There is more to Multiple Sclerosis disease activity than meets the eye

The impact of software-assisted MRI reading on outcomes and costs

Takeaway/summary

More detailed information can be found in Sima et al. 2021

• Several treatment decision paradigms exist in Multiple Sclerosis (MS) management?
• Health outcomes and associated costs in MS depend on the timely detection of disease activity and associated treatment switching
• MRI is an important tool to assess disease activity in MS, and software-assisted reading can decrease the average time on suboptimal treatment from 3.2 years to 1.3 years on average
• By using quantitative MRI, the health outcome of people with MS is improved by 68% relative to the introduction of disease-modifying therapies in the field of MS
• Based on expected health outcomes, in the US, software-assisted MRI reading can save around $1700 per MS patient, per year vs visual MRI reading

Therapeutic options in MS: check?

Multiple Sclerosis (MS) is a chronic inflammatory as well as a neurodegenerative brain disorder. When a person is diagnosed with MS, the therapy goal is to stop or slow down the natural course of disease evolution, while balancing at the same time an acceptable level of burden, risks of side effects, and costs. The development of MS treatments has brought tremendous progress. Since the launch of the first drug in 1993, the health of MS patients, expressed in quality-adjusted life years (QALYs), has increased by 66% thanks to the availability of disease-modifying treatments (DMTs) (Hult et al. 2017).

Currently, there are more than 20 FDA-approved DMTs available for MS patients (nationalmssociety.org/Treating-MS/Medications), typically subdivided into low- and high-efficacy drugs based on the clinical trial data. These DMTs are intended to reduce the disease burden, disease activity, and progression, but they do not cure the underlying disease. Hence, the current treatment guidelines state that any evidence of disease activity while on consistent treatment should prompt the consideration of an alternative regimen to optimize therapeutic benefit (Costello et al. 2019).

Currently, we don’t have access to precision medicine tools that help to select the most optimal treatment for each individual MS patient. Therefore, early switching from suboptimal treatments is the main target, thereby also taking into account the drug safety profiles and patients’ preferences.

But, MS treatment decisions are in the eye of the beholder

As early switching from suboptimal treatments is key, it’s important to evaluate how these decisions are made. Therapy selection, either immediately after diagnosis or in further follow-up, is driven by the perceived level of clinical and subclinical disease activity and progression. Evidence of clinical disease activity is monitored by new relapses and disability worsening, as often measured by the EDSS scale. Subclinical disease activity and progression are evaluated on brain (and spinal cord) magnetic resonance imaging (MRI) scans, by measuring the number of new and enlarging lesions as well as brain atrophy (Giovannoni et al. 2015). Hence, treatment decisions in MS depend on the availability of clinical and subclinical information, as shown in the Figure below.

As the aim in MS is to stop or significantly slow down the natural course of disease evolution, the treatment paradigms are focused on finding ‘no evidence of disease activity’ (NEDA). In this context, NEDA-3 refers to taking clinical and MRI-based lesion information into account, for NEDA-4 the above and brain atrophy is evaluated.?

MS treatment decision-making depends on several factors, such as patient preference, intolerability and adverse effects, the clinical and subclinical information available at the time of the decision, and the (observer) variability of the available information. Indeed, the clinical and subclinical disease activity assessments are prone to observer interpretations and measurement errors. In this context, it is known that EDSS measures, if available, can suffer from up to 20-30% inter-rater variability (Noseworthy et al. 1990, Goodkin 1991) and that almost half of the relapses are missed by people with MS (Duddy et al. 2014). In addition, when looking at the subclinical disease activity detection, it is known that up to 24% of radiological reports of brain scans contain discrepancies (Rosenkrantz et al. 2018).

Why it matters? The health (economic) impact of treatment decisions in MS

Early switching from suboptimal treatments matters, as being on a suboptimal treatment has the same impact as not being on treatment (Hult et al. 2017). In this context, it is known that patients stay on their unsuccessful first treatment for almost as long (3.9 years) as they stay on their successful treatment (4.2 years) (Rio et al. 2012). Given the fact that disease progression is irreversible in MS, optimized treatment decisions are crucial to impact an MS patient’s long-term health.??

