Am. Khalil et al., MAGNETIC-RESONANCE-IMAGING FINDINGS IN PULMONARY KAPOSIS-SARCOMA - A SERIES OF 10 CASES, The European respiratory journal, 7(7), 1994, pp. 1285-1289
Since chest X-ray and CT scan features of Kaposi's sarcoma (KS) are no
nspecific, we wanted to test the hypothesis that the histological comp
onents of this tumour and/or the associated haemorrhagic component, ma
y result in a characteristic signal pattern on magnetic resonance imag
ing (MRI). Thoracic MRI was performed in a prospective manner in ten p
atients with acquired immune deficiency syndome (AIDS) and pulmonary K
S. MRI examinations (1.5 Tesla) included Spin-echo Tl (SE-Tl), before
and after gadolinium injection, as well as T2-weighted sequences (SE-T
2). For each sequence the signal intensity of lesions was measured and
compared with each other as well as with the signal intensity of musc
le. Results were compared to the hemosiderin content of macrophages in
the bronchoalveolar lavage (BAL) in all patients and with histologica
l findings in three. The results were compared to values obtained in a
control group of seven patients with pneumocystis carinii pneumonia.
SE-Tl showed focally increased signal intensity in the pulmonary paren
chyma (n=5). Signal enhancement in parenchymal lesions (n-10) and alon
g peribronchovascular trees (n=5) was observed after gadolinium inject
ion. The second echo of SE-T2 showed a markedly reduced signal intensi
ty in pathologic areas (n-10). This last finding was not observed in t
he control group. In conclusion, we have identified a pattern of MRI s
ignal abnormalities suggestive of Kaposi's sarcoma. The MRI signal int
ensity of KS lesions may be related to the angiomatous and fibrous com
ponents of the tumour. A correlation between hemosiderin deposits and
signal intensity was not demonstrated in this study, perhaps because o
f a poor correlation between BAL results and the actual occurrence of
an haemorragic component in parenchymal lesions.