Management of intracranial tuberculous mass lesions: how long should we treat for? [version 2; peer review: 1 approved, 2 approved with reservations]

Tuberculous intracranial mass lesions are common in settings with high tuberculosis (TB) incidence and HIV prevalence. The diagnosis of such lesions, which include tuberculoma and tuberculous abscesses, is often presumptive and based on radiological features, supportive evidence of TB elsewhere and response to TB treatment. However, the treatment response is unpredictable, with lesions frequently enlarging paradoxically or persisting for many years despite appropriate TB treatment and corticosteroid therapy. Most international guidelines recommend a 9-12 month course of TB treatment for central nervous system TB when the infecting Mycobacterium tuberculosis (M.tb) strain is sensitive to first-line drugs. However, there is variation in opinion and practice with respect to the duration of TB treatment in patients with tuberculomas or tuberculous abscesses. A major reason for this is the lack of prospective clinical trial evidence. Some experts suggest continuing treatment until radiological resolution of enhancing lesions has been achieved, but this may unnecessarily expose patients to prolonged periods of potentially toxic drugs. It is currently unknown whether persistent radiological enhancement of intracranial tuberculomas after 9-12 months of treatment represents active Open Peer Review


v2 Abstract
Tuberculous intracranial mass lesions are common in settings with high tuberculosis (TB) incidence and HIV prevalence. The diagnosis of such lesions, which include tuberculoma and tuberculous abscesses, is often presumptive and based on radiological features, supportive evidence of TB elsewhere and response to TB treatment. However, the treatment response is unpredictable, with lesions frequently enlarging paradoxically or persisting for many years despite appropriate TB treatment and corticosteroid therapy. Most international guidelines recommend a 9-12 month course of TB treatment for central nervous system TB when the infecting Mycobacterium tuberculosis (M.tb) strain is sensitive to first-line drugs. However, there is variation in opinion and practice with respect to the duration of TB treatment in patients with tuberculomas or tuberculous abscesses. A major reason for this is the lack of prospective clinical trial evidence. Some experts suggest continuing treatment until radiological resolution of enhancing lesions has been achieved, but this may unnecessarily expose patients to prolonged periods of potentially toxic drugs. It is currently unknown whether persistent radiological enhancement of intracranial tuberculomas after 9-12 months of treatment represents active disease, inflammatory response in a sterilized lesion or merely

Disclaimer
The views expressed in this article are those of the author(s). Publication in Wellcome Open Research does not imply endorsement by Wellcome.

Introduction
Neurological tuberculosis (TB) manifests as meningitis, radiculomyelitis, bony spinal disease and tuberculoma/tuberculous abscess that may occur intracranially or within the spinal space 1 . Similar to the other neurological TB manifestations, tuberculous mass lesions are common in settings with high TB incidence 2,3 , and high HIV prevalence 4-7 , where this diagnosis accounts for a significant proportion of intracranial space occupying lesions. The diagnosis of intracranial tuberculoma is most often presumptive and based on radiological features, supportive evidence of TB elsewhere and response to TB treatment. However, the treatment response of tuberculomas is unpredictable and lesions may persist for many years despite appropriate TB treatment and adjunctive corticosteroid therapy 8-12 . The optimal duration of TB treatment is unknown and clinical practice varies. In this manuscript we highlight current divergent clinical practice, benefits and disadvantages of different TB treatment durations and the need for prospective clinical trial data to determine the optimal treatment duration in patients with intracranial tuberculous mass lesions.

Pathogenesis and pathology
Hematogenous seeding after the primary infection is one proposed mechanism of central nervous system ( Occasionally, tubercles may coalesce or continue to progress to form a tuberculous abscess, which is a large pus-filled encapsulated lesion containing bacilli 20,21 . Histopathologically, the tuberculous abscess wall shows chronic vascular granulation tissue whilst lacking the granulomatous reaction of a tuberculoma.

