Machine Learning, Large Language Models, and Multimodal AI for Diagnosing Pediatric Rare Diseases: Scoping Review

Jungang Zhao; Jiawei Luo; Qiu Li

doi:10.2196/96169

Machine Learning, Large Language Models, and Multimodal AI for Diagnosing Pediatric Rare Diseases: Scoping Review | ResearchHub

Abstract

Abstract Background Pediatric rare diseases often cause a prolonged diagnostic odyssey. AI, including machine learning, deep learning, large language models (LLMs), and multimodal systems, may support diagnosis, but these applications in children have not been systematically mapped. Objective The aim of the study is to map diagnostic applications, data modalities, validation strategies, and evidence maturity of AI methods for pediatric rare diseases. Methods We conducted a scoping review following Joanna Briggs Institute methodology and reported it according to PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews). On June 26, 2026, we searched PubMed, Scopus, Web of Science Core Collection, Embase, China National Knowledge Infrastructure (CNKI), Wanfang Data, and the Cochrane Library for records published from January 1, 2015, through June 1, 2026. We additionally searched medRxiv and arXiv and hand-searched the reference lists of included studies and relevant reviews. Eligibility was defined using the population-concept-context framework: pediatric rare diseases, diagnostic AI, and any clinical or research setting. JZ and JL independently screened titles and abstracts and assessed potentially eligible full-text reports. JZ charted the data, and JL verified every field. Findings were synthesized descriptively according to disease focus, AI technology, input modality, diagnostic task, validation strategy, and evidence maturity. Results Database searches identified 2557 records; 2063 remained after deduplication. Of 106 full-text reports assessed, 77 database studies and 4 studies from hand searching and preprint servers were included, yielding 81 studies. Studies were published from 2016 through 2026, with 55 of 81 (67.9%) published from 2024 through 2026. Using a mutually exclusive primary technology classification, classical machine learning accounted for 38 (46.9%) studies, facial AI for 18 (22.2%), deep learning for 15 (18.5%), LLMs for 6 (7.4%), and multimodal AI for 4 (4.9%). Electronic health records, claims, clinical text, or structured clinical vignettes were used in 25 (30.9%) studies, facial images in 17 (21%), and other medical imaging in 13 (16%). Evidence remained mainly retrospective and internally validated: 62 (76.5%) studies included a retrospective component and 76 (93.8%) reported internal validation, whereas 21 (25.9%) included external validation and 12 (14.8%) included a prospective component. Conclusions Research on AI-assisted diagnosis of pediatric rare diseases has expanded rapidly, but evidence maturity has not kept pace. Most studies established technical feasibility rather than generalizable clinical benefit, and performance should be interpreted by task, inputs, reference standard, and validation design rather than used to rank technologies. Evidence for LLMs and multimodal AI remains limited. Future research should prioritize multicenter validation, reproducible task-specific benchmarks, prospective evaluation, and assessment of incremental clinical value.

External reference records

Time to diagnosis and determinants of diagnostic delays of people living with a rare disease: results of a rare barometer retrospective patient survey

10.1038/s41431-024-01604-z

Estimating cumulative point prevalence of rare diseases: analysis of the Orphanet database

10.1038/s41431-019-0508-0

Identifying facial phenotypes of genetic disorders using deep learning

10.1038/s41591-018-0279-0

Large language models for disease diagnosis: a scoping review

10.1038/s44387-025-00011-z

Deep learning for rare disease: a scoping review

10.1016/j.jbi.2022.104227

The use of machine learning in rare diseases: a scoping review

10.1186/s13023-020-01424-6

Artificial intelligence in the diagnosis of pediatric rare diseases: from real-world data toward a personalized medicine approach

10.3390/jpm15090407

Updated methodological guidance for the conduct of scoping reviews

10.11124/jbies-20-00167

PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation

10.7326/m18-0850

Scoping studies: towards a methodological framework

10.1080/1364557032000119616

China has officially released its first national list of rare diseases

10.5582/irdr.2018.01056

Machine learning-based prediction of inborn errors of immunity in patients with low IgE levels

10.1111/pai.70283

Few shot learning for phenotype-driven diagnosis of patients with rare genetic diseases

10.1038/s41746-025-01749-1

Functional data strengthen clinical validation of PhenoScore phenotype-guided AI for ANKRD11 missense variants

10.1111/cge.70159

Genomic DNA methylation signatures enable concurrent diagnosis and clinical genetic variant classification in neurodevelopmental syndromes

10.1016/j.ajhg.2017.12.008

Evaluation of a new inclusive next-generation synthetic face tool for dysmorphology

10.1002/ajmg.a.64298

10.64898/2026.03.25.26349289

Feasibility of machine learning analysis for the identification of patients with possible primary ciliary dyskinesia

