patients with MRI scans available before and after GSF showed changes consistent with red marrow reconversion. These results agree with those of previous studies examining the MRI appearance of red marrow reconversion. Areas of red marrow showed decreased T1 relaxation times resulting in areas of decreased T1 signal on T1-weighted scans [2, 3, 7, 8, 10]. Slightly more than half of the patients in our study exhibited higher T2 signal in the areas of reconversion. Unlike previous studies, which used STIR as a fluid-sensitive sequence, most (7/10) of our patients had fat-suppressed T2-weighted FSE sequences on the postchemotherapy images. The red marrow reconversion occurred primarily in the pelvis or proximal long bones, consistent with previous reports that showed that the areas that first undergo red marrow reconversion in adults are those areas that were the last to undergo fatty conversion. The signal of healthy marrow most closely approximates the signal of fat. Areas of red marrow reconversion have lower T1 signal intensity that approximates the signal characteristics of muscle. In our practice, fat-suppressed T2-weighted FSE sequence is preferred to STIR. Because of the small number of patients with STIR sequences, no attempt was made to compare the merits of one technique over the other.
The WBC count (normal range, 3.5–10.5 x 109/L) at the time of the post-GSF MRI was normal in half (5/10) of the patients with red marrow reconversion. However, the counts were elevated, sometimes markedly, in 40% of those with red marrow reconversion. Even those patients with WBC counts in the healthy range were experiencing some stimulative effect from the GSF, because they would normally be expected to be neutropenic as a result of their chemotherapy. In contrast, only two patients in the study group with elevated WBC counts at the time of the post-GSF MRI did not show changes of red marrow reconversion. One of these patients had been treated for an embryonal cell sarcoma, and only abdominal images were obtained. The bones visualized in this patient were limited to parts of the spine and ribs, where red marrow is normally found in an adult. Reconversion changes may not have been appreciated because the pelvis, femora, or humeri were not included. It is conceivable that our estimation of marrow reconversion is on the low side because MRI was limited to the anatomic site of the primary malignancy. The only PET scan in our study (Fig. 4E) showed disseminated skeletal activity consistent with red marrow reconversion.
The challenge for a radiologist seeing follow-up MRI scans in a patient being treated with GSF and chemotherapy for a known musculoskeletal malignancy is the differentiation of bone metastases or pathologic marrow infiltration from benign physiologic red marrow reconversion. On the basis of our findings, abnormally low T1 intensity in the bones of these patients most commonly represents areas of red marrow reconversion, especially when it is encountered in the proximal long bones and pelvis. Focal low T1 intensity could be mistaken for metastasis. When low T1 signal intensity is diffuse and contiguous with a primary bone tumor, it could be misinterpreted as extension and growth of the primary malignancy. If it is diffuse and at a site removed from the primary malignancy, low T1 signal intensity could be mistaken for a new infiltrative osseous malignancy in patients with known soft-tissue sarcoma or mistaken for dissemination of disease in patients with a primary bone malignancy. The location of these signal changes, the presence of an elevated WBC count, and mostly awareness of the changes in marrow that can be caused by GMCSF and GCSF should permit appropriate interpretation of the MRI findings. In eight of our 10 patients with red marrow reconversion, the signal intensity on T2-weighted sequences was mild relative to that of muscle. The distinction between the mild increased signal intensity of reconversion from the moderate to marked hyperintensity of tumor with or without necrosis (Figs. 4A, 4B, 4C, 4D, and 4E) was an aid in distinguishing between the two entities. When these signal intensities are confluent (Figs. 5A, 5B, 5C, 5D, 5E, 5F, 5G, 5H, and 5I), a distinction must be made preoperatively when planning limb-sparing surgery. In our only case in which the T2 prolongation was judged to be moderate in intensity, edema from the nearby biopsied lesion may have contributed to the increased T2 prolongation.
We used FDG PET scanning for confirmation of physiologic reconversion in only one patient whose T2 signal changes were significantly brighter over a larger area than expected. The FDG PET findings were negative in this area but showed diffuse increased uptake in the rest of the skeleton, as has been reported in patients undergoing treatment with GSF [11, 12].
The one patient who had a technetium bone scan in the presence of MRI-confirmed red marrow reconversion showed no abnormal uptake in the regions of red marrow reconversion.
The differentiation of tumor from red marrow reconversion is essential because the course of management differs greatly. Surgical excision can still occur in patients experiencing red marrow reconversion; alternatively, progression of tumor or metastases can alter management and therapeutic approaches and could preclude definitive therapy. In determining the source of signal changes in these patients, one should evaluate the pre-GSF and post-GSF images in combination. The pre- and post-GSF scans should be obtained with parameters as closely matched as possible to aid the comparison. If the radiologist or clinician is uncertain whether the scans show red marrow reconversion or tumor, it would be preferable to reimage the area after a short interval rather than biopsy—especially if the changes are in the expected sites for marrow reconversion and a corresponding WBC response is present. If, however, the MRI findings and clinical picture remain confounding, MRI of the femora (if bone abnormality is found in another site) or the contralateral femur (if abnormality is in the femur) may resolve the problem.
Red marrow reconversion in patients receiving GSF with their chemotherapy for primary musculoskeletal tumors reflects a desired clinical response and appears to correlate moderately well with the WBC count. The use of hematopoietic growth factors would appear to be a part of the armamentarium in general medical oncologic practice for a broad range of malignancies and is not confined to tertiary orthopedic oncologic centers or is not only for musculoskeletal malignancies. We believe that all radiologists interpreting MRI studies of patients with known primary malignant musculoskeletal tumors should be familiar with its effect on bone marrow and the spectrum of MRI appearances.
Limitations of our study include its retrospective nature, lack of a control group, and interpretation of images by consensus rather than by blinded review. Histologic proof of red marrow reconversion was only available in one patient, and that was only because the changes were juxtaposed to a resected bone tumor (Figs. 5A, 5B, 5C, 5D, 5E, 5F, 5G, 5H, and 5I). Because most patients (7/10) with marrow changes caused by GSF had soft-tissue sarcomas, a bone biopsy would have been required for histologic proof of the benign nature of the osseous changes. These patients did not undergo biopsy because a watchful clinical approach was favored on the basis of each individual patient's clinical evaluation and course at the time of the MRI findings. Our retrospective study would appear to confirm the appropriateness of that approach, however dramatic the MRI findings.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. S. Steinbach "MRI in the Detection of Malignant Infiltration of Bone Marrow"--A Commentary Am. J. Roentgenol., June 1, 2007; 188(6): 1443 - 1445. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
