AI-AIF: artificial intelligence-based arterial input function for quantitative stress perfusion cardiac magnetic resonance

Abstract

Purpose One of the major challenges in the quantification of myocardial blood flow (MBF) from stress perfusion cardiac magnetic resonance (CMR) is the estimation of the arterial input function (AIF). This is due to the non-linear relationship between the concentration of gadolinium and the MR signal, which leads to signal saturation. In this work, we show that a deep learning model can be trained to predict the unsaturated AIF from standard images, using the reference dual-sequence acquisition AIFs (DS-AIFs) for training. Methods A 1D U-Net was trained, to take the saturated AIF from the standard images as input and predict the unsaturated AIF, using the data from 201 patients from centre 1 and a test set was comprised of both an independent cohort of consecutive patients from centre 1 and an external cohort of patients from centre 2 (n = 44). Fully-automated MBF was compared between the DS-AIF and AI-AIF methods using the Mann-Whitney U test and Bland-Altman analysis. Results There was no statistical difference between the MBF quantified with the DS-AIF (2.77 ml/min/g (1.08)) and predicted with the AI-AIF (2.79 ml/min/g (1.08), p = 0.33. Bland-Altman analysis shows minimal bias between the DS-AIF and AI-AIF methods for quantitative MBF (bias of -0.11 ml/min/g). Additionally, the MBF diagnosis classification of the AI-AIF matched the DS-AIF in 669/704 (95%) of myocardial segments. Conclusion Quantification of stress perfusion CMR is feasible with a single-sequence acquisition and a single contrast injection using an AI-based correction of the AIF.

A simplified method to correct saturation of arterial input function for cardiac magnetic resonance first-pass perfusion imaging: validation with simultaneously acquired PET

“…Another noticeable advantage of this AIF correction method is that it is theoretically independent of the MRI system used, and the pulse sequence settings, as long as sampling rate is sufficiently high. Recently, an artificial intelligence-based AIF correction method was reported to estimate stress MBF by training data from perfusion imaging with the dual-bolus approach [ 22 ]. This fully automatic method demonstrated an excellent agreement in stress MBF between corrected AIF and dual-bolus methods.…”

High-resolution quantification of stress perfusion defects by cardiac magnetic resonance

“… 13 , 14 More recently, fully automated solutions for quantifying perfusion CMR are being integrated into the clinical workflow, with Artificial Intelligence (AI)–based image analysis 15–17 and AI-based quantification. 18 , 19 However, none of these works have addressed the reporting and interpretation of the quantitative values.…”

Assessment of Microvascular Disease in Heart and Brain by MRI: Application in Heart Failure with Preserved Ejection Fraction and Cerebral Small Vessel Disease

“…Myocardial perfusion imaging has evolved from a qualitative visual assessment in which reduced MBF was appreciated as a hypointense, to advanced validated quantitative perfusion sequences that automatically segment the myocardium and quantify MBF (in mL/min/g) on a pixel-by-pixel basis [ 78 , 79 ]. With quantification, not only are absolute stress and rest MBF values obtained but the myocardial perfusion reserve (MPR) is calculated as a ratio between stress MBF/rest MBF.…”

“…The above studies have examined global values for MBF and MPR; however, the latest iteration of quantitative perfusion sequences [ 78 , 79 ] utilises CMRs superior spatial resolution to automatically segment the heart into not just the regional 16-segment model but further subdivide each segment into an endocardial and epicardial component. This allows the quantification of an endocardial/epicardial ratio that can quantify the degree of inducible subendocardial ischaemia, which has long been suggested as a risk factor for HF and MACE, not just in HFpEF but across cardiovascular diseases [ 71 ].…”