Quantitative In Silico Reality Check on the Antiparasitic-Cancer-Cure Hypothesis: A Multi-Cancer PBPK and Network-Pharmacology Evaluation of Ivermectin and Fenbendazole Across Seven Solid and Hematologic Malignancies
Year: 2026
Type: article
Abstract: Background.Following anecdotal reports popularised on social media, oral self-administration of the veterinaryantiparasitics ivermectin (IVM) and fenbendazole (FBZ) is being adopted by a growing patientpopulation as putative cancer therapy. Genuine peer-reviewed in vitro evidence supports anti-tumouractivity for both compounds: IVM modulates the WNT/TCF, PAK1, MDR1/P-gp and KPNB1axes, while FBZ destabilises microtubules at the colchicine binding site of β-tubulin and inhibitsglucose uptake via GLUT4 and HK-II. However, no controlled human phase II/III data exist, andthe quantitative gap between in vitro potency and clinically achievable plasma exposure has neverbeen formally evaluated across cancer types.Methods.We adapted the open-source CureNet eight-layer computational pharmacology pipeline [1] toevaluate IVM and FBZ across seven malignancies: glioblastoma (GBM), acute myeloid leukaemia(AML), triple-negative breast cancer (TNBC), colorectal cancer (CRC), non-small-cell lung cancer(NSCLC), pancreatic ductal adenocarcinoma (PDAC) and cutaneous melanoma (MEL). Eachcancer was modelled with disease-specific gene sets, protein-protein interaction (PPI) topology,target proteins (with PDB structures), pathway enrichment, fold-change priors and standard-of-carechemotherapeutics. The CureNet pipeline executes ADMET filtering, target-disease overlap, networkpharmacology, empirical and (where feasible) AutoDock-GPU physics-based molecular docking,100 ns equivalent molecular dynamics, Bliss/Loewe synergy analysis, an eight-variable ODE pathwaymodel and transcriptomic herb-drug-interaction analysis. We extended this pipeline with threeadditional analytical rungs: (i) a physiologically-based pharmacokinetic (PBPK) plasma-versus-IC50 reality check parameterised by published human Cmax data, (ii) an allometric oral-dose extrapolationcomparing IC50-equivalent dose against documented safety ceilings, and (iii) a composite plausibilityscore combining in-silico signal strength with pharmacokinetic feasibility.Results.Across all seven cancers, the in vitro IC50 values for both drugs (IVM 3–15 μM; FBZ 0.8–3.0 μM)lie 0.6–30-fold above the maximum total-plasma Cmax achievable at the safety ceiling, and aftercorrection for plasma protein binding (IVM 93%, FBZ 95%) the free-drug fold-deficit ranges from12.3× (FBZ in NSCLC) to 439.6× (IVM in CRC). The composite plausibility score classifies zeroof seven cancers as "promising", five as "worth further preclinical study" (TNBC, NSCLC, CRC,MEL, AML; scores 42–51 of 100) and two as "implausible at safe doses" (PDAC, GBM; both 33).The strongest in-silico signals — IVM–PAK1 (–5.2 kcal/mol) and FBZ–β-tubulin (–4.9 kcal/mol)— recapitulate the established mechanisms but cannot be reached at safe oral doses without newformulations. The triple-combination Layer-7 ODE predicts 87.9% cell death uniformly acrosscancers when chemotherapy is combined with sub-IC50 IVM and FBZ, supporting a chemosensitiserhypothesis rather than a monotherapy cure hypothesis.Conclusions.The anti-cancer biology of IVM and FBZ is mechanistically real, but the social-media "oral antiparasiticcures cancer" narrative is not supported by quantitative pharmacokinetics at any testedcancer. The free-drug plasma concentration achievable within the human safety envelope sits oneto three orders of magnitude below the in vitro killing concentration. Two specific paths preservescientific honesty: (a) intravenous or liposomal reformulation that raises Cmax 10–30-fold (mostpromising in AML, where IC50 is lowest and the BCL-2/MCL-1 axis converges with venetoclax),and (b) sub-IC50 chemosensitiser trials with established standard-of-care chemotherapy doing thecytotoxic work (most promising in CRC with FOLFOX, NSCLC with osimertinib, and MEL withvemurafenib). Until such data exist, oral self-medication at doses sufficient to match in vitro IC50 ismathematically impossible without exceeding documented hepatotoxicity and neurotoxicity ceilings
Authors: Rai, Ankush, Jain, Kamal
Institution: Indian Institute of Technology Roorkee
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7224644910