Sorafenib Tosylate

Other protein kinase inhibitors · given by oral · on the market since 2005

Leads

71.5M

People

Makers

How abandoned

What it's made of

Each functional group on the molecule does a specific, predictable job.

SETS HOW LONG IT LASTS

Trifluoromethyl

What it does. Blocks metabolism hard and makes the molecule substantially greasier and more membrane-permeable at the same time.

What it opens up. Long half-life plus good tissue and brain penetration — a common signature in drugs designed to be taken once daily.

SETS HOW LONG IT LASTS

Aromatic halogen

What it does. Slows liver breakdown and increases membrane crossing. Also fills space in a binding pocket, which often raises potency several-fold.

What it opens up. Better tissue penetration and longer duration, at the cost of higher fat storage.

Where it can get to

Predicted from molecular weight, lipophilicity, polar surface area and hydrogen-bond donor count. A drug can only act where it distributes, so this constrains which conditions are addressable before any biology is considered. derived

Easily

Systemic circulation

Absorbed from the gut and distributed in plasma.

Barely

Central nervous system

Crosses the blood–brain barrier.

Intestinal lumen only

Poorly absorbed. Acts locally in the gut with minimal systemic exposure.

Barely

Skin

Crosses the stratum corneum. Viable as a topical.

Barely

Urinary tract

Cleared unchanged in urine. Local concentration exceeds plasma.

What it binds — and whether we can reach it · 8 of 7 reachable

Binding affinity is only meaningful against the plasma concentration reached at the approved dose. A target requiring higher concentrations cannot be engaged in a patient, whatever the in vitro potency. Most repurposing hypotheses fail at this step, and it is rarely checked before a trial is proposed. measured

Values are pAct — the negative logarithm of the molar concentration at which the drug half-occupies the target. Each whole step is a tenfold difference in potency: 9.0 = 1 nanomolar, 6.0 = 1 micromolar, 5.0 = 10 micromolar. Oral drugs typically reach 0.1–10 micromolar in plasma, so roughly 6.0 is the threshold below which a target cannot be reached at a tolerated dose. Figures come from DrugCentral’s curated binding data.

KDR

Kinase · inhibitor

10.68

Engaged easily

FLT3

Kinase · inhibitor

8.35

Engaged easily

RAF1

Kinase · inhibitor

8.22

Engaged easily

RET

Kinase · inhibitor

8.13

Engaged easily

FLT4

Kinase · inhibitor

7.8

Engaged

KIT

Kinase · inhibitor

7.8

Engaged

BRAF

Kinase · inhibitor

7.66

Engaged

Target	Kind	Action	Affinity (pAct)	Can it be engaged?
KDR	Kinase	inhibitor	10.68	Engaged easily
FLT3	Kinase	inhibitor	8.35	Engaged easily
RAF1	Kinase	inhibitor	8.22	Engaged easily
RET	Kinase	inhibitor	8.13	Engaged easily
FLT4	Kinase	inhibitor	7.8	Engaged
KIT	Kinase	inhibitor	7.8	Engaged
BRAF	Kinase	inhibitor	7.66	Engaged

New uses, and who is already trying them

One card per candidate indication: the reasoning link by link, the evidence tier that governs how strongly it can be stated, and any registered trials.

5 of the 216 conditions linked to this drug’s targets have been checked against trial records so far. The rest are not ruled out — they are unchecked.

Not yet used for these — 5

gastrointestinal stromal tumor

Tried in people

It showed an effect in a human trial that was not followed up.

The condition

Gastrointestinal stromal tumor (GIST) is the most common mesenchymal neoplasm of the gastrointestinal (GI) tract, typically presenting in adults over the age of 40 (mean age 63), and only rarely in children, in various regions of the GI tract, most commonly the stomach or small intestine but also less commonly in the esophagus, appendix, rectum and colon. GISTs can be asymptomatic or present with various non-specific signs, depending on the location and size of tumor, such as loss of appetite, anemia, weight loss,

KIT · KIT proto-oncogene, receptor tyrosine kinase

Tyrosine-protein kinase that acts as a cell-surface receptor for the cytokine KITLG/SCF and plays an essential role in the regulation of cell survival and proliferation, hematopoiesis, stem cell maintenance, gametogenesis, mast cell development, migration and function, and in melanogenesis. In response to KITLG/SCF binding, KIT can activate several signaling pathways. Phosphorylates PIK3R1, PLCG1, SH2B2/APS and CBL. Activates the AKT1 signaling pathway by phosphorylation of PIK3R1, the regulatory subunit of phospha

Open Targets scores each kind of evidence from 0.00 to 1.00 for this exact pairing — KIT and gastrointestinal stromal tumor. The score reflects how much evidence of that kind exists and how consistent it is. It is not a probability that the drug works. Rows are ordered by how much each kind counts toward the evidence figure used in the ranking, not by its raw score, because the two disagree: literature co-occurrence saturates near 1.00 on almost any well-studied pair, while genetic association rarely exceeds 0.70 even when causal.

