An imaging method that reliably differentiates between toxoplasmosis and lymphoma would allow prompt treatment of lymphoma and obviate the need for brain biopsy. Positron emission tomography can measure regional metabolic activity within tissue with a sensitivity and specificity similar to that of radioimmunoassay when used with specific labeled substrates [10]. By measuring the accumulation of fluorodeoxyglucose F-18, it can distinguish malignant brain tumors from surrounding normal tissue [11]. Experience at our institution and elsewhere [12] indicates that positron emission tomography can accurately differentiate between lymphoma and nonmalignant causes of central nervous system lesions in patients with AIDS.

We did a study to determine whether evaluating contrast-enhancing central nervous system lesions in patients with AIDS by using Toxoplasma serologic testing and positron emission tomography can accurately differentiate between lymphoma and nonlymphoma diagnoses.

Methods

Top Methods Results Discussion Author & Article Info References

Our prospective clinical study was approved by the Vanderbilt University Committee for the Protection of Human Subjects. All patients with AIDS at Vanderbilt University Medical Center who had contrast-enhancing central nervous system lesions shown by computed tomography or magnetic resonance imaging between May 1991 and March 1993 were eligible. Patients were excluded only for failure to give informed consent. Eligible patients received a history; a physical examination; a computed tomographic or magnetic resonance imaging scan, or both; a positron emission tomographic scan; a Toxoplasma serologic test; a cerebrospinal fluid examination (if clinically indicated); a serum cryptococcal antigen test; and a CD4⁺ T-cell count. Autopsy consent was obtained prospectively to enable us to confirm diagnoses that remained in question.

Computed tomography was done on a Siemens Somatom Plus (Siemens Medical Systems, Iselin, New Jersey) using 10-mm thick transaxial slices before and after infusion of intravenous contrast material. Magnetic resonance imaging was done on a 1.5-Tesla Siemens Magnatom SP (Siemens Gammasonics, Hoffman Estates, Illinois) using standard T₁-weighted, proton density, and T₂-weighted sequences before contrast and T₁-weighted images after infusion of gadolinium-diethylenetriamine pentaacetic acid. Positron emission tomographic scans were obtained using a Siemens emission computed axial tomograph 933-08-16 scanner (Siemens Gammasonics) with inplane resolution of 6.5 mm and axial resolution of 8 mm. This scanner provides 15 simultaneously acquired tomographic frames over a 12.8-cm view. Images were obtained 45 minutes after the administration of 10 mCi of 2-fluorodeoxyglucose F-18 using two bed positions, 30 scans each, 4 mm apart. All patients had transmission scans for a measured attenuation correction. Counts were normalized to a homogeneous mixture of brain tissue. Lesions were considered hypometabolic if their measured activity was roughly equal to or less than the activity of white matter. Hypermetabolic lesions had at least 1.5 times the activity of white matter. (Figures 1 and 2) show the magnetic resonance and positron emission tomographic images of two patients in our study. Figure 1 is taken from patient 4, who had toxoplasmosis; Figure 2 is taken from patient 12, who had lymphoma. The magnetic resonance imaging scans are nearly identical in appearance, but the positron emission tomographic scans show hypometabolic and hypermetabolic lesions, respectively.

View larger version (72K): [in this window] [in a new window]   Figure 1. Central nervous system toxoplasmosis associated with the acquired immunodeficiency syndrome. Top. A typical ring-enhancing mass lesion of toxoplasmosis with surrounding edema in the inferior left temporal lobe is seen in this coronal, gadolinium-enhanced, T1-weighted magnetic resonance image. Bottom. Coronal positron emission tomographic image showing decreased glucose metabolism in the region of the enhancing lesion shown in the top panel.

View larger version (59K): [in this window] [in a new window]   Figure 2. Lymphoma associated with the acquired immunodeficiency syndrome. Top. A ring-enhancing mass lesion in the right cerebellar hemisphere on this axial, gadolinium-enhanced, T1-weighted magnetic resonance image, which has the same appearance as the image in Figure 1, top, but is due to lymphoma. Bottom. A similar positron emission tomographic image showing a ring of increased glucose metabolism in the same region as the lesion in the top panel.

Toxoplasma (IgG) serologic testing was done by immunofluorescent assay, and results were considered positive if the titer was 1:4. Autopsies were done by one of us using standard human immunodeficiency virus (HIV) precautions. After 2 weeks fixation in 10% formalin, lesions were evaluated with hematoxylin-eosin, Giemsa, Gomori methenamine silver, periodic acid-Schiff, and methyl green-pyronine stains and appropriate immunohistochemistry.

