Different Views On The Problem Of Tuberculoma

Solitary and multiple intracranial ring enhancing lesions are frequent diagnostic conundrum faced by radiologists. The differentials are very wide and include both neoplastic and non-neoplastic causes. In developing countries tuberculomas and NCC make up for majority of infectious cause [Shetty G].

Tuberculomas are caused by microbe mycobacterium tuberculosis . They usually result from hematogenous spread from a primary focus or dormant elsewhere in the body [Pillenahalli]. They typically involve the gray-white junction. Most common location in adults are cerebral hemispheres and basal ganglia, in children its cerebellum in [Bomanji, Wetzel]. Most of the times tuberculomas are solitary, however multiple lesions can occur in 15 to 34% cases [Wetzel S, Abuhamed M]. CNS tuberculosis is more severe in toddlers compared to older children [Newton SM].

The diagnosis is often delayed due to absence of sputum production and difficulty in getting biopsy of CNS lesions. CT and conventional magnetic resonance imaging (MRI) features of tuberculoma mimic many infectious lesions [Ky Santy]. On magnetic resonance imaging (MRI) tuberculomas are usually iso to hypointense relative to gray matter on both T1- and T2-weighted images and to appear as a conglomerated ring-enhancing mass on gadolinium-enhanced T1-weighted images [Tae Kyoung].

Neurocysticercosis (NCC) of the central nervous system (CNS) is caused by the larval stage of the pork tapeworm Taenia solium. It is the most common parasite causing CNS infection worldwide. The gray-white junction is most commonly affected site in NCC [Eric T]. Epilepsy is the most frequent symptom and seen in 50 to 70 % of cases of NCC [Jing-Long]. NCC has four different stages the vesicular, colloidal vesicular, granular nodular and nodular calcified which are recognized on imaging. CT and MR imaging both are useful tool in the detection of different stages of NCC [Jing-Long, Suss RA, Zee CS, Martinez HR]. The hallmark feature of vesicular stage is a cyst with eccentrically located scolex [Min, Z]. The nodular calcified stage presents as a calcific focus with no perilesional edema. CT rather than MRI better depicts this stage. It is the colloidal vesicular and granular nodular stages which present as ring enhancing lesions with perilesional edema poses diagnostic challenge with tuberculomas [Del Brutto, LT Lucato].

In this study we attempt to distinguish the tuberculomas from the colloidal vesicular and granular nodular stages of NCC using MRS and DWI.Our study included 30 patients with newly diagnosed inflammatory granulomas (NCC and tuberculoma). They were subjected to conventional MR imaging mainly with DW sequence and MR Spectroscopy.

Spectroscopic measurement also obtained from uninvolved healthy brain tissue in affected patients and healthy volunteers.Maximum number of cases was in age group of 10-20 years. Most of the cases were in the age range of 10-30 years (70 %). There were 21 male patients and 9 female patients suggesting significant male predilection. Age range of patients was between 7-50 years and mean age was 21.8 years. Supratentorial lesions (96.66%) were significantly greater than infratentorial (3.33%). Parietal lobe was the most common location in supratentorium closely followed by frontal lobe.

Only 1 lesion was seen predominantly in the posterior fossa (left cerebellum). In 4 patients, lesions were scattered all over the brain parenchyma. Most common mode of presentation was seizure. Next in frequency was headache. Less common modes of presentations were Focal neurological deficit, Altered sensorium and Cerebellar symptoms.Inflammatory granulomas in particular the single and multiple enhancing lesions are diagnostic dilemma to both clinician and radiologist alike. They have been found to persist, increase in size and number or even disappear spontaneously. Their etiology remains in doubt, speculative at best in the absence of any definitive diagnostic test. In cases which are biopsied, histopathology can also be inconclusive finding as only chronic inflammation. However in cases were HPE has been definitive, etiology has commonly been a tuberculoma or cysticercus granuloma among patient population in India (Epilepsia 1994).

The main basis for deciding whether the lesions are inflammatory or not, are the CT or MRI picture. This radiological diagnosis is sometimes supported by the evidence of tuberculosis in some other parts of the body (commonly lung or lymph nodes). In which case evidence like fever, raised ESR may go along with radiological picture and suggest an inflammatory lesion. This evidence would only point to it being inflammatory and not being very specific on confirmatory. In such a scenario, techniques such as magnetic resonance spectroscopy that can provide additional biochemical information and diffusion imaging that is based on restriction of water molecules by lesion would increase the possibility of a correct diagnosis in vivo.