Genetic association0.91 / 1.00

Inherited variants in this gene alter disease risk in human populations. The datatype most predictive of later clinical success.55% of evidence score

Somatic mutation0.78 / 1.00

This gene is recurrently mutated in this tumour type.15% of evidence score

Animal model0.48 / 1.00

Altering this gene in animals reproduces or relieves the disease phenotype.15% of evidence score

Pathway0.90 / 1.00

The gene sits in a pathway with established disruption in this condition.8% of evidence score

Curated genetic literature0.82 / 1.00

Gene–disease links curated from published genetic studies.5% of evidence score

Literature co-occurrence0.99 / 1.00

Gene and condition appear together in published text. Reflects publication volume, not causal evidence.2% of evidence score

Clinical0.98 / 1.00

Derived from clinical or trial records.not counted

Phase 2 done, then stopped

A Phase 2 trial finished and was not followed by Phase 3. Dosing and safety in this population are already established.

NCT00265798 Sorafenib in Treating Patients With Mali · Ph 2 · 38 people

+1 more on ClinicalTrials.gov

What would kill it. A null trial, or pharmacokinetic data showing the drug does not reach KIT in the relevant tissue at tolerated doses.

cardiofaciocutaneous syndrome

Mechanism implies it

It could plausibly help.

The condition

Cardiofaciocutaneous (CFC) syndrome is a RASopathy characterized by craniofacial dysmorphology, congenital heart disease, dermatological abnormalities (most commonly hyperkeratotic skin and sparse, curly hair), growth retardation and intellectual disability.

BRAF · B-Raf proto-oncogene, serine/threonine kinase

Protein kinase involved in the transduction of mitogenic signals from the cell membrane to the nucleus (Probable). Phosphorylates MAP2K1, and thereby activates the MAP kinase signal transduction pathway (PubMed:21441910, PubMed:29433126). Phosphorylates PFKFB2 (PubMed:36402789). May play a role in the postsynaptic responses of hippocampal neurons (PubMed:1508179). {ECO:0000269|PubMed:1508179, ECO:0000269|PubMed:21441910, ECO:0000269|PubMed:29433126, ECO:0000269|PubMed:36402789, ECO:0000305}.

Open Targets scores each kind of evidence from 0.00 to 1.00 for this exact pairing — BRAF and cardiofaciocutaneous syndrome. The score reflects how much evidence of that kind exists and how consistent it is. It is not a probability that the drug works. Rows are ordered by how much each kind counts toward the evidence figure used in the ranking, not by its raw score, because the two disagree: literature co-occurrence saturates near 1.00 on almost any well-studied pair, while genetic association rarely exceeds 0.70 even when causal.

Genetic association0.94 / 1.00

Inherited variants in this gene alter disease risk in human populations. The datatype most predictive of later clinical success.55% of evidence score

Animal model0.45 / 1.00

Altering this gene in animals reproduces or relieves the disease phenotype.15% of evidence score

Somatic mutation0.00 / 1.00

This gene is recurrently mutated in this tumour type.15% of evidence score

Pathway0.83 / 1.00

The gene sits in a pathway with established disruption in this condition.8% of evidence score

Curated genetic literature0.86 / 1.00

Gene–disease links curated from published genetic studies.5% of evidence score

Literature co-occurrence0.53 / 1.00

Gene and condition appear together in published text. Reflects publication volume, not causal evidence.2% of evidence score

Nobody has tried this

No registered trial of this drug in this condition.

What would kill it. A null trial, or pharmacokinetic data showing the drug does not reach BRAF in the relevant tissue at tolerated doses.

Noonan syndrome

Mechanism implies it

It could plausibly help.

The condition

Noonan Syndrome (NS) is characterized by short stature, typical facial dysmorphism and congenital heart defects.

