Gary Birnbaum, MD
Predicting Progression: Identifying Risks for Transition to Progressive Disease
A novel prognostic score to assess the risk of progression in
relapsing−remitting multiple sclerosis patients
Anna Isabella Pisani, Antonio Scalfari, Francesco Crescenzo, Chiara Romualdi,
and Massimiliano Calabrese
Eur J Neurol. 2021;28:2503–2512
Predicting the future can be challenging, but for persons with MS, knowing how their illness may evolve is of utmost importance. Historically, about one-half of persons with relapsing forms of multiple sclerosis transition to a course of gradual progression (secondary progressive MS). While use of potent disease-modifying therapies reduces the risk of developing secondary progressive MS, progressive MS can still develop. Unfortunately, only nominally effective therapies are currently available for this phase of disease.
While the underlying tissue changes found in progressive forms of multiple sclerosis appear very early in relapsing MS not all persons evolve to the progressive phase. It thus would be of great value to identify persons with relapsing forms of multiple sclerosis who are at highest risk for developing secondary progression. The above paper describes a new and novel way to accurately identify individuals at high risk.
The authors used a combination of clinical evaluations, MRI imaging of the central nervous system and a machine learning program (the Random Survival Forest model) to assess risks of developing progressive MS. Test results were combined into a “secondary progressive risk score” (SP-RiSc). This score was able to predict progression 88% of the time (accuracy), with a specificity of 87% (ability to identify true negatives) and a 92% sensitivity (percentage of correct predications). This powerful tool thus could: a) provide insights into brain changes that lead to progression, b) allow health care providers to offer high-risk individuals treatment with the most effective disease-modifying therapies in an effort to delay transition to progressive disease, and c) identify persons most suited to participate in trials developing new treatments for progressive disease.
Key Points Related to Above Article:
1. Certain clinical characteristics are associated with a more disabling course of MS and with a greater risk of developing secondary progressive disease. Some such characteristics are older age at disease onset, being male, multiple clinical attacks with poor recovery, and multiple inflammatory and structural changes in central nervous system grey matter and meninges on MRI imaging.
2. Treatment options for persons with progressive forms of multiple sclerosis are very limited. While some drugs are approved for this phase of disease they provide, at best, only modest benefit.
3. Treatment with high-efficacy disease-modifying therapies reduces, but does not eliminate, the risk of developing secondary progressive MS. Treatment with high-efficacy disease-modifying therapies also carries an increased risk of toxicities and adverse events. Thus, such drugs may not be suitable for all persons with MS. The early identification of individual risk for developing secondary progressive MS could influence the choice of initial disease-modifying therapy.
4. Identifying persons at high risk for progressive disease would also be of great value in identifying potential participants for clinical trials of drugs for progressive disease.
5. The authors followed 262 persons with MS from onset of diagnosis for a mean time of almost ten years. All patients were initially treated with either interferon-beta or glatiramer acetate. When “breakthrough” disease was noted, individuals were switched to either fingolimod or natalizumab.
6. Twelve different variables were studied. These included age, gender, disability status (EDSS), extent of brain tissue loss (atrophy), especially in the cerebral cortex and cerebellum, numbers of cortical and cerebellar MRI lesions, and whether the spinal cord was involved.
7. Using a machine learning program, “Random Survival Forest,” the authors ranked variables according to their effect on the risk of developing progressive disease. Each variable was given a dimensionless “minimal depth,” and these were included in the risk score, SP-RiSc.
8. Seven of the twelve variable had the strongest predictive values. These included loss of tissue in the cortex (“cortical thinning”), loss of brain tissue (atrophy), especially in the cerebellar cortex over two years, the initial numbers of lesions at diagnosis and the accumulation of new lesions of the next two years, especially in the grey matter. Age at diagnosis and total numbers of white matter lesions were only moderately predictive of transition to progression. The other variables noted above did not improve prediction.
9. When the seven most predictive variables for risk were combined into the SP-RiSc, persons with MS could be divided into three groups: low risk (110 persons), medium risk (54 persons), and high risk (55 persons).
10. Over the almost 10-year follow-up, sixty-nine persons with MS (26%) developed secondary progressive MS, most in the high-risk group (46 individuals or 85.5%), only 9 in the medium risk group (1`7.7%), and none in the low-risk group. One hundred fourteen individuals (43.5%) switched their disease-modifying therapy to either fingolimod or natalizumab due to “breakthrough” disease.
11. The authors conclude that their technique has the potential to be of great value in predicting conversion to secondary progressive MS in persons with relapsing forms of multiple sclerosis, and that their data are supported by observations from other research groups, “Our model highlighted the early accumulation of focal and diffuse grey matter damage as the most important determinants of the conversion to the progressive phase, supporting the crucial role played by the grey matter pathology in the development of late severe disability.”
Despite the advent of large numbers of disease-modifying therapies for relapsing forms of multiple sclerosis, there are no therapies of equal effectiveness for progressive forms of multiple sclerosis. Several drugs are approved for progressive MS, but their benefit is, at best, modest and greatest in those individuals that also have evidence of acute central nervous system inflammation.
The features of progressive multiple sclerosis do not suddenly appear following transition from relapsing multiple sclerosis to secondary progressive MS. They are present very early in the disease, becoming predominant as the acute inflammatory phase of MS subsides. Treatment with high efficacy disease-modifying therapies reduces the risk of transitioning from relapsing multiple sclerosis to secondary progressive MS, especially if started early. The practical issue is that these drugs have increased risks for toxicities and adverse reactions, and not all persons with relapsing forms of multiple sclerosis transition to secondary progressive MS. It is thus essential to only expose persons at high risk for developing secondary progressive MS to high-efficacy drug-related risks. The paper by Pisani and colleagues provides a means of doing just that. It is additionally reassuring that their technique using a machine-learning program is in line with previous observations showing that inflammation and loss of grey matter in the central nervous system, especially the cerebral and cerebellar cortices area hallmark of progressive MS .
Health care providers, using the techniques described in the Pisani et al paper, will be able to recommend the appropriate disease-modifying therapies to persons early in their course of MS in a manner consistent with their risks and potential benefits (high-efficacy drugs for individuals at high-risk for transition to progressive disease and less toxic, but possibly less effective drugs for low-risk individuals). In addition, categorizing persons with MS according to their risks for developing progressive disease will allow more appropriate and selective recruitment of participants into therapeutic trials of drugs for this phase of the illness.