487
Downloaded from www.ajronline.org by 114.27.48.47 on 10/16/15 from IP address 114.27.48.47. Copyright ARRS. For personal use only; all rights reserved
Intraabdominal Mycobacterium tuberculosis vs Mycobacterium avium-intracellulare Infections Patients with AIDS: Distinction
in Based
on CT Findings
D. Randall
Radin1
Abdominal CT scans of 71 patients with AIDS who had proved disseminated infection due to Mycobacterium tuberculosis (27 patients) or Mycobacterium avlum-intracellulare
(44 patients) used
were reviewed
to distinguish
the two
retrospectively
to identify radiologic
infections.
findings
CT
in patients
features with
that can be
disseminated
M.
included focal lesions in the liver (1 1%), spleen (30%), kidneys (19%), pancreas (7%), and gastrointestinal tract (15%) and lymph nodes with central or diffuse low attenuation (93%). CT findings in patients with disseminated M. avium-intracellulare included marked hepatomegaly (20%); marked splenomegaly (14%); focal lesions in the
tuberculosis
liver (9%), spleen (7%), and kidneys (2%); diffuse jejunal wall thickening (18%); lymph nodes with central low attenuation (14%); and enlarged lymph nodes exclusively of homogeneous soft-tissue density (55%). The presence of focal visceral lesions and low-attenuation lymph nodes suggests disseminated M. tuberculosis, whereas marked hepatic and splenic enlargement, diffuse jejunal wall thickening, and enlarged soft-tissue-density lymph nodes suggest dissemmated M. avium-intracellulare. Recognition of these CT features can lead to a tentative diagnosis so that appropriate therapy can be instituted before the results of mycobacterial cultures become available.
AJR 156:487-491,
March
1991
Disseminated infections due to Mycobacterium avium-intracellulare (MAI) [1 ] and Mycobacterium tuberculosis (MTB) [2] in patients with AIDS were first reported in 1 982 and 1983, respectively. In 1987, the Centers for Disease Control included disseminated extrapulmonary infection by both MAI and MTB in a list of diseases that indicate a diagnosis of AIDS in patients with laboratory evidence of infection with human immunodeficiency virus (HIV) [3]. Distinction between MTB and MAI infection is important because of the therapeutic implications [1 4]. Standard antimycobacterial therapy is far more effective in AIDS patients with MTB infection than in those with MAI infection. Findings in one large series suggest that clinical features and findings on chest radiographs may help to distinguish between MTB and MAI infection in HIV-infected patients [4]. Abdominal CT findings in AIDS patients with disseminated mycobactenal infection have been described in several studies and review articles [5-i 0]. However, a comparative study of abdominal CT findings in MTB and MAI infection has not yet been reported. ,
Received July 25, 1990: September 13, 1990.
accepted
after revision
Department of Radiology, University of Southem California School of Medicine, L. A. CountyUSC Medical Center, 1200 N. State St., Los Angeles, CA 90033-1084. Address reprintrequests to
D. R. Radin. 0361 -803x/91/1
c
American
563-0487
Roentgen
Ray Society
Materials
and Methods
According to mycobactenology laboratory records at our hospital, cultures of specimens from lymph node biopsies, bone marrow biopsies, or blood yielded MTB in 173 patients and MAI in 373 patients between February 1 985 and February 1990. An abdominal CT examination done during the same hospital admission during which the biopsy with positive mycobacterial results was performed was available for 33 patients with MTB infection and
45 patients
with
MAI
infection.
Medical
records
of these
78 patients
were
reviewed.
Six
RADIN
488
TABLE
1: Clinical
Features
of 71 Patients
AJR:156,
with AIDS and Disseminated
Number (%)
Feat u re
with
MTB
Mycobacterial
of Patients (n = 27)
Number with
March
1991
Infection
(%) of Patients (n = 44)
MAI
HIV risk factor 1 8 (67)
Homosexuality
Downloaded from www.ajronline.org by 114.27.48.47 on 10/16/15 from IP address 114.27.48.47. Copyright ARRS. For personal use only; all rights reserved
IV drug use
Homosexuality Other
and IV drug use
Unknown
illness
Concurrent AIDS-defining illness Disseminated mycobacterial infection
as only AIDS-defining Note.-MTB
patients
with MTB infection
1 1 patients
(with
Mycobacterium
=
and one patient
AIDS-defining
known
MAI
tuberculosis;
with
MAI infection
were
illness
extracutaneous
were
Kaposi
involvement
sarcoma
in
in only one pneumonia
patient); lymphoma (three patients); Pneumocystis carinll (24 patients); cytornegalovirus infection (nine patients); esophageal candidiasis (four patients); cryptococcosis (two patients); and histoplasmosis, toxoplasmosis, and cryptospondiosis (one patient each). CT examinations of the 71 patients were performed on a Picker 1 200 SX scanner. Slices 1 0 mm thick 15-mm intervals. IV contrast medium
were obtained at 1 0-mm or was administered in all but
seven cases. The studies were reviewed retrospectively and evaluated for liver and spleen size; the presence of hepatic, splenic, renal, pancreatic, adrenal, and gastrointestinal lesions and ascites; and the location, size, and density of abdominal lymph nodes. Of the 27 patients with MTB infection, involvement of abdominal lymph nodes was proved by percutaneous fine-needle aspiration in nine patients and autopsy in three patients. Follow-up CT examination after antituberculous therapy showed resolution of abdominal adenopathy in one patient. In seven of the 14 patients without direct evidence of abdominal nodal involvement, MTB was cultured from peripheral lymph node biopsy specimens. Splenic lesions in three patients and renal and pancreatic lesions in one patient each were confirmed to be due to MTB by fine-needle aspiration or autopsy. Of the 44 patients with MAI infection, involvement of abdominal lymph nodes was proved by percutaneous fine-needle aspiration of softtissue-density nodes in three patients and necrotic nodes in three patients. Involvement of nodes less than 10 mm in diameter was found at autopsy in one patient. Biopsy specimens of peripheral nodes yielded MAI in two patients. Hepatic lesions were confirmed at autopsy in one patient.
