RF105. American Association for Thoracic Surgery Quality Gateway Procedure Outcomes Models for General Thoracic Surgery Quality Assurance

Objective: To establish a general thoracic surgery database and develop quality assurance models with modern analytics for postoperative adverse events and procedural and oncologic quality after lung resection and esophagectomy for cancer.
Methods: Outcome models were separately developed based on 92 preoperative patient, cancer, induction therapy, and intended surgery variables of 16,430 lung resections for non-small cell cancer and 89 similar preoperative variables of 3,068 esophagectomies for adenocarcinoma and squamous-cell cancer in adults performed in 9 hospitals/hospital systems. Random forest quantile classification for imbalanced (rare-event) data and model-free variable priority screening identified predictors with high information content. 9 models were produced for each cancer, 5 for individual outcomes and 4 for consensus-driven composite models.
Results: From 24 postoperative outcomes after lung resection and esophagectomy considered, 6 models of individual outcomes were developed for primary reporting (Table 1A), 5 in common between lung and esophageal resections and 1 specific to each. A pulmonary composite of 8 complications plus a 9th (prolonged air leak) for lung resection was formed (Table 1B) as was a composite of 6 infections (Table 1C). A procedural composite was formed of 9 outcomes plus 1 each unique for lung and esophageal resections (Table 1D). An oncologic quality composite for both lung and esophageal resections included positive margins and non-compliance with lymph node sampling plus, for lung resection, use of thoracotomy during resection of a clinical Stage I cancer (Table 1E). The most common preoperative risk factors identified for outcomes after both lung and esophageal resections were older age, larger patient size, lower FEV1 (% of predicted) and DLCO (% of predicted), higher ECOG score, more pack-years of smoking, lower hemoglobin, and higher creatinine. Models incorporated between 14 and 32 preoperative variables in all to predict well-calibrated probabilities of outcomes with good discrimination.
Conclusion: Models for predicting operative mortality and major morbidity, composites of pulmonary and infection complications, and composites of procedural and oncologic quality have been developed for lung resection and esophagectomy for cancer using advanced machine learning technology. They provide surgeons, patients, hospitals, and hospital systems with modern tools for quality assurance in general thoracic surgery.

Eugene Blackstone (1), Sudish Murthy (1), Daniel Raymond (1), David Jones (2), James Isbell (2), Jonathan Yeung (3), Hemant Ishwaran (4), (1) Cleveland Clinic, Cleveland, OH, (2) Memorial Sloan Kettering Cancer Center, New York, NY, (3) University Health Network, Toronto, ON, (4) University of Miami, Miami, FL

Sudish Murthy

Rapid Fire Abstract Presenter

Dr. Sudish Murthy holds the Daniel and Karen Lee Endowed Chair in Thoracic Surgery, is the Section Head of General Thoracic Surgery, and Surgical Director of the Center of Major Airway Disease. He is a Thoracic Surgeon in the Department of Thoracic & Cardiovascular Surgery at The Sydell and Arnold Miller Family Heart & Vascular Institute at Cleveland Clinic and a Professor of Surgery at the Cleveland Clinic Lerner College of Medicine of Case Western Reserve University. Dr. Murthy is over 10,000 operations throughout his lengthy and distinguished career.

Education and Training: Dr. Murthy earned his medical degree from Columbia University College of Physicians and Surgeons and a PhD in pathology from the University of British Columbia, Vancouver, BC. At Columbia, he received the Janeway Prize for top achievement in his graduating class, the Robert F. Loeb Award for Excellence in Clinical Medicine, the Merck Award for outstanding scholarship, and several other awards.

Dr. Murthy completed an internship and a residency in surgery at Brigham and Women’s Hospital of Harvard University, Boston. He continued there for a residency in cardiothoracic surgery. As a Harvard University Clinical Fellow, he was selected as the Whitman Traveling Scholar and served as the surgical emissary for Harvard Medical School in the Department of Esophageal Surgery at Queen Mary Hospital Medical Center at the University of Hong Kong.

Publications and Speaking: Dr. Murthy is a dedicated researcher and prolific writer. He has authored or co-authored more than 250 scientific articles in leading peer-reviewed medical journals and more than a dozen chapters in medical textbooks. His research interests include lung cancer, lung transplant and emphysema

Specialties: Congenital, Multi-Specialty, Thoracic