10.1186/s13023-025-03966-z

Leveraging machine learning to identify acute myeloid leukemia patients and their chemotherapy regimens in an administrative database

10.1002/pbc.30260

Phenotype driven molecular genetic test recommendation for diagnosing pediatric rare disorders

10.1038/s41746-024-01331-1

RareArena: a comprehensive benchmark dataset unveiling the potential of large language models in rare disease diagnosis

10.1016/j.landig.2025.100953

Patient-patient similarity-based screening of a clinical data warehouse to support ciliopathy diagnosis

10.3389/fphar.2022.786710

SABPI-Net: a structure-aware bidirectional proxy interaction network for infantile retinal disease diagnosis

10.1109/tmi.2026.3650968

Identification of atypical scoliosis patterns using X-ray images based on fine-grained techniques in deep learning

10.1177/21925682251349999

Genomic answers for children: dynamic analyses of >1000 pediatric rare disease genomes

10.1016/j.gim.2022.02.007

Epidemiological analysis to identify predictors of X-linked hypophosphatemia (XLH) diagnosis in an Italian pediatric population: the EPIX project

10.1007/s12020-024-03793-5

Diagnostic value of machine-learning using conventional magnetic resonance imaging markers for pediatric idiopathic intracranial hypertension: a retrospective study

10.1007/s00247-026-06638-7

10.64898/2026.03.03.26346563

Enhancing TSH-based congenital hypothyroidism screening using machine learning and resampling algorithms

10.1186/s12911-025-03312-0

Artificial intelligence enables comprehensive genome interpretation and nomination of candidate diagnoses for rare genetic diseases

10.1186/s13073-021-00965-0

AI-Based facial phenotyping supports a shared molecular axis in PACS1-, PACS2-, and WDR37-related syndromes

10.3390/ijms26167964

PhenoScore quantifies phenotypic variation for rare genetic diseases by combining facial analysis with other clinical features using a machine-learning framework

10.1038/s41588-023-01469-w

Population-specific facial traits and diagnosis accuracy of genetic and rare diseases in an admixed Colombian population

10.1038/s41598-023-33374-x

Success of face analysis technology in rare genetic diseases diagnosed by whole-exome sequencing: a single-center experience

10.1159/000505800

Explainable AI-driven graph-based neural networks for mucopolysaccharidoses diagnosis

10.1186/s13040-026-00523-7

IESS-FusionNet: physiologically inspired EEG-EMG fusion with linear recurrent attention for infantile epileptic spasms syndrome detection

10.3390/bioengineering13010057

Next generation phenotyping using narrative reports in a rare disease clinical data warehouse

10.1186/s13023-018-0830-6

A weakly supervised transformer for rare disease diagnosis and subphenotyping from EHRs with pulmonary case studies

10.1038/s41746-026-02406-x

A machine learning model accurately identifies glycogen storage disease Ia patients based on plasma acylcarnitine profiles

10.1186/s13023-025-03537-2

Machine learning models reveal distinct disease subgroups and improve diagnostic and prognostic accuracy for individuals with pathogenic SCN8A gain-of-function variants

10.1242/bio.060286

Additional external references

An AI-based approach driven by genotypes and phenotypes to uplift the diagnostic yield of genetic diseases

10.1007/s00439-023-02638-x · doi-reference

ADHTransNet-based radiomics on multimodal pituitary MRI for non-invasive hormone prediction in children

10.1016/j.cmpb.2026.109235 · doi-reference

An agentic system for rare disease diagnosis with traceable reasoning

10.1038/s41586-025-10097-9 · doi-reference

Digital-tier strategy improves newborn screening for glutaric aciduria type 1

10.3390/ijns10040083 · doi-reference

Study on the method of using facial features to improve the accuracy of rare disease prediction in small sample scenarios [Article in Chinese]

10.1016/j.heliyon.2024.e39479 · doi-reference

Diagnostic accuracy of a custom large language model on rare pediatric disease case reports

10.1002/ajmg.a.63878 · doi-reference

Finding commonalities in rare diseases through the undiagnosed diseases network

10.1093/jamia/ocab050 · doi-reference

Using deep learning and electronic health records to detect Noonan syndrome in pediatric patients

10.1016/j.gim.2022.08.002 · doi-reference

The significance of machine learning in neonatal screening for inherited metabolic diseases

10.3389/fped.2024.1366891 · doi-reference

Establishment of an auxiliary diagnosis system of newborn screening for inherited metabolic diseases based on artificial intelligence technology and a clinical trial

Machine Learning, Large Language Models, and Multimodal AI for Diagnosing Pediatric Rare Diseases: Scoping Review

Referenced works

Cited by

External reference records

Additional external references