Precision-guided treatment in high-risk pediatric cancers

Pediatric precision oncology has revealed potentially targetable molecular alterations in over 65% of children with high-risk cancers, yet clinical uptake of matched therapies has often been low (10–33%), partly due to clinician uncertainty about the efficacy and benefit–risk of precision-guided treatment (PGT). It remains unclear which pediatric patients are most likely to benefit and whether PGT improves survival. Prior studies suggested potential benefit but frequently lacked objective response assessments or extended follow-up. The present study addresses these gaps by evaluating both objective response and survival outcomes in a large cohort from the PRISM trial and by identifying prognostic factors to guide which patients should receive PGT and when.

Earlier pediatric precision oncology efforts reported high rates of actionable findings but modest uptake of matched therapy (10–33%). The INFORM registry observed improved outcomes limited to patients with high-evidence targets. The GAINS study suggested responses might be confined to treatments targeting activating fusions. MAPPYACTS showed improved response rates with higher-tier evidence but did not assess survival. Many prior studies lacked comprehensive objective response evaluations or long-term outcome data, limiting conclusions about survival benefit and patient selection for PGT.

Design: Multicenter prospective observational cohort (PRISM; ZERO Childhood Cancer Program), ClinicalTrials.gov NCT03336931. Recruitment: September 2017–December 2020 across eight pediatric oncology centers in Australia; data cutoff 30 June 2022. Eligibility: Patients <21 years (and select adults with pediatric-type cancers) with high-risk malignancy (estimated cure rate <30%) at diagnosis, relapse, or refractory status. Samples: Tumor and germline; fresh-frozen preferred; FFPE accepted with approval. Molecular profiling: Paired tumor–germline whole-genome sequencing (WGS) whenever possible; whole-transcriptome sequencing (WTS) when RNA adequate; DNA methylation profiling for CNS and selected sarcomas; targeted panels when WGS/WTS unavailable. Variant curation and analytical pipelines as previously described. MTB process: Fortnightly national molecular tumor board reviewed reportable/actionable alterations and evidence; five-tier recommendation system (Tier 1: clinical evidence in same cancer; Tier 2: clinical evidence in other cancer; Tier 3: preclinical same cancer; Tier 4: preclinical other cancer; Tier 5: consensus). Recommendations required pediatric safety data and plausible drug access in Australia (on-label, trial, compassionate, off-label). Outcomes and analyses: Treatments categorized as PGT (matched to MTB recommendation) or non-PGT (standard of care, SOC; or unguided targeted/new therapy, UGT). Inclusion for outcome analyses required treatment duration ≥4 weeks, no progression within first 4 weeks, and evaluable response data. Response assessment: RECIST v1.1 or PERCIST for solid tumors, RANO for CNS, and NCCN criteria for leukemia. Objective response rate (ORR) for measurable disease defined as CR+PR; evaluable response combined measurable and non-measurable. Objective clinical benefit (OCB) defined as CR/PR or sustained SD (≥24 weeks), or disease-free ≥24 weeks if in CR at start. Progression-free survival (PFS): from treatment start to progression/death; overall survival (OS): from enrollment or first post-MTB treatment (per comparison). PFS ratio: PFS on PGT divided by PFS on immediate prior regimen in the same patient; benefit defined as >1.3 and ≥4-week absolute gain. Statistics: Kaplan–Meier with log-rank tests; chi-squared for proportions; multivariable Cox regression for prognostic factors. Ethical oversight: Conducted per GCP and Declaration of Helsinki; ethics approval by Hunter New England HREC (2019/ETH00701); informed consent obtained.

