A Clinically and Biologically Based Subclassification of the Idiopathic Inflammatory Myopathies Using Machine Learning
Bibliographic record
Abstract
Objective Published predictive models of disease outcomes in idiopathic inflammatory myopathies (IIMs) are sparse and of limited accuracy due to disease heterogeneity. Computational methods may address this heterogeneity by partitioning patients based on clinical and biological phenotype. Methods To identify new patient groups, we applied similarity network fusion (SNF) to clinical and biological data from 168 patients with myositis (64 adult polymyositis [PM], 65 adult dermatomyositis [DM], and 39 juvenile DM [JDM]) in the Rituximab in Myositis trial. We generated a sparse proof‐of‐concept bedside classifier using multinomial regression and identified characteristics that distinguished these groups. We conducted χ2 tests to link new patient groups with the myositis subtypes. Results SNF identified five patient groups in the discovery cohort that subdivided the myositis subtypes. The sparse multinomial regressor to predict patient group assignments (areas under the receiver operating characteristic curve = [0.78, 0.97]; areas under the precision‐recall curve = [0.55, 0.96]) found that autoantibody enrichment defined four of these groups: anti–Mi‐2, anti–signal recognition peptide (SRP), anti–nuclear matrix protein 2 (NXP2), and anti‐synthetase (Syn). Depletion of immunoglobulin M (IgM) defined the fifth group. Each group was associated with one subtype, with adult DM being associated with anti–Mi‐2 and anti‐Syn autoantibodies, JDM being associated with anti‐NXP2 autoantibodies, and adult PM being associated with IgM depletion and anti‐SRP autoantibodies. These associations enabled us to further resolve the current myositis subtypes. Conclusion Using unsupervised machine learning, we identified clinically and biologically homogeneous groups of patients with IIMs, forming the basis of an integrated disease classification based on both clinical and biological phenotype, thus validating other approaches and what has been previously described.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".