Artificial intelligence in rare pediatric solid tumor research and clinical care: A scoping review

Clinical Snapshot

Key Findings

• P Value: Not applicable — scoping review with descriptive synthesis only

• Effect Size: Descriptive only; no pooled effect size calculated. Many models reported AUCs >0.85 across included studies. Supervised learning was the predominant approach (20/23 studies)

• Primary Outcome: Characterisation of AI/ML applications in rare pediatric extra-cranial solid tumors across diagnosis (12/23 studies), prognosis (10/23), and risk stratification (9/23)

• Nnt Or Sensitivity: AUC >0.85 reported descriptively for many models; sensitivity, specificity, and NNT not synthesised. In 17/23 comparative studies, AI/ML performance equalled or exceeded expert consensus or traditional models. External validation reported in only 8/23 studies (35%)

• Confidence Interval: Not reported at the review level; individual study-level confidence intervals not synthesised

Clinical Application

Clinical deployment of AI/ML tools identified in this review is not currently feasible for routine practice. The majority of models lack external validation, prospective testing, or regulatory clearance. Implementation would require prospective multicentre validation studies, standardised data infrastructure, and clinician training. Interpretability of models — essential for clinical trust and medicolegal accountability — is not adequately addressed in the current evidence base In Australia, rare pediatric solid tumors are managed through the Children's Oncology Group (COG)-affiliated centres and the Australian and New Zealand Children's Haematology/Oncology Group (ANZCHOG). The small Australian pediatric population further amplifies the rarity challenge, making AI/ML tools trained on international datasets potentially valuable but requiring local validation. No AI/ML tools for rare pediatric solid tumors are currently listed on the TGA's ARTG as Software as a Medical Device (SaMD). The PBS does not currently subsidise AI-assisted diagnostic or prognostic tools in this context. RACGP and paediatric oncology guidelines do not yet incorporate AI/ML-based risk stratification for these tumors. Australian participation in international data-sharing consortia (e.g., SIOPE, COG) would be essential to generate datasets of sufficient size for robust model development and validation Pediatric patients (≤19 years) with rare extra-cranial solid tumors, particularly hepatoblastoma and pediatric thyroid cancers. Findings are most relevant to tertiary paediatric oncology centres with access to high-dimensional clinical, imaging, or genomic data

Abstract

BACKGROUND: Clinical and research advances for children with solid tumors are limited by their rare nature. Artificial intelligence (AI) and machine learning (ML) hold potential for advancing diagnosis, risk stratification, and treatment in rare diseases where individual patients contribute high-dimensional data. This review characterizes current AI/ML applications in rare pediatric solid tumors. METHODS: PubMed, Embase, and Web-of-Science Core Collection databases were queried for relevant articles published before February 10, 2025. Eligible studies applied AI/ML methods to study rare tumors in pediatric populations (≤19 years). Articles were evaluated for tumor diagnosis, AI/ML methods, clinical applications, comparators, effectiveness, and interpretability. RESULTS: Twenty-three studies (2009-2025) were included. Hepatoblastoma (12/23) and pediatric thyroid cancers (4/23) were most frequently studied. Supervised learning predominated (20/23), followed by unsupervised (9/23), deep learning (4/23), and hybrid/ensemble models (6/23). Applications included diagnosis (12/23), prognosis (10/23), and risk stratification (9/23). Twenty studies reported effectiveness measures, with many models achieving AUCs >0.85. In comparative analysis (17/23), AI/ML often equaled or exceeded expert consensus, traditional models, or alternative algorithms. Eight studies reported external validation. CONCLUSIONS: Current AI/ML research in rare pediatric extra-cranial solid tumors focuses on diagnosis, risk stratification, and prognosis, often outperforming traditional methods. Future work should prioritize external validation and clinical applicability.

References

1. 1.Happ, M. F., Lindsay, C. V., Onyeji, C., Soto, A. L., Harb, M. G., Kaplan, S., Gomez, S. M., Blinman, T., & Tracy, E. T. (2026). Artificial intelligence in rare pediatric solid tumor research and clinical care: A scoping review. Journal of Pediatric Surgery. Advance online publication. https://doi.org/10.1016/j.jpedsurg.2026.163140