• Cohort: 470 consented; 384 eligible and presented at MTB; 3-year OS 34% (95% CI 29–40%). Median follow-up among survivors 33.7 months (range 18.2–56.9). Tumor types: 146 CNS, 183 solid, 56 hematologic malignancies.
• Molecular profiling and recommendations: 67% (256/384) received ≥1 PGT recommendation (510 total recommendations). Support levels: 53% Tier 1–2, 43% Tier 3–4, 1% Tier 5. Most common targets/pathways: PI3K/mTOR (20%), MAPK (15%), PARP (10%), CDK4/6 (8%); RTKs included FGFR (28%), VEGF/VEGFR (20%), EGFR/ERBB (16%).
• Uptake and access: 110 patients (29% of cohort; 43% of those recommended) received 117 PGTs; median return of results 6.6 weeks; 70% started within 3 months (median 9 weeks). Access routes: compassionate 36%, institutional funding 33%, clinical trials 16%, government PBS 9%, cost-sharing 4%, self-funded 2%.
• Response and clinical benefit: Among 70 measurable-disease PGTs, CR 9%, PR 27%, SD 34%, PD 30% (ORR ≈35%). Including evaluable non-measurable disease (n=90), CR 9%, PR 21%, SD 38%, PD 32% (ORR 30%). OCB observed in 55% (53/97) PGTs. PFS ratio >1.3 in 42% (95% CI 25–61%) of 31 evaluable patients; associated with improved PFS (2-year PFS 36% vs 0%; P=0.02) and trend to better OS (46% vs 8.3%; P=0.10).
• Comparative effectiveness: • PGT vs non-PGT (SOC + UGT): 2-year PFS 27% vs 11% (P=0.01); OS 38% vs 24% (P=0.08). • PGT vs UGT: 2-year PFS 26% vs 5.2% (P=0.003); OS 38% vs 20% (P=0.15); ORR 30% vs 2.3% (P<0.0001); OCB 55% vs 23% (P=0.0003). • PGT vs SOC: 2-year PFS 26% vs 12% (P=0.049); OS 38% vs 23% (P=0.11).
• Prognostic factors and subgroups: • Evidence tier: Tier 1 had highest ORR (39%) and OCB (74%) and superior survival vs Tier 2 and Tiers 3–5 (e.g., 2-year PFS 42% for Tier 1 vs 22% Tier 2 and 13% Tiers 3–5). • Target type: Fusion/SV targets had ORR 60% and 2-year PFS 68%, OS 69%; superior to SNV (ORR 32%; 2-year PFS 29–30%), high RNA-only (ORR 15%; 2-year PFS 5.9%), and CNV (ORR 14%; 2-year PFS 7.7%). • Timing: Initiating PGT before progression improved outcomes vs after progression (ORR 40% vs 20%, P=0.04; OCB 74% vs 36%, P=0.0001; 2-year PFS 42% vs 12%, P<0.0001; 2-year OS 53% vs 29%, P=0.0002). • Treatment type: Targeted mono/dual therapy showed better survival than targeted agents combined with chemotherapy (2-year PFS 32% vs 0%, P=0.03; 2-year OS 42% vs 15%, P=0.048). • Hematologic malignancies had poorer outcomes with PGT; uptake was low (22%) and many were heavily pretreated.
• Multivariable Cox for PFS: Tier 1 evidence (HR 0.43, 95% CI 0.24–0.77, P=0.005), fusion/SV (HR 0.42, 0.18–0.98, P=0.045), no PD since enrollment (HR 0.50, 0.31–0.82, P=0.006), and non-hematologic malignancy (HR 0.21, 0.06–0.79, P=0.02) independently favorable.
• Composite favorable factors (Tier 1 + fusion/SV + no PD): Increasing number associated with stepwise gains; with 3 factors, ORR 75%, OCB 100%, and 2-year PFS 88%.

This study provides comprehensive objective response and survival data demonstrating that PGT improves clinical outcomes in children with high-risk cancers. Compared with non-molecularly guided therapies (SOC and UGT), PGT significantly prolonged PFS and produced higher response and OCB rates, indicating that selecting therapies based on genomic drivers confers tangible benefit. The strongest benefits were observed when treatments were supported by high-level clinical evidence (Tier 1), targeted oncogenic fusions/structural variants, and were initiated before disease progression. These factors independently predicted improved PFS and together identified a subgroup with exceptional outcomes. The superiority of PGT over UGT underscores the importance of biomarker-driven therapy selection and supports incorporating comprehensive molecular profiling and MTB-guided recommendations into routine care and trial designs. Although OS benefits did not reach statistical significance within the current follow-up, likely due to multiple subsequent therapies and salvage regimens, the consistent PFS advantages and response improvements suggest meaningful clinical benefit that may translate into longer-term survival with extended follow-up and optimized sequencing of therapies.

Comprehensive molecular profiling with MTB-guided PGT significantly improves antitumor activity and PFS in children with high-risk cancers compared with unguided targeted therapies and standard cytotoxic regimens. The greatest benefit occurs with Tier 1 evidence, fusion/SV targets, and when PGT is initiated before progression. These data support early integration of precision medicine into standard pediatric oncology pathways and prioritization of biomarker-driven clinical trials. Future work should focus on enhancing access to targeted agents, developing therapies for pediatric-specific drivers, optimizing combination strategies and timing, and evaluating broader outcomes including quality of life and toxicity.

Nonrandomized observational design with potential clinician selection bias; heterogeneity in treatments and salvage therapies may confound OS analyses; challenges with drug access influenced treatment choices; relatively small numbers for some subgroup analyses (e.g., hematologic malignancies); limited local clinical testing availability for many targets; OS differences did not reach statistical significance within current follow-up; generalizability may be affected by healthcare system-specific drug access pathways.