Paediatrician in London

Personalised Medicine for Children: Could Doctors Predict Which Treatment Will Work Best?

If your child has been diagnosed with a condition, you may naturally wonder why a treatment that works well for one child does not always work in exactly the same way for another. Even children with the same diagnosis can respond differently because factors such as age, genes, metabolism and overall health can influence how a medicine works.

Personalised medicine aims to use information about your child, such as their genetic results, medical history and clinical features, to make treatment decisions more individualised. It does not mean doctors can predict the perfect treatment every time, but it may help them choose options that are more appropriate for your child’s particular needs.

What Does Personalised Medicine Mean for a Child?

Personalised medicine does not mean creating a unique medicine for every child. Instead, your child’s doctors may use genetic results, biomarkers and medical history to help identify which treatment is most appropriate.

When you discuss treatment options for your child, you may hear doctors consider age, weight, puberty, organ function and previous medicines alongside genetic results. This means personalised paediatric care combines molecular information with what your child’s team already knows about their development, health and treatment history.

Why Doesn’t the Same Medicine Work Equally Well for Every Child?

You may wonder why a medicine that works well for another child is not automatically right for your child. Age, weight, development and organ function can all affect how your child processes a medicine, while genetic differences may change how quickly it is broken down.

Disease biology can also change how your child responds to treatment, even when two children have the same diagnosis. Your child’s doctors therefore need to consider their age, genetics, overall health and condition when choosing the safest and most effective treatment.

What Is Pharmacogenomics and Could It Affect Your Child’s Medicine?

If your child is offered pharmacogenomic testing, you can think of it as a way of asking whether their DNA could affect how a medicine works. The result may show that your child is likely to process a medicine faster or slower, helping their doctor decide whether the dose or treatment needs adjusting.

Pharmacogenomic testing has a different purpose from genetic testing used mainly to diagnose a condition. It looks for genetic differences that may affect how a medicine is processed or the risk of certain adverse effects, helping doctors decide whether a dose or treatment approach may need adjusting.

Is Pharmacogenomic Testing Already Used in Children?

Yes, but only in selected situations. In England, TPMT and NUDT15 testing is available through the NHS National Genomic Test Directory for patients starting 6-mercaptopurine for acute lymphoblastic leukaemia (ALL). For other thiopurine indications, this combined genomic testing is not currently routinely funded through the Test Directory.

If testing is used for your child, the result may show that a lower starting dose or different approach is safer. You should still expect blood tests, clinical review and monitoring, because genetic information is only one part of safe prescribing.

Evidence Note

If your child has TPMT or NUDT15 variants, you may be told that standard thiopurine doses could carry a higher risk of serious bone-marrow toxicity. The practical point for you is that pharmacogenomic results can guide safer starting treatment, but your child will still need blood tests and clinical monitoring.

Does a Pharmacogenomic Result Tell Doctors the Exact Dose?

Usually not by itself. Your child’s genetic result can show how they may process a medicine, but their doctor must also consider blood tests, organ function, other medicines and your child’s response to treatment.

You should therefore see TPMT and NUDT15 results as one part of the prescribing decision. They can guide dosing, but they cannot predict every response, so your child’s doctor will still need to consider clinical findings and monitoring.

Clinical Tip

If a genetic test suggests that your child may respond differently to a medicine, do not change or stop their dose yourself. You can ask their doctor what the result means, whether it changes the recommended starting dose and what blood tests or other monitoring will still be needed.

Can Genetics Help Guide Your Child’s Treatment?

Sometimes. In cystic fibrosis, your child’s CFTR gene variants can help determine whether they may be eligible for specific CFTR-modulator medicines. Your child’s clinical team can use this genetic information alongside their age and clinical circumstances to decide whether a targeted treatment is suitable.

This goes beyond treating symptoms such as infections or mucus build-up. Your child still needs appropriate standard cystic fibrosis care, while relevant genetic results may help their clinical team determine whether a targeted CFTR-modulator treatment is suitable.