It would also point towards a tuberculoma or a cysticercus granuloma i.e. a specific diagnosis instead of general diagnosis of inflammatory lesion. In our study, significant markers for NCC were raised amino acid peaks In 8 patients of which 4 patients had raised acetate peak (1.9), 5 had raised succinate (2.4) and 2 patients had both acetate and succinate peaks. PN Jayakumar et al, (2004) [PN Jayakumar] reported the MRS findings in racemose cysticercus cyst and found raised peak at 2.4 ppm (succinate) in these lesions. Which were similar to the results in our study. Mishra et al, (2004) [Mishra] reported the same result in their study by concluding raised acetate, lactate and succinate in neurocysticercosis.

Andreia V Faria et al, (2004) [Faria] studied the pattern of MRS in non-neoplastic encephalitic lesions (like NCC, MS infarct, heterotopias, neurofibromatosis, encephalitis) and found raised succinate peak (2.4 ppm) in NCC and total regression of the lesion was observed after treatment with albendazole. M. Agrawal et al, (2004) [M Agarwal] evaluated 3 large intraparenchymal isolated degenerating cysticerci in which diagnosis was primarily based on in vivo MRS and subsequently confirmed histologically and reported the similar kind of result i.e. presence of succinate alone or more succinate than acetate indicating the presence of degenerating cysticerci and differentiates them from anaerobic brain abscess which showed acetate alone or acetate in higher concentration than succinate.

Peruvamba N Jayakumar et al, (2003) [Peruvamba] reported the similar kind of result in their study on 16 patients with intracranial cestodal cyst (3 hydatid and 13 neurocysticercosis) and found large resonance at 2.4 ppm confirmed as succinate found in all cestodal cyst and concluded succinate as a metabolite in cestodal cyst infecting human CNS. In four patients our study showed extremely low level of metabolites with poor signal noise ratio (SNR) likely because the lesions are either cystic (viable) or calcified (dead).

Rama Jayasunder et al, (1999) [Rama Jayasunder] showed results similar to our study i.e. extremely low level of metabolites with poor SNR can itself act as a marker for NCC.In our study, all the lesions diagnosed as NCC were hypointense on DW i.e., no restricted diffusion.

Lucinana S Rafin er al, (2001) [Lucinana] showed similar results in their study on DW findings in diagnosed NCC and concluded all lesions had hypointense signal on DW and similar ADC values as CSF.

However our results were contradictory to study done by Balakrishna et al, (2005) [Balakrishna] to analyzed the diffusion finding in granulomatous lesions and concluded central diffusion restriction seen in vesicular stage of NCC and peripheral restriction in tuberculoma.Intracranial tuberculoma are characterized by MR spectra pattern predominantly involving lipids. Lipid peaks are seen at 0.9, 1.3, 2.0, 2.8 ppm. A phosphoserine peak at 3.7 ppm is also commonly found. Lipid resonances at 0.9, 1.3 ppm are assigned to methylene groups and terminal methyl groups of fatty acids founds within the centre of tuberculoma. Lipids as marker for tuberculoma were first suggested by Gupta et al [Gupta, Gupta].

The major part of tubercle bacillus is known to contain lipids since saturated fatty acid are known to be involved in the production of tubercles. In vitro proton MRS studies on perchloric acid and lipid extracts of tubercular granuloma have shown resonances similar to those observed from mycobacterium bacilli. For the above mentioned reasons and also since lipid resonance are not seen in proton spectrum from normal human brain, they have been noted as a possible marker of tuberculosis.MRS has also been shown to be useful in differentiating tubercular from pyogenic abscess, while tubercular abscess reveals predominant lipid peak, pyogenic abscess reveal in addition amino acid at 0.9 ppm with variable combination of acetate, succinate and alanine glycine related to breakdown caused by proteolytic enzyme (Gupta et al, 2001 [1st Gupta]).

Although most tubercular lesions reveal lipid peak on MRS a similar pattern of dominant lipid peak with near depletion of normal metabolites may be seen in toxoplasmosis (Chang L et al). Moreover, the whole presence of lipids is universally reported in focal cerebral tubercular lesion. An associated increase in choline has also been reported. The presence of choline in non-neoplastic lesions can be seen when the margin of lesions that contains dense inflammatory cell reaction are included within the voxel showed for evaluation (Sudhakar K et al).

