Researchers recently used high-precision hierarchical clustering to identify disease subtypes in patients with intermediate-to-high risk pulmonary embolism upon admission, a new study showed.
“Prognostic risk assessment and stratification of PE patients is of critical importance in guiding PE diagnosis and treatment strategies to reduce PE mortality,” study researchers wrote. “However, current intermediate-high risk PE includes heterogeneous subgroups with varying prognosis, raising controversy regarding thrombolysis.”
A retrospective study was designed to apply machine learning algorithms to further risk stratify patients with intermediate-to-high risk pulmonary embolism. This study analyzed data from 79 patients at two clinical centers. Cluster analysis was performed based on 10 continuous variables including age, gender, chronic lung disease, and Pulmonary Embolism Severity Index (PESI) score.
Using this method, three clusters were identified. one main cluster containing 67 cases (cluster 2), and two additional clusters containing 6 cases each (clusters 1 and 3). There were statistically significant differences between the clusters regarding age, chronic lung disease, chronic heart disease, and diabetes. Specifically, cluster 3 (six cases) was “characterized by older age, chronic heart disease, and diabetes.”
Importantly, there was also a significant difference in PE mortality at 1 and 3 years between groups. For example, the 1-year PE mortality rate was 100% in cluster 1 (6 cases) and 97.0% in the predominant cluster, whereas it was 50% in cluster 3. Cluster 3 also had significantly longer ICU and hospital lengths and higher total costs than the other two groups.
“In this study, hierarchical clustering uncovered a subgroup with poor prognosis in patients with intermediate-to-high risk PE at admission,” the researchers wrote. “Further research is therefore needed to qualify the large sample size with a prospective study and validate the results with external datasets.”
