SPECT/CT Imaging of Intramyocardial Blood Volume with 99mTc-RBC for Detection of Doxorubicin-induced Microvascular Cardiotoxicity in Dogs

2019 
237 Objectives: Doxorubicin (DOX) is an effective chemotherapy agent, but induces cardiotoxicity in up to 26% of treated patients, and toxicity often manifests as heart failure. Clinically, a reduction in the left ventricular (LV) ejection fraction (EF), which reflects global systolic function, is used to determine cardiotoxicity. Recently, it has been suggested that DOX also induces microvascular damage. 99mTc-Red Blood Cell (RBC) scans can be used to measure changes in the intramyocardial blood volume (IMBV), which reflects microvascular density and/or function. Patients with microvasculature injury are expected to have decreased resting IMBV as compared to normal subjects. Therefore, we investigated IMBV non-invasively using 99mTc-RBCs and SPECT imaging in canines chronically receiving DOX, as an indicator of DOX-induced microvascular injury. To provide accurate quantification of IMBV, respiratory and cardiac motion correction, attenuation correction and scatter correction were performed. Methods: Seven dogs were treated weekly with DOX HCl (1 mg/kg) for 12-14 weeks. Four serial SPECT 99mTc-RBC scans, attenuation CT scans and contrast-enhanced CT (CTA) scans were performed in anesthetized animals under rest conditions using a GE Discovery 570c SPECT/CT at baseline and at cumulative DOX doses of 3-5, 7-9, and 12-14 mg/kg. 6 of the 7 dogs completed all 4 scans, while 1 animal only completed three scans due to experimental complications. SPECT list-mode data were rebinned into 5 respiratory and 8 cardiac gates using an offline rebin algorithm. Dual-motion correction was performed by summing the motion-free end-expiration respiratory gates to yield a high-count projection data for each cardiac phase. In this study, we focused on the end-diastolic (ED) and end-systolic (ES) cardiac phases. SPECT images were reconstructed using the maximum likelihood expected-maximization algorithm (MLEM, maximum iterations 80) with attenuation and scatter correction. A body-contour (BC) was delineated using the CT attenuation map to constrain the spatial support for image reconstruction. Forward and back projections were implemented with a precomputed system matrix incorporated with collimator-detector response (GE Healthcare). IMBV was estimated as: IMBV=Cm/Cb, where Cm is the mean activity in the LV myocardium and Cb is the mean activity in the LV blood pool. The LV myocardium and LV blood pool ROIs were segmented based on co-registered CTA and applied to SPECT images. In addition, transthoracic echocardiography (TTE, iE33 Philips ) was performed in awake animals at the same imaging time points as SPECT/CT imaging to assess LVEF. End-diastolic and end-systolic IMBV, as well as LVEF, were compared across different cumulative DOX doses. Results: In general, both ED and ES IMBV decreased in a step-wise fashion with increasing cumulative dose of DOX. The reductions in IMBV at 3-5 mg/kg of DOX were minor and not significant. The ES IMBV showed a significant reduction of 9% (p=0.02) from baseline at 7-9 mg/kg of DOX, and thereafter was reduced by 13% (p=0.03) from baseline at 12-14 mg/kg of DOX. On the other hand, the ED IMBV only showed a significant reduction from baseline (10%, p=0.01) at cumulative DOX doses of 12-14 mg/kg. The LVEF measured by echo was significantly decreased from 64% at baseline to 56% at 3-5mg/kg of DOX (p=0.003) and 58% at 7-9mg/kg (p=0.03), although remained within normal limits. However, LVEF dramatically fell to 46% at 12-14 mg/kg of DOX (p=2.6×10-5). Conclusions: Serial SPECT/CT imaging with 99mTc-RBC was able to detect step-wise decrements in IMBV in a progressive, chronic canine model of DOX-induced cardiotoxicity. The application of this method to other cardiovascular diseases and during the administration of chemotherapy agents that primarily influence the microvasculature (e.g. trastuzumab) is needed to realize the full potential of SPECT quantification of IMBV.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    1
    Citations
    NaN
    KQI
    []