99mTc-hexamethylpropyleneamine oxime leukocyte scintigraphy and C-reactive protein levels in the differential diagnosis of brain abscesses

Transcription

1 J Nenrosurg 77: , mTc-hexamethylpropyleneamine oxime leukocyte scintigraphy and C-reactive protein levels in the differential diagnosis of brain abscesses IVAR AMUND GRIMSTAD, M.D., PH.D., HENaY HIRSCHBERG, M.D., PH.D., AND K JELL ROOTWELT, M.D., PH.D. Department of Clinical Chemislry, Institute of Clinical Biochemistry, Section of Nuclear Medicine, and Department of Neurosurgery, Rikshospitalet, The National Hospital, Oslo, Norway u- The demonstration and accurate localization of intracerebral mass lesions are commonly performed with computerized tomography (CT), which often cannot determine the nature of the lesion. As an aid in the differential diagnosis between brain abscess and neoplasm, the authors have evaluated both 99mTc-hexamethylpropyleneamine oxime (99mTc-HMPAO) leukocyte scintigraphy and the serum C-reactive protein level. Of 23 patients with intracranial mass lesions, 22 individuals showed ring-like contrast enhancement on CT scans', the one exception was a patient treated for a meningioma who had a negative CT scan despite clinical suspicion of intra- or extracranial abscess. The final diagnosis was invariably established by microscopic examination of tissue specimens. In 10 patients the final diagnosis was brain abscess; the other 13 patients harbored a brain neoplasm (glioma in nine, astrocytoma in one, and metastasis in three). The ~9~"Tc-HMPAO leukocyte scintigraphy detected all cases of abscess. There were no false-positive results. An elevated C-reactive protein level (> 13 mg/liter) was found in all but one patient with abscess and in three patients with neoplasm; two of these three patients had dental root infections which could account for the elevation of C-reactive protein. It is concluded that 99mTc-HMPAO leukocyte scintigraphy should be performed when there is a possibility that a brain abscess may exist. Any steroid treatment should be discontinued for 48 hours prior to leukocyte scintigraphy. Also, C-reactive protein determination should be performed and is useful even when steroids are given. KEY WORDS abscess brain neoplasm 9 C-reactive protein. isotope labeling 9 leukocyte 9 radionuclide 9 technetium T HE differential diagnosis of intracerebral mass lesions remains difficult despite the great value of computerized tomography (CT). Many brain tumors 22 and occasionally infarction ~9 and hematoma 4 can mimic the CT appearance of a brain abscess. ~''~ Therefore, supplementary diagnostic modalities are required to establish the nature of the lesion so that adequate therapy can be given promptly. 722 In acute or chronic inflammation, C-reactive protein is synthesized by the liver in increased amounts. The concentration of C-reactive protein in plasma has been found useful in the differential diagnosis of brain abscesses. 7 However, inflammation elsewhere would also raise the C-reactive protein level, although no such cases were identified in our previous study. 7 Leukocyte scintigraphy involves labeling autologous leukocytes in vitro with a radioactive isotope and reinjecting them into the patient so that inflammatory foci where leukocytes accumulate may be identified by scintigraphy. One label used for this purpose is '~ qn-oxine/ tropolonate, which has proven useful in the diagnosis of brain abscesses. 2'15'1~'2~ However, ~' qn has disadvantages related to radiation dose, image quality, cost, and availability (it is a cyclotron product with a halflife of 2.5 days). A newer agent, 9gmTc-hexamethylpropyleneamine oxime (99mTc-HMPAO), was recently introduced for studying cerebral blood flow 5'~3 and was subsequently found to be very well suited for leukocyte scintigraphyj 7'2~ The advantage of 99mTc-HMPAO-labeled leukocyte scintigraphy is that all necessary reagents can be in stock and readily available in any nuclear medicine department. Thus, scintigraphy can be provided whenever needed. Moreover, the near ideal physical properties ef the 99mTC label give higher photon flux, superior resolution, reduced radiation dose, and better image quality with current gamma cameras than '~ tin. 732 J. t~2,urosurg. / Volume 77/November, 1992