RAF1 · Raf-1 proto-oncogene, serine/threonine kinase

Serine/threonine-protein kinase that acts as a regulatory link between the membrane-associated Ras GTPases and the MAPK/ERK cascade, and this critical regulatory link functions as a switch determining cell fate decisions including proliferation, differentiation, apoptosis, survival and oncogenic transformation. RAF1 activation initiates a mitogen-activated protein kinase (MAPK) cascade that comprises a sequential phosphorylation of the dual-specific MAPK kinases (MAP2K1/MEK1 and MAP2K2/MEK2) and the extracellular s

Open Targets scores each kind of evidence from 0.00 to 1.00 for this exact pairing — RAF1 and Noonan syndrome. The score reflects how much evidence of that kind exists and how consistent it is. It is not a probability that the drug works. Rows are ordered by how much each kind counts toward the evidence figure used in the ranking, not by its raw score, because the two disagree: literature co-occurrence saturates near 1.00 on almost any well-studied pair, while genetic association rarely exceeds 0.70 even when causal.

Genetic association0.93 / 1.00

Inherited variants in this gene alter disease risk in human populations. The datatype most predictive of later clinical success.55% of evidence score

Animal model0.49 / 1.00

Altering this gene in animals reproduces or relieves the disease phenotype.15% of evidence score

Pathway0.94 / 1.00

The gene sits in a pathway with established disruption in this condition.8% of evidence score

Curated genetic literature0.86 / 1.00

Gene–disease links curated from published genetic studies.5% of evidence score

Literature co-occurrence0.48 / 1.00

Gene and condition appear together in published text. Reflects publication volume, not causal evidence.2% of evidence score

Nobody has tried this

No registered trial of this drug in this condition.

What would kill it. A null trial, or pharmacokinetic data showing the drug does not reach RAF1 in the relevant tissue at tolerated doses.

medullary thyroid gland carcinoma

Tried in people

It showed an effect in a human trial that was not followed up.

The condition

A neuroendocrine carcinoma arising from the C-cells of the thyroid gland. It is closely associated with multiple endocrine neoplasia syndromes. Approximately 10% to 20% of medullary thyroid carcinomas are familial. Patients usually present with a thyroid nodule that is painless and firm. In the majority of cases nodal involvement is present at diagnosis. Surgery is the preferred treatment for both primary lesions and recurrences. This carcinoma is generally not very sensitive to radiation and almost unresponsive to

RET · ret proto-oncogene

Receptor tyrosine-protein kinase involved in numerous cellular mechanisms including cell proliferation, neuronal navigation, cell migration, and cell differentiation in response to glia cell line- derived growth family factors (GDNF, NRTN, ARTN, PSPN and GDF15) (PubMed:20064382, PubMed:20616503, PubMed:20702524, PubMed:21357690, PubMed:21454698, PubMed:24560924, PubMed:28846097, PubMed:28846099, PubMed:28953886, PubMed:31118272). In contrast to most receptor tyrosine kinases, RET requires not only its cognate ligan

Open Targets scores each kind of evidence from 0.00 to 1.00 for this exact pairing — RET and medullary thyroid gland carcinoma. The score reflects how much evidence of that kind exists and how consistent it is. It is not a probability that the drug works. Rows are ordered by how much each kind counts toward the evidence figure used in the ranking, not by its raw score, because the two disagree: literature co-occurrence saturates near 1.00 on almost any well-studied pair, while genetic association rarely exceeds 0.70 even when causal.

Genetic association0.94 / 1.00

Inherited variants in this gene alter disease risk in human populations. The datatype most predictive of later clinical success.55% of evidence score

Somatic mutation0.61 / 1.00

This gene is recurrently mutated in this tumour type.15% of evidence score

Animal model0.35 / 1.00

Altering this gene in animals reproduces or relieves the disease phenotype.15% of evidence score

Curated genetic literature0.80 / 1.00

Gene–disease links curated from published genetic studies.5% of evidence score

Literature co-occurrence0.99 / 1.00

Gene and condition appear together in published text. Reflects publication volume, not causal evidence.2% of evidence score

Clinical0.94 / 1.00

Derived from clinical or trial records.not counted

Phase 2 done, then stopped

A Phase 2 trial finished and was not followed by Phase 3. Dosing and safety in this population are already established.