Treatment

The therapeutic decisions for each patient were ultimately made by the attending physicians, but therapy was recommended by the protocol on the basis of the schema given in Table 1. Patients with positive Toxoplasma serologic test results and positron emission tomographic scans that showed hypometabolic lesions were assumed to have toxoplasmosis, and antitoxoplasmosis therapy was recommended. Negative serologic test results and a positron emission tomographic scan showing hypermetabolic lesions were considered indicative of lymphoma, and empiric cranial irradiation was recommended. Patients with positive serologic test results and positron emission tomographic scans that showed hypermetabolic lesions were considered to have lymphoma, and empiric cranial irradiation was recommended. Empiric antitoxoplasmosis therapy was allowed in this situation at the discretion of the attending physician. Patients with negative serologic test results and a positron emission tomographic scan showing hypometabolic lesions were considered to be candidates for brain biopsy if an alternative diagnosis was not established. Although this schema was developed to guide therapy, in some cases the actual treatment received was different from that recommended. For example, patient 9 Table 2 had positive serologic test results and hypermetabolic lesions on a positron emission tomographic scan. The protocol would have allowed both antitoxoplasmosis therapy and radiation therapy. In fact, the patient received only radiation therapy. Similarly, patient 12 Table 2 had hypermetabolic lesions on a positron emission tomographic scan and negative serologic test results. He received a brain biopsy even though this was not recommended by the protocol. The differences between the recommended treatment and the actual treatment received do not bias the results because confirmation of diagnoses was dependent on the actual treatment received.

Patients were followed longitudinally, and a repeated computed tomographic or magnetic resonance imaging scan, or both, was obtained after approximately 2 weeks and then at varying intervals, depending on clinical response. Clinical and radiologic findings were evaluated to determine response to therapy. Diagnoses were confirmed by clinical and radiologic improvement or resolution in response to a single type of therapy (such as antibiotics compared with radiation); by autopsy; or, in one case, by brain biopsy. Patients who had clinical and radiologic improvement in response to more than one type of therapy were only considered to have a confirmed diagnosis if it was confirmed by autopsy or brain biopsy.

Results

Top Methods Results Discussion Author & Article Info References

Twenty patients with focal, contrast-enhancing central nervous system lesions were enrolled in the study between May 1991 and June 1993. Two patients were not evaluable because their diagnoses could not be confirmed. Table 2 lists the confirmed diagnoses and major diagnostic findings in the 18 evaluable patients.

Confirmation of Diagnoses

Eight patients had toxoplasmosis; six of these diagnoses were confirmed by clinical and radiologic response to therapy and two were confirmed by autopsy. Six patients had lymphoma: One diagnosis was confirmed by clinical and radiologic response, one by stereotactic biopsy, and four by autopsy. Two patients with progressive multifocal leukoencephalopathy and one with a cerebral cryptococcoma had their diagnoses confirmed by autopsy. One patient had clinical and radiologic responses to therapy directed at both toxoplasmosis and tuberculosis; he remains alive at 9 months.

Results of Serologic Testing and Positron Emission Tomographic Scanning

Toxoplasma serologic testing was done in seven of the eight patients with toxoplasmosis, and all results were positive (range, 1:10 to 1:32 768). Three of the six patients with lymphoma also had positive serologic test results.

Seven patients with toxoplasmosis each had a positron emission tomographic scan, and all scans showed hypometabolic lesions. In contrast, the six patients with lymphoma all had hypermetabolic lesions on positron emission tomographic scans. The difference between these two sets of results was highly significant (P < 0.001, Fisher exact test, two-tailed). Four patients had other diagnoses. Patient 17 had a cryptococcoma, negative Toxoplasma serologic test results, and hypometabolic lesions on a positron emission tomographic scan. Two patients had progressive multifocal leukoencephalopathy; the positron emission tomographic scans showed hypometabolic lesions in one and hypermetabolic lesions in the other. In patient 18, who had clinical and radiologic responses to therapy directed at both toxoplasmosis and tuberculosis, the positron emission tomographic scan showed hypometabolic lesions.

