The economic costs of precision medicine for clinical translational research among children with high-risk cancer

Cancer is the leading cause of disease-related death in children in high-income countries, including Australia. High-risk paediatric malignancies (expected overall survival <30%) account for roughly 20–25% of diagnoses annually. Conventional therapies cause substantial acute and chronic toxicities, leading to significant long-term burden. Precision paediatric oncology aims to tailor treatment by characterizing each child’s tumour at a molecular level, potentially improving outcomes and reducing harms and costs by avoiding ineffective treatments. Multiple programmes globally have implemented precision approaches using multi-omics and preclinical models, but adoption as standard care requires evidence of clinical response and cost-effectiveness. Despite various studies of paediatric precision medicine, there is limited evidence on the costs of implementing comprehensive testing. This study’s purpose is to systematically measure the costs of employing precision medicine for high-risk childhood cancers within Australia’s Zero Childhood Cancer Precision Medicine Programme (ZERO), which integrates multi-omics profiling and, where feasible, preclinical drug testing, to inform diagnosis, prognosis, and management.

The paper notes numerous paediatric precision oncology programmes and studies globally, demonstrating feasibility and potential clinical utility of genomic profiling. However, it highlights limited evidence regarding economic feasibility and the cost of implementing comprehensive precision testing in paediatrics. Prior cost analyses have often focused on single technologies or narrower test ranges. The authors found no previous studies matching the breadth of ZERO’s combined multi-omics and preclinical platform. As a comparator, a micro-costing study by Schwarze et al. reported costs for adult matched tumour and germline genome sequencing (~£6841; ~AUD $12,742 in 2020), which is comparable to ZERO’s overall access cost for a far more comprehensive platform, supporting observed cost reductions over time in comprehensive approaches.

Design and setting: Sub-study (EPIC-CC PRISM Cohort) within ZERO’s national, multicentre, prospective clinical trial PRISM (NCT03336931). Participants: Children and adolescents (0–21 years; adults with paediatric-type cancers) with histologically confirmed high-risk malignancy (expected overall survival <30%). Ethics approval by Hunter New England Human Research Ethics Committee, NSW, Australia. Consent for EPIC-CC obtained at PRISM enrolment. Testing platform: Comprehensive multi-omics for all enrolled patients with adequate tumour nucleic acids: tumour and germline whole-genome sequencing (WGS), whole-transcriptome sequencing (RNASeq), and methylation profiling. Where sufficient viable tumour was available, preclinical high-throughput in vitro drug screening (HTS) and in vivo patient-derived xenograft (PDX) drug efficacy testing were attempted. Bioinformatic analysis, manual curation, and interpretation were performed, and actionable findings assessed and recommendations made by a national Multidisciplinary Tumour Board (MTB) (therapy change, diagnosis change, or cancer predisposition finding) and reported to clinicians. Data sources and costing approach: Costs and labour data were obtained from ZERO’s laboratory information management system and expenditure records (2021 AUD). Sequencing was outsourced at fixed prices per test; thus, labour/consumables for WGS/RNASeq/methylation were not itemized separately. The costed period was from sample receipt to MTB report issuance (average 8.7 weeks). A micro-costing approach identified specific resources per assay, measured resource use, and assigned monetary values per patient, reflecting variability due to tissue quality/quantity and feasible tests. To reflect case complexity, a weighted average time-on-activity per patient of 0.4 simple:0.6 complex was applied. Economic outcomes and scenarios: Three outcomes were modeled: Outcome A—access cost per patient to perform multi-omics and preclinical assays; Outcome B—cost per identification of a molecular cause of cancer (given Outcome A data); Outcome C—cost per potentially actionable MTB recommendation (including additional MTB preparation, meeting, and reporting labour). Each outcome was evaluated under three scenarios: high-cost (circa 2020; low volume ~130/year), base case (circa 2022; actual study costs), and low-cost (forecast circa 2025; high volume ~1000/year). Included costs spanned laboratory consumables, labour, shipping, data computation/transfer/storage, analysis, curation, and reporting; Outcome C additionally included MTB-related labour. Exclusions: National/site trial coordination; tumour sampling (done in usual care); capital equipment and maintenance; infrastructure (e.g., benchtops, fridges); staff training; long-term data storage (>5 years); discovery research; biobanks/databanks; national laboratory accreditations; broader clinical care and treatment costs. Statistical analysis: Mean per-patient costs were reported with 95% uncertainty intervals (UI) via bootstrapping (1000 replicates). Base-case access costs were stratified by cancer category and stage; group differences were assessed using linear regression with robust standard errors and post-hoc pairwise f-tests with Bonferroni correction (p<0.005). Analyses used SAS 9.4.

