An Algorithmic Approach to Intracranial Mass Lesions in HIV/AIDS

By Smego, Raymond A Jr; Orlovic, Dragana; Wadula, Jeanette

Summary: We developed a diagnostic and therapeutic algorithm for intracranial mass lesions in patients with HIV/AIDS that obviates the need for neurosurgical intervention. The approach is based upon CD4+ lymphocyte count, serum toxoplasma immunoglobulin G (IgG) serology, chest X-ray, routine lumbar puncture studies, cerebrospinal fluid (CSF) cytology, CSF adenosine deaminase or Mycobacterium tuberculosis polymerase chain reaction testing, single positron emission-computed tomography (SPECT) scanning for intracranial enhancing lesions, and limited therapeutic trials. Over a 12-month period involving 26 patients, we found that the algorithm correctly identified the aetiology of focal intracranial lesions in all 23 evaluable patients. Costs for SPECT scanning for the entire study cohort were more than offset by the savings achieved by reduced hospital stays for the four patients with lymphoma alone. An algorithmic approach can accurately identify the cause(s) of central nervous system (CNS) mass lesions in HIV-infected patients, and SPECT scanning can replace stereotactic brain biopsy in most cases where opportunistic malignancy is suspected.

Keywords: HIV/AIDS, intracranial mass lesions, opportunistic infections, diagnostic algorithm

Introduction

Neurological signs and symptoms are among the commonest clinical presentations for HIV-infected patients. The detection of intracranial mass lesions on computed tomography (CT) or magnetic resonance imaging (MRI) scans is a frequent and often difficult diagnostic and management problem in patients with HIV/AIDS. Microbiologie agents such as Toxoplasma gondii, Mycobacterium tuberculosis, Cryptococcus neoformans, Treponema pallidum, and others, as well as primary central nervous systems (CNS) lymphoma can produce radiographically indistinguishable lesions on neuroimaging studies.1,2 Determining the specific cause of these lesions is crucial in order to prescribe specific and appropriate therapy, and in optimal situations can usually be accomplished by conducting a structured and systematic laboratory and radiologie work-up. A diagnostic approach that minimizes empiricism and the need for brain biopsy is the most desirable. However, financial constraints in poorer countries often demand presumptive clinical diagnosis and empirical treatment, with observation of the response to therapy. In this article, we report the results of a prospective study designed to determine the aetiology of intracranial mass lesions in patients with HIV/AIDS, and present a simplified diagnostic and therapeutic algorithm for these patients that is relevant and cost effective for both developed and resource-limited countries.

Patients and methods

This prospective clinical study was conducted at the Johannesburg Hospital, a 1700-bed public hospital and one of the three major teaching institutions of the University of the Witwatersrand, and at the Sizwe Tropical Hospital, a 500-bed provincial hospital in Rietfontein, South Africa. Approval for this research was granted by the University Committee for Ethics in Human Research. We developed a diagnostic and therapeutic algorithm for HIV-seropositive patients that included the following features: CD4+ lymphocyte count, serum toxoplasma immunoglobulin G (IgG) serology, chest X-ray, lumbar puncture studies including India ink smear, cryptococcal antigen assay, stain and culture for acid-fast bacilli and fungi, cerebrospinal fluid (CSF) cytology, CSF adenosine deaminase (ADA) determination or polymerase chain reaction (PCR) for M, tuberculosis, serum and CSF syphilis serologies (i.e., rapid plasma reagin, fluorescent treponemal antibody absorption, T. pallidum haemagglutination), single positron emission-computed tomography (SPECT) for selected patients, and limited therapeutic trials (Figure 1). If CSF was not available for cryptococcal antigen testing, a serum cryptococcal antigen test was substituted.

Figure 1 Enhancing lesions on CT or MRI scan. ADA=adenosine deaminase; AFB=acid-fast bacilli; CSF=cerebrospinal fluid; CXR=chest X-ray; MRI=magnetic resonance imaging; PCR=polymerase chain reaction; PET=positron-emission tomography; PML=progressive multifocal leukencephalopathy; Rx=treatment; SPECT=single positron- emission computed tomography (CT); TB=tuberculosis; toxo=toxoplasma; toxo ab=toxoplasma antibodies

From April 1998 through March 1999, 26 HIV-infected individuals were identified with focal intracranial mass lesion(s) on CT scan, with or without surrounding oedema and mass effect, and were prospectively enrolled into the study. A correct diagnosis was assumed if patients demonstrated: (1) substantial radiologie improvement or resolution of lesion(s) on repeat CT scanning, (2) a clinically significant response to presumptive therapy, or (3) characteristic CT (for progressive multifocal leukoencephalopathy [PML]) or SPECT scan features (for lymphoma).