Clinical presentation
The clinical features of tuberculomas depend on their anatomic location in the brain, related to local mass effect, obstruction of cerebrospinal fluid pathways, and/or seizures. Supratentorial lesions are common in adults while infratentorial involvement is slightly more common in children 22

Imaging findings
Neuroimaging is essential for identifying intracranial tuberculous mass lesions with findings determined by the composition of the lesion. Tuberculomas have classically been categorized as non-caseating, caseating solid, and caseating liquid, that can be differentiated on computed tomography (CT) and magnetic resonance imaging (MRI) 21 . Multiple lesions are seen more often than isolated lesions though the latter is still common 27,28 . Perilesional edema can be present or absent.
CT is the most frequent modality used to identify tuberculomas due to its wide availability though it has limitations in resolution. Tuberculomas typically appear as round or lobulated nodules that are hypodense or isodense to the brain parenchyma.
CT with contrast most commonly shows rim enhancement of lesions but nodular or homogeneous enhancement can also be seen 12 . The presence of a "target sign" on CT which consists of a rim enhancing lesion with central calcification is highly suggestive of a tuberculoma but uncommon 29 .
MRI is the preferred modality for the identification of tuberculomas due to superior resolution and better visualization of the posterior fossa relative to CT. Non-caseating granulomas are hypointense or isointense on T1-weighted imaging (T1WI) and hyperintense on T2-weighted imaging (T2WI, "T2-bright") with Any further responses from the reviewers can be found at the end of the article REVISED homogeneous contrast enhancement 21 . Caseating solid granulomas are hypointense or isointense on T1WI and hypointense on T2WI ("T2-black") with rim enhancement. Caseating liquid granulomas, which are rare, are hypointense on T1WI and hyperintense on T2WI with rim enhancement. Tuberculous abscesses may be indistinguishable from tuberculomas with a liquid center on standard MRI settings, but they are usually larger (>3 cm in diameter) and thin-walled in appearance 21 . Miliary tuberculomas appear as multiple, small (2-3 mm), scattered lesions that typically rim enhance with contrast administration and lack perilesional edema 30 .
Evidence of a satisfactory radiological response on serial brain imaging after TB treatment initiation includes a reduction in perilesional edema, decrease in lesion size and calcification (seen on CT). Other findings supportive of improvement of liquified tuberculomas and abscesses on MRI are a decrease in T2 brightness and, subsequently, loss of T2 signal. Evolution of TB abscesses from early-stage "T2-bright" with edema to "T2-black" lesions may represent a marker for cure 31 . In our experience, the resultant homogeneous "T2-black" tuberculoma (with rim T1 contrast enhancement) may persist for many months in asymptomatic patients without relapse off TB treatment. CT of such lesions usually shows gradual calcification, which most often involves the capsule.

Paradoxical reactions
Paradoxical enlargement or the development of new intracranial tuberculomas or abscesses in patients with CNS or extraneural TB on appropriate treatment is well-described 8,32-44 . Such reactions typically occur within the first six months after TB treatment initiation 35,41,42,45 , but may rarely be delayed for a year or more 10,45-47 . Paradoxical reactions are often identified when patients present with neurological deterioration during TB treatment, prompting brain imaging. In case series of predominantly HIV-uninfected patients with CNS TB, clinical deterioration due to paradoxical tuberculoma reaction has been described in 6-29% 8,32-39 . However, many of these patients are asymptomatic during these episodes and the frequency of detecting paradoxical tuberculoma development or enlargement increases substantially (from 29% to 65%) if surveillance brain imaging is performed during the first six months of TB treatment 35 . Paradoxical TB reactions are more common in HIV-infected patients, particularly in those who commence antiretroviral therapy (ART) after starting TB treatment, in which case it is referred to as paradoxical TB-immune reconstitution inflammatory syndrome (TB-IRIS) 48-51 . The influence of HIV on the frequency of paradoxical tuberculoma reactions (separate from the effect of ART) has rarely been reported. One recent study of 47 HIV-infected and 14 HIV-uninfected adults with tuberculomas found no difference in the frequency of paradoxical reactions by HIV status (36% in each group) 12 . The majority of HIV-infected patients were receiving ART prior to tuberculoma presentation or did not start ART after diagnosis, precluding the development of TB-IRIS in this group. The pathogenesis of paradoxical reactions (including IRIS) remains unclear but is likely related to an aberrant immune response to TB antigens rather than failure of TB treatment 42,52 . Clinical findings supporting this view are the observation that new or enlarging tuberculomas in TBM patients frequently appear in those known to be infected with drug-susceptible strains who show clinical and radiological improvement of other aspects of TBM (Figure 1) 43 . Another argument is that anti-inflammatory drugs (corticosteroids and thalidomide) are effective in the prevention and management of paradoxical TB reactions, including tuberculomas 43,53-55 .

