Atrophy in MS: How icometrix ensures accurate brain volumetric analysis

May 19, 2020

The value of brain atrophy in outcome prediction and treatment monitoring for multiple sclerosis (MS) is undisputed. However, caution is advised when including brain volume change measurements into your clinical practice since both physiological and technical factors need to be taken into account. In this blog, we walk you through these barriers towards clinical implementation and how our icobrain solution can help you break these barriers down.

Even in healthy people, brain volume is dependent on several physiological factors. Brain volume loss occurs as part of the normal aging process, starting during the adolescent phase (age 20-30: -0.05%) and continuing at a more rapid pace as we get older (age 60-70: -0.3%) (Battaglini et al. 2019)1. Secondly, on average, women have smaller brain volumes than men (Ruigrok et al. 2014)13. Because of this age- and sex-dependency all quantified volumes in the icobrain reports are normalized for age and gender.

Brain atrophy is part of the normal aging process but is accelerated in patients with MS, up to 0.5-1.3% per year (Barkhof et al. 20093; Fotenos et al. 20089; Simon et al. 200616). In literature, a pathological cutoff of 0.4% has been proposed as a reference value to distinguish normal aging (De Stefano et al. 2016)6. As this difference between brain atrophy in MS patients and healthy subjects is subtle, a high accuracy or small measurement error is necessary to draw meaningful conclusions in clinical practice. We have shown that the icobrain software can be reliably used in clinical practice, with a test-retest measurement error for whole-brain volume changes of 0.13% across different scanners (Smeets et al. 2016)17.

Diurnal fluctuations and hydration changes are additional confounding factors. Not only is our brain volume larger in the morning (Nakamura et al. 2015)11 but de- and rehydration, as well as vitamin deficiencies, can have an impact on brain volumes (Duning et al. 2005)7. Likewise, lifestyle choices such as cigarette and alcohol consumption have been proven to be associated with brain atrophy rates (Enzinger et al. 2005)8. Nevertheless, it is known that these confounding factors only result in minor volume changes.

By measuring such daily variations over a period of time, we have been able to show that the icobrain software is robust towards such physiological changes (Smeets et al. 201617; Sima et al. 201614; Sima et al. 201715).

Aside from these confounding factors, there are also technical barriers that make the adoption of atrophy in clinical practice challenging. These barriers include the acquisition protocol, distortion differences, and scanner variability. By combining built-in quality controls with a visual quality check, icobrain can guide you in the correct interpretation and acquisition optimization in order to enable you to confidently incorporate brain volumetrics into your clinical practice.

Acquisition optimization:

• Longitudinal follow-ups require a minimum time span of four months. In doing so, the effects of physiological processes and measurement errors are minimized, relative to the pathological change we aim to measure within this period. icobrain automatically applies this 4-month threshold to ensure reliable results.

• Multiple data points can visualize a robust trend. When evaluating MRI scans of MS patients in a clinical routine setting, it should be noted that one or two scans might not be sufficient to properly reflect the disease progression. Hence, it is advised to scan MS patients once a year during the course of their disease. This also ensures robustness against physiological processes as well as variation in the acquisition and analyses. By plotting multiple time points on a reference population graph, icobrain allows you to identify trends relative to age-and-gender matched controls.

  
  

  

  
  

• Pseudo-atrophy can be mitigated with a re-baseline MRI. This process, linked to the anti-inflammatory effect of steroids and disease-modifying treatments, causes a reduction in brain volume 6 to 12 months after treatment initiation. Gray matter volume is less affected by pseudo-atrophy, more substantial during the MS disease course, and more clinically relevant than white matter changes. (Calabrese et al., 20074; Filippi et al., 201310; Tiberio et al., 200518). By focussing on whole brain and gray matter, icobrain guides you in the use of the most clinically relevant and robust brain volumes, tailored to a specific disease population.

Method adjustments:

• Registration-based atrophy quantification ensures high precision. Measuring brain volume changes can be done by comparing volumes at different time points (segmentation-based technique) or by directly quantifying changes over time (registration-based). icobrain uses a registration-based quantification to reduce the measurement error for whole brain volume changes to 0.13% (Beadnall et al. 2019; Wang et al. 201619; Smeets et al. 201617).

• Lesion filling is crucial to identify true change. The lack of lesion filling leads to the misclassification of T1 black holes as gray matter or cerebrospinal fluid (CSF), whereas not accounting for lesion change can confound measured brain volume changes. The impact of lesion filling has been reported to range from 0.3-2.5% (depending on the nature and intensity of the hypo-intense T1 regions) when WM hypo-intensities on T1 are not filled (Battaglini et al. 20122; Chard et al. 20105; Popescu et al. 201412). By combining the information of both T1 and FLAIR sequences, icobrain is able to fill the T1 hypo-intensities with the image intensity of normal-appearing WM on T1.