2 99mTc-HMPAO leukocyte scintigraphy of brain abscesses The present study was undertaken to determine the usefulness of 9~ leukocyte scintigraphy in the differential diagnosis of patients harboring intracranial mass lesions. This method proved to be both sensitive and reliable. Clinical Material and Methods Patient Population A total of 23 consecutive patients, 11 females and 12 males ranging in age from 15 to 72 years, were included in the study. With one exception, an intracranial mass lesion with ring-like contrast enhancement was seen on each patient's CT scan. Clinical findings were compatible with brain abscess or neoplasm. Many of the patients had received various antibiotic agents or steroid therapy (dexamethasone, maximum dose 16 mg/day) prior to evaluation with 99mTc-HMPAO leukocyte brain scintigraphy. Whenever steroids had been used, this treatment was discontinued for 48 hours prior to leukocyte scintigraphy. Leukocyte Scintigraphy Autologous mixed leukocytes were isolated and labeled in vitro by 99mTc-HMPAO as recommended by the manufacturer.* The labeling yield was approximately 55%. Leukocytes labeled with 250 MBq 99mTc- HMPAO were reinjected intravenously. Starting 4 hours after reinjection of the labeled cells, 5- to 10- minute planar images of the head were obtained in anterior, posterior, and both lateral views on a gamma camerat equipped with a low-energy high-resolution parallel-hole collimator using a 20% window centered around the 140-keV photopeak of 99rnTC. The scintigrams were assessed visually; no quantitative criteria were employed. Studies that demonstrated focally increased activity were considered positive for infection. Serum C-Reactive Protein Measurement Levels of C-reactive protein in blood sera were measured by immunoturbidimetry using an analyzer and reagent kit.z~ At a C-reactive protein concentration of 36.3 mg/liter, the standard deviation of the method was 2.0 mg/liter (including variation between series); thus, the coefficient of variation was 5.5%. In a previous study of similar patients with brain abscesses or tumors, we have found that no patient with intracranial neoplasm (and without abscess) had a C-reactive protein level of over 13 rag/liter. 7 Therefore, this value was chosen as the discriminatory level between brain tumor and abscess. * Ceretec obtained from Amersham International, Buckinghamshire, England. t Gamma Maxicamera, Model 300 Autotune ZS, manufactured by General Electric Co., Milwaukee, Wisconsin. lmmunoturbidimetry analyzer, Model RA-1000, manufactured by Technicon Instrument Corp., Tarrytown, New York; reagent kit obtained from Orion Diagnostica, Espoo, Finland. FIG. 1. Computerized tomography scan w~th contrast enhancement (left) and a leukocyte scintigram in the posteroanterior projection (right) from a patient with brain abscess. The lesion in the right occipital part of the brain is seen on both studies. Final Diagnosis A specimen for cytological or histological examination was obtained by stereotactic puncture or at surgery. This gave the final diagnosis in all patients. Final Diagnosis Results The final diagnosis was brain abscess in 10 of the 23 patients studied. In one of these 10 patients (described below), the abscess was colocalized with a glioblastoma multiforme. Since the primary purpose of the present study was assessment of 99mTc-HMPAO leukocyte scintigraphy (and C-reactive protein levels) in the differential diagnosis of brain abscesses, this patient was classified in the abscess group. The 13 remaining patients had a malignant brain tumor without any abscess. There were nine cases of glioma (eight glioblastomas multiforme and one gliosarcoma), one case of anaplastic astrocytoma, and three cases of cancer metastases. Leukocyte Scintigraphy and C-Reactive Protein Value Figure 1 shows a contrast-enhanced CT scan and a 9*mTc-HMPAO leukocyte scintigram from a patient with brain abscess. Figure 2 shows similar images from a tumor patient. The positive values of C-reactive protein ranged from 15 mg/liter to more than 200 mg/liter. In one patient the C-reactive protein value was 112 mg/liter and a brain abscess was diagnosed by 99mTc-HMPAO leukocyte scintigraphy. Stereotactic puncture confirmed the presence of pus; however, cavity loading with gentamicin and systemic antibiotics was not followed by clinical improvement according to expectation and, at surgery 2 weeks later, a glioblastoma multiforme was found in the same location. Another patient had been surgically treated for a parietal calcified meningioma. From the 7th postoperative day she experienced fever, local swelling, and pain, J. Neurosurg. / Volume 77/November,

3 I. A. Grimstad, H. Hirschberg, and K. Rootwelt TABLE 1 C-reaclive protein level~' in the d([l~,rential diagnosis of brain abscess FIG. 2. Computerized tomography (CT) scan with contrast enhancement (left) and a leukocyte scintigram in the posteroanterior projection (right) from a patient with brain tumor. The lesion in the left occipital part of the brain is seen only on the CT scan. and an abscess was suspected. Apart from postoperative changes, no positive findings were made on CT. On Day 13 the C-reactive protein value was 160 mg/liter and a 9~Tc-HMPAO leukocyte scintigraphy showed an extracraniai abscess at the operation site. This was verified by the aspiration of pus. The performance of the C-reactive protein test (assuming that a level over 13 mg/liter means that the patient's lesion is an abscess) is shown in Table 1 and the findings of the 99~Tc-HMPAO leukocyte scintigraphy are shown in Table 2. Three patients had a falsely positive C-reactive protein test with respect to the presence of a brain abscess, and one brain abscess patient had a falsely negative C-reactive protein test. No mistakes in diagnosis were made with 99mTc-HMPAO leukocyte scintigraphy. None of the five parameters assessed (Tables 1 and 2) showed a statistically significant difference between the C-reactive protein test and the 99~'Tc-HMPAO leukocyte scintigraphy (p _ 0.11, Fisher's exact test, two-tailed). Discussion The presence and location of intracerebral mass lesions can be established with CT. However, the nature of the lesions frequently remains unknown, and accurate preoperative diagnosis is important in these patients. Preoperative steroid therapy, appropriate in the case of a suspected brain tumor, is not desirable in the management of a brain abscess, which should be treated preoperatively with antibiotics. Furthermore, the timing of surgical intervention and the procedure employed, and even the decision whether to operate depend on the preoperative diagnosis. Therefore, noninvasive, reliable, and commonly accessible methods of differential diagnosis are necessary for optimum treatment. C-Reactive Protein Our experience (data not given) agrees with previous C-Reactive Protein Factor Value (mg/liter) Totals > 13 _< 13 diagnosis abscess tumor I 0 13 sensitivity specificity 9/I 0 (90%) 10/13 (77%) predictive value positive finding 9/12 (75%) negative finding 10/11 (91%) diagnostic efficiency 19/23 (83%) TABLE 2 '~'ntc-hmpao leukoc.vte scintigraphy in the differential diagnosis (?[brain abscess* 9~mTc-HMPAO Leuko- Factor cyte Scintigraphy TotaLs Positive Negative diagnosis abscess tumor sensitivity 10/10 (100%) specificity predictive value 13/13 (100%) positive finding 10/10 (100%) negative finding 13/13 (100%) diagnostic efficiency 23/23 (100%) *'~mtc-hmpao = technetium-99m-hexamethylpropyleneamine oxime. findings that body temperature, white cell count, or erythrocyte sedimentation rate are not useful in determining the nature of intracerebral mass lesions. On the other hand, the present study confirms our previous observations that the C-reactive protein level is very useful for this purpose. However, three patients with brain tumor had an elevated C-reactive protein level. Among these three patients, two had dental root infections which could account for the C-reactive protein elevations. Why the C-reactive protein level was elevated (although only marginally) in the third patient is not known. There was one case of falsely negative C- reactive protein, but the reason is not known. Leukocyte Scintigraphy The basis of 99~Tc-HMPAO leukocyte scintigraphy is the accumulation of radioactively labeled leukocytes in inflammatory foci. This accumulation is inhibited in a dose-dependent manner by glucocorticoids/steroids have been found to interfere with many processes involved, as they reduce granulocyte adherence, TM the production of chemotactic factors by macrophages, ~6 the binding of chemotactic factors to granulocytes, ~3 734 J. Neurosurg. / Volume 77/November, 1992

4 99mTc-HMPAO leukocyte scintigraphy of brain abscesses the release of granulocyte activators from granulocyte granules,~4 and the endothelial permeability? For these reasons, steroid administration was discontinued for 48 hours prior to leukocyte scintigraphy. The labeling procedure employed tags both granulocytes and mononuclear cells. In theory, pure granulocyte labeling would be optimum since granulocytes accumulate more rapidly in inflammatory foci. Since granulocytes normally occur in greater numbers in the circulation and seem to be preferentially labeled by 99mTc-HMPAO, granulocyte isolation is by no means mandatory. Isolation and labeling of pure granulocytes add time to the labeling procedure and increase the risk of neutrophil activation. Granulocyte separation therefore does not seem worthwhile and has, in general, not increased the detection rate of occult inflammatory processes. The main disadvantage of 99mTc-HMPAO-labeled leukocytes compared with ~ qn-oxine/tropolonate-labeled leukocytes is that the 99mTc label is eluted from the leukocytes at a rate of approximately 5% 9 hour -~, whereas the ~In label is not eluted. The eluted 99rnTC activity is carried on a hydrophilic metabolite that is unable to pass the normal blood-brain barrier, but has the potential to accumulate in intracranial mass lesions whether or not these are of an infectious nature. Therefore, 99mTc'HMPAO leukocytes might not be as effective as ~ qn-oxine/tropolonate leukocytes for discriminating between abscesses and tumors. However, the 99mTc-HMPAO method gave a diagnostic accuracy of 100%, and the elution of labeled ieukocytes was thus not important in practice. Initially we performed tomographic registration in addition to obtaining planar images. Tomography has the advantage of better sensitivity than planar registration in the case of deeply situated lesions. The sensitivity of tomography, however, also creates problems because the slight leukocyte accumulation commonly seen in neoplasms is frequently detected by tomography and confused with abscesses. For this reason, we do not routinely use tomography; only planar scintigrams were employed in the present evaluations. Comparison With mln Leukocyte Scintigraphy In the present study, there were no false-positive or false-negative cases of 9~mTc-HMPAO leukocyte scintigraphy, and the localization of the lesions agreed well with the CT findings. The extraordinary diagnostic accuracy of 100% is probably a chance finding that is bound to be reduced with future clinical experience. Nevertheless, our results compare favorably with similar studies performed with ~qn-labeled leukocytes. Rehncrona, et al., z~ studied 16 patients with intracerebrat lesions and inconclusive CT scans; their final diagnoses were five abscesses and 11 tumors. They reported a sensitivity, specificity, and accuracy of labeled leukocyte imaging of 80%, 91%, and 88%, respectively. Bellotti, et al., 2 found the sensitivity, specificity, and accuracy of labeled leukocyte imaging to be 100 %, 94 %, and 96%, respectively. Schmidt, et al., 22 found 100% sensitivity in seven patients with abscesses (all off steroid therapy), whereas the specificity was a poor 60% ( 12 of 20 cases). All abscesses showed intense accumulation, whereas the scintigraphically positive tumors showed only weak or moderate uptake. Balachandran, et al., ~ also reported poor specificity with uptake of labeled leukocytes in four (57%) of seven cases of intracerebral tumors, all of them metastatic. The uptake of labeled leukocytes in acute cerebral infarction/9 as well as in an acute subdural hematoma, 4 has also been reported. The best results with ~l qn leukocyte scintigraphy have been obtained by Palestro, et al., ts who retrospectively reviewed 16 patients with contrast-enhanced intracerebral lesions identified on CT scans and found two positive labeled leukocyte studies; both were cerebral abscesses. No labeled leukocyte activity was identified in primary brain tumors (six cases), metastases (four cases), dermoid