|
|
||||||||
Clinical Observations |
1 Department of Imaging, Royal Marsden Hospital, Downs Rd., Sutton, Surrey, SM2
5PT, United Kingdom.
2 Department of Histopathology, Royal Marsden Hospital, London, United
Kingdom.
3 Department of Surgery, Royal Marsden Hospital, London, United Kingdom.
Received July 23, 2006;
accepted after revision March 26, 2007.
Address correspondence to S. D. Allen
(steven.allen{at}rmh.nhs.uk).
Abstract
|
|
|---|
CONCLUSION. Although adult rhabdomyosarcomas have certain imaging appearances in common with other soft-tissue sarcomas, features at presentation such as tumor heterogeneity, site, regional lymphadenopathy, and pulmonary metastasis should make the radiologist consider this important diagnosis.
Keywords: cross-sectional imaging CT imaging-histopathologic correlation MRI rhabdomyosarcoma
|
|
|---|
The different subtypes of rhabdomyosarcoma are embryonal, alveolar, and pleomorphic (Figs. 1A, 1B, 1C, 1D, 2A, 2B, 2C, 2D, 3A, 3B, 3C, 3D). Overall, the embryonal subtype is the most common subtype, accounting for up to 49% of all rhabdomyosarcomas [3]. It arises most frequently in the head and neck and most often occurs during the first decade of life. Some embryonal rhabdomyosarcomas exhibit pure spindle cells at histology and have been termed "spindle cell rhabdomyosarcomas" [4, 5]. The alveolar subtype accounts for approximately 30% of all rhabdomyosarcomas and most commonly affects adolescents; however, the age range of patients with this subtype is 10-25 years [3]. It is most commonly an intramuscular tumor in the soft tissues of the extremities [2]. The pleomorphic subtype is the least common subtype and almost exclusively affects patients older than 45 years. It arises most commonly in the skeletal muscles of the thigh. It is histologically similar to a malignant fibrous histiocytoma; in fact, many pleomorphic rhabdomyosarcomas have been reclassified as fibrous histiocytomas, making the diagnosis of pleomorphic rhabdomyosarcoma an even rarer entity [2]. These two entities to our knowledge have not been compared with or differentiated from one another in any specific imaging study, and distinction is based on immunohistochemical findings.
|
|
|
|
|
|
|
|
|
|
|
|
For this study, we reviewed the clinicoradiologic features of rhabdomyosarcomas in adults from the Royal Marsden Sarcoma Unit, which has one of the largest databases of soft-tissue tumors worldwide. Our results are compared with the imaging data for pediatric and adult rhabdomyosarcomas, although the imaging data for adults are limited to case reports and small case series [6, 7]. MRI and CT parameters were assessed to characterize this tumor and its subtypes, and each case was correlated with histopathology results.
|
|
|---|
18 years) with biopsy-proven
rhabdomyosarcoma presented to the Royal Marsden Hospital Sarcoma Unit over a
5-year period between January 2001 and December 2005. Fourteen patients with
inadequate imaging records and five patients with recurrent disease were
excluded. Another four patients were excluded because of inadequate
histopathologic records. Thus, the final study population was composed of 26
patients with adequate pretreatment MR images (n = 16) or CT scans
(n = 10) and with histopathology results for retrospective
correlation with imaging findings. A large number of the patients in our study
group had been referred to our tertiary center for management, so they often
brought images obtained at outside institutions with them. Hence, we were
often not able to influence the technique or protocols of these studies.
Patient demographics such as age at presentation and sex were obtained from the hospital's electronic patient-record system. Features of the tumor recorded included site; size; margin; local extent; and presence of calcification, hemorrhage, or necrosis. Sites of metastatic disease were also noted. CT and MRI signal intensity and heterogeneity were recorded. Tumor size was recorded from the images as the mean of three orthogonal measurements. Two dedicated sarcoma radiologists reviewed each MRI and CT examination, with findings reached by consensus. The data were analyzed for the entire patient group and also for the three subgroups organized by tumor subtype.