Cohort: 423 patients received multi-omics results between Aug 2019 and Dec 31, 2021; 55.8% male (n=236), 44.2% female (n=187). By study end, 50.4% (n=213) were deceased. Mean age at sampling 10.3 years (SD 6.5; median 10.0; range 1 day–39.7 years). Base case (current costs, 2021 AUD): Outcome A—mean access cost per patient $12,743 (95% UI: $12,254–$13,284), comprising Omics $11,136 (95% UI: $10,891–$11,393) and Preclinical $1,607 (95% UI: $1,188–$2,055). Outcome B—mean cost per molecular diagnosis $14,262 (95% UI: $13,614–$14,994); increase vs Outcome A: $1,518. Outcome C—mean cost per actionable MTB recommendation $21,769 (95% UI: $20,037–$23,657); increase vs Outcome A: $9,017. Omics share for Outcome C: $19,548 (95% UI: $17,956–$21,163); Preclinical: $2,221 (95% UI: $1,624–$2,855). Preclinical testing uptake and unit costs: 81/423 (19.2%) received HTS or PDX. Among completed tests: HTS completion n=79 (18.7%); mean access cost $3,171. PDX drug efficacy completion n=16 (3.8%); mean access cost $19,757. PDX cases incurred an additional $270 labour for later MTB report re-issue to achieve Outcome C. Stratified costs: Mean access cost differed by cancer type (p=0.0156). Pairwise analysis showed significantly lower mean access cost for CNS vs haematological cancers by $2,461.88 (95% CI: $951.67–$3,972.10; p=0.0021). No significant differences by cancer stage (p > 0.10). Scenario analysis: Substantial cost reductions with time and higher volumes. Outcome A—low estimate (circa 2025) $9,122 vs high estimate (circa 2020) $21,179. Outcome B—$10,209 (low) vs $23,703 (high). Outcome C—$16,795 (low) vs $33,356 (high). Tabled components show WGS as the dominant omics cost driver, with marked decreases projected at higher volumes and over time.

The study directly addresses the evidence gap on the costs of comprehensive precision medicine implementation in paediatric oncology. It demonstrates that a national, integrated multi-omics platform with selective preclinical testing can deliver clinically significant, reportable results in a feasible timeframe (average 8.7 weeks to MTB report), with quantifiable per-patient costs for access (Outcome A), molecular diagnosis (Outcome B), and actionable recommendations (Outcome C). The findings indicate that multi-omics constitutes the largest share of costs and is broadly feasible across cancer types, while preclinical platforms (HTS/PDX) are limited by tissue availability, technical requirements, and longer timelines, leading to lower uptake and stable per-patient platform costs. Cost differences by cancer type (lower for CNS vs haematological) likely reflect feasibility constraints and lower PDX utilization in CNS tumours. Importantly, costs have decreased substantially over recent years, and further reductions are anticipated with higher sample volumes, process automation, improved computational pipelines, and efficiencies in data compression, storage, and compute (e.g., leveraging cloud services). The comprehensive approach, integrating WGS, RNASeq, and methylation profiling, is positioned to yield higher rates of identifying molecular causes compared to narrower testing, reinforcing potential clinical utility. The comparison to prior micro-costing of narrower genomic tests suggests that comprehensive platforms can now approach similar costs to earlier, less comprehensive approaches, supporting scalability and integration into clinical workflows. The results are relevant for policymakers planning the adoption of precision medicine into public health systems and for designing economic evaluations that incorporate both costs and downstream clinical impacts.

This study provides the first systematic micro-costing of a comprehensive paediatric precision medicine platform in Australia, quantifying per-patient costs to access testing, achieve molecular diagnosis, and obtain actionable MTB recommendations. It shows feasible turnaround times and highlights that multi-omics dominates costs while preclinical testing should be used judiciously. Costs have decreased significantly over time and are projected to fall further with higher volumes and process efficiencies, supporting the potential for broader implementation. Future research should evaluate costs and benefits in routine clinical settings, including accreditation-related costs, test consolidation, and real-world uptake of recommendations, and measure clinical outcomes and cost-effectiveness. Further work should assess optimal test suites across risk strata, the role and value of preclinical testing as evidence evolves, and the potential transition toward computational drug response predictions to reduce reliance on labour-intensive biological models. Expanded longitudinal studies will help generalize costs across cancer types and stages and inform policies for integrating precision medicine into publicly funded care.

The analysis covers costs only from sample receipt to MTB report issuance; it excludes broader clinical care costs, treatment uptake, and patient outcome measures. Several cost categories typical of research set-up or institutional infrastructure were excluded (e.g., trial coordination, tumour sampling, capital equipment and maintenance, infrastructure, staff training, long-term data storage beyond 5 years, discovery research, biobanks/databanks, and national accreditations). Results reflect a research setting and may differ in routine diagnostic or commercial contexts due to volume changes, accreditation, and workflow familiarity. Heterogeneity in cancer types and stages may limit generalizability; significant cost differences were observed between some cancer types (e.g., CNS vs haematological), while stage differences were not detected. Preclinical testing feasibility constraints (e.g., viable tissue requirements, long timelines) limit uptake and may influence cost distributions.