Medical treatment
The mainstay of treatment of intracranial tuberculomas is similar to that of TBM and includes TB therapy and corticosteroids. The World Health Organization, Centers for Disease Control and Prevention of America and the British Thoracic Society recommend a 9-12 month course of TB treatment for CNS TB when the M.tb strain is sensitive to all drugs 56-58 . However, these guidelines are based on expert opinion rather than randomized controlled trials. Specifically, no studies have compared different treatment durations in patients with intracranial tuberculomas. The morphology of the lesion plays an important role in response to therapy and a one-size-fits-all approach may therefore be inappropriate in the decision regarding tuberculoma treatment duration. This is suggested by the almost invariably good response of miliary tuberculomas to TB treatment (presumably non-caseous) and the frequent persistence of caseous and liquified TB lesions (e.g. abscesses) despite TB treatment 43,59 .
Some guidelines suggest adjunctive systemic corticosteroids in all forms of CNS TB, including those in whom a strong suspicion of tuberculoma exists 58 . Corticosteroid therapy may be of particular value when there is significant perilesional edema (resulting in symptomatology) and in cases where there is paradoxical enlargement despite optimal TB therapy 60 . Corticosteroid duration should be tailored according to the radiological response of the tuberculoma and clinical wellbeing of the patient and balanced against side effects.
TB abscesses are often unresponsive to standard TB therapy with corticosteroids. Although no clinical trials exist, adjuvant thalidomide therapy (3-5 mg/kg/day) has been shown to be beneficial in patients who develop enlarging TB abscesses 43 . In our experience, thalidomide can be stopped without relapse when clinical improvement is optimal or reached a plateau, regardless of whether radiological resolution has been achieved.

Surgical management
There are no controlled studies to determine the role of surgery in patients with intracranial tuberculous mass lesions. However, there are general principles from clinical practice and the existing literature that can be summarized 10,61 . Biopsy for diagnosis is considered: 1) at the outset if the definitive diagnosis is unclear, and 2) for persistence or paradoxical growth of a presumed tuberculoma despite medical treatment (for diagnostics and drug sensitivity testing). Resection of the lesion may be considered: 1) to relieve symptomatic or potentially lifethreatening mass effect and/or hydrocephalus, and 2) to treat medically refractory seizures. Drainage of abscesses is considered for symptomatic mass effect or hydrocephalus, especially when large and/or in the posterior fossa. However, surgery for tuberculous mass lesions is rarely performed in TB endemic settings as the clinical and imaging information is usually sufficient to make the diagnosis. Furthermore, risks associated with surgery, especially if the lesion is located in an eloquent or difficult to access brain area, and inadequate neurosurgical facilities usually combine to preclude surgical management.

