Phenoconversion in Thiopurine and Methotrexate Pharmacotherapy: A Systematic Framework for Integrating Germline Pharmacogenomics with Dynamic Phenotype Markers

Background: Pharmacogenomics (PGx) has improved thiopurine and methotrexate (MTX) safety by identifying inherited variants in drug-metabolizing enzymes and transporters, including TPMT, NUDT15, and folate-pathway genes. Genotype-guided dosing is established for thiopurines, whereas MTX PGx evidence in immune-mediated disease remains weaker, without comparable consensus recommendations. Germline genotype alone cannot fully predict the expressed pharmacological phenotype, and phenoconversion—discordance between genotype-predicted and actual metabolic capacity—remains insufficiently addressed during inflammation, infection, hospitalization, and co-medication changes. Methods: This systematic review was conducted using prespecified eligibility and evidence-appraisal criteria and reported in accordance with PRISMA 2020 guidelines. Results: Thiopurines and MTX were selected as contrasting evidence cases. Thiopurines have established TPMT/NUDT15-guided starting doses, routine safety monitoring, and erythrocyte 6-TGN/6-MMP measurements. In contrast, MTX pharmacogenetic associations remain inconsistent in immune-mediated disease; erythrocyte polyglutamate thresholds lack standardization, and concentration-guided interventions are established primarily in high-dose oncology. Across 45 core reports, the most reproducible dynamic modifiers included inflammation-mediated ADME suppression, drug–drug interactions, organ dysfunction, and exposure changes detectable through serial laboratory or metabolite measurements. Epigenetic, translational, post-translational, and microbiome mechanisms remain less clinically established. The operational thiopurine panel is therefore limited to TPMT/NUDT15 status, treatment context, CBC, liver/renal chemistry, disease-activity markers as reassessment triggers, and erythrocyte metabolites; DNA-TG and TPMT-locus methylation remain investigational. Conclusions: The Phenotype-Predictive Score is a theoretical, hypothesis-generating framework, not a validated clinical score, calculator, or dosing algorithm. Its equations and module weights require future drug-, indication-, endpoint-, and time-horizon-specific estimation in multicenter cohorts using penalized regression or dynamic pharmacokinetic models. PPS must not reassign TPMT/NUDT15 metabolizer phenotypes or guide treatment until external validation establishes calibration, transportability, safety, and clinical net benefit.