Access to the PDF text

Free Article !

Journal de radiologie
Vol 88, N° 1-C1  - janvier 2007
pp. 47-51
Doi : JR-01-2007-88-1-C1-0221-0363-101019-200701086
gastrointestinal imaging
MR features of hepatic macronodular mycobacteriosis

T Caramella [1], F Fuerxer [2], P Chevallier [1], E Cua [3], MC Saint Paul [4], F Dausse [1], S Novellas [1], C Gueyffier [5], P Brunner [2], JN Bruneton [1]
[1] Service d’imagerie médicale (Pr Bruneton), Centre Hospitalier Universitaire de Nice, Hôpital Archet II.
[2] Service d’IRM (Dr Cucchi, Dr Mourou), Centre Hospitalier Princesse Grâce, Av Pasteur, 98000 Monaco.
[3] Service des maladies infectieuses (Pr Dellamonica),
[4] Service d’anatomopathologie (Pr Michiels), Centre Hospitalier Universitaire de Nice, Hôpital Archet II, 151 route de Saint Antoine de Ginestiere, BP 3079, 06202 Nice.
[5] Service de gastroentérologie, Centre Hospitalier Régional de Cannes, Hôpital Les Broussailles, 15 av des Boussailles, 06400 Cannes.

To cite the present paper, use exclusively the following reference. Caramella T, Fuerxer F, Chevallier P et al. IRM des atteintes macronodulaires hépatiques à mycobactéries. (full text in english on click Here). J Radio 2007;88:47-51.

Tirés à part : T Caramella



Hepatic macronodular mycobacteriosis is rare. Its diagnosis is challenging and is most often proposed on the basis of histological analysis. Final diagnosis, except for germ-proven cases, is made in conjunction with clinical, biological, and radiological arguments. We retrospectively report the MR features of ten hepatic lesions discovered on five patients. MRI is sensitive but has a low specificity in demonstrating pseudotumoral lesions most often exhibiting hypointensity on the T1-weighted sequence, hyperintensity on the T2-weighted sequence, and a slight rim enhancement after gadolinium-enhanced T1-weighted sequences.

IRM des atteintes macronodulaires hépatiques à mycobactéries

L’atteinte mycobactérienne macronodulaire hépatique isolée est rare. Son diagnostic, difficile et méconnu, est le plus souvent évoqué après l’analyse histologique de biopsies ou de pièces opératoires. Le diagnostic de certitude, en dehors des rares cas où le germe est isolé, repose sur un faisceau d’arguments à la fois cliniques, biologiques et radiologiques. Nous rapportons, de manière rétrospective, l’aspect en IRM de dix lésions hépatiques développées chez cinq patients. L’IRM paraît sensible mais peu spécifique, avec des lésions pseudo tumorales présentant le plus souvent un hyposignal spontané en T1, un hypersignal en T2 et un très faible rehaussement périphérique de leur intensité aux phases portales et parenchymateuses d’une injection intraveineuse de complexes de Gadolinium en T1.

Mots clés : Imagerie par résonance magnétique , Foie , Infection à mycobacterium

Keywords: Magnetic Resonance Imaging , Liver , Mycobacterium infection

Mycobacteria are a rare cause of liver parenchyma infection. Few publications, or only studies with small samples, have described the imaging features of hepatic macronodular mycobacteriosis, particularly with MRI [1], [2], [3], [4], [5]. Here we report the MRI features of ten liver lesions found in five patients.

Material and methods

We conducted a retrospective study over the period from January 1998 to December 2005, based on our institution’s Pathological Anatomy Laboratory’s database, of all patients for whom a histological diagnosis of hepatic mycobacteria involvement was found. With the objective of describing the MRI features of this disease, our study included all patients from this first list who had also had a liver MRI examination in the month preceding or following the histological diagnosis. Each of the patients included in the study also had to have a complete medical file including clinical observation and the patient’s immune and pulmonary status.