Histopathologic specimens were analyzed for clear resection margins, degree of necrosis, lymphovascular invasion, number of mitoses, and overall tumor grade.
|
|
|---|
|
The imaging findings of the tumors organized by the different subtypes are summarized in Table 2, and it should be noted that active hemorrhage, recent hemorrhage, and calcification were not findings in any of the cases. Seven patients had metastases at presentation with the most common site being the lungs (six patients).
|
On T1-weighted MRI, 13 of 16 rhabdomyosarcomas (81%) were isointense to skeletal muscle, and three (19%) were slightly hyperintense. Ten rhabdomyosarcomas (63%) returned a homogeneous signal, and six (38%) returned a slightly heterogeneous signal.
On T2-weighted MRI, 11 of 16 rhabdomyosarcomas (69%) were moderately hyperintense to skeletal muscle, and four (32%) were very hyperintense. Ten rhabdomyosarcomas (63%) returned a moderately heterogeneous signal, and six (38%) returned a very heterogeneous signal.
A STIR sequence was included in the examination of only seven patients. In all seven patients, the rhabdomyosarcomas were heterogeneous and hyperintense, six of which were very hyperintense.
After injection of IV gadolinium, all tumors showed either moderate (n =8) or avid (n = 8) heterogeneous enhancement.
On CT, the tumors were either isodense (n = 7) or slightly hyperdense (n = 3) to skeletal muscle. Four cases were homogeneous; three, slightly heterogeneous; and three, very heterogeneous.
Twelve patients had primary surgical treatment (limb, n = 5; pelvis, n = 4; skull base, n = 3), eight of whom had a clear resection margin at pathology. Three patients had macroscopic margin involvement, and one had microscopic involvement. Histologic analysis of a radiologic or surgical biopsy specimen or of a surgical resection specimen was performed in all 26 cases. However, if only biopsy samples were available, the degree of pathologic information available was often restricted. The histologic characteristics of the rhabdomyosarcomas in the 26 study patients are summarized in Table 3.
|
Histologic analysis of the embryonal subtype revealed variable proportions of loose myxoid foci containing primitive cells and regions of dense spindle cells. No lymphovascular invasion was identified, only scant necrosis. The tumors were various grades, but all showed a high mitotic rate. Alveolar rhabdomyosarcomas were composed of large round or oval cells separated by fibrous septa into alveoli. Central cells were frequently loose and necrotic. Lymphovascular invasion was infrequently identified, although necrosis was common, and tumors were intermediate or high grade with a high mitotic rate. Pleomorphic rhabdomyosarcomas were highly cellular and were composed of irregularly arranged round and pleomorphic cells with bizarre mitoses. Lymphovascular invasion was frequently identified, necrosis was present in all cases, and tumors were invariably high grade with a high mitotic rate.
Disease has recurred to date in 16 patients (62%), with a mean time from imaging to recurrence of 8.25 months. Fourteen patients have died, with a mean time from diagnosis to death of 8.3 months. Seven of the patients treated with primary surgery underwent follow-up for 2 years after surgery; in six of those seven patients, tumors recurred at a mean time of 9.3 months after surgery.
Sixteen patients were treated with chemotherapy, and 17 were treated with radiation therapy.
|
|
|---|
The principal role of imaging in the management of rhabdomyosarcoma in adults is preoperative planning and detection of recurrence or of disease progression in nonsurgically managed patients. In terms of surgical staging, imaging findings are of vital importance because complete surgical resection bears the most favorable prognostic outcome [8]. However, metastases are frequent, being present in up to 44% of cases at diagnosis [9-12].
Sites of metastasis in adult rhabdomyosarcoma are, in decreasing order of frequency, the lungs, the bone marrow, and the lymph nodes [2]. The presence of metastases at presentation along with tumor size, tumor resectability, and patient age at presentation are also prognostic factors. Histologic subtype also has a bearing on prognosis: Embryonal tumors are usually associated with the most favorable outcome and pleomorphic, the least favorable.