Duration of TB treatment: what happens in practice?
There is variation in opinion and practice with respect to the duration of TB treatment in patients with intracranial tuberculomas or tuberculous abscesses. A major reason for this is the lack of prospective clinical trial evidence. In rare cases where a microbiological diagnosis is achieved, it is not feasible to access repeated clinical specimens from the site of disease to ascertain whether and when culture conversion has occurred, unlike pulmonary TB where sputum M.tb culture can be monitored and treatment duration adjusted accordingly. Monitoring is performed clinically and with brain imaging.
The routine duration of TB treatment in intracranial tuberculoma cases include periods of 6 62 , 9 24,32,40,63 , 12 64 , 15 39 and 18 9,10,23,35,65,66 months depending on the clinician's preference. Although some studies describe radiological resolution of tuberculoma in more than 80% of patients after 6-12 months of TB treatment 33,62,64,69,70 , others have reported persistently enhancing lesions in the vast majority (71-82%) of cases after 9-12 months of treatment 23,32 . Even after 24 months of therapy, tuberculomas may persist in 22%-46% of cases 9,12,23 (Figure 1). Larger lesions (>2.5 cm) are significantly more likely to persist after 18 to 24 months of treatment 9,12 . The medical management of patients with persistent intracranial tuberculoma after a "complete treatment course" (6-18 months) is particularly controversial. Some experts suggest continuing treatment until radiological resolution of enhancing lesions has been achieved 23,73 , which may unnecessarily expose patients to potentially toxic drugs for many years 8,9,11,12,23,64,67,74 ; in a study from South Africa, more than 50% of tuberculoma patients followed for 9 months or more (31/57) received TB treatment for more than 18 months (range 19-46 months) 12 . Others are of the opinion that lesional persistence beyond 18 months does not reflect treatment failure, but rather represents a persistent immune response at the disease site that has been sterilized, hence extending TB treatment beyond this period will not add any benefit 75 .

Rationale for using longer versus shorter regimens
It is currently unknown whether persistent radiological enhancement of intracranial tuberculomas after 9-12 months of appropriate treatment represents active disease, inflammatory response in a sterilized lesion or merely revascularization. The consequences of stopping TB treatment prior to complete radiological resolution of these lesions has rarely been explored. These important issues were discussed at the 3 rd International TBM Consortium meeting. Most clinicians were of the opinion that the continued enhancement does not necessarily represent treatment failure and that prolonged TB therapy (beyond 9-12 months) is not warranted in patients suspected of infection with or with proven M.tb strains susceptible to first-line drugs. This position is supported by the asymptomatic The definition varies between studies and include descriptions such as "complete recovery", "no neurological disability", "asymptomatic" and unspecified "good clinical recovery". Several studies included patients with co-existing tuberculous meningitis that might have influenced clinical outcomes.
3 Including 30 patients with available data 4 Including patients followed up for at least 9 months 5 Including patients treated medically without surgical intervention 6 Excluding 1 patient who died during therapy 7 "32" refers to number of meningeal tuberculomas in 25 patients 8 Including patients followed up for at least 12 months state of many patients and the paucity of AFB on staining and sterility of tuberculoma biopsy samples obtained prior to and following TB treatment initiation 23,68 . Immunohistochemical staining of excised tuberculomas also demonstrates high expression of vascular endothelial growth factor (VEGF) in the lesions with intense positivity of inflammatory mononuclear cells as well as reactive astrocytes and fibrocytes 76 . The VEGFinduced angiogenesis in the granuloma capsule may therefore contribute, in addition to inflammation, to the persistent and prolonged contrast enhancement frequently seen on serial brain imaging. Furthermore, one trial reports no clinical or radiological deterioration at 24 months follow-up in 20 patients with persistent intracranial tuberculomas after completion of 9 months' TB therapy 24 .
A theoretical argument in favor of continuing treatment longer than 9-12 months is that drug penetration into the CNS is suboptimal and is likely even more suboptimal into the tuberculoma or tuberculous abscess. Drug penetration into cerebrospinal fluid is poor for rifampicin, the key sterilizing drug 77 . Tuberculous abscesses that, unlike tuberculomas, are teeming with bacilli may potentially act as an immune sanctuary protecting the bacilli from immune effector cells within pus 21 . The consequence of these factors may be that sterilization is not always achieved with 9-12 months treatment and that a longer duration may be required. The inability to obtain specimens to confirm sterilization make this an area of uncertainty. Pertinent, too, is that relapse of CNS TB could have catastrophic consequences. Furthermore, some patients need late re-initiation of immunomodulatory treatment and this should ideally be done while on TB treatment to avoid relapse resulting from iatrogenic immunosuppression. However, if treatment is continued because of residual lesions, when does the clinician stop therapy? Should this be until all contrast enhancing lesions have resolved -which can take years -or some arbitrary timepoint before then?