The MRIs (Gyroscan NT-10, 1 tesla, Philips Medical Systems, Eindoven, Netherlands) comprised T2-weighted axial Turbo spin-echo sequence (TR/TE = 1800/80) with fat saturation (Spoiled Inversion Recovery), in T1-weighted in-phase and opposed-phase gradient echo (TR/TE in-phase/opposed-phase TE/tilt angle = 245/6.9/3.4/80°), and in T1-weighted gradient echo sequence (TR/TE = 4/1.9) after IV injection of 0.1 mmol/kg of gadolinium complexes (Dotarem®, Guerbet, France) at arterial, portal, and parenchymal times. For two patients, we also had reconstructions in maximum intensity projection of cholangiopancreatography sequences taken in 3D TSE (TR/TE eff = 2000/700).

The MRIs were analyzed in consensus reading by two senior radiologists with experience in liver diseases.

Diagnosis was confirmed by operative specimens or liver biopsies analyzed by a pathologist and/or a senior biologist specialized in liver diseases.

Clinical data

The mean age of the population studied was 67 years (range, 40-82 years). The entry symptoms corresponded to fever (two cases), pain in the right hypochondrium (three cases), and/or a deterioration of the general condition (three cases). For a patient treated for chronic liver disease, a focal point liver mycobacterium infection was discovered incidentally, and another patient presented with a surgical picture of acute angiocholitis with an enclosed lithiasis at the bottom of the bile duct. One patient had already been hospitalized 10 years before for pulmonary tuberculosis. None showed a congenital or acquired immunodeficiency. All patients improved clinically on specific antibiotic treatment within a few weeks and the lesions shown on imaging disappeared between 2 and 6 months after beginning treatment.

The biological results showed that transaminases were normal in two cases (40%) and high in three cases (eight times the normal level in one patient and 1.5 times the normal level for the two others). Gamma-glutamyl transferases were only normal for one patient and increased for the four others, between three and ten times the normal level. Alkaline phosphatases were increased in three cases, between three and four times the normal level. Total and conjugated bilirubin were only high in the acute angiocholitis case because of the bile duct obstruction. Finally, inflammation as demonstrated by the reactive protein C values was present in only two patients.

All patients had a thoracic CT, a bronchial fibroscopy examination, and several bacterial serology tests. This workup was negative in all cases. The tuberculin intradermal reaction test was positive for two patients, including the patient with a history of pulmonary tuberculosis.


The mean delay between MRI examination and histological confirmation was 29.8 days (range, 1-71 days).

Ten lesions, measuring a mean 48 mm (range, 30-70 mm), were identified. The results before biopsy are reported in table I and illustrated in figures 1 - 3 . The lesions exhibited T2-weighted hyperintensity in seven cases and T1-weighted hypointensity in all cases. Contrast uptake was progressive, predominating in the portal time and in the periphery of the lesion in seven cases. There was no signal modification on T1-weighted opposition-phase sequences compared to in-phase sequences. The lesions were located in the left lobe three times and in the right lobe seven times. Patient no. 4 presented a central hepatic lesion inducing dilatation of the intrahepatic bile ducts predominating on the left.

Histological results

The histological analyses were done for one patient on a segmentectomy specimen and on liver biopsies for the others. For all patients, histology demonstrated giant cell granuloma lesions. In four patients, this was associated with necrosis, described as caseous for two of them. Ziehl-Nielsen staining was negative in all cases. Peliosis was noted in three patients and fibrosis in three cases. For the two patients for whom hepatic tuberculosis was suggested before biopsy, a complementary sample was sent for bacteriological analysis. The culture came back positive for Mycobacterium fortuitum after 90 days for one patient. The PCR method was not used to look for mycobacteria.


Hepatic tuberculosis is usually associated with active pulmonary tuberculosis or miliary tuberculosis. There is also an abdominal association in 15% of tuberculosis cases, the most often with a lymph node variant (25%–93%) [6]. In a study of 92 cases of liver biopsies done in patients with pulmonary tuberculosis, Morere et al. [7] identified granulomatous lesions in 20 cases. Isolated liver involvement of a mass or tuberculoma is rare. This diagnosis is most often unrecognized and confused with primary or secondary liver tumor. The liver infection probably stems from the para-aortic or portal lymph nodes and may reach the liver through the portal system or the superior mesenteric artery [8].