Treatment of patients with rhabdomyosarcoma is usually primarily surgery if excision is attainable, followed by radiation therapy and chemotherapy. The prognosis for patients with rhabdomyosarcoma has been typically very poor, with a 5-year survival rate of less than 50% even in the most recent series [10-12]. However, this survival rate is improved compared with the rates reported in older series, which were as low as 21% [13]. Nevertheless, such figures should be viewed with caution because all of these data are from retrospective series and were collected over long time periods and because treatment regimens have clearly evolved during this time.
Rhabdomyosarcomas are common solid tumors in children, and affected children tend to have a better prognosis than adults with rhabdomyosarcoma [14]. In addition, rhabdomyosarcomas occur at different sites in pediatric patients compared with adults. The embryonal subtype is the predominant subtype in children with rhabdomyosarcoma and is most commonly in head and neck sites, followed by the pelvis. This subtype was the least frequent one in our series, although all cases were within the head, neck, or pelvis. The alveolar subtype in children occurs most commonly in extremity sites [2], whereas the adults in our series with the alveolar subtype had tumors that were in the head or neck. The pleomorphic subtype of rhabdomyosarcoma is exceedingly rare in children and is also most commonly sited in the extremities [2]. This site was a common one for this subtype in our series, along with the trunk.
Lymphadenopathy varies in frequency in children with rhabdomyosarcoma and is dependent on the organ of origin of the primary tumor. Retroperitoneal nodal disease with pelvic tumors is most common but is present in fewer than 20% of cases at presentation [1, 15]. In adults with rhabdomyosarcoma, lymphadenopathy appears to be more common: Specifically, 46% of patients had lymph node involvement at diagnosis in one series [11]. Our series confirms this increased prevalence of lymphadenopathy in the adult form, and although our data did not show a specific predilection for a primary tumor site, the head and neck were the most common sites. Lymphadenopathy was most prominent in alveolar tumor subtypes and was least common in extremity tumors and pleomorphic tumor subtypes. Pulmonary metastases have been reported to be the most frequent site of distant disease [16], and this was also the case in our study.
In children with rhabdomyosarcoma, calcification is rare; similarly, calcification was not present radiologically or pathologically in any of the patients in our series. Other series of adults with rhabdomyosarcoma have also confirmed calcification to be a rare finding [7]. Necrosis was frequently observed in our study patients, and this finding is consistent with those of other studies [7]. Bone destruction was infrequently identified in our series and was seen exclusively in head and neck sites, as has been reported in previous studies [17].
At MRI, rhabdomyosarcomas typically show nonspecific low signal intensity on T1-weighted pulse sequences and high signal intensity on T2-weighted sequences. Heterogeneous signal on T2-weighted imaging is common and has been reported previously to result from hemorrhage [15]. None of the cases in our series had evidence of recent or active hemorrhage either radiologically or histopathologically. In a previous study, investigators identified a pseudocapsule as a low-signal-intensity outer border to the lesion that was most conspicuous on T2-weighted sequences [18]. We did not identify this feature in any of the cases in our series. Consistent with the CT appearance of the tumors in our series, pediatric rhabdomyosarcomas are typically bulky masses with varying degrees of heterogeneous attenuation on CT.
Adult rhabdomyosarcomas have certain imaging appearances in common with a number of other soft-tissue sarcomas (i.e., large mass, irregular enhancement, and necrosis), although because lymphadenopathy is so frequently observed, aggressive lymphoma is in the differential diagnosis. Differential diagnoses are also site-specific, although the exact site of origin is often difficult to determine. For example, primary prostate or vaginal lesions invading the bladder base can be indistinguishable from a primary bladder base tumor. In general terms, a differential diagnosis based on imaging is difficult, and radiologists should, instead, initially incorporate findings at clinical presentation; patient age; site of disease; tumor heterogeneity and signal; presence of lymph node and metastatic disease; and, if the tumor is at the head or neck, whether bone destruction is present. Pathology is required in all cases to make a definitive diagnosis.