NCT01736878 Efficacy and Safety Study of Sorafenib t · Ph 2

+5 more on ClinicalTrials.gov

What would kill it. A null trial, or pharmacokinetic data showing the drug does not reach RET in the relevant tissue at tolerated doses.

multiple endocrine neoplasia type 2A

Tried in people

It showed an effect in a human trial that was not followed up.

The condition

An autosomal dominant tumor predisposition disorder caused by pathogenic variants in the RET gene, characterized by an increased risk of medullary thyroid carcinoma, pheochromocytoma, and hyperparathyroidism.

RET · ret proto-oncogene

Open Targets scores each kind of evidence from 0.00 to 1.00 for this exact pairing — RET and multiple endocrine neoplasia type 2A. The score reflects how much evidence of that kind exists and how consistent it is. It is not a probability that the drug works. Rows are ordered by how much each kind counts toward the evidence figure used in the ranking, not by its raw score, because the two disagree: literature co-occurrence saturates near 1.00 on almost any well-studied pair, while genetic association rarely exceeds 0.70 even when causal.

Genetic association0.94 / 1.00

Inherited variants in this gene alter disease risk in human populations. The datatype most predictive of later clinical success.55% of evidence score

Curated genetic literature0.84 / 1.00

Gene–disease links curated from published genetic studies.5% of evidence score

Literature co-occurrence0.46 / 1.00

Gene and condition appear together in published text. Reflects publication volume, not causal evidence.2% of evidence score

Phase 2 done, then stopped

A Phase 2 trial finished and was not followed by Phase 3. Dosing and safety in this population are already established.

NCT00390325 Sorafenib Tosylate in Treating Patients · Ph 2 · 21 people

What would kill it. A null trial, or pharmacokinetic data showing the drug does not reach RET in the relevant tissue at tolerated doses.

Why nobody will fund it

5 companies make this across 5 approved products. With that many manufacturers the price has fallen to production cost. A company funding a trial could not recover it — competitors could market the same molecule for the new indication immediately, without contributing.

How abandoned it is3 / 100

0 to 100, calculated from manufacturer count, remaining patent or exclusivity, and whether any reformulation route to exclusivity is still open. 100 means no route to commercial return exists. The weights behind it are unsourced.

THE ONE WAY A COMPANY COULD STILL OWN IT

injectableinhaledmodified-release

A new route or dosage form qualifies for a 505(b)(2) application, which carries three years of exclusivity for that formulation. It is the only remaining route to exclusivity on a generic molecule.

What it would cost to find out

Derived, not quoted. Events needed is 4(z_α+z_β)²/ln(HR)²; enrolment is that divided by the event rate and follow-up. The 1/ln(HR)² term dominates: halving the detectable effect size roughly quadruples enrolment. derived

Effect sought (hazard ratio)0.75

Lower = bigger effect = smaller trial

How often the event happens per year8.0%

Follow-up2 yr

Cost per patient$32K

Unsourced placeholder

379

Events needed

Patients

$93M

Trial cost

assumed

$3.71

Per person reached

assumed

Cost per person reached is the relevant metric for a non-commercial funder, since there is no revenue to set against the outlay.

Ask for ideas nobody has listed

Reasons from the functional groups and distribution above, excluding conditions already listed. Output is hypotheses for falsification, not findings. Not medical advice.

Who might pay for the trial

Trials of off-patent drugs are rarely funded by venture capital or pharma. They are paid for by government grants and disease foundations. Here is who has funded work on this molecule before, what federal mechanisms are open now, and which foundations fund trials in the disease areas it touches. measured

Federal grants that name this medicine

668 grants on NIH RePORTER

5R01CA278956-03(NCI · National Cancer Institute)Active

$356K (2026)

Transcatheter Intra-Arterial Delivery of Oriented Anti-PD-L1 Immune checkpoint inhibitors Immobilized Nanocarriers for Local Combination Immunotherapy of Hepatocellular Carcinoma

PI: Kim, Dong-Hyun · Northwestern University, Chicago, IL · through 2029-03

View on NIH RePORTER ↗Apply to same mechanism (PA-20-185) ↗

5U54MD019970-03(NIMHD · National Institute on Minority Health)Active

$382K (2026)

Rio Grande Valley Cancer Health Disparity Research Center

PI: Chauhan, Subhash C. · University Of Texas Rio Grande Valley, Edinburg, TX · through 2029-03

View on NIH RePORTER ↗Apply to same mechanism (RFA-MD-24-001) ↗

5R01CA269140-04(NCI · National Cancer Institute)Active

$480K (2026)

Targeting the LIFR-LCN2 pathway to improve liver cancer therapy

PI: Ma, Li · University Of Tx Md Anderson Can Ctr, Houston, TX · through 2027-12

View on NIH RePORTER ↗Apply to same mechanism (PA-20-185) ↗

See all 668 projects on NIH RePORTER →

Open federal programmes that would fund this work

Open notices with an explicit mandate to fund work on approved or off-patent molecules.