Results Abdominal
disseminated
CT findings mycobacterial
5 (11) 0 (0)
(4)
0
(0)
2 (7)
31 (70)
2 (7) 23 (85)
5 (11) 8 (18)
=
Mycobacterium
avium-intracellulare:
HIV
=
human
immu-
virus.
.
nonmycobactenal
2 (7) 2 (7)
illness
excluded because of a lack of serologic proof of HIV infection. All but one of the 71 patients with AIDS who had disseminated mycobactenal infection were men. The 27 patients with MTB ranged in age from 21 to 54 years (average, 35 years). The 44 patients with MAI ranged in age from 25 to 53 years (average, 35 years). HIV risk factors and the temporal relationship of mycobacterial infection to nonmycobacterial AIDS-defining illness are indicated in Table 1 Kaposi sarcoma was the only previous or concurrent disease in the four MTB patients with nonmycobacterial AIDS-defining illness. Extracutaneous Kaposi sarcoma was documented in only one of these four patients. Previous and concurrent diseases in the 36 MAI patients with
35 (80) 4 (9)
1
Previous AIDS-defining
nodeficiency
4 (1 5)
in the 71 patients with infection are presented
AIDS and in Tables
TABLE 2: Abdominal CT in 71 Patients with AIDS and Disseminated Mycobacterial Infection: Extranodal Findings
(%) of Patients
Number
CT Findin
g
Number
with MTB
with MAI
(n=27) Hepatomegaly Mild (1 9-20
cm cephalocau-
dal diameter) Marked (21 -24 cm cephalocaudal diameter) Splenomegaly Mild (1 4-1
5 cm cephalocau-
(%)
of Patients (n=44)
5 (1 9)
20 (45)
5 (1 9)
1 1 (25)
0
(0)
9 (20)
7
(26)
10 (23)
7 (26)
4
(9)
0
(0)
6 (14)
12 (44) 3(11) 8 (30)
6 (14) 4 (9) 3 (7)
dal diameter) Marked (1 6-21 cm cephalocaudal diameter)
Focal lesions in solid viscera Liver Spleen Kidney
5 (1 9)
1
Pancreas Adrenal abnormalities
2 (7) 0 (0)
0 (0) 0 (0)
Gastrointestinal
4 (15)
8 (18)
5 (1 9)
5 (11)
tract abnormali-
(2)
ties
Ascites Note.-MTB
=
Mycobacterium
tuberculosis
; MAI
=
Mycobacterium
av-
ium-intracellulare.
2 and 3. Although hepatomegaly and splenomegaly were seen in a minority of patients in each group, marked enlargement occurred only in patients with MAI. Because only one third of these MAI patients had other diseases commonly associated with hepatosplenomegaly (lymphoma, cytomegalovirus infection, histoplasmosis), it is likely that MAI infection was responsible for the marked hepatic and splenic enlargement in most instances. CT evidence of focal involvement of solid abdominal viscera was far more frequent in patients with MTB. CT showed one or more round or oval hypodense lesions between several millimeters and 2 cm in diameter in the liver, spleen, kidneys, and/or pancreas in 1 2 (44%) of the 27 MTB patients (Fig. 1) but in only six (1 4%) of the 44 MAI patients (Fig. 2). These parenchymal lesions had a similar appearance in both groups of patients but tended to be more numerous in the patients with MTB. A striking finding was the
AJR:156,
March
MYCOBACTERIAL
1991
TABLE 3: Abdominal Nodal Findings
CT in 71 Patients
CT Findin Solid
nodes
g
Downloaded from www.ajronline.org by 114.27.48.47 on 10/16/15 from IP address 114.27.48.47. Copyright ARRS. For personal use only; all rights reserved
node nodes