Virtual AI-Assisted Music Therapy for Anxiety Symptoms in People with Advanced Cancer (VIRTUOSO)

As oncological treatments improve, people with advanced cancer are living longer but are increasingly burdened by physical and psychological symptoms. Anxiety affects 1 in 2 people with advanced cancer and is linked to worse health outcomes and poor quality of life. Despite existing treatments, anxiety is under-treated in 40% of this population and is one of the highest unmet needs. Music therapy (MT) is a non-pharmacological intervention in which therapists guide patients through music experiences, ranging from receptive (music listening) to active (songwriting). Most MT studies for anxiety used receptive techniques, which produce short- term anxiety reduction, but not long-term reduction. Research in non-cancer populations suggests collaborative songwriting can foster social connection and provide outlets for processing difficult experiences, both of which are critical for long-term anxiety reduction. However, studies had key methodological limitations that make it impossible to determine the specific efficacy of collaborative songwriting. While songwriting has historically required extensive time, effort, equipment, and expertise, the growth of song creation platforms, powered by artificial intelligence (AI), has democratized songwriting. Despite their widespread availability, these AI-powered platforms have not been integrated into MT interventions or rigorously evaluated in clinical trials. We conducted a pilot RCT in which N=30 were randomized to 8 weekly Zoom sessions of an AI-assisted collaborative songwriting (ARTIST) intervention or a therapist attention-music listening (TAME) control. ARTIST participants reported clinically meaningful, durable anxiety reduction at week 16, whereas TAME did not. Moreover, in AI- powered facial expression analyses, ARTIST participants demonstrated significantly higher engagement, an objective marker of emotional expressivity, compared to TAME. Our work challenges the conventional MT paradigm of using receptive approaches to treat anxiety and suggests collaborative songwriting may produce greater long-term improvement by providing a safe outlet for expressing emotions and processing cancer- related experiences. However, its specific efficacy needs to be confirmed in a rigorous trial. We bring together a multidisciplinary team to propose the Virtual AI-Assisted Music Therapy for Anxiety Symptoms in People with Advanced Cancer (VIRTUOSO) study. We will conduct an RCT randomizing N=240 equally to the ARTIST intervention or a TAME control. Interventions will be delivered via Zoom over 8 weeks. Anxiety and social- cognitive processes will be assessed for up to 26 weeks. We will use AI-powered facial expression analyses as objective indices of emotional engagement. The specific aims are: 1) to determine the long-term efficacy of ARTIST on anxiety, 2) to assess the social-cognitive mechanisms of ARTIST, and 3) to explore associations between AI-powered facial expression analyses and anxiety outcomes in the context of MT. The VIRTUOSO study has the potential to expand treatment options for anxiety, deepen mechanistic understanding of MT, and integrate AI technologies into MT practice and research to benefit millions of people impacted by cancer. Project Number: 1R37CA296338-01A1 | Fiscal Year: 2026 | NIH Institute/Center: National Cancer Institute (NCI) | Principal Investigator: Kevin Liou | Institution: SLOAN-KETTERING INST CAN RESEARCH, NEW YORK, NY | Award Amount: $743,809 | Activity Code: R37 | Study Section: Biobehavioral Medicine and Health Outcomes Study Section[BMHO] View on NIH RePORTER: https://reporter.nih.gov/project-details/11296984