Lee et al. [7] evaluated 11 adult patients with head and neck rhabdomyosarcomas, most of which were the embryonal subtype. On CT, these lesions were mainly homogeneous and had ill-defined margins [7]. Our study confirmed that lesions at this site had ill-defined margins and were often homogeneous, especially those that were the embryonal subtype. Otherwise, to our knowledge, no previous imaging study has compared and differentiated the imaging features of the subtypes of adult rhabdomyosarcomas with pathologic correlation.
Although our study numbers are relatively small, certain features of rhabdomyosarcomas in adults suggest that differentiation among the subtypes is possible. Specifically, embryonal tumors were usually composed of sheets of spindle cells with various proportions of loose myxoid foci containing primitive cells. They closely resemble various stages in the embryogenesis of skeletal muscle but can be poorly differentiated and difficult to diagnose [2]. In our series, on imaging, tumors of this subtype tended to be the most homogeneous and isodense to skeletal muscle, with low- or intermediate-grade tumors possibly reflecting uniform sheets of spindle cells that resemble fetal muscle. However, with only four embryonal subtype tumors included in our series, we cannot derive any conclusions because this number is too small.
In general terms, the lack of identifiable lymphovascular invasion and variable grade cannot be directly correlated to imaging but may, in part, explain that three of four cases were suitable for primary surgery, a ratio much higher than for any other group. On imaging, necrosis was identified in two of four cases, but necrosis was detected in only one case at pathology. We may have interpreted an area of loose myxoid foci as necrosis on imaging, but this theory cannot be confirmed.
We found that alveolar rhabdomyosarcomas were composed of large round or oval cells separated by fibrous septa into alveoli. Central cells were frequently loose and necrotic. These loose central cells may explain the frequent heterogeneity and necrosis observed on imaging.
Pleomorphic rhabdomyosarcomas were highly cellular and were composed of irregularly arranged round and pleomorphic cells with bizarre mitoses. Such irregular, haphazard cell arrangement may explain the frequent heterogeneity on imaging. Necrosis was present in all cases histologically and in all but one on imaging, presumably reflecting the highly aggressive nature of this subtype. In addition, pleomorphic subtype tumors were generally of very high signal on both T2-weighted and STIR sequences. We are unsure whether there is a histopathologic explanation for this distinguishing feature but, nonetheless, it is of interest. Again, the higher prevalence of lymphovascular invasion and the higher tumor grade seen in the alveolar and pleomorphic subtypes likely reflect the more aggressive nature of and worse prognosis associated with these subtypes, but these findings cannot be directly correlated to the imaging findings.
Histologically, all rhabdomyosarcomas in our series had a very high mitotic rate, which explains their aggressive nature and the poor prognosis.
A history of local radiation therapy is an interesting factor to consider in adults with rhabdomyosarcoma. Only two patients in our series had this relevant history, although both patients had tumors that were small at presentation (< 5 cm) and were completely surgically resected, and both were alive 2 years after surgery. It was not possible given the low number of cases to distinguish specific imaging features in this group.
To our knowledge, this study is the largest imaging study to date of adult rhabdomyosarcoma with pathologic correlation. This study provides new information about the imaging appearances of these tumors—in particular, about differences between adult and pediatric rhabdomyosarcomas—and unique data about the appearances of the different cell subtypes. The larger previously reported case series of adult patients have included almost exclusively tumors in head and neck sites and have not been sufficient in number to comment on differences among subtypes on imaging [7].
Our results have shown tumor heterogeneity to be prominent in alveolar and pleomorphic subtypes and extremely high signal on T2 and STIR MRI to be very prominent in the pleomorphic subtype. Moreover, on the basis of our results, we have identified features associated with certain subtypes of rhabdomyosarcomas or with rhabdomyosarcomas in general such as tumor site; tumor heterogeneity; and the presence of necrosis, associated regional lymphadenopathy, and pulmonary metastasis at presentation. Despite its rarity, it is important to recognize the radiologic appearances of adult rhabdomyosarcoma because prompt diagnosis and staging of this tumor are imperative so that the narrow curative treatment window may be exploited.
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |