Joint Program in Nuclear Medicine

Scintigraphic Evaluation of FUO in Patients with Tumors

Justine Arthur, BMBS, BMedSci, MRCP 
Annick Van den Abbeele, MD

October 24, 1995

Presentation

A 64-year old man presented with a pathological fracture of the proximal left humerus secondary to high grade chondrosarcoma. Staging workup including abdominal and chest computed tomography and bone scan revealed a small nodule in the left mid lung but no other evidence of metastasis. He underwent successful resection of the tumor, repair of the defect with an allograft and latissimus dorsi flap, followed by 5000 rad of radiation treatment to the tumor bed for local control and palliation of his pain.

Three months later he presented with left lateral chest wall pain following a fall. Chest CT and plain films showed a left lateral pleural based mass and bilateral small lung nodules. Bone scan (48k bytes) showed three discrete foci of increased tracer uptake in the left 8th, 9th and 10th ribs anteriorly. Correlative radiographs did not reveal any bony lesions; specifically, no fractures were seen. In view of the aggressive progressive nature of his disease, it was decided to proceed with palliative MAID chemotherapy (mesna, adriamycin, ifosfamide and dacarbazine) for the sarcoma.

He had a history of fever for four weeks. On admission, his temperature was 101 degrees F with no localizing signs or symptoms. The infection screen including blood, urine and sputum cultures, hepatobiliary scintigraphy (to rule out acute cholecystitis) and echocardiography (to rule out bacterial endocarditis) was negative. Laboratory results included a hemoglobin of 9.3g/dl and WBC count of 6,600 cells/microlitre, with 13%% lymphocytes, 19%% monocytes and 58%% polymorphonuclear cells. Subsequently, a Tc-99m-labeled white cell study was ordered to look for an occult source of infection prior to commencing chemotherapy.

Imaging Technique

Routine in vitro labelling of autologous white cells was performed with Tc-99m hexamethylpropyleneamine oxine (HMPAO). The patient was injected with 9.7 mCi of Tc-99m HMPAO labeled leukocytes.

Imaging Findings

Planar images of the Tc-99m-labeled white cell study (81k bytes) obtained one hour after injection revealed markedly abnormal patchy uptake throughout the bone marrow. There was normal hepatic and splenic uptake and no collections of tracer to suggest a focus of infection outside the reticuloendothelial system. Residual bowel activity in the right lower quadrant of the abdomen due to the Tc-99m-DISIDA hepatobiliary scan performed 20 hours earlier was noted prior to reinjection of the autologous labeled white cells. A suggestion was made that this pattern may reflect bone marrow replacement by tumor cells and MRI correlation was recommended.

MRI study (70k bytes) revealed diffusely decreased T1 signal in all the thoracic spine vertebral bodies, ribs and sternum and foci of increased T2 signal intensity scattered throughout the vertebral bodies consistent with bone marrow expansion, myeloproliferative disease or diffuse metastatic involvement. Together, the Tc-99m-labeled white cell study and the MRI suggested diffuse metastatic marrow involvement.

Clinical Course:

Without an identifiable source of infection, the patient was started on the planned palliative MAID chemotherapy regime. During his treatment however, his aggressive disease continued to progress with the development of brain metastases. The patient died before completing his therapy.

Diagnosis

Bone Marrow Replacement

Discussion

Tc-99m-HMPAO labeled leukocyte scintigraphy is an effective way of identifying sites of infection and has an overall accuracy for detection of infection similar to In-111 labeled leukocytes (92%% accuracy for Tc-99m-HMPAO and 88%% for In-111)(1). The theoretical advantages of labeling with Tc-99m are:

Radiopharmaceuticle:

Tc-99m-HMPAO is lipophilic and readily crosses the white cell membrane where it becomes trapped by binding to the mitochondria and nucleus. As it has a predilection for granulocytes, an almost pure granulocyte preparation is obtained without the need for separation of granulocytes from the buffy coat layer. The normal biodistribution of Tc-99m-HMPAO-labeled leukocytes includes:

Fever of Unknown Origin:

Assessment of a fever of unknown origin (defined as the cyclical persistence of a fever over a period of 3 weeks or more), often leads to "infection imaging" to identify the occult source of infection. Techniques available for imaging infection include gallium-67 citrate, indium-111 labeled leukocytes and technetium-99m labeled leukocytes. Other promising agents currently still in clinical trials include radiolabelled polyclonal and monoclonal antibodies, and labeled chemotactic peptides. In patients who have not had recent surgery, Ga-67 may be the most sensitive test for identifying the source of fever as Ga-67 may uncover not only acute and chronic infection, but also granulomatous disease (including tuberculosis) and tumor. In fact, 20%% of patients who have fever of unknown origin have occult neoplasms(3).

