Cardiothoracic surgery AI scribe 2026: op notes for CABG, valve, lung resection, and the high-stakes documentation reality

May 8, 2026 · 6 min read

Cardiothoracic surgery (CT surgery) generates the most documentation-heavy operative notes in medicine. CABG, valve replacement, lung resection, esophagectomy, transplant — each procedure has structured CPT-driven elements, STS database fields, and downstream ICU handoff requirements. Mortality and morbidity data are submitted to the STS National Database, which feeds quality reporting and surgeon-level outcomes data. The documentation has to be exhaustive and accurate.

The 2026 CT-surgery-aware AI scribe stack handles four things general scribes miss: structured op note format aligned with STS database fields, ICU handoff documentation with hemodynamic detail, complication tracking aligned with STS adverse event definitions, and the multi-specialist coordination (cardiology, pulmonology, anesthesia, intensive care) that defines CT surgery practice.

STS database alignment

The Society of Thoracic Surgeons (STS) National Database collects detailed data on every adult cardiac, pediatric cardiac, and general thoracic surgical case at participating institutions:

• STS Adult Cardiac Surgery Database: ~5 million records
• STS Congenital Heart Surgery Database
• STS General Thoracic Surgery Database

Documentation that maps directly to STS fields (specific risk factors, procedure detail, complications by category) reduces the manual abstraction burden and improves data quality for benchmarking. AI scribes producing structured output can populate STS fields directly via integration.

The CT surgery op note structure

You are documenting a cardiothoracic surgery operative note.

INPUT:
- Op dictation transcript (with surgeon, anesthesia, perfusion, scrub voices)
- Pre-op data: dx, indications, risk factors (ejection fraction, prior MI/CABG, COPD/FEV1, eGFR, diabetes, prior stroke)
- Implant inventory (prosthetic valve types/sizes, conduit, stapler cartridges)

OUTPUT structured op note for STS Database alignment:
1. Pre-op diagnosis with anatomic detail (e.g., 3-vessel CAD with LM disease, severe AS with mean gradient 48)
2. Procedure performed with CPT codes (CABG x N grafts, AVR, mitral repair vs replacement, lobectomy, etc.)
3. Surgeon, assistants, anesthesiologist, perfusionist (for cardiac)
4. Anesthesia type, ETT placement, lines (PA catheter, art line, central line)
5. Pre-bypass: position, prep, incision approach (full sternotomy / mini / robotic), pleurae opened
6. Cannulation: aortic, venous (bicaval vs single), additional cannulas
7. Bypass times: total CPB time, cross-clamp time, circulatory arrest time if used
8. Cardioplegia: type (blood/crystalloid), antegrade/retrograde, dose intervals, temperature
9. Procedure detail per anatomic step:
   - For CABG: each graft with target vessel, conduit (LIMA/RIMA/SVG/radial), anastomosis technique, distal anastomosis order
   - For valve: prosthesis type/size/lot, suture technique (interrupted vs continuous), annular preservation
   - For lung resection: anatomic vs sub-anatomic, stapler use, lymphadenectomy stations, frozen section results
10. Weaning from bypass: rhythm, hemodynamic, inotrope use, transfusion
11. Closure: chest tubes (location, size), wires, sternum reapproximation, layer closure
12. Estimated blood loss + transfusion
13. Intra-op complications by STS category
14. Disposition: ICU with hemodynamic profile

For each clinical fact, cite transcript or implant inventory. For STS-relevant fields (bypass time, cross-clamp time, conduit type), ensure values are precise.

ICU handoff documentation

CT surgery patients go directly to ICU with active hemodynamic management. The handoff documentation should structure:

• Hemodynamic profile at transfer: HR, BP, CVP, PA pressures, cardiac output if available, urine output
• Inotrope/vasopressor regimen with current doses
• Anesthesia-relevant events during case (hypotension, arrhythmia, transfusion)
• Bleeding status: chest tube output trend, coagulation parameters
• Ventilation: FiO2, PEEP, PaO2/FiO2 ratio
• Specific concerns: graft patency, valve function on TEE, etc.
• Plan: extubation timing target, expected ICU course

Complication tracking aligned with STS

STS adverse event definitions include specific criteria for: stroke (with NIHSS), AKI (KDIGO), prolonged ventilation (> 24 hr), reoperation for bleeding, deep sternal wound infection, pneumonia, sepsis. Documentation should capture each event with the qualifying detail (timestamp, severity, intervention).

Vendor and DIY paths

For larger CT surgery practices embedded in IDN, vendor enterprise integration with Epic / Oracle Health is standard. For smaller practices or IDN-employed surgeons wanting personal-control documentation, the DIY stack with structured op note schema works — but the multi-clinician audio capture in the OR is hardware-intensive (multiple mics, separate channels for surgeon / perfusion / anesthesia).

BAA chain

Hospital + EHR + multi-channel audio capture vendor + transcription vendor + LLM vendor + STS Database submission infrastructure.

When to start

For CT surgery practices participating in STS National Database (almost all do for accreditation and benchmarking), structured AI scribe output that maps to STS fields directly reduces abstraction burden. The surgeon-level outcomes reporting accuracy benefits from structured documentation. The ROI is in time saved + outcomes reporting quality.