Surgical Segmentation
Computer vision for anatomy, workflow, and procedure understanding across laparoscopic and robotic surgery.
Projects
Surgical AI systems across the full surgical journey.
A-STAR organizes active systems around practical clinical questions, model-card discipline, and staged validation across planning, guidance, education, and outcomes.
Interactive surgical journey
Four phases, one validation path.
Select a phase to see the focus, methods, example systems, and validation outputs.
Research directions
Focus areas for a growing surgical AI lab.
Dissection Zone Recognition
Go and No-Go safety-zone modeling for surgical coaching, intraoperative review, and decision-support research.
Generative Surgical AI
Multimodal and generative systems for simulation, education, workflow synthesis, and surgical intelligence.
External Validation
Prospective and multi-site validation so model claims stay measurable, bounded, and clinically honest.
Clinical Ecosystem Integration
Collaboration with surgeons, trainees, engineers, and clinical platforms to fit tools into real workflows.
Active projects / systems
Specific systems with intended use, validation state, and readiness.
Status · validation
MOSI
Metabolic & Obesity Staging Index
A staging algorithm and clinical audit platform for bariatric surgery, validated on 3,097 patients. MOSI scores patients across BMI, comorbidities, and severity to recommend procedures and target weight-loss tiers.
- Intended use
- Pre-operative decision support for bariatric surgery candidate evaluation. Advisory only; clinical judgment remains with the treating surgeon.
- Clinical phase
- Risk stratification, planning, and outcomes validation
- Validation
- Derivation and publication complete; prospective external multi-site validation is in progress.
- Readiness
- Validation phase. Available as a clinical decision-support aid, not for autonomous use.
Status · deployed
SIRIS
Surgical-IRIS Education
SIRIS helps patients ask specialty-focused education questions, review Mayo Clinic resources, and prepare questions for their care team — bridging the information gap between booking and operation.
- Intended use
- Patient-facing education support for bariatric surgery preparation, question generation, and resource navigation.
- Clinical phase
- Patient journey and education
- Validation
- Deployed with formal clinical evaluation and user-satisfaction study design in progress.
- Readiness
- Public education deployment. Not a diagnostic tool and not a replacement for clinician consultation.
Status · validation
GoNoGoNet
AI-guided Go/No-Go surgical safety zone detection
Computer vision model for identifying surgical dissection safety zones and supporting intraoperative decision-making.
- Intended use
- Research decision-support for identifying laparoscopic dissection regions that may represent Go and No-Go safety zones. Advisory only; it is not a replacement for surgeon judgment.
- Clinical phase
- Intraoperative surgical guidance and education
- Validation
- Published retrospective and experimental validation records are linked; broader translational validation and generalizability testing remain required.
- Readiness
- Research and validation phase. Not designed for autonomous clinical decision-making.