Automatic segmentation of myocardium at risk from contrast enhanced SSFP CMR: validation against expert readers and SPECT

BACKGROUND

Efficacy of reperfusion therapy can be assessed as myocardial salvage index (MSI) by determining the size of myocardium at risk (MaR) and myocardial infarction (MI), (MSI = 1-MI/MaR). Cardiovascular magnetic resonance (CMR) can be used to assess MI by late gadolinium enhancement (LGE) and MaR by either T2-weighted imaging or contrast enhanced SSFP (CE-SSFP). Automatic segmentation algorithms have been developed and validated for MI by LGE as well as for MaR by T2-weighted imaging. There are, however, no algorithms available for CE-SSFP. Therefore, the aim of this study was to develop and validate automatic segmentation of MaR in CE-SSFP.

METHODS

The automatic algorithm applies surface coil intensity correction and classifies myocardial intensities by Expectation Maximization to define a MaR region based on a priori regional criteria, and infarct region from LGE. Automatic segmentation was validated against manual delineation by expert readers in 183 patients with reperfused acute MI from two multi-center randomized clinical trials (RCT) (CHILL-MI and MITOCARE) and against myocardial perfusion SPECT in an additional set (n = 16). Endocardial and epicardial borders were manually delineated at end-diastole and end-systole. Manual delineation of MaR was used as reference and inter-observer variability was assessed for both manual delineation and automatic segmentation of MaR in a subset of patients (n = 15). MaR was expressed as percent of left ventricular mass (%LVM) and analyzed by bias (mean ± standard deviation). Regional agreement was analyzed by Dice Similarity Coefficient (DSC) (mean ± standard deviation).

RESULTS

MaR assessed by manual and automatic segmentation were 36 ± 10% and 37 ± 11%LVM respectively with bias 1 ± 6%LVM and regional agreement DSC 0.85 ± 0.08 (n = 183). MaR assessed by SPECT and CE-SSFP automatic segmentation were 27 ± 10%LVM and 29 ± 7%LVM respectively with bias 2 ± 7%LVM. Inter-observer variability was 0 ± 3%LVM for manual delineation and -1 ± 2%LVM for automatic segmentation.

CONCLUSIONS

Automatic segmentation of MaR in CE-SSFP was validated against manual delineation in multi-center, multi-vendor studies with low bias and high regional agreement. Bias and variability was similar to inter-observer variability of manual delineation and inter-observer variability was decreased by automatic segmentation. Thus, the proposed automatic segmentation can be used to reduce subjectivity in quantification of MaR in RCT.

CLINICAL TRIAL REGISTRATION

NCT01379261. NCT01374321.

FLASH proton density imaging for improved surface coil intensity correction in quantitative and semi-quantitative SSFP perfusion cardiovascular magnetic resonance

BACKGROUND

A low excitation flip angle (α < 10°) steady-state free precession (SSFP) proton-density (PD) reference scan is often used to estimate the B1-field inhomogeneity for surface coil intensity correction (SCIC) of the saturation-recovery (SR) prepared high flip angle (α = 40-50°) SSFP myocardial perfusion images. The different SSFP off-resonance response for these two flip angles might lead to suboptimal SCIC when there is a spatial variation in the background B0-field. The low flip angle SSFP-PD frames are more prone to parallel imaging banding artifacts in the presence of off-resonance. The use of FLASH-PD frames would eliminate both the banding artifacts and the uneven frequency response in the presence of off-resonance in the surface coil inhomogeneity estimate and improve homogeneity of semi-quantitative and quantitative perfusion measurements.

METHODS

B0-field maps, SSFP and FLASH-PD frames were acquired in 10 healthy volunteers to analyze the SSFP off-resonance response. Furthermore, perfusion scans preceded by both FLASH and SSFP-PD frames from 10 patients with no myocardial infarction were analyzed semi-quantitatively and quantitatively (rest n = 10 and stress n = 1). Intra-subject myocardial blood flow (MBF) coefficient of variation (CoV) over the whole left ventricle (LV), as well as intra-subject peak contrast (CE) and upslope (SLP) standard deviation (SD) over 6 LV sectors were investigated.

RESULTS

In the 6 out of 10 cases where artifacts were apparent in the LV ROI of the SSFP-PD images, all three variability metrics were statistically significantly lower when using the FLASH-PD frames as input for the SCIC (CoVMBF-FLASH = 0.3 ± 0.1, CoVMBF-SSFP = 0.4 ± 0.1, p = 0.03; SDCE-FLASH = 10 ± 2, SDCE-SSFP = 32 ± 7, p = 0.01; SDSLP-FLASH = 0.02 ± 0.01, SDSLP-SSFP = 0.06 ± 0.02, p = 0.03). Example rest and stress data sets from the patient pool demonstrate that the low flip angle SSFP protocol can exhibit severe ghosting artifacts originating from off-resonance banding artifacts at the edges of the field of view that parallel imaging is not able to unfold. These artifacts lead to errors in the quantitative perfusion maps and the semi-quantitative perfusion indexes, such as false positives. It is shown that this can be avoided by using FLASH-PD frames as input for the SCIC.

CONCLUSIONS

FLASH-PD images are recommended as input for SCIC of SSFP perfusion images instead of low flip angle SSFP-PD images.