Downloaded from www.ajronline.org by 58.220.253.162 on 10/21/15 from IP address 58.220.253.162. Copyright ARRS. For personal use only; all rights reserved
109
Metastases Melanoma CT Study
Randall M. Patten1 William P. Shuman2 Sharlene Teefe?
from Malignant to the Axial Skeleton:
of Frequency
A
and Appearance
Radiologic detection of bone lesions from malignant melanoma is reported to be uncommon or infrequent. To ascertain the characteristics and frequency of detection of melanoma metastases to the axial skeleton by CT, we retrospectively reviewed 464 body CT studies of 125 consecutive melanoma patients for presence, appearance, and site of skeletal metastases. Results were correlated with patient’s age, sex, clinical course, and both the Clark and Breslow classifications of the primary lesion. Of 98 patients with metastatic disease evident on their CT studies, 17 (17%) had bony metastases; two (12%) of these 17 patients had skeletal lesions as the only CT evidence of metastatic disease. Metastatic bony lesions were predominantly osteolytic, slightly expansile, and commonly located in the spine. Associated soft-tissue masses were frequent, but periosteal reaction and identifiable tumor matrix were not seen. Skeletal metastases were found only in those patients with thick or intermediate primary melanom&(Breslow) classified as Clark level III or greater, and the CT demonstration of osseous metastases was a poor prognostic sign. The data suggest that CT detection of skeletal melanoma metastases is not uncommon. When CT is performed to evaluate for metastatic melanoma, the axial skeleton should be carefully examined, especially in those patients with more advanced primary lesions. AJR 155:109-112,
July
1990
Plain film detection of bony lesions from primary melanoma has been reported to be uncommon or infrequent [1 2]. However, skeletal metastases from malignant ,
melanoma are found during autopsy in 23-49% of patients who die of the disease [3, 4]. In many patients with melanoma, CT is used to stage the primary lesion and to monitor
therapeutic
response,
yet the
CT appearance of osseous melanoma to determine the frequency of
metastases has not been well described. In order CT detection of metastases to the axial skeleton reviewed
CT studies
We wished metastases Received January 4, 1990: accepted after revision February 12, 1990. 1 Department of Radiology(ZA-65), University of
Washington, Harborview Medical Center, 325 Ninth Ave., Seattle, WA 981 04. Address reprint requests to R. M. Patten. 2 Department of Radiology(SB-05), University of Washington,
St., 3
University
Hospital,
WA 98195. Department of Radiology
1959 N. E. Pacific
Seattle,
Administration
Medical
Center,
(ZB-20), Seattle,
0361-803X/90/1551-0109 © American Roentgen Ray Society
Veterans WA 98108.
wished
performed
to determine was greater
to determine
melanoma (according predict the occurrence Materials
on 1 25 consecutive
and their CT appearance, patients
with known
we
melanoma.
whether the frequency of CT detection of axial skeletal than that reported for plain film radiology. Further, we
if the
thickness
and
depth
to the Breslow and Clark of skeletal metastases.
of penetration
classifications)
of the
primary
can be used to
and Methods
We reviewed 464 body CT studies performed during a 4-year period on 125 patients at risk for metastases from malignant melanoma. There were 74 men and 51 women with an average age of 52 years (range, 25-76 years). All patients had a histologic diagnosis of primary malignant melanoma, with the primary lesion identified on the trunk (43 patients), head/neck (23 patients), upper extremity (1 9 patients), and lower extremity (28 patients). The anatomic site of origin of the primary tumor was unknown for 12 patients.
Downloaded from www.ajronline.org by 58.220.253.162 on 10/21/15 from IP address 58.220.253.162. Copyright ARRS. For personal use only; all rights reserved
110
PATTEN
CT studies were performed either to assist in clinical staging of the primary melanoma or to evaluate known metastatic disease for response to therapy. The anatomic regions studied included the abdomen (41 %), chest (27%), pelvis (25%), and neck (7%). All CT studies were performed on a GE 8800 or 9800 (General Electric Medical Systems, Milwaukee, WI). Five-millimeter axial collimated images were obtained through the neck at 5-mm intervals, Axial scans with 1 -cm collimation were obtained through the chest, abdomen, and pelvis with either 1 .0-, 1 .5-, or 2.0-cm spacing. In the neck and chest, a 150- to 1 80-mI bolus of IV contrast material was administered during imaging; IV contrast material was administered in a similar fashion for imaging of the abdomen and pelvis. Standard lung, liver, mediastinal, and abdominal window and level settings were used for photography. Bone windows (W = 2000 H) were obtained if requested by the physician monitoring the examination and were performed for 23 studies (5%). All studies were reviewed by three radiologists experienced in body
CT for the presence or absence of skeletal metastases.
