Measuring Brain Volume Extra bits

ProfG forgot to tell me he was posting on this one so here are some pictures from the actual paper

Smeets D, Ribbens A, Sima DM, Cambron M, Horakova D, Jain S, Maertens A, Van Vlierberghe E, Terzopoulos V, Van Binst AM, Vaneckova M, Krasensky J, Uher T, Seidl Z, De Keyser J, Nagels G, De Mey J, Havrdova E, Van Hecke W. Reliable measurements of brain atrophy in individual patients with multiple sclerosis. Brain Behav. 2016;6(9):e00518. eCollection 2016 

Techniques to measure brain volume or brain volume loss can be subdivided into two main categories, that is, cross‐sectional and longitudinal methods.

Cross‐sectional methods use a single MRI scan to segment specific tissues or structures. As a result, the volume of these tissue types and/or structures is calculated. Well‐known and validated examples of these are BPF, SIENAX and Freesurfer. In contrast to cross‐sectional approaches, longitudinal methods take into account two MRI scans of the same subject from different time points to calculate brain volume changes or atrophy. Longitudinal methods typically try to match the two MRI scans using warping techniques and directly extract small changes in brain volume from this process. A longitudinal method that is frequently used in clinical trials is SIENA, while a longitudinal processing pipeline that can take more than two time points into account is included in FreeSurfer 

When applying brain atrophy measures for individual MS patients, the measurement error of the method and thus the reliability becomes of paramount importance. It is indeed known that the average atrophy rate in MS patients is approximately 0.5%–1.3% per year, compared with 0.1%–0.4% per year in healthy individuals.

The measurement error of the brain atrophy measure therefore needs to be very low, in order to draw meaningful conclusions in individual patients. This includes a robustness of the method toward daily physiological processes that might affect brain volume

                                                MSmetrix  lesions = red Grey Matter in blue-green

                                                                    SIENA Grey mater in blue-green

This study measures different ways of measuring brain volume change and there is good correlation between the two method

On a longitudinal data set of patients with MS , the correlation between whole‐brain atrophy measurements obtained with MSmetrix‐long and SIENA is relatively high, with a Pearson correlation coefficient equal to 0.91

When brain atrophy measures are introduced in clinical practice for individual MS patients, interpretation of these results should be done with caution. In this context, it is indeed known that there are many confounding factors that can affect the measurement of brain atrophy and therefore the interpretation of the results. However this study suggests that the MSmetrix may have some utility,