How Could Personalised Medicine Change Your Child’s Cancer Treatment?

If your child has cancer, you may hear the team talk about molecular or genomic testing. Instead of looking only at where the tumour started, these tests can examine genetic changes that may help your child’s cancer team identify possible treatment targets.

Genomic testing can also help your child’s doctors understand prognosis or identify suitable clinical trials. Not every genetic change has a matching medicine, but this information may help your child’s team develop a more detailed treatment plan.

UK Guidance Note

In England, genomic testing is delivered through the NHS Genomic Medicine Service. The National Genomic Test Directory sets out which genomic tests are commissioned and which patients are eligible, so you should not assume that every genetic or pharmacogenomic test is routinely needed or available for every child.

What Are Biomarkers and How Could They Guide Your Child’s Treatment?

Clinician talking with a mother and child

If your child’s team mentions a biomarker, you can think of it as a measurable sign that gives information about health, disease or treatment response. It might be a gene, protein, hormone level or tumour feature that helps the team understand what is happening.

In personalised medicine, validated biomarkers may help clinicians estimate whether a treatment is likely to work or show whether your child is responding. Only biomarkers with sufficient clinical evidence should influence treatment decisions.

Could Medical Imaging Help Predict Which Treatment May Work for Your Child?

If your child has an MRI or CT scan, you may hear researchers discuss radiomics. This involves extracting measurable features such as shape, texture and intensity from medical images. AI may support some radiomics workflows, but radiomics itself is not the same as AI.

In paediatric sarcoma research, radiomics is being studied as a way to help estimate tumour characteristics, prognosis or treatment response. However, much of the evidence is still retrospective and methodologically variable, so radiomics should not currently be relied on by itself to choose an individual child’s treatment.

How Could Predictive Models Help Your Child’s Doctors?

A predictive model can combine information such as your child’s age, test results and scans to estimate the likelihood of an outcome, such as treatment response or side effects. Its result is a probability rather than a guarantee, and its reliability depends on how well the model has been validated for the children in whom it will be used.

If you are shown a prediction from a model, you should treat it as one piece of information rather than a final answer. Your child’s doctors still need to interpret it alongside examination findings, test results and clinical evidence.

Could AI Choose the Best Treatment for Your Child?

Clinician talking with a mother and daughter at a desk

AI and predictive models are being studied for tasks such as risk prediction, diagnosis and clinical decision support in children. Some systems may eventually help clinicians bring together different types of information, but real-world implementation of predictive AI in paediatric care remains limited.

If AI is used to support your child’s care, you should expect it to assist rather than replace clinical judgement. Your child’s doctors still need to weigh the evidence, expected benefits, side effects and what matters to you and your family.

Research Insight

A 2025 systematic review examined paediatric predictive-AI implementation studies published from 2010 to October 2021. Of 126 articles reviewed in full, only 17 met the inclusion criteria for AI that had actually been implemented in a paediatric clinical setting. This does not mean that only 17 paediatric AI systems exist today; instead, it shows how limited real-world implementation evidence was within the period covered by the review.

Could a Digital Twin One Day Help Plan Your Child’s Treatment?

If you hear the term “digital twin”, it refers to a developing virtual model that can bring together information such as genetics, medical history, scans and changing health data. In principle, such a model could help your child’s team explore how different treatments or doses might affect them.

Digital twins are still mainly a research concept in healthcare. Importantly, the 2025 review cited here was not limited to children: across 149 human healthcare studies, only 18 fully met the National Academies’ criteria for a true digital twin. This highlights how early the field remains, particularly when considering routine paediatric care.

What Information Could a Future Digital Twin Use?

Father and daughter using a tablet at home

A future digital twin could combine your child’s scans, blood pressure, oxygen levels, genetics and previous treatments into one personalised model. In future, clinical and research teams could potentially use such models to explore how different operations, medicines or doses might affect an individual child.

The model could also use wearable data and update as your child grows and their condition changes. This technology is still developing, and fully functioning digital twins are not yet routine tools for guiding paediatric treatment.

