AI in Oncology: Extending Capacity Responsibly at UCLA Health

AI in Oncology –  Real-World Deployment at UCLA Health

The oncology clinic faces an impossible equation. Rising patient volumes, complex maintenance therapies, limited staffing, and the pressure to improve outcomes and trial access have created chronic strain. Clinicians know patients need support between visits. They know trials sit unfilled because eligible patients never hear about them. They know referral processes are stuck in email chains from another decade. But adding another tool to an already overwhelmed workflow feels impossible. 

Yet at UCLA Health, oncology teams are solving this problem not by adding burden, but by reducing it.

The principle guiding this work is straightforward: AI should augment clinical care, not create additional administrative burden. This isn’t aspirational thinking—it’s what’s actually happening in two concrete deployments that show a different path forward.

How AI in Oncology Is Reducing Operational Burden

The AskGRACE Study: Supporting Patients Between Clinic Visits

The Challenge

Start with the patient experience. Women initiating maintenance therapy for advanced ovarian, fallopian tube, or primary peritoneal cancer face significant symptom burden and anxiety between their scheduled clinic visits. They need guidance on nausea management, clarification about whether a side effect is expected or serious, understanding about their prognosis. They experience what clinicians call “fear of recurrence”—anxiety about disease progression that significantly impacts quality of life.

What do they currently have? An option to call the triage line during business hours and hope someone picks up. The internet, which amplifies uncertainty with unreliable information. Silence until the next clinic appointment.

What they need: Evidence-based guidance available when the anxiety hits—often at night or on weekends. Clarity on the difference between expected side effects and something that requires immediate attention. Continuity of information between clinic visits.

The Deployment

AskGRACE—Guided Recommendation Assistant for Cancer Engagement—is a partnership between Botco.ai and UCLA’s Department of Gynecologic Oncology, in collaboration with City of Hope and sponsored by GSK. It addresses this gap through a 24-month prospective, multicenter randomized controlled trial (VIRGO2 Study), with principal investigator Dr. Dana Chase, MD from the David Geffen School of Medicine at UCLA.

The study enrolls women ages 18 and older with advanced stage (III-IV) or recurrent ovarian, fallopian tube, or primary peritoneal cancer who are initiating maintenance therapy, or in active surveillance within 90 days of enrollment. Here’s how it works from a patient perspective: Available via SMS (310-347-4850) or QR code, patients can ask AskGRACE questions about symptom management (“How do I manage my nausea?”), drug safety (“I experienced a new side effect, is that expected?”), prognosis education (“Help me understand my prognosis and survival rates”), or general guidance about next steps.

The system responds with evidence-based tips verified by UCLA clinicians. Patients receive reminders and outreach designed to keep them engaged with their care plan. But critically: if a patient reports something concerning, the system doesn’t just reassure—it flags the case immediately for clinical review.

Why This Matters Clinically

From the care team perspective, this translates to three concrete changes:

1. Better-prepared patient visits. Clinicians have visibility into what patients have been experiencing between visits—symptom patterns, medication concerns, questions about prognosis. Instead of using precious clinic time to gather this history, it’s already documented.

2. Reduction in low-acuity phone triage. Many calls to the triage line are questions AskGRACE can handle—clarification about expected side effects, symptom management guidance, medication questions. The system handles these, freeing triage staff for higher-acuity issues.

3. Earlier intervention for serious issues. Because the system is in contact with patients 24/7, it detects adherence challenges, severe symptoms, or psychological distress faster than a monthly clinic visit would catch them. Escalation happens automatically.

The infrastructure is built on Botco.ai’s HIPAA-verified, SOC 2 Type II compliant architecture with field-level masking and encryption for all protected health information, role-based access controls, and complete logging of all AI interactions for audit requirements.

This isn’t a black-box AI in Oncology project. It’s embedded in clinical governance—UCLA clinicians verify the content, set escalation protocols, and oversee the entire process.

OncoTrial Match™: Closing the Trial Access Gap in 90 Seconds

The Problem

One of the biggest barriers to clinical trial participation is awareness. Many eligible cancer patients never learn about relevant trials.

Behind that are operational realities. Trial listings are fragmented across institutions and change daily. Busy oncologists lack time to manually track evolving eligibility criteria. Cross-institution referrals rely on slow email communication or phone tag. When a patient might be eligible for a potentially life-extending trial, the referral process is so cumbersome that it often doesn’t happen at all.

The Solution

OncoTrial Match™, developed in collaboration with UCLA’s Department of Gynecologic Oncology, addresses this with remarkable simplicity. A physician texting a designated number initiates an easy workflow that takes less than 90 seconds from start to automated referral.

What Makes This Different

The referral email goes directly to the trial site coordinator with structured patient information. 

This is closed-loop communication that previously took days happening in under 90 seconds.

OncoTrial Match integrates real-time data from three sources: ClinicalTrials.gov (the primary federal repository), the Sandy Rollman clinical trial database, and the UC5 cancer center portfolios. It’s powered by Botco.ai’s GenAI platform with secure, HIPAA-compliant SMS technology as the interface.

The goal is streamlining referral communication and patient navigation, and reducing the friction that keeps eligible patients from learning about trials.

The impact includes:

• Faster identification (matching and referral under 90 seconds)
• Cross-campus visibility of recruiting trial portfolios across all five UC cancer centers
• Improved access for both patients and physicians by removing administrative barriers

The Operational Philosophy

Both deployments reflect a mature approach to clinical AI: the goal isn’t implementing cutting-edge technology for its own sake. It’s reducing friction in already-strained workflows while maintaining clinical governance, privacy, and safety.

This shows up in concrete design choices. SMS-based access (no apps, no logins, no new systems to learn). Escalation pathways that route concerning symptoms immediately to human clinicians. Complete audit trails and role-based access ensuring institutional privacy requirements are met.

“As Rebecca Clyde, CEO of Botco.ai, puts it:

‘The opportunity is extending capacity responsibly.’”

 

For health systems watching this work, the takeaway is clear: responsible AI deployment in oncology is happening now, not in the future. The question isn’t whether to implement it, but how to do it in ways that reduce burden, maintain governance, and respect clinical expertise.

About Botco.ai

Botco.ai builds enterprise-grade agentic AI platforms for regulated industries. The AskGRACE and OncoTrial Match deployments represent proven implementations in oncology care. To learn more, visit botco.ai or contact [email protected].