Results

Enrolled study subjects had the following definitive or presumptive aetiologies: tuberculous brain abscess(es) – 14 cases, primary CNS lymphoma – four cases, toxoplasma brain abscesses -four cases, and syphilis, cryptococcal brain abscess, and PML – one case each. One patient (patient 15) with six CNS lesions, a diffuse interstitial chest X-ray pattern, detectable serum toxoplasma antibodies, and a CSF ADA level of 30 U/L responded to treatment directed at both M. tuberculosis and T. gondii and was thought to have CNS tuberculosis and/or toxoplasmosis. A summary of study patients’ clinical, laboratory, and radiographie features is shown in Table 1. Intracranial lesions were solitary in 15 instances and multiple in 11 patients, ranging from two to >10 lesions. For tuberculous abscesses, nine patients had single lesion and five had more than one lesion (mean, three lesions). Toxoplasmosis lesions were multiple in three of four instances (mean, 4.5 lesions). Patients with tuberculous brain abscesses had a mean CD4+ count of 121 cells/mm^sup 3^, while for patients with primary CNS lymphoma or toxoplasmosis the mean CD4+ counts were lower at 61 cells/mm^sup 3^ (P = 0.0001) and 83cells/mm^sup 3^ (P = 0.0001), respectively.

Acute or chronic chest X-ray changes were noted in 11/14 (79%) patients with tuberculous abscesses, compared with 3/4 (75%) patients with toxoplasmosis and 3/4 (75%) with lymphoma. Five subjects with tuberculosis had a diffuse reticulonodular (interstitial) radiographic pattern, and three patients demonstrated unilateral or bilateral infiltrates, pleural effusion, or a pulmonary nodule suggesting active infection. In three individuals the chest radiograph showed only changes of healed, inactive infection (e.g., pleural or parenchymal fibrosis and scarring; calcified thoracic adenopathy). CSF ADA levels were elevated in three of eight (38%) individuals with tuberculous brain abscesses but in none of five patients with a nontuberculous aetiology (P = 0.38). Patient 15 with suspected tuberculosis and/or toxoplasmosis had a high CSF ADA level of 30U/L. Where performed, CSF culture for M. tuberculosis yielded growth in five of seven (71%) persons with suspected tuberculosis. Tuberculous cultures were positive in two additional patients (pleural fluid – patient 3; blood and sputum – patient 9), while one patient with CNS tuberculosis (patient 12) had concomitant pericarditis presumptively due to M. tuberculosis.

Table 1 Laboratory and radiographic features in 26 HIV- seropositive patients with intracranial mass lesions on computed tomography (CT)

Where serum toxoplasma serology was performed, the seroprevalence in this study was 32% (six of 19 patients). Active CNS toxoplasma infection was suspected in four and possibly five of these individuals (one with concomitant detectable CSF toxoplasma titres). Two patients had toxoplasma srologies of uncertain significance: patient 3 had detectable serum and CSF antibodies but had a follow- up CT scan that showed nearly complete resolution of his solitary abscess while receiving antituberculous therapy alone; patient 15, who was treated and had radiologie response to therapy directed at both tuberculosis and toxoplasmosis, had negative serum but detectable CSF toxoplasma antibody titres. Six subjects had reactive serum syphilis srologies, and inactive infection was presumed in five of these individuals. One patient, however, had active neurosyphilis as determined by positive CSF serology as well, and this was believed to be the aetiology of his cerebral infarction seen on CT scan. Patient 21 was given a diagnosis of PML based upon a characteristic CT scan appearance of multiple non-enhancing lesions without associated oedema, located exclusively in cerebral white matter and some in subcortical locales.?tlsb

For each of the four patients with CNS lymphoma, diagnosis was made solely on the basis of a characteristic SPECT scan showing increased 201Th uptake. Strictly according to the diagnostic algorithm, all patients with solitary mass lesions on CT scan should have undergone SPECT scanning. However, because of constrained provincial and hospital budgets these scans were not always performed as suggested (only eight patients had a SPECT scan). Three patients died (patients 1, 11, and 18) without consent given for an autopsy and so a definitive diagnosis could not be made in these instances. The patient with a fungal \brain abscess had his diagnosis definitively made by undergoing stereotactic needle biopsy of the lesion (after failure to respond to four weeks of empiric antituberculous therapy) with resultant visualization of yeast forms and confirmation by cultural growth of C. neoformans. Only after the biopsy was a serum cryptococcal antigen test performed; it revealed a positive result (1:1024 titre) and thus a correct presumptive diagnosis would have been made non-invasively according to the protocol if this test had been ordered earlier.

We tried to determine the potential cost-effectiveness of functional CNS imaging in our hospitals, and found that costs for SPECT scanning (R6 904) for the entire study cohort were more than offset by the savings achieved by reduced hospital stays for the four patients with lymphoma alone (R39 600). A total of 66 hospitals days would have been saved if SPECT scanning had been performed on these patients within the first week of their hospital admission. Costs for inappropriate antimicrobial drug therapy during these hospital days were not determined.