For patients who have a shorter duration of non-localizing fever, in whom more acute infection with a higher probability of neutrophil infiltration is suspected, labeled leukocytes have shown good results and allow more rapid diagnosis than with gallium. In 8 different series, the sensitivity and specificity for detecting infection with Tc-99m leukocytes varied between 90-100%% and 89-100%% respectively, whilst in 5 other series the sensitivity and specificity of Ga-67 citrate for infection varied between 81-94%% and 64-100%% respectively(4). In addition, for those cases where intra-abdominal sepsis is suspected, In-111 labeled leukocytes are preferred as, unlike Ga-67 citrate and Tc-99m labeled leukocytes, there is no bowel excretion to confound the interpretation.

Fever Assessment in Patients with Tumor:

The assessment of fever in a patient with known tumor, however, poses particular problems as the fever may be due to tumor, chemotherapy, infection or an inflammatory lesion. In addition, patients undergoing chemotherapy and/or radiotherapy are frequently leukopenic. If localizing symptoms are present, computed tomography or ultrasonography are indicated to evaluate the focus of infection. If no localizing signs are present, whole body imaging in search of an occult focus of infection is indicated particularly if, as in the present case, the patient is due to commence therapy that has the potential to significantly impair the immune response to infection. As many tumors are gallium avid, false positive results for foci of infection or inflammation may occur and so labeled-leukocyte studies are more helpful.

A study comparing Ga-67 citrate scans with In-111 labeled leukocytes in 10 febrile patients with known tumor found, not surprisingly, that specificity for infection was higher with the labeled white cells than with Ga-67 citrate and also concluded that normal findings on Ga-67 citrate and labeled leukocyte scans indicated tumor-chemotherapy fever(5). Labeled leukocyte studies however require a leukocyte count above 5000/mm3 for optimal imaging (although diagnostic imaging can be successfully performed with counts as low as 3000/mm3). Although cross matched donor leukocytes have been successfully used, Ga-67 citrate imaging is preferable in neutropenic patients.

Antibodies:

Recent studies of Tc-99m- or I-123-labeled murine monoclonal antigranulocytic antibodies and In-111- or Tc-99m-labeled polyclonal human immunoglobulin G (IgG) have yielded favorable results in infection imaging and are of particular interest as, in contrast to labeling leukocytes, these techniques do not require ex vivo labeling and handling of blood products(6-9). One disadvantage of Tc-99m-labeled monoclonal antibody, which is directed against non-specific cross-reacting antigen (NCA-95, a differentiation antigen of granulopoiesis) (10) is that the antibody appears to label bone marrow myelocytic series to a greater degree than it labels peripheral granulocytes. This particular propertyhas led to recent interest in using labeled antigranulocyte antibody (AGA) as a marrow imaging agent in the assessment for bone marrow metastases(11-14). Several studies have reported higher detection of marrow metastases with Tc-99m-labeled AGA compared with bone scans (78%% vs. 53%% in breast cancer (11); 32%% vs. 10%% for breast, 18%% vs. 12%% for prostate, 11%% vs. 5%% for lung and 3.6%% vs. 1.5%% for kidney and bladder (14) and more conventional colloid marrow scintigraphy. Using Tc-99m-labeled AGA, bone marrow scans of superior quality compared with colloid scans have been obtained, as hepatic and splenic AGA uptake is significantly lower whilst hematopoietic marrow uptake is 2-4 times higher. This may prove to be clinically useful in the early detection of bone marrow metastatic invasion.

A drawback of murine monoclonal antibodies, however, is the immunogenicity of intact murine antibodies, limiting the possibility of repeat administrations because of changes in the biodistribution secondary to immune complex formation. Transient HAMA (human antimurine antibodies) responses have been detected in between 5%% and 40%% of patients after monoclonal antigranulocytic antibody scintigraphy(15,16,17). However, no allergic reactions have been observed in patients reinjected up to 3 times(15).

