Analysis IDs 4 patient groups with distinct ALS progression patterns
Such real-world findings 'key' to better understanding disease, per study
A computer analysis using real-world data from an amyotrophic lateral sclerosis (ALS) clinic identified four groups of patients with distinct rates of disease progression, a new study by researchers in Portugal reported.
The findings continue to demonstrate the differences in ALS presentation and progression seen among people with the rare neurodegenerative disease, and may contribute to identifying specific groups of patients who may respond differently to existing and investigational treatments.
“Identifying groups of patients with similar disease progression patterns is key to understand disease heterogeneity [variability], guide clinical decisions and improve patient care,” the scientists wrote, noting that their work “unraveled four clinically relevant disease progression groups.”
“We encourage the application of our method in other ALS populations to establish its utility,” the team added.
The study, “Temporal stratification of amyotrophic lateral sclerosis patients using disease progression patterns,” was published in Nature Communications.
Scientists used computer algorithm called ClusTric on patient data
ALS is a neurodegenerative disorder marked by worsening muscle weakness, which leads to movement problems and other symptoms such as difficulties swallowing, speaking, and breathing.
After an ALS diagnosis, patients are typically monitored with standardized tools such as the Revised ALS Functional Rating Scale (ALSFRS-R). This measure tracks changes in the person’s ability to do day-to-day functions such as turning in bed, walking, grabbing a pencil, or speaking.
Collecting such data over time can identify patterns that help clinicians better understand a patient’s disease progression. However, while ALS patients generally decline over time, they may experience periods of stability with mild improvement alternating with periods of rapid decline.
This variability, or heterogeneity, in symptom severity and ALS progression makes identifying patterns challenging.
To address this problem, scientists at the University of Lisbon applied a computer algorithm called ClusTric to real-world patient data. Their goal was to classify ALS progression into patient groups based on similar rates and patterns. The algorithm allowed for variable changes in disease progression over time.
Data were collected from the Lisbon ALS Clinic, which has regularly followed ALS patients since 1995. As of 2023, the clinic database contained records and information on 1,677 ALS patients.
Real-world data from ALS clinic used to analyze progression patterns
Based on clinician-defined key ALS features collected over five clinic visits covering about one year of follow-up, ClusTric detected four patient clusters. These were identified as slow progressors, moderate bulbar progressors, moderate spinal progressors, and fast progressors.
Importantly, the muscles around the face and throat are called the bulbar muscles, and individuals who first experience symptoms involving the head and neck are said to have bulbar-onset disease. This typically is seen in about one-third of patients. The other two-thirds have spinal-onset disease, marked by initial symptoms of muscle weakness affecting the limbs.
In these clusters, slow progressors were mostly younger male patients who progressed slowly in the bulbar and respiratory domains. This cluster had better lung function and ALSFRS-R scores at the first clinic visit and a lower ALSFRS-R decline in the first six months.
Moderate bulbar progressors consisted of patients with better limb and trunk function and more symptoms affecting the face and throat muscles. This group had more female patients (51.9%), and had an older age at onset than the slower progressors group. These patients showed slightly better lung and ALSFRS-R scores than did the moderate spinal progressors but worse than slow progressors.
Moderate spinal progressors, for their part, were predominantly male patients (59.8%) and had an average onset age of 60. These individuals had better bulbar function than did bulbar progressors. Their lung function was similar to bulbar progressors, but their limb and trunk scores were comparable to fast progressors. This is consistent with the symptoms seen in people with limb-onset or spinal ALS, whose symptoms first affect the limbs.
As expected, fast progressors had the worst outcomes. This group had the highest proportion of female patients (71%), the oldest age at onset (average 70 years), and the lowest lung and ALSFRS-R scores at the first clinic visit.
ClusTric analysis was nearly 400 times faster than another algorithm
The researchers validated these outcomes against the Pooled Resource Open-Access ALS Clinical Trials (PRO-ACT) database, which holds data from 11,675 ALS patients who participated in previous clinical trials.
Similar clusters were identified with both ClusTric and PRO-ACT.
Compared with a different clustering algorithm called MoGP, which identified progression groups based on ALSFRS-R total scores, ClusTric uncovered relationships between multiple features and patients over time almost 400 times faster than MoGP.
A ClusTric survival analysis also found that fast progressors had the shortest mean survival of 24.6 months, or two years, and a two-year survival rate of 27%. In comparison, slow progressors lived the longest, with a mean survival of 53.9 months, or 4.5 years, with a two-year survival rate of 81%. Moderate progressors, both bulbar and spinal, had similar survival rates and times.
Our results show that [a computer algorithm called] ClusTric … was effective in capturing ALS disease progression by grouping a heterogeneous hospital-based population in four very distinct prognosis groups validated by clinicians.
Because ALS progression can change for certain patients at any time, the team determined whether patients remained in their cluster or changed over time. When the first six months were compared with the next six months, most patients (66.8%) remained in the same cluster during the 12 months of follow-up, whereas the remaining patients (33.2%) changed their cluster.
When slow progressors, classified in the first six months, were reclassified in the second six months as either slow progressors, moderate progressors, or moderate bulbar progressors, the new slow progressor group had a two-year survival rate of 93%.
“Our results show that ClusTric … was effective in capturing ALS disease progression by grouping a heterogeneous hospital-based population in four very distinct prognosis groups validated by clinicians,” the researchers wrote.
About the Author