Why Aren’t Digital Twins Already Being Used Routinely?

Building a reliable digital twin of your child is difficult because human biology is complex and constantly changing. Your child’s medicines, hormones, genetics, illness, sleep and activity can all interact, so reliable models depend on large amounts of high-quality, continuously relevant data.

Even if a model looks sophisticated, you would still need evidence that its predictions are accurate for real patients. Before a digital twin could guide your child’s care, researchers would need to show that it reliably improves meaningful outcomes.

How Can Different Types of Information Personalise Your Child’s Treatment?

Type of informationHow it may helpImportant limitation
PharmacogenomicsMay help predict medicine response or toxicity and guide dosing in selected situations.Genetic results do not determine the dose or outcome on their own.
Disease or tumour geneticsMay identify eligibility for a targeted treatment or clinical trial.Not every genetic change has an effective targeted treatment.
BiomarkersMay help assess disease behaviour or response to treatment.Only validated biomarkers should influence clinical decisions.
AI and predictive modelsMay combine multiple types of data to estimate treatment response or risk.Predictions depend on the quality and relevance of the data used to develop the model.
Digital twinsCould eventually simulate how an individual may respond to different options.They are currently largely a research concept and are not routine paediatric treatment tools.

Could Personalised Medicine Reduce Side Effects?

You may hope personalised medicine could reduce the chance of side effects for your child. In some situations, genetics, age and other health information may help doctors estimate which treatments are more likely to work or cause particular problems.

However, no test can predict every reaction. Your child will still need appropriate clinical review, blood tests and other monitoring to help keep treatment safe.

What Should You Know About Personalised Treatment for Your Child?

Clinician talking with a family at a desk

Personalised medicine does not mean you need every genetic test for your child. Testing is most useful when the result has the potential to change your child’s treatment, monitoring or wider care.

Your child’s age, symptoms, growth, previous treatments and health can also guide personalised care. The most useful test is not necessarily the most advanced one, but the one that can answer a relevant clinical question and influence care.

Myth vs Fact

MythFact
Children with the same diagnosis should receive exactly the same treatment.Children can respond differently because age, genetics, organ function, disease biology and previous treatment response may all influence which treatment or dose is most suitable.
A pharmacogenomic test can tell doctors the exact dose your child needs.Pharmacogenomic results can guide prescribing in selected situations, but doctors still need to consider clinical factors and monitor how your child responds.
Every genetic change has a targeted treatment.Some genetic findings can help identify a targeted treatment, but many variants do not yet have an effective therapy matched to them.
AI can choose the best treatment for your child on its own.AI may support clinical decision-making, but current paediatric implementation is limited and doctors still need to interpret the prediction in context.
Digital twins are already routinely used to test treatments virtually for children.Human digital twins remain an emerging research technology and are not yet routine tools for choosing paediatric treatment.

Key Takeaways

  • Personalised medicine uses information such as your child’s genetics, biomarkers, age, organ function and treatment history to support more individualised decisions.
  • Pharmacogenomic testing can already influence prescribing in selected situations, but genetic results do not replace clinical monitoring.
  • Disease genetics and biomarkers can sometimes help identify targeted treatments, although not every result changes treatment.
  • AI and digital twins may further personalise care in the future, but their ability to predict the best treatment for an individual child still requires much stronger clinical validation.

Frequently Asked Questions

1. Does personalised medicine mean your child needs genetic testing?
Not always. Genetics can be important, but personalised care can also use your child’s age, weight, organ function, disease features, previous treatment response and other clinical information.

2. Can your child’s treatment change as they grow?
Yes. Age, weight, puberty, organ function and metabolism can change during childhood, so the most appropriate medicine or dose may also change over time.

3. Does a pharmacogenomic result mean a medicine is safe for your child?
No. Pharmacogenomic testing may identify certain genetic risks, but your child still needs appropriate prescribing, blood tests and clinical monitoring because genetics cannot predict every adverse effect.