Discussion

We found that a structured and systematic diagnostic algorithm can accurately determine the cause of focal intracranial mass lesions in HIV-infected patients in most instances, obviating the need for neurosurgical intervention and guiding appropriate empiric therapy (Figure 1, Table 2). Using radiologie and/or clinical response to treatment rather than histopathologic confirmation as our diagnostic gold standard, or characteristic radiographie features for PML or CNS lymphoma, we found that our algorithm correctly identified the aetiology of focal intracranial lesions in all 23 evaluable patients. An empiric treatment approach for resource-limited settings where SPECT scan, CSF ADA and M. tuberculosis PCR assays, and other tests are not available is shown in Table 3.

It must be emphasized that caution is advised when considering lumbar puncture for patients with CNS space-occupying lesions. For individuals who demonstrate focal neurologic signs or papilloedema, or those who display significant cerebral oedema or mass effect on neuroimaging studies, lumbar puncture is contraindicated. While stereotactic needle biopsy of the brain can provide valuable diagnostic information in such instances, in all four patients in our study who were unable to undergo spinal tap a correct diagnosis was made without the need for brain tissue sampling.

Absolute lymphopenia (

Table 2 Laboratory investigations in the work-up of HIV-infected patients with CNS mass lesions

Table 3 Therapeutic algorithm for HIV-seropositive patients with enhancing CNS mass lesions, for use in resource-limited areas

Other clues to the aetiology of tuberculous brain abscesses include the geographic occurrence in an area of hyperendemicity, and the presence of an abnormal chest X-ray showing active disease or chronic changes in a majority of cases.7 In the present series, 79% of patients with tuberculous brain abscesses had abnormal chest X- rays. In our series, CSF ADA levels were relatively insensitive in detecting tuberculous brain abscesses but, like CSF M. tuberculosis PCR testing, may serve as adjunctive diagnostic aids in some instances. Although the sensitivity of CSF PCR is good for tuberculous meningitis, the precise sensitivity for focal lesions is unclear because the diagnostic gold standard, CSF culture, can be negative in many cases of CNS tuberculosis.10

Radiographically, primary CNS lymphoma lesions are most often solitary but can occasionally be multiple (usually only two or three).11 They are larger than toxoplasmosis lesions but have identical circumferential ring-enhancement with surrounding cerebral oedema and frequent mass effect. Their location is usually within the periventricular white matter or the basal ganglia. Non-specific CSF abnormalities are seen in approximately 80% of patients with primary CNS lymphoma and include mild lymphocytic pleocytosis and protein elevation, and occasional mild hypoglycorrachia but these are rarely helpful diagnostically. Cytology of CSF can reveal typical malignant cells in up to 25% of cases,12 and is thus a useful diagnostic test. Flow cytometry and immunocytologic staining for B-cell markers may help demonstrate the monoclonal population of these tumors and establish the neoplastic nature of the cells. Additionally, detection of CSF Epstein-Barr virus (EBV) deoxyribonucleic acid (DNA) by PCR can be used as another surrogate marker for lymphoma.13 The functional neuroimaging techniques – positron emissions tomography (PET) and SPECT – can help accurately distinguish between primary CNS lymphoma and infectious lesions like tuberculosis, toxoplasmosis, and PML, and frequently obviate the need for brain tissue sampling.14-17 Lymphoma typically presents as a hypermetabolic or ‘hot’ area on PET or SPECT scan while infectious lesions appear as ‘cold’ lesions with little tracer uptake. While these scans are relatively expensive and not widely available in South Africa, they can, as we have shown, be cost effective by shortening the time until diagnosis and reducing hospital stays and the use of inappropriate antimicrobial therapies.

Knowledge of the background seropositivity of the population is important in accurately interpreting toxoplasma serology in the individual patient with focal CNS mass lesion(s). Detectable IgG antibody in areas of extremely high seroprevalence has less predictive value than in areas of low seroprevalence. It is rare for patients with CNS toxoplasmosis to lack serum IgG; and thus a negative toxoplasma serology has good negative predictive value.1,18 Anti-toxoplasma antibodies are seen in the CSF in only 30-50% of instances and add little to the diagnostic work-up.1 The likelihood of active CNS toxoplasmosis is also significantly reduced by the prior receipt of trimethoprim-sulphamethoxazole prophylaxis. In stable patients with radiographically compatible CNS lesions and detectable serum IgG antibodies, a therapeutic trial of sulphadiazine or clindamycin plus pyrimethamine with observation of the clinical and radiographie response to therapy is recommended.19 On repeat CT or MRI scanning, lesions dramatically decrease in size or disappear after 10-14 days of treatment, thus making brain biopsy generally unnecessary.