Discussion

Top Methods Results Discussion Author & Article Info References

Cerebral toxoplasmosis is much more common than primary central nervous system lymphoma in most series [2, 13], but evidence indicates that this lymphoma is increasing [14]. The reasons for this are not entirely understood. In populations such as ours, which have a low prevalence of persons seropositive for Toxoplasma (12.5% of the 350 patients with HIV infection followed at the Vanderbilt Infectious Disease Clinic [unpublished data]), the number of cases of primary central nervous system lymphoma may equal the number of cases of cerebral toxoplasmosis. The present standard of care for patients who have AIDS, contrast-enhancing central nervous system lesions, and positive results on a toxoplasmosis serologic test is to administer empiric antitoxoplasmosis therapy for 10 to 21 days and to consider an alternative diagnosis in patients who do not respond to this therapy [6, 7]. In centers where the prevalence of positive Toxoplasma serologic test results is similar to that in ours, this approach necessarily delays the correct diagnosis and treatment of the almost 50% of patients who have contrast-enhancing central nervous system lesions due to lymphoma.

The median survival of patients with AIDS and cerebral toxoplasmosis is 8 months; the median survival of patients with AIDS and central nervous system lymphoma is 3 to 4 months [15]. Our experience and that of others [7-9] indicate that the baseline characteristics of patients with cerebral toxoplasmosis and central nervous system lymphoma are similar; thus, differences in survival probably do not stem from differences in patients at risk. Evidence indicates that patients with primary central nervous system lymphoma who receive whole-brain irradiation have substantially longer median survival times (119 days) than patients who do not receive treatment (42 days a). Treated patients tend to die of opportunistic infections rather than progressive lymphoma. We assume that most patients with cerebral toxoplasmosis receive prompt antitoxoplasmosis therapy, whereas current practice necessitates considerable delay in the treatment of those with lymphoma. Earlier therapy directed at lymphoma may improve both response and length of survival.

In our investigation, positron emission tomographic scanning accurately differentiated lymphoma from nonlymphoma diagnoses in 17 of 18 patients with AIDS and central nervous system mass lesions. In one patient, the positron emission tomographic scan showed hypermetabolic lesions because of progressive multifocal leukoencephalopathy rather than lymphoma. We have now studied four patients with autopsy-proven progressive multifocal leukoencephalopathy, and the positron emission tomographic scans have shown hypermetabolic lesions in two of these patients. This finding has also been noted by Hoffman and colleagues [12]. The reasons for this finding are unknown but may involve the enhanced use of fluorodeoxyglucose by infiltrating macrophages. In general, this will not be an important confounding factor because most of the lesions caused by progressive multifocal leukoencephalopathy are not contrast enhancing. No accepted therapy is currently available for progressive multifocal leukoencephalopathy.

This protocol markedly reduced the need for stereotactic brain biopsy. Although this was not specifically addressed in our study, the reduction in the number of brain biopsies required and the attendant reduction in hospital stay has obvious potential for saving costs. Only one patient (patient 10) required biopsy. He had positron emission tomographic findings consistent with lymphoma, but a concomitant undiagnosed pulmonary infiltrate made his primary physicians uncomfortable with empiric brain irradiation. Stereotactic biopsy done after a nondiagnostic bronchoscopy confirmed the presence of intracranial lymphoma.

Patients with AIDS and primary central nervous system lymphoma currently have poor long-term survival. Stereotactic biopsy can readily be done in these patients, but this surgical procedure is not without risk, and avoiding it without compromising care is a worthwhile goal. We have shown that evaluating contrast-enhancing central nervous system lesions in patients with advanced HIV infection by using Toxoplasma serologic testing and positron emission tomography can accurately guide therapy and obviate the need for most brain biopsies. This approach to the diagnosis of contrast-enhancing brain lesions in patients with AIDS has become standard in our institution and, if applicable to other centers, represents an advance in the management of primary central nervous system lymphoma. A larger, national, multicenter study is needed to confirm these findings and to determine the effect of earlier diagnosis and treatment on morbidity and mortality in patients with AIDS and primary central nervous system lymphoma.

Dr. Johnson: Division of Neuropathology, Vanderbilt University Medical Center, 1161 21st Avenue South, C-3321 Medical Center North, Nashville, TN 37232-2561.

Drs. Maciunas and Allen: Department of Neurosurgery, Vanderbilt University Medical Center, 1161 21st Avenue South, T-4224 Medical Center North, Nashville, TN 37232-2380.

Dr. Murray: E. C. Green Cancer Center, 1717 High Street, Hopkinsville, KY 42240.

Dr. Harbison: St. Thomas Hospital, 4230 Harding Road, Plaza West, Suite 303, Nashville, TN 37205.

Dr. Creasy: Division of Radiological Sciences, Vanderbilt University Medical Center, 1211 22nd Avenue South, 907 Vanderbilt University Hospital, Nashville, TN 37232-2675.

Dr. Kessler: Division of Radiological Sciences, Vanderbilt University Medical Center, 1211 22nd Avenue South, 1270 Medical Research Building, Nashville, TN 37232-6315.