Positive diagnosis 

The clinical manifestation of tuberculosis takes several forms and can vary from the absence of symptoms to a picture of severe hepatitis with jaundice and hepatic insufficiency [6], [9]. Positive diagnosis is difficult because it requires demonstrating resistant acid-fast bacilles. Ziehl-Nielsen staining is not positive in 0%-45% of cases, culture is long (90 days) and is only profitable in 10%-60% of series, so that the PCR search for mycobacteria would most often be contributive, in at least 57% of cases [8].

Nevertheless, for most authors [1], [2], [3], [8], [10], [11], [12], [13], the diagnosis can be reached based either on the presence of hepatic granulomas associated with documented tuberculosis in another organ, particularly the lungs, or when the clinical symptoms and radiological examinations regress after starting up antituberculous treatment, particularly if the initial antibiotic therapy failed [8].

In our study, only one culture was positive and demonstrated an atypical mycobacterium. Therefore, for the four other cases, the diagnosis was presumed based on a range of clinical and histological arguments.

Differential diagnosis

Tuberculoid granuloma is characterized by the presence of epithelioid cells, giant cells, or Langhans cells and lymphocytes in the peripheral rim. This is not itself specific and can be found in other diseases. A study by McCluggage [14] inventoried the hepatic granulomatous etiologies over 30 years and found a number of causes: infectious (mycobacterioses, brucellosis, mycoses, schistosomiasis), metabolic (copper overload/Wilson disease), inflammatory (sarcoidosis, primary biliary cirrhosis), and neoplastic (lymphoma).

Complementary tests will therefore always be necessary to eliminate each of these hypotheses, which was done for all the patients in this retrospective study.

Imaging features

The descriptions of hepatic tuberculoma in imaging are few in number and most often are sonographic and/or CT descriptions. In ultrasound, the tuberculoma is most often hypoechogenic, but hyperechogenic nodules have already been described [15], [16]. With CT, the nodules classically appear hypodense without enhancement or with low levels of peripheral enhancement after iodine injection [15].

To our knowledge, fewer than ten MRI observations of hepatic tuberculomas have been reported in the literature [1], [2], [3], [4], [5].

The classically accepted feature is a lesion exhibiting hypointensity on the T1-weighted sequence and isointensity or hypointensity on the T2-weighted sequence. After injection there is a slight rim enhancement [4], [13], [15].

In our study, enhancement was absent or low in all cases after IV injection of gadolinium. When it was present, it was most often a peripheral rim enhancement at the portal time and late after injection. None of the lesions showed signs suggesting a specific hepatic lesion and the unusual aspect had encouraged pursuing investigations in the hypothesis of a malignancy. The T2-weighted signal was most often high.

These data are in agreement with the preceding cases recently described in the literature. The spontaneous hypointensity on the T1-weighted sequence seemed constant [2], [3], [11]. In the T2-weighted sequences, the descriptions were variable, with the intensity either lower [2], [3], [11] or higher [3], [12]. Yuji et al. [1] discussed the possible reasons for a hypointensity on the T2-weighted sequences. It could be related to the presence of fibrosis, calcifications, or free radicals [1]. In this first case, the pattern could be close to the cerebral tuberculomas that classically appear in hypointensity on the T2-weighted sequences for this same reason [17]. Other authors report a lesion showing hyperintensity on the T2-weighted sequences because of a cystic component [3] or liquid necrosis [11]. In our case, and as suggested by Fan et al. [2], hyperintensity in T2-weighted images is most certainly related to the very presence of granulomatous tissue. The intensity of the signal may be modulated by inflammatory reorganizations, the associated fibrosis and necrosis, which may reflect the same physiopathological form at different stages of advancement [5]. Moreover, the T2-weighted sequences that we obtained were systematically associated with fat saturation [18], contrary to the protocols reported by other authors [2], [3], [4], [5]. Therefore, the lesions described showing T2-weighted isointensity or hypointensity in other studies may have stemmed from hyperintensity after fat saturation, with the signal description made in relation to the adjacent healthy liver.


