Leflunomide

證據等級: L5

預測適應症: 2 個

Leflunomide: Original Indication Not on File → Predicted New Indication: Brachydactyly-Syndactyly Syndrome

One-Sentence Summary

Leflunomide’s original approved indication and mechanism of action are not documented in this evidence pack (flagged as Blocking/High-severity data gaps). The TxGNN model predicts potential efficacy for brachydactyly-syndactyly syndrome, a rare congenital limb-development disorder, but this prediction is supported by zero clinical trials and zero publications, and the model’s own rationale flags the mechanistic link as biologically implausible — likely a statistical artifact from shared gene nodes in the knowledge graph rather than a genuine therapeutic signal.

Quick Overview

Item	Content
Original Indication	Not documented in evidence pack (see Data Gaps)
Predicted New Indication	Brachydactyly-Syndactyly Syndrome
TxGNN Prediction Score	99.93%
Evidence Level	L5 (model prediction only, no supporting studies)
US Market Status	✗ Not Marketed
Number of NDAs	0
Recommended Decision	Hold

Why is This Prediction Reasonable?

Detailed mechanism of action data for leflunomide is not available in this evidence pack (Data Gap DG002, High severity). Leflunomide is known in general pharmacology as a DHODH (dihydroorotate dehydrogenase) inhibitor that suppresses pyrimidine synthesis, producing anti-inflammatory/immunomodulatory effects — but this is not sourced from the evidence pack itself and cannot be treated as verified for this report.

More importantly, the model’s own repurposing rationale argues against a plausible mechanistic connection: brachydactyly-syndactyly syndrome is a congenital limb-development disorder driven by developmental gene regulation, not an inflammatory or immune-mediated pathology. There is no documented mechanistic pathway linking a pyrimidine-synthesis inhibitor to skeletal/limb morphogenesis. The evidence pack explicitly attributes the high TxGNN score to likely indirect graph associations (e.g., shared gene nodes between skeletal-development pathways and the drug’s targets) rather than a treatment-relevant relationship.

The same caveat applies to the second-ranked candidate, colobomatous microphthalmia-rhizomelic dysplasia syndrome — another rare congenital malformation syndrome (ocular and proximal skeletal dysplasia) with no plausible link to leflunomide’s known pharmacology and no corroborating trials or literature.

Clinical Trial Evidence

Currently no related clinical trials registered.

Literature Evidence

Currently no related literature available.

US Market Information

No marketing authorizations are on file for leflunomide in this jurisdiction (0 licenses; market status recorded as Not Marketed).

Safety Considerations

Please refer to the package insert for safety information. (Note: TFDA label warnings/contraindications are an unresolved Blocking data gap — DG001 — and drug-drug interaction data was not found.)

Conclusion and Next Steps

Decision: Hold

Rationale: Both predicted indications are Evidence Level L5 (model prediction only) with no supporting clinical trials or literature, and the mechanistic rationale itself suggests the high TxGNN scores likely reflect indirect knowledge-graph associations rather than true biological plausibility. In addition, a Blocking safety data gap (missing TFDA label/warnings) prevents even a preliminary safety screen.

To proceed, the following is needed:

TFDA label (warnings, contraindications) — Blocking gap, required before any safety pre-screen
Verified mechanism of action data (e.g., via DrugBank API) to properly assess mechanistic plausibility
Independent literature/preclinical validation specifically testing leflunomide in developmental/skeletal disorders, given the model’s own flag that this signal may be a graph artifact
If no independent support emerges, deprioritize both candidates rather than advancing to further evaluation stages
Disclaimer

This content is for research purposes only and does not constitute medical advice. Clinical validation is required before any clinical application.