Use of Technetium-99m–labeled Eosinophils to Detect Active Eosinophilic Inflammation in Humans

To the Editor:

The diagnosis of focal eosinophilic inflammation is notoriously difficult. Here we describe a novel technique using technetium-99m–labeled autologous eosinophils to image eosinophilic vasculitis affecting the lung and skin in a patient with granulomatosis with polyangiitis.

Our patient, a 48-year-old male, presented with cough, purulent sputum, hemoptysis, lethargy, and painful ulcers affecting his hands and feet. His thoracic computed tomography (CT) showed bilateral upper zone nodules with cavity formation. Previously he had experienced recurrent anterior uveitis and small joint polyarthropathy. His medications were hydroxychloroquine and etodolac, which he took intermittently. Examination revealed periungual erythema and multiple deep ulcers affecting his fingers and right ankle (Figure 1A). His blood eosinophil count was 0.76 × 10⁹/L, with normal renal and liver function and normal antinuclear antibody and rheumatoid factor. His antineutrophil cytoplasmic antibody (ANCA) immunofluorescence was positive with a (PR3) ANCA of 25 IU/ml (normal range < 0.2 IU/ml). He also had evidence of IgG₄-related disease associated with his ANCA-positive vasculitis with a serum IgG₄ of 4.1 g/L. Surgical lung biopsy revealed dense fibrosis, with scattered noncaseating granulomas, heavy eosinophil infiltration, a medium-small vessel vasculitis, and IgG₄-staining plasma cells. Similar findings were observed in the skin (Figure 1B).

Figure 1.

Technetium-99m–labeled eosinophil scanning in granulomatosis with polyangiitis. (A) Photograph of vasculitic ulcers affecting the patient’s right ankle. (B) The corresponding histology (×20 magnification) shows active vasculitis with fibrinoid necrosis, red cell extravasation, and surrounding eosinophil infiltration. (C) Planar γ camera images of the legs 5 hours after the injection of technetium-99m–labeled autologous eosinophils shows enhanced uptake corresponding to the area of ulceration on the right ankle. (D) Dynamic imaging of the thorax and abdomen was performed every 20 seconds for 40 minutes after injection of technetium-99m–labeled autologous eosinophils and revealed their rapid transit through both lungs followed by uptake in the liver, spleen, and bone marrow. (E–G) SPECT-CT images at 45 minutes revealed an enhanced signal focally in the lungs, corresponding to the areas of nodularity. (H) A cross section of the surgical lung biopsy (3.8 × 2.5 × 4.5 cm).

Technetium-99m–labeled eosinophils (98.5% pure) were prepared using clinical grade anti-CD16 immuno-magnetic beads as described (1) and 16 × 10⁶ cells (187 MBq) reinjected. Dynamic images were taken for 40 minutes (1) followed by single-photon emission CT (SPECT-CT) images at 45 minutes and 2, 4, and 6 hours. His legs were imaged using a γ camera (Figure 1C). The intravascular residence time (1.5 h) of the technetium-99m–labeled eosinophils was much shorter than that observed in healthy subjects (1) with only 4.9% of the injected cells remaining in the circulation at 30 minutes. This compares to an intravascular residence time of indium-111–labeled eosinophils in healthy volunteers of 28.1 ± 1.7 hours (mean ± SEM), with 14.5 ± 2.5% of radiolabeled cells remaining in the circulation at 30 minutes (1). Although the technetium-99m–labeled eosinophils accumulated as expected in the liver, spleen, and bone marrow (Figure 1D), SPECT-CT revealed additional uptake in some but not all of his lung nodules (Figures 1E–1G).

Although radiolabeled leukocyte scanning is used routinely for localizing occult infection, this approach has not been used previously to demonstrate focal eosinophilic inflammation. Advances in immuno-magnetic bead technology have permitted the isolation of ultra-pure nonactivated eosinophils from blood, which are amenable to radiolabeling and behave physiologically when reinjected in healthy subjects (1). Alternative, indirect methods of detecting eosinophilic inflammation in humans, such as exhaled nitric oxide and analysis of breath condensates, are nonspecific and can only assess the inflammatory burden in the respiratory tract. In contrast, technetium-99m–labeled eosinophils plus SPECT-CT can localize focal eosinophilic inflammation directly in vivo. Furthermore, this technique is less invasive than performing bronchoscopy and transbronchial biopsy, which may be contraindicated in some patients and carry significant risk. This novel approach therefore offers the prospect of a noninvasive method to aid in the diagnosis, localization, and assessment of disease activity.