References

1 Clough LA, Clough JA, Maciunas RJ, Haas DW. Diagnosing CNS mass lesions in patients with AIDS. AiDS Reader 1997;7:83-8, 93-7, 100

2 Cobb MA, Krist AF. Current concepts in the management of HIV- related intracranial mass lesions. Contemp Neurosurg 1995;17:1-4

3 Jentsch U, Spencer DA. Causes of mortality in HIV-seropositive patients seen at the Johannesburg Hospital, South Africa, 1991-95. 36th South African Congress for Pathology and Microbiology, Pretoria, June 1996

4 Walkden D, Patel J, Snipelski M, Heney C. HIV statistics at Baragwanath Hospital, March 1998 to June 1989. 32nd South African Congress for Pathology and Microbiology, Pretoria, May 1992

5 Whiteman M, Espinosa L, Post MJD, et al. Central nervous tuberculosis in HIV-infected patients: clinical and radiographic findings. Am J Neuroradiol 1995;16:1319-27

6 Berenguer J, Moreno S, Laguna F, et al. Tuberculous meningitis in patients infected with the human immunodeficiency virus. N Engl J Med 1992;326:668-72

7 Sacks L, Thonell L, Pendle S. Tuberculous abscesses of the central nervous system in patients with HIV infection. Sixth Joint Biennial Congress of STD and Infectious Diseases Societies of Southern Africa, Cape Town, South Africa, November 1997

8 Villoria MF, Fortea F, Moreno S, et al. MR imaging and CT of central nervous system tuberculosis in patients with AIDS. Radial Clin North Am 1995;33:805-20

9 Haas DW. Current and future applications of polymerase chain reaction for Mycobacterium tuberculosis. Mayo Clin Proc 1996;71:311- 13

10 Miyazaki Y, Koga H, Kohno S, et al. N\ested polymerase chain reaction for Mycobacterium tuberculosis in clinical samples. J Clin Microbiol 1993;31:2228-33

11 Goldstein JD, Zeifer D, Chao C, et al. CT appearance of primary CNS lymphoma in patients with acquired immuno-deficiency syndrome. J Comput Assist Tomogr 1991;15:39-41

12 So YT, Beckstead JH, Davis RL. Primary central nervous system lymphoma in acquired immune deficiency syndrome, a clinical and pathological study. Ann Neurol 1986; 20:566-72

13 Arribas JR, Clifford DB, Fichtenbaum CJ, et al. Detection of Epstein-Barr virus DNA in cerebrospinal fluid for diagnosis of AIDS- related central nervous systems lymphoma. J Clin Microbiol 1995;33:1580-3

14 Pierce MA, Mahlan D, Maciunas RJ, et al. Evaluating contrast- enhancing brain lesions in patients with AIDS by using positron emission tomography. Ann Intern Med 1995;123:594-8

15 Heald AE, Huffman JM, Barlett JA, Waskin HA. Differentiation of central nervous system lesions in AIDS patients using positron emissions tomography (PET), lnt J STD AIDS 1996;7:337-16

16 Ruiz A, Ganz WI, Post MJD, et al. Use of thallium-201 brain SPECT to differentiate cerebral lymphoma from Toxoplasma encephalitis in AIDS patients. Am J Neuroradiol 1994; 15:1885-94

17 D’Amico A, Messa C, Castagna A, et al. Diagnostic accuracy and predictive value of Th SPECT for the differential diagnosis of cerebral lesions in AIDS patients. Nucl Med Comm 1997;18:741-50

18 Beaman MH, McCabe RE, Wong S-Y, et al. Toxoplasma gondii serology in HIV-infected patients: the development of central nervous systems toxoplasmosis in AIDS. AIDS 1990; 4:519-21

19 Cohn JA, McMeeking A, Cohen W, et al. Evaluation of the policy of empiric treatment of suspected Toxoplasma encephalitis in patients with the acquired immunodeficiency syndrome. Am J Med 1989;86:521-7

(Accepted 1 January 2005)

Raymond A Smego Jr MD FRCP1, Dragana Orlovic MD2 and Jeanette Wadula MD1

1 Department of Clinical Microbiology and Infectious Diseases, University of the Witwatersrand, Johannesburg; 2 Sizwe Tropical Diseases Hospital, Rietfontein, Republic of South Africa

Correspondence to: Dr Raymond A Smego )r, Oman Medical College, PO Box 391, PC 321, AI-Tareef, Sohar, Sultanate of Oman

Email: [email protected]

Presented in part at the 30th IUATLD World Conference on Lung Health, Madrid, Spain. 15-18 September, 1999.

Copyright Royal Society of Medicine Press Ltd. Apr 2006

(c) 2006 International Journal of STD & AIDS. Provided by ProQuest Information and Learning. All rights Reserved.