Marrow Replacement:

For the index patient, the abnormal uptake of Tc-99m labeled leukocytes throughout the bone marrow was very similar to patterns documented in diffuse metastatic disease imaged with Tc-99m-labeled AGA(11). In addition, similar to findings with labeled AGA, Tc-99m-labeled leukocytes, in this case, were more sensitive than the bone scan in identifying diffuse skeletal metastatic involvement. In fact, the first indication of widespread marrow involvement, which came as a surprise to the referring clinicians in this particular patient, was the abnormal marrow uptake on the leukocyte study performed as an infection screen. Diffuse marrow invasion was subsequently confirmed by MRI, the fever was attributed to the underlying carcinomatosis and he was commenced on chemotherapy despite his poor prognosis.

Conclusions:

Tc-99m-labeled leukocytes are an effective way of imaging infection with a sensitivity ranging from 90-100%% and specificity ranging from 89-100%% in various series. In febrile patients with known tumor and with fever of unknown origin, labeled leukocyte scans are preferred over Ga-67 citrate, to help differentiate between fever due to the underlying tumor or due to infection or inflammation. In the index case, the Tc-99m labeled leukocyte scan not only helped rule out an occult infection, it also identified "occult" widespread metastatic skeletal involvement which had not been appreciated on the prior staging bone scan.

References

1. Kipper SL and Evans D. Simultaneous In-111-oxine and Tc-99m-HMPAO leukocyte imaging in a community hospital. (Abstract). J Nucl Med 1992;33:902

2. Thrall JH and Ziessman HA. Infection and inflammation. In Nuclear Medicine - The Requisites. 1995,Mosby-Year Book, Inc pps:149-169.

3. Alazraki NP Radionuclide imaging in the evaluation of infections and inflammatory disease. Radiol Clin North America 1993;31:783-794

4. Bester MJ, Van Heerden PDR, Klopper JF et al. Imaging infection and inflammation in an African environment: Comparison of Tc99m-HMPAO-labelled leukocytes and Ga-67 citrate. Nucl Med Commun 1995;16:599-607.

5. Palestro CJ, Fineman DS, Needle LB et al. Utility of In-111 labeled autologous leukocytes and Ga-67 citrate imaging in patients with tumour {abstract}. Radiology 1987;165:73.

6. Lind P, Langsteger W, Koltringer P et al. Immunoscintigraphy of inflammatory processes with a Tc-99m-labeled monoclonal antigranulocyte antibody (MAb BW 250/183). J Nucl Med 1990;31:417-423

7. Becker w, Goldenberg DM and Wolf F. The use of monoclonal antibodies and antibody fragments in the imaging of infectious lesions. Semin Nucl Med 1994;24:142-153.

8. Rubin RH, Fischman AJ, Callahan RJ et al. Indium-11-labeled nonspecific immunoglobulin scanning in the detection of focal infection. N Engl J Med 1989;321:935-940.

9. Rubin RH and Fischman AJ. The use of radiolabeled nonspecific immunoglobulin in the detection of focal inflammation. Semin Nucl Med 1994;24:169-179. 10. Reske SN. Recent advances in bone marrow imaging. Eur J Nucl Med 1991;18:203-221

11. Duncker CM, Carrio I, Berna L et al. Radioimmune imaging of bone marrow in patients with suspected bone metastases from primary breast cancer. J Nucl Med 1990;31(9):1450-14558

12. Lee KH, Chung J-K, Choi CW et al. Tc-99m-labelled antigranulocyte antibody bone marrow scintigraphy. J Nucl Med 1995;36:1800-1805.

13. Reske SN, Karsten JH, Gloeckner W et al. Radioimmunoimaging for diagnosis for bone marrow involvement in breast cancer and malignant lymphoma. Lancet Feb1989:299-301

14. Munz DL, Sandrock D and Rilinger N. Comparison of immunoscintigraphy and colloid scintigraphy of bone marrow. Lancet Jul 1990:258-259.

15. Reske SN, Sohn M, Karstens JH et al. Immunoscintigraphy of bone marrow with Tc-99m labelled NCA-95/CEA antibodies (TcNAA). Comparison with bone scanning, plain radiographs and HAMA-response (abstract). J Nucl Med 1990;31:751.

16. Joseph K, Hoffken H, Bosslet K et al. In vivo labelling of granulocytes with Tc- 99m-anti-NCA monoclonal antibodies for imaging inflammation. Eur J Nucl Med 1988; 14:367-373.

17. Abdel-Nabi H, Doerr R, Roth SC et al. Recurrent colorectal cancer detection with repeated infusions of indium-111 ZCE-025 MoAb. J Nucl Med 1989;30:748.

Click here to go to Joint Program in Nuclear Medicine home page and Copyright notice. 

J. Anthony Parker, MD PhD, Tony_Parker@bidmc.harvard.edu