Isolated hepatic macronodular mycobacteriosis is rare and MRI has a low level of specificity in its positive diagnosis. The most frequent image is one showing hypointensity on the T1-weighted sequence, hyperintensity on the T2-weighted sequence, with very low rim enhancement after IV injection of gadolinium predominating in the portal and later phases.


Yuji M, Ichiro Y, Yasushi S, Yuji S, Yoshio S. Abdominal macronodular tuberculomas : MR findings. J Comput Assist Tomogr 1996;20:643-6.
Fan ZM, Zeng QY, Huo JW et al. Macronodular multi-organs tuberculoma: CT and MR appearances. J Gastroenterol 1998;33:285-8.
Ri-Sheng Yu, Shi-Zheng Zhang, Jian-Jun Wu, Rong-Fen Li. Imaging diagnosis of 12 patients with hepatic tuberculosis. World J Gastroenterol 2004;10:1639-1642.
Murata Y, Yamada I, Sumiya Y, et al. Abdominal macronodular tuberculosis MR Findings. J comput Assist Tomogr 1996; 20:643-6.
Mercusot B, Arrivé L, Rotenberg L, et al. Imagerie du tuberculome hépatique. J Radiol 1995;76:277-9.
Akhan O, Pringot J. Imaging of abdominal tuberculosis. Eur Radiol 2002;12:312-323.
Morere P, Nouvet G, Stain JP, Paillot B, Metayer J, Hemet J. Renseignements fournis par la biopsie hépatique chez 100 malades tuberculeux. Sem Hop Paris 1975;51: 2095-2102.
Adnani A, Dafiri R. La tuberculose hépato-splénique chez l’enfant. J Radiol 2005;86:1710-1.
Huang WT, Wang CC, Chen Wj, Cheng YF, Eng HL. The nodular form of hepatic tuberculosis : a review with five additional new cases. J Clin Pathol 2003; 56:835-9.
Venkatesh SK, Tan LKA, Siew EPY, Putti TC. Macronodular hepatic tuberculosis associated with portal vein thrombosis and portal hypertension. Australasian Radiol 2005;49:322-324.
Kawamori Y, Matsui O, Kitagawa K, Kadoya M, Takashima T, Yamahana T. Macronodular tuberculoma of the liver: CT and MR findings. AJR Am J Roentgenol 1992;158:311-3.
Hayashi M, Yamawaki I, Okajima K, Tomimatsu M, Ohkawa SI. Tuberculous liver abscess not associated with lung involvement. Internal Medecine 2004;6: 521:5.
Hickey N, McNulty JG, Osborne H, Finucane J. Acute hepatobiliary tuberculosis : a report of two cases and a review of the literature. Eur Radiol 1999;9:886-9.
McCluggage WG, Sloan JM. Hepatic granulomas in Northern Ireland: a thirteen years review. Histopathology 1994; 25:219-228.
Mortelé J, Segatto E, Ros PR. The infected liver : radiologic-pathologic correlation. Radiographics 2004;24:937-955.
Tan TCF, Cheung AYC, Wan WYL, Chen TC. Tuberculoma of the liver presenting as a hyperechoic mass on ultrasound. Br J Radiol 1997;70:1293-5.
Kioumehr F, Dadsetan MR, Rooholamini SA, Au A. Central nervous system tuberculosis : MRI. Neuroradiology 1994; 36:93-6.
Hussain SM, Zondervan PE, IJzermans JN, Schalm SW, de Man RA, Krestin GP. Benign versus malignant hepatic nodules: MR imaging findings with pathologic correlation. Radiographics 2002;22:1023-1036.

© 2007 Elsevier Masson SAS. Tous droits réservés.
EM-CONSULTE.COM is registrered at the CNIL, déclaration n° 1286925.
As per the Law relating to information storage and personal integrity, you have the right to oppose (art 26 of that law), access (art 34 of that law) and rectify (art 36 of that law) your personal data. You may thus request that your data, should it be inaccurate, incomplete, unclear, outdated, not be used or stored, be corrected, clarified, updated or deleted.
Personal information regarding our website's visitors, including their identity, is confidential.
The owners of this website hereby guarantee to respect the legal confidentiality conditions, applicable in France, and not to disclose this data to third parties.
Article Outline