Evaluating the health economic impact of treatment decisions in MS is also extremely relevant, as MS is the second most costly chronic condition in the US, with more than $4 million in total lifetime costs per patient (Adelman, Rane, and Villa 2013; Chen, Chonghasawat, and Leadholm 2017; Owens et al. 2013). With several MS therapies available, it has been shown that adopting a more personalized medicine in MS, including data-driven clinical decision-making, has the potential to increase the health impact of existing treatments by over 50%, and therefore significantly reduce costs (Hult 2017).

Also, thanks to recent imaging AI innovations, reliable regulatory cleared software solutions for MRI volumetry are being increasingly used in clinical practice to enhance radiological reporting (Van Hecke et al. 2021). This technology brings potential advantages in terms of enhanced sensitivity for detecting subclinical pathologic aspects, as well as increased accuracy, reproducibility, and consistency when combined with the radiologist's visual evaluation (Erbayat Altay et al. 2013; Rosenkrantz et al. 2018; Wang et al. 2017).?

Given the significant impact treatment decisions have on the huge burden and costs associated with MS, we performed a health (economic) microsimulation study (Sima et al. 2021). Specifically, the effect of different treatment paradigms, including the use of FDA-cleared brain MRI software to quantify disease activity, on outcomes and costs was assessed. We ran the model on a simulated cohort of 1000 MS patients from the moment they were prescribed a low efficacy DMT, and evaluated the impact of the therapeutic decision-making path on outcomes, health utilities, and related costs, adopting the US healthcare perspective (Sima et al. 2021). We thereby assume that, following the guidelines, patients with detected disease activity should switch to other (and typically higher efficacy) DMTs, in order to have the best outcome.

Time is brain: effect of quantitative MRI on detecting disease activity

Based on the results of this 1000 patient microsimulation study, it is clear that using MRI and assistive software leads to benefits in terms of faster detection of disease activity and better long-term health outcomes due to faster escalation to high efficacy DMTs, as shown in the Figure below (Sima et al. 2021).

Indeed, a clear difference in the detection of true disease activity was observed of 28%, 42%, 60%, and 75% after one year for decisions based on clinical information without MRI, NEDA-3 with visual MRI reading, NEDA-3 with software assistance, and NEDA-4 with software assistance, respectively (Sima et al. 2021). The undetected disease activity for these treatment strategies after year 1 is 50%, 36%, 18%, and 4%, respectively (Sima et al. 2021).?

As more patients with active disease are detected earlier, a faster escalation to high efficacy DMT occurs in the decision-making strategies assisted by (software-assisted) MRI. This is illustrated in the DMT distribution per strategy in the Figure below, with notable differences in escalation speed between the strategies (Sima et al. 2021).

Brain is time: impact on long-term health and costs

It is clear that lowering the time MS patients are on suboptimal therapies is crucial. The average time per patient in the “undetected disease activity” state heavily depends on the information at hand and the treatment paradigm used, which also impacts the time people with MS are on suboptimal low-efficacy treatment, as shown in the Figure below.

The introduction of DMTs in the early nineties improved the lives of people with MS with about 0.5 QALYs accumulated in 15 years (Chirikov et al. 2017). We now demonstrated that the introduction of a software-supported treatment paradigm (NEDA-4 with software) has the potential to add 0.34 QALYs compared to a visual analysis of brain MRI scans in 15 years, as shown in the Table below (Sima et al. 2021).

This QALY change represents a relative improvement of 68% compared to the introduction of disease-modifying therapies. This significantly improved health status, by adopting the optimal treatment decision paradigms, will naturally impact the cost of care as well. The annualized costs, as well as incremental comparisons between the strategies, are presented in the Table below (Sima et al. 2021).

Time is now?

So what does this all mean? I guess, first of all, this research reinforces the importance of data-driven decision-making in MS, of having all relevant information at hand when decisions are made, and of being able to make these decisions timely.

