Quantitative Interpretation of FDG PET/CT with Myocardial Perfusion Imaging Increases Diagnostic Information In the Evaluation of Cardiac Sarcoidosis

Supplemental Data

*Azadeh Ahmadian MD1, *Ashley Brogan MD1, Jeffrey Berman MD2, Aaron L. Sverdlov MD PhD1, Gustavo Mercier, MD PhD3 Michael Mazzini MD1, Praveen Govender MD2, Frederick L. Ruberg MD4, Edward J. Miller MD PhD4#

Author Affiliations:

*Indicates co-first authors

1Department of Medicine, Section of Cardiovascular Medicine, Boston University School of Medicine, Boston, MA

2Department of Medicine, Section of Pulmonary, Allergy, Sleep, and Critical Care Medicine, Boston University School of Medicine, Boston, MA

3Department of Radiology, Boston University School of Medicine, Boston, MA

4Departments of Medicine and Radiology, Section of Cardiovascular Medicine, Boston University School of Medicine, Boston, MA

Supplemental Methods:

Outcomes/events were defined as follows: Ventricular tachycardia (VT): ECG/telemetry/Holter monitor/pacemaker/AICD documentation of sustained (>30 beats) VT; Sudden Cardiac Death: Sudden death presumed to be due to VT or ventricular fibrillation (VF), or appropriate ICD firing for VT/VF; Worsening atrio-ventricular block: Progression in stage of AV conduction disorder; Cardiac Hospitalization: Hospitalization for cardiac symptoms (heart failure, arrhythmia, etc); New or Worsening Heart Failure: New or increased use of diuretic for management of volume overload, signs/symptoms of volume overload on exam (S3, edema, elevated JVP, rales), and/or radiographic evidence of volume overload (pulmonary vascular congestion).

FDG PET/CT with MPI Imaging.

FDG imaging: Patients were injected with 11 mCi of 18-F-fluorodeoxyglucose I.V. and allowed to incubate for 60 minutes. Two-dimensional PET imaging was acquired using 5 frames of 5 minutes (BMI<30) or 7 minutes (BMI>30) each followed by summation. Low dose helical CT centered over the heart was performed for attenuation correction, with parameters are as follows: matrix size 512x512; kVp 140 kV, 10 mA; rotation speed 1 sec/rotation; table speed 11.24mm/rotation; pitch 0.562 with beam collimation of 16x1.25mm; reconstruction was done in 5mm slice thickness every 5mm. The CT scan was reduced to 128x128 matrix and blurred with a Gaussian filter as per the manufacturer (GE, Milwaukee, USA) specifications before using the CT scan for construction of the attenuation maps. PET images were processed using OSEM (14 subsets, 2 iterations), a 128x128 matrix, and scatter correction.

MPI Imaging: Gated SPECT imaging (30mCi Tc-99m sestamibi) was performed using a Siemens eCAM with Gd-153 line source attenuation correction as per ASNC guidelines, with the exception of use of a 128x128 matrix required for AC. Gated Rb-82 PET/CT imaging (60mCi Rb-82) was performed per ASNC guidelines using a GE Discovery STN-16 PET/CT and CT attenuation correction scanner prior to FDG scans.

Echocardiography and Cardiac MRI. Echocardiograms and/or cardiac MRI examinations were identified from the Boston Medical Center electronic medical record. Clinical echocardiogram and MRI reports were reviewed for wall motion abnormalities and ejection fraction (echo and MRI), as well as late gadolinium enhancement and abnormal T2 signal (MRI) obtained as part of standard Boston Medical Center clinical imaging protocols for the evaluation of cardiac sarcoidosis.

Quantitative analysis was performed on fused PET/CT images by measuring LV blood pool SUVmax (LVBP SUVmax), Cardiac SUVmax, Cardiac Metabolic Volume (CMV) and Cardiac Metabolic Activity (CMA) as follows (Figure 1):

1.  LV blood pool SUVmax: An approximately 300 mm2 circular region of interest (ROI) was placed in the left ventricular blood pool (LVBP) in the axial image with the largest LV chamber. The maximum SUVbw in that ROI (LVBP SUVmax) was measured, taking care to ensure the LVBP ROI did not encompass any pixels of myocardium.

2.  Cardiac SUVmax: A volume of interest was inserted on the fused axial image encompassing the entire heart (left and right ventricles), and the coronal and sagittal images were reviewed to ensure the entire myocardium and no adjacent non-cardiac FDG-avid structures were included. The maximal SUV voxel in this volume was automatically identified and visually assigned a segment number corresponding to a standard 17-segment model. The segment location of this voxel was confirmed by comparing the fused PET/CT image to the conventional cardiac display in 4DM (above). In addition, in studies with abnormal perfusion (below), the SUVmax in the myocardial segment with the maximum intensity perfusion defect was measured.

3.  Measurement of Cardiac Metabolic Volume (CMV) and Cardiac Metabolic Activity (CMA): The volume and activity of abnormal voxels in the Cardiac ROI (above) at various thresholds of detection was determined using automated analysis in the GE AW Suite. In order to determine the minimum SUV threshold value at which automated voxel volumetric and activity measurements should begin, ratios of 1x, 1.5x, and 2x LVBP ROI SUVmax/Cardiac ROI SUV max were employed

Supplemental Figure Legends

Supplemental Figure 1. Study Design. CS=Cardiac Sarcoidosis. *CS Diagnostic Criteria defined in Table 2. ‘Rev. JMHW’ refers to the addition of FDG and Rb82 to the 2006 JMHW criteria; ‘Isolated CS’=isolated cardiac sarcoidosis; ‘Tx Responsive’=steroid responsive cardiomyopathies with CMR LGE and a high clinical likelihood of CS.

Supplemental Figure 2. Clinical Characteristics of Isolated Cardiac Sarcoidosis Patients. ND=Study not performed. *Note: Patient #3 meets 2006 JMHW and revised clinical criteria due to positive endomyocardial biopsy, but is included in the Isolated Cardiac Sarcoidosis Group due to the lack of other organ involvement. Patient #4 did not have a positive pulmonary lymph node biopsy for sarcoid at the time of his index FDG PET/CT with MPI study, which defined his inclusion/classification in the study. The lymph node was identified at that examination and subsequently biopsied.

Supplemental Figure 3. Linear regression of correlation between measurements by two blinded observers.



Ahmadian et al. Online Supplement. p. 7