A chatbot-based communication tool significantly improved understanding of cancer risk among adolescents and young adults (AYAs) with hereditary cancer predisposition syndromes, according to findings from a randomized clinical trial published in JAMA Network Open.

The study, led by Junne Kamihara, MD, PhD, and colleagues at Dana-Farber Cancer Institute and collaborating institutions, found that the AYAs who used the digital tool following a routine clinical appointment demonstrated greater improvements in their knowledge of cancer risk compared with those receiving standard clinical care alone.

The findings suggest that using personalized digital communication tools could complement traditional genetic counseling and cancer surveillance discussions, allowing younger patients to better understand their risks without increasing any psychological distress.

“Young people with cancer predisposition have an elevated risk of future cancers, but many are unaware of the magnitude of that risk,” corresponding author Jennifer W. Mack, MD, MPH, a pediatric hematologist-oncologist at the Dana-Farber Cancer Institute and Professor of Pediatrics at Harvard Medical School, Boston, explained to ASCO AI in Oncology. “Unfortunately, clinical visits seemed to have limited impact on cancer risk knowledge in the patients we studied. Additional tools could help young people to better recognize their risks, which in turn may help them to plan for the future and more fully engage in their care.”

Evaluating a Chatbot-Based Intervention

The multicenter trial enrolled 106 patients aged 12 to 24 years with one of nine hereditary cancer predisposition syndromes, including Li-Fraumeni syndrome, Lynch syndrome, and hereditary breast and ovarian cancer associated with BRCA1 and BRCA2 pathogenic variants. Conducted across four U.S. cancer risk clinics at academic medical centers between December 2022 and May 2025, the study randomly assigned 54 participants to receive standard clinical care supplemented by the AYA Risk Information and Screening Education (AYA-RISE) chatbot and 52 to receive standard care alone. The mean age of participants was 18.3 years and 64% were girls or women. 

Developed with input from the patients and their families, AYA-RISE provided information tailored to each syndrome regarding cancer risk, recommended screening, and prevention strategies. Its interactive format allowed users to explore information at their preferred level of detail, alongside resources addressing family communication, psychological support, and future planning.

Importantly, the chatbot did not use generative AI—all educational content was developed by the research team using published literature and clinical screening guidelines, with predefined conversational pathways allowing patients to navigate relevant topics. Of this decision, Dr. Mack noted that the team wanted to ensure that all information provided by the chatbot was “accurate, up to date, and based on existing national guidelines.”

Improvements in Cancer Risk Knowledge

Before their clinical appointments, 52% of participants assigned to AYA-RISE and 38% assigned to standard care accurately understood their cancer risk before turning 30. Following the intervention, these proportions increased to 78% in the chatbot group but remained at 37% in the control group (odds ratio [OR] = 3.50; 95% confidence interval [CI] = 1.45–9.19; P = .005). The intervention was associated with significantly greater improvement in knowledge. Improvements in understanding lifetime cancer risk and recommended screening also favored the intervention arm but did not reach statistical significance (OR = 2.17; 95% CI = 0.85–5.59; P = .11).

Crucially, the increased understanding generated by the chatbot did not appear to increase psychological distress, among other psychological aspects of hereditary cancer that were studied. Distress scores decreased slightly among chatbot users relative to controls, although the difference was not statistically significant (beta = –0.50; 95% CI = –1.09 to 0.08; P = .09).

“We felt that it was critical to assess distress related to the intervention, because we were putting sensitive, sometimes difficult, information directly into the hands of young people. We worried that we might cause distress and have limited opportunity to address it given that the intervention was delivered remotely,” said Dr. Mack. “However, we also knew from our previous work that this knowledge can be empowering and has the potential to relieve distress by ameliorating uncertainty, acknowledging challenges rather than hiding them, and giving young people tools to manage their concerns.”

Digital Oncology Care Implications

The researchers acknowledged several limitations with their study, including a relatively small sample size, recruitment of only 25% of eligible patients they approached, limited racial and ethnic diversity, and the revision of enrollment goals to match the study’s funding timeline. Nevertheless, the findings highlight the potential of integrating carefully designed digital interventions into oncology care, particularly for patients navigating complex hereditary cancer information.

Rather than replacing clinical interactions, the AYA-RISE model demonstrates how personalized, evidence-based chatbot technology can supplement clinician-led discussions and provide patients with greater autonomy over how they access sensitive health information. “Clinical consultations are extremely important for this young population. At the same time, we know that adolescents and young adults do not always feel that communication in medical visits meets their needs,” Dr. Mack explained. “They may also not feel comfortable asking questions or may feel anxious during visits and have a hard time processing the information they receive. We felt that this hybrid approach would allow young people to have in-person attention from a clinician, and also the opportunity to explore topics independently on their own time.”

As such AI-driven patient communication tools become increasingly prominent in oncology, the study also provides evidence supporting structured, clinically validated conversational systems that deliver individualized information within established care pathways.

“We think that interventions like this one could be a helpful adjunct to care for young people with cancer risk across a range of settings,” said Dr. Mack. “They have associated costs to build, maintain, deploy, and update material, and so wider dissemination would require health care systems or foundations to support these tools, including the technology and staff needed to maintain them. This has strong potential to benefit young people who are facing a challenging diagnosis, and who greatly need our support and ongoing attention in research and clinical care.”

DISCLOSURES: Dr. Kamihara reported that their spouse has received consulting fees, honoraria, and research support from multiple pharmaceutical and biotechnology companies, and holds equity, founder, and advisory roles in several firms. Dr. Schienda and Dr. Maese reported NIH grants during the study, with Dr. Maese also receiving personal fees from Jazz Pharmaceuticals and Servier Pharmaceuticals; Dr. Vanderwall reported NCI grants and employment-related compensation from Labcorp, while Dr. Simmons received institutional support from Dana-Farber Cancer Institute. Dr. Schiffman reported employment and stock options in Peel Therapeutics, Dr. Desai reported consultancy fees, advisory roles, stock ownership, and institutional clinical trial support from multiple pharmaceutical companies, and Dr. Savage reported compensation and stock ownership in Invitae. Dr. Mack reported NCI grants unrelated to the study.

ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO^®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.