Why pharmacogenomic biomarkers for chemotherapy-induced peripheral neuropathy fail: a systematic review of genetic associations, replication, and clinical translation.

Class-specific candidate germline variants: CEP72 and ETAA1 for vincristine, EPHA5, FGD4 and FZD3 for taxanes, ABC transporter genes and GSTP1 for platinum agents; replication overall poor.

Systematic review (PROSPERO CRD42025635757) of 72 genetic association studies of chemotherapy-induced peripheral neuropathy, covering four neurotoxic drug classes: vincristine, taxanes, platinum agents and bortezomib. Meta-analysis was not possible because of heterogeneity in study design and outcome measures, and cohorts were predominantly of European ancestry. The most consistently replicated loci were CEP72 and ETAA1 for vincristine, EPHA5, FGD4 and FZD3 for taxanes, and ABC transporter genes and GSTP1 for platinum agents; for bortezomib, associations involving TRPV1, DNA repair pathways and inflammatory signalling remained emerging and largely unreplicated. Marked heterogeneity in neuropathy definitions, phenotyping instruments and effect allele reporting limits comparability and reproducibility.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

This is a methodological post-mortem rather than a biomarker review, and that is exactly its value: it explains why twenty years of studies have produced no routinely usable test to predict chemotherapy-induced neuropathy. The practical message is clear: none of these variants, CEP72 included, should be offered outside research. The proposed roadmap — harmonised phenotyping, multi-ancestry GWAS, functional validation — is the right one, but the absence of meta-analysis leaves the reader without a quantitative ranking of signals.

Analysis by Dr Thibaut Benquey