Criteria
for skeletal metastases on CT included a focal bony lesion or a region of bony permeative infiltration. Sites of bony lesions were recorded along with the presence or absence of any associated soft-tissue mass or adjacent reactive bone. Pathologic verification of osseous metastases was available in six patients, either from biopsy or autopsy. In 11 additional patients, bony metastases were verified by follow-up CT that showed growth in size or number of bony lesions
over time. histology
Skeletal
lesions
or follow-up
in six other
examination
and
patients
were
excluded
were
Correlative bone imaging studies had been performed in our population of patients. In those patients with
metastases, correlation
bone scintigraphy was
provided
in only
was performed eight
benign
from
on
the study.
infrequently documented
in only four. Plain film
patients.
Description of the Clark level (the vertical extent of the lesion as measured histologically) was available in 92 of the original 125 patients;
the
distribution
of lesion
depth
was
level
I in three
patients,
level II in six patients, level Ill in 15 patients, level IV in 44 patients and level V in 24 patients. Results of the Breslow classification of tumor thickness (the depth of tumor invasion from the skin surface) were available for 81 of the 1 25 patients; 1 4 patients had thin lesions (4 mm). Patients with CT evidence of skeletal metastases had an average interval of 3.5 years from initial diagnosis of melanoma to the diag-
nosis of bone metastases
(range, 15 months-6
survival
bone
of 4 months
1 -1 8 months).
after
metastases
were
years), and a mean diagnosed
(range,
ET
AL.
AJR:155, July 1990
Results Ninety-eight (78%) metastatic melanoma.
of the 1 25 patients had CT evidence of Of these, 1 7 (1 7%) had skeletal metas-
tases shown on CT. One patient had diffuse infiltrative ment
of the spine.
The
other
1 6 patients
had
involve-
50 separate
sites of involvement of the axial skeleton, with the majority of the lesions identified in the spine (27), and the rest seen in the pelvis (1 1), ribs (1 0), sternum (one), and scapula (one). The CT appearances were nonspecific and paralleled the plain film findings, which have been reported before [5-8] (Figs. 1 and 2). The lesions were osteolytic in all but one patient and slightly expansile in five patients. Osteolysis was geographic at 32 sites, moth-eaten or permeative at nine sites, and mixed at eight sites. Significant periosteal reaction was absent. In
one patient, a solitary, irregularly marginated, blastic vertebral body metastasis was identified (Fig. 3). Associated soft-tissue mass was seen in nine patients (Fig. 4). In three patients with osteolytic vertebral body lesions and pathologic compression fractures, soft-tissue mass compromised the spinal canal (Fig. 5). In five additional patients, the soft-tissue mass appeared to be the primary lesion, with bone involvement caused by
invasion or erosion
(Fig. 6).
In most cases of bony metastases, CT evidence of widespread metastatic disease was present elsewhere in the body. However, in two (1 2%) of the 1 7 patients, skeletal metastases were the only CT evidence of metastatic disease. In addition to CT, both of these patients were studied with bone scintigraphy. In one of these cases, findings on the bone scan were normal despite biopsy-proved melanoma metastasis to the
spine. All bony metastases were found in patients with a thick or intermediate primary tumor (Breslow) classified as Clark level III or deeper. No age or sex predominance patients with metastatic bone involvement.
was seen for
Discussion Malignant melanoma accounts for approximately 1 % of cancer and the prevalence of the disease is reported to be increasing [9, 1 0]. Melanoma is an unpredictable tumor with a tendency to metastasize widely by direct extension or by lymphatic or hematogenous spread [1 1 ]. Prognosis is most
dependent
on the location
and depth
of the primary
tumor
Fig. 1.-CT scan of sternal metastasis from melanoma shows mixed osteolytic destruction of sternum with associated soft-tissue mass rows). Diagnosis was proved by biopsy.
(ar-
Fig. 2.-CT scan of melanoma metastases to U vertebral body shows discrete “punched out” osteolytic lesions (arrows), which increased in size and number on follow-up studies. Abdominal CT was
performed
to evaluate
response
of
patient to gamma interferon therapy. Osseous metastases were clinically unsuspected.
