Leukocyte SPECT/CT for Detecting Infection of Left-Ventricular-Assist Devices: Preliminary Results

Patients

Between February 2007 and September 2008, 13 imaging procedures were performed on 8 consecutive men (mean age ± SD, 56 ± 11 y) who had an implanted LVAD. The mean duration that the device was in use was 666 d (range, 188–868 d). Patients were referred for leukocyte scintigraphy if device-related infection was suspected or to assess the outcome of antibiotic therapy. Five patients underwent 2 imaging procedures. Bacteriologic samples were taken from all patients (repeated blood cultures and local sample collection at the cable exit site). All patients were implanted with a HeartMate II (Thoratec) metallic rotary blood pump (130 × 45 mm) just below the diaphragm. The inflow cannula is inserted into the left ventricular apex, and the outflow polyester graft is inserted into the ascending aorta. A percutaneous electric cable connects the pump, through the skin of the right flank, to an external system controller linked to a power source (2,10).

Leukocyte Labeling and Imaging Protocol

Leukocyte scintigraphy was performed immediately after intravenous injection of 714 ± 58 MBq of ^99mTc-exametazime-labeled autologous granulocytes, according to a validated protocol (11). The mean average labeling yield was 90% ± 4%. All trypan blue tests gave positive results (≥99% viable cells).

All images were acquired using a SPECT/CT system (Symbia T2; Siemens Medical Solutions). Planar scans (128 × 128 matrix) were obtained over the chest and abdomen at hours 1, 4 and 24. SPECT/CT was performed over the same areas at hours 4 and 24. The CT current was adapted for the patient's weight (110 kV and 40–60 mAs), with dose reduction (CareDose; Siemens Medical Solutions). SPECT data were acquired over 360° (matrix size, 128 × 128; 128 projections; 30 s/projection at hour 4 and 60 s/projection at hour 24), and images were corrected for attenuation, scatter, and collimator depth dependence. Transverse, sagittal, and coronal slices of corrected and uncorrected SPECT data were merged with CT and interpreted on a dedicated workstation (Leonardo; Siemens Medical Solutions). The mean effective dose for an entire imaging procedure was 14 mSv.

Interpretation Criteria

Foci of abnormal uptake by leukocytes were analyzed on planar and SPECT/CT acquisitions by 2 experienced nuclear medicine physicians who worked in consensus and had no knowledge of the clinical context (i.e., diagnosis of suspected infection or antibiotic therapy assessment). Findings were considered positive when foci of labeled leukocyte uptake were greater than background activity and increased over time. Foci of nonphysiologic abnormal uptake that decreased at hour 24 were considered inflammatory.

Planar Scans

Foci of abnormal uptake considered positive for device-related infection were detected in 6 of 13 scans (46%). In these patients, planar analysis did not reveal any additional foci indicating distant infection and could not distinguish between deep infection and subcutaneous tissue infection at the cable exit site (Fig. 2A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Planar and SPECT/CT images of multifocal device-related infection in patient 1. Patient had persistent infection of percutaneous site and fistulized epigastric abscess 225 d after implantation despite prolonged antibiotic therapy. (A) Planar scan demonstrates 2 foci of abnormal uptake by leukocytes. (B) Axial SPECT/CT with volume-rendering technique shows precise location and extent of infection. (C and D) First infectious focus (white arrow) was percutaneous (C), and second focus (yellow arrow), 9 cm in length, surrounded cable close to fistulized abscess (D).

SPECT/CT Scans

Foci of abnormal uptake considered positive for device-related infection were detected in 8 of 13 scans (62%), which included the 6 cases of infection detected on planar scans and 2 additional foci that were due to physiologic uptake (bone marrow and digestive excretion). In all cases, SPECT/CT provided additional information on the extent and location of infection (Figs. 2B–2D). In 5 cases (patient 3, scan 1; patient 4, scan 1; patient 6, scan 1; patient 7, scan 2; and patient 8, scan 1), SPECT/CT demonstrated a superficial infection limited to the percutaneous site (Supplemental Fig. 1; supplemental materials are available online only at http://jnm.snmjournals.org). In the 3 other cases (patient 1, scan 1; patient 1, scan 2; and patient 5, scan 1), SPECT/CT demonstrated deep infectious foci (Fig. 2).

In 1 of the 8 patients who had a device-related infection (patient 5, scan 1), SPECT/CT revealed 2 additional foci of abnormal distant uptake corresponding to previously unsuspected left basal pneumonitis and to an epigastric abscess near the device. The pneumonitis was near the splenic physiologic uptake, and the epigastric abscess was behind the sternum; both were therefore missed on the planar scan. These abnormal foci were confirmed by bacteriologic cultures and the patient's clinical course.

In 5 of 13 scans (38%), no device-related infection was observed. Both planar scans and SPECT/CT scans showed normal findings in 1 patient (patient 4, scan 2), who remained free from infection for several months. Distant cervical infectious foci were revealed in 2 patients by additional planar scans (no cervical SPECT/CT was performed): tonsillitis (patient 3, scan 2) associated with inflammatory uptake at the cable exit site on planar and SPECT/CT acquisitions (Supplemental Fig. 2), and an abscessed tooth (patient 5, scan 2) without any other scintigraphic abnormality (Supplemental Fig. 3). Because of the scintigraphic findings, these 2 patients received only short-term oral antibiotic treatment, and they were free of relapse for several months.

In the 2 remaining scans, isolated foci of inflammatory uptake were observed: the first (patient 2, scan 1) was a moderate focus at the cable exit site, close to physiologic digestive excretion, and detected only on SPECT/CT. This patient was weaned off the device after 692 d because the patient's heart recovered. All bacteriologic samples harvested during surgery were sterile. The second scan (patient 7, scan 1) presented foci of abnormal uptake on the shoulders, related to subcutaneous injection of antibiotics, and the patient was free of any clinical infection for more than 1 y. In these 5 of 13 scans, the scintigraphic results were considered true-negative on the basis of the patient's evolution (patient 3, scan 2; patient 4, scan 2; patient 5, scan 2; and patient 7, scan 1) or bacteriologic analysis (patient 2, scan 1).