Conclusion
Intracranial tuberculoma represents a major health concern in developing countries. Routine practices often include prescription of TB therapy until lesional enhancement has resolved, which may expose some patients to an unnecessarily prolonged treatment course. Because of the lack of evidence-based guidelines and equipoise with respect to shorter versus longer duration regimens, further research is needed. In the first instance, a multi-country audit of existing practice and outcomes in terms of cure and relapse would help in defining the spectrum of current practice. Ultimately, a randomized controlled trial comparing a standardized duration of TB treatment with duration based on brain imaging would provide a definitive answer to this question.

Ethics statement
Images presented in Figure 1 were obtained during a retrospective study of patients who presented with intracranial tuberculoma to Inkosi Albert Luthuli Central Hospital in Durban, South Africa. The Biomedical Research Ethics Committee (BREC) of the University of KwaZulu-Natal (UKZN) approved the study (BREC class approval number BCA325/15). As this was a retrospective folder review, and data were analyzed anonymously outside of the clinical setting, the ethics committee of UKZN waived the requirement for informed consent and informed consent was not obtained.

Data availability Underlying data
No data is associated with this article.
The article presents the current clinical approach to the management of tuberculous intracranial mass lesions, which is variable because of limited data that inform treatment interventions that might optimize outcomes.
Overall, the article generally presents a helpful review of the status of this problem and I agree with the recommendations for management, but could be improved by the following: While corticosteroids remain the mainstay of immunomodulatory therapy, other agents besides thalidomide may improve outcomes. There are few reports of other agents improving outcomes (e.g. infliximab -Blackmore (2008) 1 , interferon gamma -Coulter (2007) 2 ), and suggestion that other agents (anakinra) may be beneficial, but the future of immunomodulatory therapy with more targeted agents is an exciting new area. A line or two with references to this handful of case reports as another avenue of investigation would make the article more complete.
Please include recommendations regarding monitoring of patients post treatment (imaging frequency and duration of follow up) -and frequency of recurrent paradoxical response events or symptoms after treatment that are not thought to be treatment failure. Other clinicians with much experience treating CNS TB who have anecdotally describe patients having focal neurological deficits that don't represent relapse of TB, but immune response years after treatment completion. If this is something that the panel of experts have noted, even if infrequent, is worth mentioning.

RESPONSE:
We have added a brief description of some additional immunomodulatory agents that have shown anecdotal benefit in paradoxical TB reactions affecting the CNS and emphasized the need for future studies investigating the utility of host-directed therapies in these patients.
Please include recommendations regarding monitoring of patients post treatment (imaging frequency and duration of follow up) -and frequency of recurrent paradoxical response events or symptoms after treatment that are not thought to be treatment failure. Other clinicians with much experience treating CNS TB who have anecdotally describe patients having focal neurological deficits that don't represent relapse of TB, but immune response years after treatment completion. If this is something that the panel of experts have noted, even if infrequent, is worth mentioning.
leaders in the field. For example, inclusion/exclusion, how to address HIV/ART status, imaging and clinical sub-group analyses, and controlling for corticosteroid use. This would be in the discussion as opposed to the conclusion.

Does the article adequately reference differing views and opinions? Yes
Are all factual statements correct, and are statements and arguments made adequately supported by citations? Yes

Is the Open Letter written in accessible language? Yes
Where applicable, are recommendations and next steps explained clearly for others to follow? Partly Competing Interests: No competing interests were disclosed.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

Author Response 21 Feb 2020
Suzaan Marais, Inkosi Albert Luthuli Central Hospital and University of KwaZulu-Natal, Durban, South Africa In this excellent and thorough overview, the authors accurately summarise the clinical presentation, neuroimaging findings, and pathogenesis of CNS TB, and focus on the data (and lack thereof) to guide the duration of TB therapy.

RESPONSE:
We thank the reviewer for this comment.
The manuscript would be strengthened if, in addition to the serial MRI scans in one patient, a figure was included to demonstrate the specific neuroimaging findings described in the text (Figure allowance permitting). This is of particular interest given that they describe, at least theoretical, rationale for cessation/continuing TB therapy based on the nature of the lesions.

RESPONSE:
We have added additional images showing various stages of intracranial tuberculous mass lesions (Figures 1 and 3).
The manuscript (space permitting) would also benefit from a more descriptive outline of the