4. What happens if your child has a genetic result but no targeted treatment exists?
The result may still help confirm a diagnosis, guide monitoring or inform future research and clinical-trial options. However, many genetic findings do not currently have a matching targeted medicine.

5. Is pharmacogenomic testing routinely available for every medicine through the NHS in England?
No. In England, genomic tests are commissioned for defined clinical indications through the National Genomic Test Directory. Availability depends on the medicine, condition and clinical criteria, so your child’s clinical team can explain whether testing is relevant in their situation.

6. Does every biomarker change treatment?
No. A biomarker is useful for personalised treatment only when there is sufficient evidence that it can meaningfully inform diagnosis, prognosis, monitoring or treatment decisions.

7. What happens if an AI prediction disagrees with your child’s doctor?
An AI prediction is only one source of information. Your child’s clinical team needs to consider the underlying evidence, symptoms, test results and potential limitations of the model before making a treatment decision.

8. Is radiomics already routine in children’s healthcare?
Not generally. Radiomics is an active research area, particularly in cancer imaging, but paediatric models still require stronger validation before they can routinely guide individual treatment choices.

9. Can your child have a digital twin created today?
True healthcare digital twins are not routine paediatric tools. Most current systems described as digital twins remain in research or early development rather than everyday clinical practice.

10. Can personalised medicine guarantee that your child will avoid side effects?
No. Personalised approaches may help reduce some predictable risks, but no genetic test, biomarker or AI model can predict every adverse reaction. Clinical monitoring remains essential.

Final Thoughts: The Future of Personalised Medicine for Children

As a parent, you may already see personalised medicine influencing parts of your child’s care, especially when a genetic result or biomarker changes prescribing or identifies a targeted treatment. You should still view AI, radiomics and digital twins as developing tools rather than replacements for clinical judgement.

If you are wondering whether genetic or other personalised testing could help your child, the most useful next step is to discuss the specific clinical question with their paediatric team. If you need a trusted paediatrician in London, you can get in touch with us at London Paediatric Clinic.

References:

  1. NHS England (no date) ‘The National Genomic Test Directory’. Available at: https://www.england.nhs.uk/genomics/the-national-genomic-test-directory/
  2. NHS England Genomics Education Programme (2025) ‘Thiopurines’. GeNotes Knowledge Hub. Last reviewed 2 September 2025. Available at: https://www.genomicseducation.hee.nhs.uk/genotes/knowledge-hub/thiopurines/
  3. National Institute for Health and Care Excellence (NICE) (2024) Ivacaftor–tezacaftor–elexacaftor, tezacaftor–ivacaftor and lumacaftor–ivacaftor for treating cystic fibrosis. Technology appraisal guidance TA988. Published 24 July 2024. Available at: https://www.nice.org.uk/guidance/ta988
  4. Mangum, R., Lin, F.Y. and Parsons, D.W. (2024) ‘Recent advancements and innovations in pediatric precision oncology’, Journal of Pediatric Hematology/Oncology, 46(5), pp. 262–271. Available at: https://pubmed.ncbi.nlm.nih.gov/38857189/
  5. Aghakhanyan, G., Filidei, T., Febi, M. et al. (2024) ‘Advancing pediatric sarcomas through radiomics: A systematic review and prospective assessment using Radiomics Quality Score (RQS) and Methodological Radiomics Score (METRICS)’, Diagnostics, 14(8), article 832. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC11049622/
  6. Tudor, B.H., Shargo, R., Gray, G.M. et al. (2025) ‘A scoping review of human digital twins in healthcare applications and usage patterns’, npj Digital Medicine, 8(1), article 587. Available at: https://pubmed.ncbi.nlm.nih.gov/41028914/
  7. Kandaswamy, S., Knake, L.A., Dziorny, A.C. et al. (2025) ‘Pediatric Predictive Artificial Intelligence Implemented in Clinical Practice from 2010 to 2021: A Systematic Review’, Applied Clinical Informatics, 16(3), pp. 477–487. Available at: https://pubmed.ncbi.nlm.nih.gov/39837545/