Sulindac

證據等級: L5

預測適應症: 10 個

Sulindac: From NSAID Anti-Inflammatory Use to Acromesomelic Dysplasia (Hunter-Thompson Type)

One-Sentence Summary

Sulindac is an NSAID (indole acetic acid class); this evidence pack does not confirm a specific approved indication in Taiwan, as the drug currently holds no Taiwan marketing license. The TxGNN model’s top-ranked prediction is Acromesomelic Dysplasia, Hunter-Thompson Type, but the evidence pack’s own analysis flags this as likely a knowledge-graph artifact rather than a genuine signal — 0 clinical trials and 0 publications support it, and no other candidate among the 10 predicted indications has any direct trial or literature evidence either.

Quick Overview

Item	Content
Original Indication	Not available — no Taiwan license on file; drug is a known NSAID (indole acetic acid class) per class references in this evidence pack, but no specific original indication text is provided
Predicted New Indication	Acromesomelic Dysplasia, Hunter-Thompson Type (top TxGNN score; see caveat below)
TxGNN Prediction Score	99.92%
Evidence Level	L5 (model prediction only, no actual studies)
Market Status (Taiwan)	Not Marketed (未上市)
Number of NDAs	0
Recommended Decision	Hold

Why is This Prediction Reasonable?

Currently, detailed mechanism of action data is not available for Sulindac in this evidence pack (flagged as a High-severity data gap, DG002). Based on general pharmacological class knowledge referenced within this evidence pack’s own candidate rationales, Sulindac is an NSAID that works primarily through COX-1/COX-2 inhibition, with some COX-independent, pro-apoptotic effects noted in the literature for other tissues.

For the top-ranked candidate, acromesomelic dysplasia (Hunter-Thompson type), the evidence pack’s own mechanistic assessment is explicit: this is a skeletal developmental disorder caused by GDF5/CDMP1 mutations, not an inflammatory condition, and there is no plausible pathophysiological link to Sulindac’s COX-inhibition/pro-apoptotic mechanism. The evidence pack states this high score is likely graph-embedding noise rather than a real pharmacological signal. The same pattern — high TxGNN score, but no mechanistic plausibility — applies to most of the top-ranked candidates (rank 1–8: acromesomelic dysplasia, brachyolmia-amelogenesis imperfecta syndrome, brachyolmia, myosclerosis, pseudoachondroplasia, brachydactyly-syndactyly syndrome, colobomatous microphthalmia-rhizomelic dysplasia syndrome, WHIM syndrome), all of which are rare structural/genetic disorders with no inflammatory component and are scored “Hold.”

By contrast, two lower-scoring but mechanistically more coherent candidates were staged further along in the internal review process (S1, “Research Question” rather than “Hold”):

Rank	Disease	TxGNN Score	Evidence Level	Decision Stage	Recommendation
1	Acromesomelic dysplasia, Hunter-Thompson type	99.92%	L5	S0	Hold
2	Brachyolmia-amelogenesis imperfecta syndrome	99.90%	L5	S0	Hold
3	Brachyolmia	99.90%	L5	S0	Hold
4	Myosclerosis	99.89%	L5	S0	Hold
5	Pseudoachondroplasia	99.85%	L5	S0	Hold
6	Brachydactyly-syndactyly syndrome	99.82%	L5	S0	Hold
7	Colobomatous microphthalmia-rhizomelic dysplasia syndrome	99.81%	L5	S0	Hold
8	WHIM syndrome	99.74%	L5	S0	Hold
9	Rheumatoid vasculitis	99.63%	L5	S1	Research Question
10	Hypermobility of coccyx	99.56%	L5	S1	Research Question

Rheumatoid vasculitis (rank 9) is a severe inflammatory vascular complication of rheumatoid arthritis; NSAIDs including sulindac have long-standing clinical use for RA symptom control, giving a theoretical (though unproven for this specific indication) mechanistic basis via COX-mediated anti-inflammatory activity — though standard treatment for vasculitis itself is steroids/immunosuppressants, with NSAIDs typically an adjunct at most. Hypermobility of coccyx (rank 10) commonly presents with sacrococcygeal joint inflammation and mechanical pain; NSAIDs are empirically used for coccydynia-type pain in clinical practice, giving this candidate a defensible pharmacological rationale despite the absence of direct supporting studies.

Clinical Trial Evidence

Currently no related clinical trials registered

Literature Evidence

Currently no related literature available

US Market Information

No marketing authorization is currently registered for Sulindac in Taiwan (market status: 未上市 / Not Marketed; 0 licenses on file).

Safety Considerations

Please refer to the package insert for safety information.

Conclusion and Next Steps

Decision: Hold

Rationale: All 10 TxGNN-predicted indications for sulindac are supported only by model score, with zero clinical trials or literature (Evidence Level L5) across the board. The drug has no Taiwan marketing license, and both mechanism-of-action data and TFDA label warnings/contraindications are missing — the latter being a Blocking-severity gap (DG001) that prevents any safety pre-screening (S1). The single highest-scoring candidate is explicitly identified within the evidence pack itself as a likely knowledge-graph artifact rather than a genuine signal, so it should not be treated as the strongest lead by default.

To proceed, the following is needed:

TFDA label (warnings/contraindications) — Blocking gap, required before any S1 safety review (DG001)
Confirmed mechanism of action from DrugBank or primary literature (DG002)
Direct clinical trial or literature evidence, prioritizing the two mechanistically plausible candidates already advanced to “Research Question” stage: rheumatoid vasculitis and hypermobility of coccyx
Independent review of the top-ranked “acromesomelic dysplasia” prediction to confirm/rule out embedding artifact before allocating further review resources to it
Disclaimer

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