Use of Bone Scintigraphy in The Diagnosis of Cardiac Amyloidosis

Abstract

Cardiac amyloidosis is a rare and progressive condition caused by the buildup of amyloid in the heart. There are two common forms of cardiac amyloidosis: light-chain amyloidosis (AL) and transthyretin amyloidosis (ATTR). The deposition of amyloid in the extracellular matrix of the myocardium leads to heart failure over time, and if left untreated, it can even be fatal. For this reason, early diagnosis is essential for both prognosis and improving patients' quality of life. Since cardiac amyloidosis is a potentially treatable condition, early diagnosis is key to improving patient survival and quality of life. There is now compelling evidence showing that nuclear imaging plays a fundamental role in the non-invasive diagnosis of transthyretin cardiac amyloidosis. Due to its high sensitivity and specificity, radiotracer compounds that target the bones are considered sufficient for establishing the diagnosis, avoiding the need for endomyocardial biopsy. In this study, we analyzed data obtained from examinations conducted on patients referred to the Cardiology Department of the "San Giovanni di Dio e Ruggi d'Aragona" University Hospital, who presented with suspected cardiac amyloidosis. These patients underwent scintigraphy with 99mTc-HMDP (hydroxymethylenediphosphonate), using a large-field computerized gamma camera with a parallel-hole collimator and SPECT. After performing early whole-body planar acquisitions at 5 minutes and late acquisitions at 3 hours after intravenous administration of approximately 700 MBq of Tc-99m HMDP, myocardial uptake was observed. Subsequently, targeted acquisitions and SPECT tomographic studies were performed on the myocardial uptake areas. All acquired images were subsequently subjected to quantitative and qualitative analysis, allowing us to extract a large number of parameters reflecting morphological and predictive characteristics using radiomics and more or less automated analysis algorithms. This analysis enabled us to obtain quantitative information that is not apparent in a qualitative image analysis. The ability to extract hidden information from digital medical images is of particular interest as it can enhance the predictive capabilities of existing automatic segmentation algorithms. Extracting new information that was previously hidden can be utilized for automatic image segmentation.