Genotype-Phenotype Predictive Gap in Pre-Emptive Screening of DPD Deficiency: Improving Sensitivity

DPYD

Unclassified rare DPYD variants and co-occurrence of the common SNPs rs1801160 and rs2297595, contributing to enzyme deficiency missed by standardized markers

Fluoropyrimidine toxicity is mainly driven by reduced dihydropyrimidine dehydrogenase activity, but standardized DPYD pharmacogenomic markers are rare and leave a substantial share of at-risk individuals undetected. This retrospective genotype-phenotype association study included 687 Italian subjects who underwent phenotyping by the 5-fluorouracil degradation assay and full exon sequencing of DPYD, grouped into 94 deficient subjects (more than 30% activity reduction) and 593 with normal activity. With phenotype as the reference standard, three binary tests were compared: standardized pharmacogenomic markers (Test 1), rare missense, nonsense, frameshift or splice-site variants (Test 2), and co-occurrence of rs1801160 and rs2297595 (Test 3). All three were significantly associated with the deficient phenotype (ORs of 6, 4.85 and 4.9 respectively), with minimal overlap between the subjects positive on each. Combining all three raised screening sensitivity from 12.77% to 35.11%, with specificity falling from 97.64% to 91.74%.

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

A sensitivity of 12.77% for guideline markers is the real message here: current pre-emptive screening misses most biochemical deficiencies, which enzyme phenotyping makes visible. Tripling sensitivity at the cost of six points of specificity would be a defensible trade-off against a potentially lethal toxicity, but the study is retrospective and uses an enzymatic phenotype, not observed clinical toxicity, as its reference. Before changing a genotyping protocol, one would need to confirm that the patients thus captured are indeed those who develop severe toxicity.

Analysis by Dr Thibaut Benquey