RFA-CA-25-033National Institutes of HealthForecast — not open yet

Glioblastoma Therapeutics Network (GTN; U19 Clinical Trial Required)

Grants.gov (RFA-CA-25-033) ↗Direct listing ↗Who already holds this grant ↗

RFA-FD-25-020Food and Drug Administrationcloses 05/16/2028

Reissue of RFA-FD-23-001- Clinical Studies of Orphan Products Addressing Unmet

Read announcement ↗Grants.gov (RFA-FD-25-020) ↗Direct listing ↗Who already holds this grant ↗

FOR-FD-25-020Food and Drug AdministrationForecast — not open yet

Clinical Studies of Orphan Products Addressing Unmet Needs of Rare Diseases (R

Grants.gov (FOR-FD-25-020) ↗Direct listing ↗Who already holds this grant ↗

FOR-FD-25-001Food and Drug AdministrationForecast — not open yet

Clinical Trials Addressing Unmet Needs of Rare Neurodegenerative Diseases (R01

Grants.gov (FOR-FD-25-001) ↗Direct listing ↗Who already holds this grant ↗

RFA-TR-25-002National Institutes of Healthcloses 07/02/2027

Preclinical Proof of Concept Studies for Rare Diseases (R21 Clinical Trial Not

Read announcement ↗Grants.gov (RFA-TR-25-002) ↗Direct listing ↗Who already holds this grant ↗

PAR-25-374National Institutes of Healthcloses 05/07/2028

Translational Bioinformatics and Experimental Approaches to Advance Drug Repos

Read announcement ↗Grants.gov (PAR-25-374) ↗Direct listing ↗Who already holds this grant ↗

HT942526PRCRPCTADefense Health Agency Contracting Activity -closes 10/05/2026

DoW Peer Reviewed Cancer, Clinical Trial Award

Grants.gov (HT942526PRCRPCTA) ↗Direct listing ↗Who already holds this grant ↗

PAR-26-049National Institutes of HealthForecast — not open yet

Establishing Pediatric CNS Pharmacodynamic Measures as Tools to Enable Psychia

Grants.gov (PAR-26-049) ↗Direct listing ↗Who already holds this grant ↗

Showing 8 of 990 currently open under this category on Grants.gov, matched to the disease areas this drug touches. derived Browse all 990 on Grants.gov ↗

Funders beyond the federal system

Grants.gov lists US federal money. Many off-patent drug trials are supported by disease foundations and nonprofits. These are matched to the body systems this drug touches. assumed

Matched by disease area, not by molecule — a foundation appearing here funds work in a field this drug touches.

Who manufactures it

Companies holding an approved abbreviated application for this molecule in the United States, from the FDA Orange Book. Supply already exists at scale, so a trial would not need a manufacturing programme — but any of these could also be a partner, or an objector.

Dr Reddys Laboratories LtdMylan Pharmaceuticals IncTeva Pharmaceuticals Usa Inc

Manufacturers

companies with their own approval

Approved products

distinct strengths and forms

Routes

oral

2005

Oldest product still sold

approval year of the earliest product still on the market

ALSO SOLD AS

NEXAVAR

Nothing on this page is medical advice. Candidate uses are unproven hypotheses derived from public data; established uses are listed for verification only. Figures carry their provenance inline — anything marked assumed or derived is not a measurement.

← Back to the full ledger

What it's made of

Trifluoromethyl

Aromatic halogen

Where it can get to

What it binds — and whether we can reach it · 8 of 7 reachable

New uses, and who is already trying them

gastrointestinal stromal tumor

cardiofaciocutaneous syndrome

Noonan syndrome

medullary thyroid gland carcinoma

multiple endocrine neoplasia type 2A

Why nobody will fund it

What it would cost to find out

Ask for ideas nobody has listed

Who might pay for the trial

Who manufactures it

Closely related medicines