Artificial intelligence is transforming how doctors approach cancer treatment, particularly for patients with specific genetic mutations. This article features insights from leading oncologists and researchers who are using AI to identify and treat bladder cancers with FGFR3 alterations. These experts share practical strategies for implementing personalized medicine approaches that improve patient outcomes.
- Use AI to Target FGFR3-Altered Bladder Cancer
- Support Biomarker Tools With Clinical Review
- Deploy Medication Alerts Against Dangerous Interactions
- Demand Fairness Audits for Medical Algorithms
- Promote Shared Decisions Amid Predictive Uncertainty
- Adopt Real-Time Monitoring With Strong Oversight
Use AI to Target FGFR3-Altered Bladder Cancer
AI has transformed personalized medicine by helping us interpret large amounts of clinical, molecular, and genomic data much more efficiently. In oncology, the challenge is often not obtaining the genetic information, but determining which alterations are clinically meaningful and how they should influence treatment. AI can support this process by integrating tumor genomics with pathology, imaging, previous treatment response, and published evidence, allowing physicians to identify potentially actionable patterns more quickly.
A representative example in uro-oncology would be a patient with advanced urothelial bladder cancer whose molecular profiling identifies an FGFR3 alteration. AI-supported genomic interpretation can help flag this alteration as therapeutically relevant, assess available evidence, and place it in the context of the patient’s other clinical characteristics. This may lead the multidisciplinary team to consider an FGFR-targeted treatment rather than relying solely on a conventional one-size-fits-all treatment sequence.
Importantly, AI does not make the treatment decision. Its greatest value is as a decision-support tool: it can organize complex information, identify associations that might otherwise be overlooked, and help clinicians arrive at a more individualized treatment strategy while the final decision remains based on clinical judgment, validated evidence, and the patient’s preferences.
Dr. Martina Ambardjieva, MD, PhD
Urologist
Martina Ambardjieva, Medical expert, Invigor medical
Support Biomarker Tools With Clinical Review
AI can study patient-specific biomarkers, such as genes, proteins, and blood results, to estimate how a person may respond to treatment. This approach helps doctors move beyond broad treatment plans that are based on the average patient. A cancer drug, for example, may work well for people with one genetic marker but not another.
AI can find these patterns in large amounts of medical data more quickly than manual review alone. Its predictions should still be checked by qualified clinicians and discussed with the patient. Support responsible biomarker-based AI tools in clinical care.
Deploy Medication Alerts Against Dangerous Interactions
Decision-support systems can scan a patient’s medication record for possible harmful drug interactions. They can also consider allergies, kidney function, age, and other factors that affect medication safety. This is especially useful for people who take several medicines from different doctors.
An alert may help a clinician spot a dangerous combination before a prescription is filled. Too many low-value alerts can create alarm fatigue, so the system must focus on serious and relevant risks. Ask care providers to use well-designed medication safety tools.
Demand Fairness Audits for Medical Algorithms
AI can help uncover unfair patterns in medical recommendations that may otherwise remain hidden. A model can compare suggested treatments across groups while accounting for relevant health needs and access to care. This may reveal that certain patients are less likely to receive specialist referrals, pain treatment, or advanced therapies.
Finding the problem is only the first step, because biased data can also lead an AI system to repeat old inequities. Regular testing, diverse data, and human review are necessary to make these tools fairer. Demand transparency and fairness checks in every medical AI system.
Promote Shared Decisions Amid Predictive Uncertainty
Predictive AI models can help clinicians choose treatments when the outcome is uncertain. They may estimate the likely benefits, side effects, and risks of several options by using medical records and research data. This can support difficult decisions in areas such as intensive care, heart disease, and rare conditions.
A prediction is not a guarantee, because each patient has needs and values that data cannot fully measure. Doctors should explain uncertainty in plain language and make decisions with patients rather than relying on a score alone. Promote shared decision-making whenever AI predictions guide treatment.
Adopt Real-Time Monitoring With Strong Oversight
Machine learning can adjust treatment guidance as new patient data becomes available. Information from wearable devices, home monitors, lab tests, and symptom reports can show whether a therapy is helping or causing problems. For someone with diabetes, an AI system may identify changing glucose patterns and alert a care team to review the treatment plan.
This makes personalized care more responsive instead of waiting until the next scheduled visit. Clear safety rules are needed so that automated suggestions do not replace medical judgment. Encourage health systems to use real-time AI monitoring with strong clinical oversight.