Although choline levels are primarily related to amount of cell membrane turnover, increase in choline has also been reported due to increased cellularity of the lesion.From our study, we found very specific markers for tuberculoma. Out of 17 patients with tuberculoma, 16 had raised lipid peak at 0.9, 1.3 and 2.0 ppm. 12 had peak at 1.3 and 3 had 2.0 and 1 had peak at 0.9 and 1.3 ppm. All the lesions of tuberculoma in which spectra were acquired within the voxel including variable part of wall show elevated normalized choline/creatinine ratio ranging from (1.034 to 2.473), mean ± SD (1.621 ± 0.403) and normalized NAA/creatinine ratio ranged from (0.252 to 0.990), mean ± SD (0.544 ± 0.296).PC Khanna et al,(2006) [PC Khanna]attempted differentiating giant extra axial tuberculoma & masquerading meningioma in one pregnant patient and concluded raised predominant lipid peak at 1.33 ppm raised choline and barely perceptible N-acetylaspartate (NAA) in tuberculomas.

Pretell et al, (2004) [Pretell] studied the roles of MRS in differentiating tuberculoma and NCC and found high peak of lipid, more choline, less NAA and creatinine. Choline/creatinine ratio was >1 in all tuberculoma but none of the cysticerci, which were similar to our results. Mishra et al, (2004) [Mishra AM] reported the same result in their study by showing raised lipid lactate peak in tuberculoma.

Kaminogo M et al (2002) [Kaminogo] studied the MRS findings in 2 cases of tuberculoma and reported similar kind of results i.e. raised lipid choline in 1 case and raised lipid lactate in another. Gupta R.K. et al, (2001) [1st Gupta] studied the same results in their study on 33 cases of tuberculoma and concluded raised lipid at 0.9, 1.3, 2.0 and 2.8 ppm in 26 patients while lipid along with choline at 3.2 ppm in remaining 7 patients.

Rama Jayasunder et al, (1999) [Jayasundar R] showed results similar to our study, concluding the presence of lipid as a marker for differentiating tuberculoma from nonspecific inflammatory granuloma and NCC.Since necrotic regions of the lesion can also give rise to signals in lipid region, some caution has to be exercised in interpretation when only a lipid peak is seen from a suspected tuberculoma especially if it is a large one. Another aspect which calls for caution is when lesions are relatively superficial since this would result in contamination from subcutaneous tissue causing resonance from lipids to be in protons spectrum.Our study showed significant DW findings in tuberculoma. Out of the 17 patients, in 14 patients lesions had diffusion restriction (hyperintense) and 3 didn’t give diffusion restriction (hypointense).

Balakrishna et al, (2005) [Balakrishna S] showed peripheral diffusion restriction in tuberculoma. Mishra et al, (2004) [Mishra] reported the same results by concluding restricted diffusion with reduced ADC in tuberculoma. Kamingo M et al, (2002) [Kamingo] studied the diffusion findings in 2 histologically proven tuberculoma and found that bright signal intensity in core of the lesion on DW.

Ozen Kacmaz Basoglu et al (2002) [Ozen] reported the same result in their study on conventional and DW imaging of tuberculomas in concluded multiple ring enhancing lesions with restricted diffusion and on follow-up normal DW and ADC values were obtained.

Our results were contradictory to the results by Vasudev et al, (2007) [MK Vasudev] to conclude in their study as mostly no restriction of diffusion as a diagnostic parameter of tuberculoma. Out of the 3 lesions which didn’t give tuberculoma (hypointense on DW), 2 lesions had hypointense centre on T2; this was probably due to the caseation necrosis seen in tuberculoma.

A Batra et al, (2004) [A Batra] showed similar kind of the results in the study by concluding that those lesions which had hyperintense centre on T2 was hypointense on DW and those lesions which had hypointense centre on T2 was hypointense on DW. There was no relation to the presence of lipid peak on MRS with signal intensity on center of lesion. A lipid peak in MR spectra was seen in both lesions with increased and decreased signal intensity centre as seen on DW or T2 weighted imaging.

The presence of lipid was seen in both lesions with increase and decreased signal intensity in centre which amplifies both solid caseation and liquefaction necrosis in tuberculoma can yield lipid peak on MRS. The presence of lipids peak in liquefied necrosis is related to presence of lipid in tubercular bacilli as well as breakdown products of gray-white matter. While in solid caseation necrosis infiltration with lipid-laden macrophage probably contributes to lipid signal.