• Correction for distortion differences. Since the magnetic field of an MRI scanner varies over time, the distortions due to the non-linearity of gradient fields also vary. To compensate for small distortion differences, icobrain applies a skull-based affine registration (Smeets et al. 2016)17.

Interpretation guidance:

• Age-dependent thresholds allow the identification of pathological change. To distinguish MS-related atrophy from normal aging and physiological factors, icobrain provides age- and gender-matched reference values. These reference rates for brain volume loss are calculated from a database including nearly 2000 healthy subjects and are in line with the reported age-related rates by Battaglini et al. 20191.

• Thorough quality control ensures reliable results. Artifacts, head coverage, low contrast, or low signal-to-noise ratio can influence the reliability of the analysis. By incorporating a quality control based on Hansen et al. (2013), icobrain flags data with incomplete head coverage, insufficient contrast to noise ratio, or distortions between sequences (T1 And FLAIR) or time points. By doing this in a fully automated fashion, and flagging only outlying cases for visual control by a trained operator, icobrain ensures a fast turnaround within the clinical practice with maximum control over its quality. As an outcome, the icobrain report includes a quality control remark and a statement on the impact of the quality on the interpretation of the results.

  
  

In conclusion, icobrain guides you in the use of clinically relevant and robust brain measurements for multiple sclerosis. By implementing state-of-the-art techniques, as well as built-in quality controls, we ensure reliable measurements in routine clinical practice. Doing so, icobrain helps you to confidently incorporate brain atrophy into your clinical reporting and decision-making for MS patients.

References

1. Battaglini, M. et al. Lifespan normative data on rates of brain volume changes. Neurobiol. Aging 81:30-37 (2019)

2. Battaglini, M., et al. Evaluating and reducing the impact of white matter lesions on brain volume measurements. Human Brain Mapping, 33:2062–2071 (2012).

3. Barkhof, F. et al. Imaging outcomes for neuroprotection and repair in multiple sclerosis. Nat Rev Neurol. 5:256-266 (2009).

4. Calabrese M, et al. Cortical atrophy is relevant in multiple sclerosis at clinical onset. J Neurol. 254:1212–1220 (2007).

5. Chard, D. T., et al. Reducing the impact of white matter lesions on automated measures of brain gray and white matter volumes. Journal of Magnetic Resonance Imaging, 32:223–228 (2010).

6. De Stefano, N. et al. Establishing pathological cut-offs of brain atrophy rates in multiple sclerosis. J. Neurol. Neurosurg. Psychiatry 87:93-99 (2016).

7. Duning, T. et al. Dehydration confounds the assessment of brain atrophy. Neurology 64: 548-550 (2005).

8. Enzinger, C. et al. Risk factors for progression of brain atrophy in aging: six-year follow-up of normal subjects. Neurology 64:1704-1711 (2005).

9. Fotenos, A. et al. Brain volume decline in aging; Evidence for a relation between socioeconomic status, preclinical Alzheimer disease, and reserve. Archives of Neurology 65:113–120 (2008).

10. Filippi, M. et al. Gray matter damage predicts the accumulation of disability 13 years later in MS. Neurology 81:1759-1767 (2013).

11. Nakamura, K. et al. Diurnal Fluctuations in Brain Volume- Statistical Analyses of MRI From Large Populations. Neuroimage 118:126-132 (2015).

12. Popescu, V. et al. Accurate GM atrophy quantification in MS using lesion-filling with co-registered 2D lesion masks. NeuroImage Clinical 4:366–373 (2014).

13. Ruigrok, A. N. et al. A meta-analysis of sex-differences in human brain structure. Neurosci. Biobehav. Rev. 39:34-50 (2014).

14. Sima, D. et al. MSmetrix validation of normative brain volume population graphs to serve as a reference in the clinical follow up of MS patients, ECTRIMS 2016, abstract and paper poster presentation.

15. Sima, D. et al. Age-dependent whole brain and grey matter annual atrophy rates in healthy adults. MSParis2017 - 7th Joint ECTRIMS-ACTRIMS

16. Simon, J. H. Brain atrophy in multiple sclerosis: What we know and would like to know. Multiple Sclerosis 12:679–687 (2006).

17. Smeets, D. et al. Reliable measurements of brain atrophy in individual patients with Multiple Sclerosis. Brain and Behaviour 6:e00518 (2016).

18. Tiberio M, et al. Gray and white matter volume changes in early RRMS: a 2-year longitudinal study. Neurology 64:1001–1007 (2005).

19. Wang, C. et al. Automated brain volumetrics in multiple sclerosis: A step closer to clinical application. J Neurol Neurosurg Psychiatry 87:754–757 (2016).