Validation of the Carotid Intima-Media Thickness Variability (IMTV)

Abstract

This Chapter (The contents of this chapter build upon the paper: K.M. Meiburger, F. Molinari, J. Wong, L. Aguilar, D. Gallo, D.A. Steinman, U. Morbiducci, “Validation of the carotid intima-media thickness variability (IMTV): Can manual segmentations be trusted as Ground Truth?”, In: Ultrasound in Medicine and Biology, http://dx.doi.org/10.1016/j.ultrasmedbio.2016.02.004) continues the study on morphological vascular studies for quantitative imaging techniques with clinical B-mode ultrasound images and specifically focuses on the validation of the intima-media thickness variability parameter. Recent studies have shown that the irregularity of the IMT along the carotid artery wall has a stronger correlation with atherosclerosis than the IMT itself, so in this chapter, the Intima-Media Thickness Variability (IMTV), a parameter defined to assess the IMT irregularities along the wall, is studied and validated. In particular, whether or not manual segmentations of the lumen-intima and media-adventitia can be trusted as ground truth in the calculation of the IMTV parameter is analyzed. A total of 60 simulated ultrasound images with a priori IMT and IMTV values were used. The images were simulated using the Fast And Mechanistic Ultrasound Simulation (FAMUS) software and presented 5 different morphologies, 4 nominal IMT values and 3 different levels of variability along the carotid artery wall (no variability, small variability, and large variability). Three experts manually traced the lumen-intima (LI) and media-adventitia (MA) profiles and two automated algorithms were used to obtain the LI and MA profiles. One expert also re-traced the LI and MA profiles to test intra-reader variability. The average IMTV measurements of the ground truth profiles used to simulate the longitudinal B-mode images were equal to $0.002\pm 0.002$ mm, $0.149\pm 0.035$ mm, and $0.286\pm 0.068$ mm for the cases of no variability, small variability, and large variability. The IMTV measurements of one of the automated algorithms showed statistically similar values ($p>0.05$, Wilcoxon signed rank) when considering small and large variability, but non-significant values when considering no variability ($p<0.05$, Wilcoxon signed rank). The second automated algorithm showed statistically similar values in the small variability case. Two Readers’ manual tracings, however, produced IMTV measurements with a statistically significant difference considering all three variability levels; the third Reader, on the other hand, showed a statistically significant difference in both the no variability and large variability case. Moreover, the error range between the Reader and automatic IMTV values was approximately 0.15 mm, which is on the same order of small IMTV values, showing how manual and automatic IMTV readings should be not used interchangeably in clinical practice. Thanks to the results found in this study, it can be concluded that expert manual tracings should not be considered reliable in the IMTV measurement and therefore should not be trusted as Ground Truth. On the other hand, the first automated algorithm was found to be more reliable, showing how automated techniques could therefore foster the analysis of the carotid artery intima-media thickness irregularity.