We all aim for precision medicine, digital twins, and ways to predict disease outcomes. And at the speed technology is evolving, one day we’ll be able to fulfill these aspirations. But let’s first start with what we can do now. Let’s start with measuring what can be measured. Let’s start with integrating these measures into treatment decision paradigms for each individual person with MS.

Of course, treatment decisions go beyond the availability and usage of MRI and clinical information. Intolerability, adverse effects, patient preference, and convenience also play an important role in deciding therapy (dis)continuation. Taking such aspects into account would increase the percentage of patients that stop or switch DMTs, but these factors weren’t taken into account in this study, as it would affect all treatment strategies considered (Sima et al. 2021).

I realize that it always takes time to adopt new technology. And that not all hospitals have access to quantitative MRI data currently. But time is now. Every 8 seconds, somewhere in this world, an MS patient sits in front of a clinician to evaluate how things are going. Though we can’t predict disease courses on an individual level yet, let’s just make sure that all information is available so that neurologists can detect disease activity and progression, and that the time people with MS are on suboptimal treatment is minimized.

?Together, let’s try to make an impact. Every 8 seconds.

References

• Adelman ET AL. (2013) The Cost Burden of Multiple Sclerosis in the United States: A Systematic Review of the Literature. Journal of Medical Economics 16 (5): 639–47.?
• Chen et al. (2017) Multiple Sclerosis: Frequency, Cost, and Economic Burden in the United States. Journal of Clinical Neuroscience: Official Journal of the Neurosurgical Society of Australasia 45 (November): 180–86.?
• Costello K, et al. (2019). The use of disease-modifying therapies in multiple sclerosis: principles and current evidence. A consensus paper by the Multiple Sclerosis Coalition.
• Duddy M, et al. The UK patient experience of relapse in Multiple Sclerosis treated with first disease modifying therapies. Mult Scler Relat Disord . 2014 Jul;3(4):450-6.
• Erbayat et al. (2013) Reliability of Classifying Multiple Sclerosis Disease Activity Using Magnetic Resonance Imaging in a Multiple Sclerosis Clinic. JAMA Neurology 70 (3): 338–44.
• Giovannoni, G., et al. (2015). Brain health: time matters in multiple sclerosis. https://doi.org/10.21305/msbh.001
• Goodkin (1991). EDSS reliability Neurology, 41(2, Pt 1):332
• Hult KJ et al. (2017). Measuring the Potential Health Impact of Personalized Medicine: Evidence from MS Treatments. National Bureau of Economic Research. doi:10.3386/w23900
• Noseworthy et al. (1990). Interrater variability with the Expanded Disability Status Scale (EDSS) and Functional Systems (FS) in a multiple sclerosis clinical trial. The Canadian Cooperation MS Study Group. Neurology, 40(6):971-5
• Owens et al. (2013) Perspectives for Managed Care Organizations on the Burden of Multiple Sclerosis and the Cost-Benefits of Disease-Modifying Therapies. Journal of Managed Care Pharmacy: JMCP 19 (1 Suppl A): S41–53.?
• Río J, et al. (2012). Change in the clinical activity of multiple sclerosis after treatment switch for suboptimal response. Eur J Neurol . 19(6):899-904.
• Rosenkrantz et al. (2018). Discrepancy Rates and Clinical Impact of Imaging Secondary Interpretations: A Systematic Review and Meta-Analysis. Journal of the American College of Radiology: JACR 15 (9).?
• Sima et al. (2021). Health Economic Impact of Software-Assisted Brain MRI on Therapeutic Decision-Making and Outcomes of Relapsing-Remitting Multiple Sclerosis Patients—A Microsimulation Study. Brain Sci. 2021, 11(12), 1570
• Van Hecke, Costers et al. (2021) A Novel Digital Care Management Platform to Monitor Clinical and Subclinical Disease Activity in Multiple Sclerosis. Brain Sciences 11 (9). https://doi.org/10.3390/brainsci11091171.?
• Wang et al. (2017) Neuroradiologists Compared with Non-Neuroradiologists in the Detection of New Multiple Sclerosis Plaques. AJNR. American Journal of Neuroradiology 38 (7): 1323–27.?