Much of the recent success of “precision medicine’ in Oncology has been the identification of somatic mutations and development of “targeted” therapies. However, inherited (germline) differences can affect the safety or effectiveness of drugs by altering the metabolic elimination of a drug or by altering the activation of prodrugs. The use of SNP-based genetic biomarkers (pharmacogenetics) to predict metabolism and therapeutic outcome has many advocates. However, there are many layers of regulatory control of genes, as well as environmental and clinical factors which can alter phenotype and influence therapeutic outcomes. My talk will highlight the role of phenotype-based approaches to help understand and clarify whether inherited (pharmacogenetic) risk factors are associated with poor bioactivation of cyclophosphamide and the excessive toxicity of 5-fluorouracil.

Fluoropyrimidines (FPs), including 5-fluorouracil (5FU), have been widely used since the 1950s, with over two million cancer patients treated annually. Despite their efficacy, severe or fatal toxicities occur in approximately 10% of patients. Dihydropyrimidine dehydrogenase (DPD) inactivates ~85% of administered 5FU, making it a key determinant of systemic drug exposure and toxicity. Four common DPYD gene variants account for ~20% of early-onset FP-related toxicity, leading to the implementation of routine genotyping and genotype-guided dosing in several European countries, including Switzerland, as recommended by the European Medicines Agency (EMA).
To address the remaining, largely non-genetic variability and to avoid underdosing, the Swiss Group of Pharmacogenomics and Personalised Therapy recommends the integration of therapeutic drug monitoring (TDM). Similar recommendations were issued by the International Association of Therapeutic Drug Monitoring and Clinical Toxicology (IATDMCT) in 2019. However, despite reimbursement by the Swiss healthcare system, clinical uptake of 5FU TDM remains limited due to practical barriers.
Currently, individualized dosing is based on body surface area (BSA), a method that has proven inadequate in preventing FP-related toxicities. Additionally, venous blood sampling for 5FU TDM is considered cumbersome and a barrier to routine implementation. Here, I present initial findings from two Swiss studies: one validating a dried blood spot (DBS)-based TDM method as a more patient-friendly and logistically feasible alternative; the other investigating demographic, biochemical, and anthropometric factors contributing to interindividual variability in 5FU exposure to support more accurate, personalized chemotherapy dosing.

Pharmacogenomically-guided dose personalisation is one of the advancing strategies of cancer medicine. Genotyping of DPYD is leading the way to refinements in fluoropyrimidine (FP; 5-fluorouracil and capecitabine) dosing to reduce the risk of severe toxicity and death. Four variants in the DPYD gene, present in 7-8% of Caucasian (European) patients, are known to contribute to 30-60 % of severe toxicity when treated with standard doses of FP, which are commonly prescribed for advanced or adjuvant therapy for a range of malignancies of the GI tract, head and neck, breast and other sites. PGx screening of patients can lead to personalized dose reduction of the first cycle of FP chemotherapy with reduced high-grade toxicity risk. A strategy of increasing subsequent doses in patients who do not develop toxicity can maintain anti-cancer benefit. One challenge in Asia-Pacific countries is the diverse ethnicities, different to Europe, including north-Asian, south-Asian and south-east Asian, many indigenous peoples, Pacifika, and Maori. There is a relative lack of data about the range of deleterious DPYD variants in these populations, their effects on DPYD activity, and the risk of severe toxicity. In particular there is limited data on safe dosing of heterozygote variant carriers; it is uncertain whether homozygotes can be treated with FP at all. In Australia, the population is ethnically diverse with a unique indigenous people, and limited genomic data. We are undertaking a national study to describe the spectrum of deleterious DPYD variants in Australians, and to determine the enablers and barriers to DPYD genotyping and FP dose personalisation in our communities. The longterm aim is to implement DPYD genotyping equitably throughout our diverse communities to minimise severe toxicity from FPs in all Australian cancer patients without compromising the benefits. An understanding of the predominant genetic variants causing severe toxicity will facilitate equity. Inclusion of community champions in the conduct of research will be required to ensure safe and culturally sensitive conduct of pharmacogenomics research to achieve these aims. This presentation will discuss the challenges in implementing this research study throughout Australia, to ensure equity and inclusion, and to maximize its value.

We conducted a genome-wide association study (GWAS) on steady-state Z-endoxifen levels in 636 hormone-receptor positive breast cancer patients from Singapore, Lebanon, and Germany who have received 20mg tamoxifen daily for at least 8 weeks prior to their pharmacokinetic (PK) blood sampling. Significant associations (P<5×10-8) were followed up in a validation cohort comprising 859 participants from the UK, France, and Singapore. We also evaluated the association between genetic markers surpassing genome-wide significance and CYP2D6 metabolizer status on the survival outcomes of 1326 non-metastatic hormone-receptor positive breast cancer patients treated with adjuvant tamoxifen . The outcomes evaluated were disease-free survival, relapse free survival and distant relapse free survival.
In the PK GWAS, genome-wide significant association with steady-state Z-endoxifen levels was observed on chromosome 22 at CYP2D6 and at TCF20 rs932376 A>G. Scrutiny of TCF20 rs932376 A>G and CYP2D6 metabolizer status using multivariable Gaussian generalized linear model with log link function showed that both are independent predictors of steady-state endoxifen levels. Each additional copy of the TCF20 rs932376-G allele was associated with a 10.79nM reduction (95% CI -13.39 to -8.20) in mean endoxifen level (P=3.26×10-16). Adjustment for CYP2D6 metabolizer status, age, ethnicity, weight, and menopausal status still resulted in a significant association between TCF20 rs932376-G and serum levels of endoxifen, with each copy of the G allele associated with a 7.48nM reduction (95% CI -10.58 to -4.39, P=2.16×10-6) in mean endoxifen level. This finding was validated (adjusted 6.30nM reduction in mean endoxifen level per-copy of the G allele [95% CI -9.51 to -3.09], P=1.00×10-4) in independently ascertained collections. Neither TCF20 rs932376 nor CYP2D6 metabolizer status was significantly associated with breast cancer outcomes in a model adjusting for age, ethnicity, body mass index,tumour size, nodal status, and tumour grade.