Chi-Fu Yang

M.D.

Stanford

Fellow

NLST (Learn more about this study)

NLST-607

Nov 25, 2019

Evaluating the Impact of Preoperative Characteristics on Outcomes of Patients Undergoing Curative-intent Treatment in the NLST

Many fundamental questions studied in lung cancer surgery have been answered by thorough, well designed analyses of national clinical and administrative databases. These analyses have used data from the major national cancer databases (e.g. the National Cancer Data Base, the Surveillance, Epidemiology, and End Results (SEER) Program, and the National Surgical Quality Improvement Program). The results from these studies have been used to inform national guidelines and recommendations regarding surgical treatment.

However, a key limitation to all of these studies noted above is that these databases contain very little to no data regarding the smoking history of patients. This is an important limitation, given that active smoking at the time of lung cancer surgery is known to adversely affect outcomes. One strength of the NLST is that not only does it contain detailed surgical and perioperative data, it also contains detailed data regarding smoking.

In our proposed project, we plan to study key, oft debated thoracic surgical questions, using the NLST. Through the strengths of the NLST, we hope that our analyses will be able to address major limitations of past studies that have used national, multicenter databases and provide results that can be used to improve decision making for surgeons and patients.

Aim #1: We plan to study the outcomes of minimally invasive vs open thoracic surgery, lobectomy vs sublobar resection and SBRT vs surgical resection for patients in the NLST who underwent curative-intent treatment.

Aim #2: We plan to study the perioperative morbidity, mortality and long-term outcomes (including recurrence-free and overall survival) for patients who are active smokers who undergo surgery versus patients who underwent surgery but quit smoking prior to surgery.

Aim #3: By analyzing the results from Aim #1 and Aim #2, and in combination with machine learning techniques, we plan to develop and validate accurate clinical prediction models to help physicians determine which patients undergoing major thoracic surgery are at highest risk of developing a perioperative complication.

Aim #4: We plan to study the optimal timing of surgery following smoking cessation and also following the diagnosis of lung cancer, using restricted cubic spline analysis and other statistical modeling.