Non-MRI biomarkers should be accepted to guide treatment escalation in MS: Yes

Combined assessments of disease activity are established practice when neurologists consider escalating disease-modifying treatment (DMT) in people with multiple sclerosis (pwMS). Currently, there is a strong focus on clinical relapses or brain magnetic resonance imaging (MRI) showing active inflammatory activity (i.e. new T2 or contrast-enhancing lesions), while markers of neurodegeneration (i.e. atrophy) receive less attention in daily practice. ¹ We argue that non-imaging biomarkers succeed in providing additional information: most notably neurofilament light chain (NfL) which reflects neuroaxonal injury and can be readily measured in blood. For pwMS on a new DMT with stable imaging and clinic but continuously elevated NfL, the paradox is striking: we have evidence of ongoing damage but no clear mandate to act. Here, we argue that non-MRI biomarkers should guide treatment escalation.

Technology readiness

A novel class of biomarkers has emerged in the past decade: blood-based markers have been extensively studied for their ability to capture disease-associated processes in the central nervous system, while offering the advantages of being minimally invasive, inexpensive, and amenable to repeated assessment. ² In numerous large cohort studies and clinical trials in MS, serum NfL (sNfL) and glial fibrillary acidic protein (sGFAP) are associated with disease activity, severity, treatment response and progression.^{3 –6} sNfL is the most thoroughly investigated biomarker and first to receive a CE-IVDR-mark for predicting MS disease activity, ⁶ next to a FDA Letter of Support for use as a trial outcome measure and Breakthrough Device Designations. As a cytoskeletal protein subunit of the axon, NfL elevations in body fluids reflect neuroaxonal breakdown. Despite a lack of etiologic specificity, sNfL has proven to be a strong indicator of acute disease activity and treatment response in MS.^{4 –6} Most consistently, significant reductions in sNfL have been seen in pwMS when initiating DMTs, with low-efficacy drugs but even more so with high-efficacy drugs. ⁴ Glial fibrillary acidic protein, on the other hand, is the key intermediate filament of astrocytes, which are abundant in the CNS and integral for function and homeostasis. Multiple cohort studies have found elevated sGFAP concentration in MS and an association with the severity of disability. ⁷ Growing evidence favours sGFAP over sNfL for prognosticating disease progression. ³ How to make sense of sGFAP levels of an individual still requires further research. Regarding treatment response, sGFAP shows inconsistent behaviour and conflicting results, ⁷ therefore rendering it a less suitable marker to monitor immediate need for escalation. For these reasons, we focus on sNfL to guide treatment escalation in MS.

From group level to the individual patient

For blood-based markers, we face several challenges when moving from cohort and trial data to individual monitoring of patients: First, we need to use robust assays and understand variation. Next to relying on a stable analytical variation, we need to know more about intra-individual (biological) variation. So far, little data has been reported on daily or weekly variation in pwMS. ⁴ A reference change value, combining both analytical variation and intra-individual variation, has recently been derived from a memory clinic cohort, ⁸ but has not yet been established for NfL in pwMS. This is especially important when defining an ‘informative change’ from baseline that can be applied to the longitudinal measurements of one individual to detect disease activity. ⁴ Second, biological confounders jeopardize correct interpretation of absolute sNfL levels. Known factors like age and body weight have been addressed by setting values in relation to a reference population (i.e. through Z-scores/percentiles). However, any comorbidities affecting the nervous system (i.e. diabetes or trauma) have to be taken into consideration when interpreting sNfL. Third, sensitivity of sNfL to detect inflammatory disease activity has recently been questioned: 71% of pwMS in the RESTORE trial showing gadolinium-enhancing (Gd+) lesions did not exhibit sNfL elevations. ⁹ The low sensitivity of 28% might have been linked to a very conservative threshold (95th percentile). Or sNfL was not elevated yet and missed: there was a median delay of approximately 8 weeks for sNfL to peak after appearance of first Gd+ lesions and the study lacked measurement time points beyond 12–16 weeks. However, given the accessibility and cost-effectiveness of a blood marker, frequent (i.e. 3-monthly) sNfL measurements could potentially still detect relevant disease activity earlier than 6-monthly or yearly MRI. Most importantly, the idea behind sNfL is not to serve as a mere surrogate for Gd+ lesions. Demonstrated by prognostic studies,^{4 –6} sNfL flags those inflammatory events substantial enough to cause measurable axonal injury and relevant to long-term outcomes, whether inflammation can be detected on MRI or not. Especially in people not showing clinical and MRI disease activity, this additional information is very relevant to detect suboptimal therapy response.

Real-world adoption

Clinical adoption requires both physicians’ openness to the new technology and clear, validated guidelines. In a tertiary clinic, physicians were recently surveyed both when requesting sNfL and after the disclosure of the measurements. Overall, clinical certainty changed after disclosure in 50% of the cases and there was an association of decision changes and higher sNfL levels, indicating both trust in the technology and actionable information gained. ¹⁰ To make guidance explicit, a group of international experts set out to review personalized algorithms taking into account sNfL. ¹¹ Overall, consensus was highest for escalation where sNfL was considered as an extra piece of information next to clinical or MRI findings. But even for scenarios with no evidence of clinical or MRI disease activity, 67% of experts agreed on escalating DMTs based on sNfL. This Delphi study forms the basis for a pragmatic clinical trial to assess 6-monthly sNfL monitoring on patient-relevant outcomes. It will hopefully greatly improve our understanding of the value of personalized treatment using sNfL.

To summarize, we strongly believe in the complementary value of non-MRI biomarkers, specifically sNfL, to guide treatment escalation in MS next to clinical and MRI parameters. The evidence bar has clearly been met and the technology is fit for purpose. Interpretive challenges at the individual level can be mitigated. Grey-zone cases will persist, but perfect should not be the enemy of good.