Downloaded from www.ajronline.org by 58.220.253.162 on 10/21/15 from IP address 58.220.253.162. Copyright ARRS. For personal use only; all rights reserved
AJR:155,July
CT
1990
OF
Fig. 3.-Axial CT scan of upper abdomen shows melanoma metastasis to Ti 1 vertebral body. Note irregularly marginated blastic lesion (arrows). Diagnosis was confirmed at autopsy.
BONE
METASTASES
FROM
111
MELANOMA
Fig. 4.-A and B, CT scans at soft-tissue window settings (A, W settings (B, W = 2000, L = 200) of melanoma metastasis to vertebral (black
arrows)
in body
of L2 with associated
soft-tissue
mass
(white
=
500, L = 50) and bone window body show an osteolytic lesion arrows). Diagnosis was proved
by biopsy.
Fig. 5.-CT
scan of metastatic melanoma to shows expansile osteolytic destruction and compression fracture. Left podide is destroyed, and thecal sac filled with contrast material is mildly effaced by associated soft-tissue mass (arrows). T10
vertebral
body
Fig. 6.-CT tending invasion
scan of melanoma
of vulva cx-
into right superior pubic ramus shows and destruction of pubis (arrowheads)
by a large contiguous pelvic soft-tissue mass (M). Follow-up CT studies showed subsequent growth
of mass
and additional
bone destruction.
Autopsy was refused.
(Clark and Breslow classifications), but depends also on the presence and extent of metastatic disease [1 2-1 5]. Because CT facilitates the simultaneous examination of multiple sites at risk for metastatic disease [9] and may allow earlier detection of metastatic lesions, CT is being used with increasing frequency to assist in clinical staging of melanoma and to
monitor use
response
of body
of patients
CT
to therapy
in melanoma
[1 6, 17]. Widespread
patients
could
theoretically
[3, 4], the radiologic
detection
of osseous
melanoma
lesions has been described as infrequent or uncommon [1 2]. Two large series of 1677 and 1 870 melanoma patients documented bony metastases by radiologic or scintigraphic survey in only 7% and 8% of patients, respectively [5, 6]. The reason for the discrepancy between autopsy figures and radiologic detection of osseous metastases is the relative ,
insensitivity
detecting
of both
plain film radiology
intramedullary
lesions
and scintigraphy
[1 8, 19]. Osseous
surveys for melanoma aggressive radiologic with evidence
improve the detection of clinically silent metastases. Despite autopsy studies showing melanoma metastases to bone in 23-49% of patients who died from malignant melanoma
tases may remain invisible on plain films until cortical destruction has occurred, and scintigraphy may miss those lesions that are growing either too aggressively or too slowly to incite significant reactive bone formation. Additional reasons for this discrepancy include the infrequent use of routine skeletal
for
metas-
of metastatic
The increased tumors
patients and the reluctance to perform investigation in the moribund patient
and monitor
disease
use of body patients
elsewhere.
CT in our institution with malignant
to stage
melanoma
led to
an increased recognition of skeletal metastases. Although a selection bias is present in our study that favors patients with more aggressive or advanced primary tumors, 14% of all melanoma
patients
studied
and
1 7%
of patients
with
CT
evidence of metastatic disease had metastatic lesions in the axial skeleton. This prevalence is higher than those reported in most studies of skeletal melanoma metastases based on plain films alone and more closely approximates autopsy statistics. Additionally, the true prevalence of skeletal melanoma metastases detectable by CT is certainly underesti-
112
mated
PATTEN
in our study.
As our retrospective
review
was
limited
Downloaded from www.ajronline.org by 58.220.253.162 on 10/21/15 from IP address 58.220.253.162. Copyright ARRS. For personal use only; all rights reserved
to body CT studies of the neck, chest, abdomen, and pelvis, lesions in the peripheral skeleton and skull were not detected. Each patients’ scans were reviewed from film, rather than from tape at the imaging window and level settings
console, where manipulation might have led to recognition
of of
additional foci of bone destruction. As bone windows were not routinely used in filming and were available in only 5% of studies, subtle lesions may have been missed on review of images photographed at abdominal window and level settings. Finally, as contiguous slices were not performed in all
instances, small bony lesions may have been missed in the skipped intervals. Other authors have reported a higher frequency of CT detection of metastatic disease in melanoma patients with more advanced primary lesions [9]. This observation appears equally valid for skeletal melanoma metastases, because none of the patients with thin (