cyst (one case), or cerebral infarcts (three cases). Conclusions We recommend prompt 99mTc-HMPAO leukocyte scintigraphy whenever there is a suspicion from clinical or CT findings that the patient may have a brain abscess, provided the patient has been off suppressive steroid doses for 48 hours. Determination of the serum C-reactive protein level is also useful in these patients; this test can be performed immediately even in steroidtreated patients and should be done in addition to leukocyte scintigraphy. References 1. Balachandran S, Boyd CM, Husain MM, et al: Indium- 111 leukocyte uptake in neoplasms. Clin Nucl Med 12: , Bellotti C, Aragno MG, Medina M, et al: Differential diagnosis of CT-hypodense cranial lesions with indium- I I 1-oxine-labeled leukocytes. J Neurosurg 64: , Clark RAF, Gallin JI, Fauci AS: Effects of in vivo prednisone on in vitro eosinophil and neutrophil adherence and chemotaxis. Blood 53: I, Dudiak CM, Ali A, Dickerson M, et al: Acute tentorial subdural hematoma as a false-positive in indium- l 11 leukocyte scintigraphy. Clin Nucl Med 10: , Ell PJ, HockneU JML, Jarrit PH, et al: A 99Tcm-labelled radiotracer for the investigation of cerebral vascular disease. Nucl Med Commun 6: , Fauci AS, Dale DC, Balow JE: Glucocorticosteroid therapy: mechanisms of action and clinical considerations. Ann Intern Med 84: , Hirschberg H, Bosnes V: C-Reactive protein levels in the differential diagnosis of brain abscesses. J Nenrosurg 67: , Holt,s S, T6rnquist C, Cronqvist S: Diagnostic difficulties in computed tomography of brain abscesses. J Comput Assist Tomogr 6: , Kadurugamuwa JL, Hengstler B, Zak O: Effects of antiinflammatory drugs on arachidonic acid metabolites and cerebrospinal fluid proteins during infectious pneumococcal meningitis in rabbits. Pediatr Infect Dis 6: t54, 1987.I. Neurosurg. / Volume 77/November,

5 I. A. Grimstad, H. Hirschberg, and K. Rootwelt 10. Kandalaft N, Diehl J, Neuwelt EA: Nonneoplastic intracranial lesions simulating neoplasms on computed tomographic scan. Excellent sensitivity with limited specificity. JAMA 248: , MacGregor RR, Spagnuolo P J, Lentnek AL: Inhibition of granulocyte adherence by ethanol, prednisone, and aspirin, measured with an assay system. N Engl J Med 291: , Masucci EF, Sauerbrunn BJ: The evolution of a brain abscess the complementary roles of radionuclide (RN) and computed tomography (CT) scans. Clin Nucl Med 7: , Nowotnik DP, Canning LR, Cumming SA, et al: Development of a ~Tcm-labelled radiopharmaceutieal for cerebral blood flow imaging. Nucl Med Commun 6: , Oseas RS, Allen J, Yang HH, et al: Mechanism of dexamethasone inhibition of chemotactic factor induced granulocyte aggregation. Blood 59: , Palestro CJ, Swyer AJ, Kim CK, et al: Role of In-lll labeled leukocyte scintigraphy in the diagnosis of intracerebral lesions. Clin Nucl Med 16: , Pennington JE, Harris EA: Influence of immunosuppression on alveolar macrophage chemotactic activities in guinea pigs. Am Rev Respir Dis 123: , Peters AM, Danpure H J, Osman S, et al: Clinical experience with 99mTc-hexamethylpropylene-amineoxime for labelling leukocytes and imaging inflammation. Lancet 2: , Peters AM, Lavender JP, Macdermot J: Diagnosing cerebral abscess with indium-111 labelled leucocytes. Lancet 2: , 1980 (Letter) 19. Pozzilli C, Lenzi GL, Argentino C, et al: Imaging of leukocytic infiltration in human cerebral infarcts. Stroke 16: , Rehncrona S, Brismar J, Holtfis S: Diagnosis of brain abscesses with indium-11 l-labeled leukocytes. Neurosurgery 16:23-26, Roddie ME, Peters AM, Danpure H J, et al: Inflammation: imaging with Tc-99m HMPAO labeled leukocytes. Radiology 166: , Schmidt KG, Rasmussen JW, Fredriksen PB, et al: Indium-11 l-granulocyte scintigraphy in brain abscess diagnosis: limitations and pitfalls. J Nucl Med 31: , Skubitz KM, Craddock PR, Hammerschmidt DE, et al: Corticosteroids block binding of chemotactic peptide to its receptor on granulocytes and cause disaggregation of granulocyte aggregates in vitro. J Clin Invest 68:13-20, 1981 Manuscript received January 2, Accepted in final form May 1, Address reprint requests to: Kjell Rootwelt, M.D., Ph.D., Department of Clinical Chemistry, Rikshospitalet, 0027 Oslo, Norway. 736 J. Neurosurg. /.Volume 77~November, 1992