24. Successful implementation of Artificial Intelligence (AI) to Support a Lung Nodule Program in Driving Early Detection of Lung Cancer

Objective: The purpose of the study was to assess the influence of integrating an AI solution into a nodule clinic on patient management, workflow, and outcomes.
Methods: We integrated an FDA-cleared AI solution for patient identification, tracking and risk stratification into our nodule program and assessed changes to patient management workflow, nodule clinic visits and outcomes over a period of 3 months. The AI automatically identified CT radiology reports mentioning pulmonary nodules across the hospital system. Discovered patients were reviewed and contacted by a dedicated Thoracic nurse navigator and offered a nodule clinic appointment if appropriate. We compared this workflow to our previous best practice initiative, which involved emergency clinicians manually identifying nodule patients to the nurse navigator. Patients seen in the nodule clinic were assessed with the help of the AI malignancy risk score from 1 to 10, with 10 indicating the highest risk of malignancy. Clinicians recorded whether the score assisted them in their clinical management decision. Those patients who received biopsies or surgeries after being assessed in the nodule clinic with the assistance of the AI malignancy score, we determined the outcomes in the form of histological diagnosis and cancer stage. We compared the stage distribution in the 3 months after implementation of the AI to the stage distribution obtained from registry data for the year 2021, as more recent registry data are not available yet.
Results: Throughout the 3-month period after AI implementation, 51 new patients were identified by the AI and seen in the nodule clinic and 45 of those were assessed with the support of the AI malignancy score. We obtained feedback data for 32 patients, on how the score helped. In 97% of the patients, the surgeon reported that the score aided in the discussion with the patient. In most cases, the clinician reported the score confirmed their decision, in 3 cases the score convinced them to change to a less invasive clinical management decision and in 1 case to change to a more invasive one. 12 surgeries and 10 biopsies were scheduled for those patients. The lung cancer staging information of those patients is summarized in table 1. Of the 13 confirmed lung cancer patients, 11 (85%) were diagnosed in Stage 0-I, compared to 61 out of 133 (46%) in 2021. The software reduced the per-patient assessment times for the nurse navigator by eliminating 4 of the 13 steps from the workflow.
Conclusions: Integrating AI-based automatic identification and tracking of lung nodule patients into our nodule clinic simplified the workflow for clinical staff and helped treat lung cancer patients at early stages, where patients have the greatest opportunity for long-term survival.

K Lee (1), Jerome Feldon (2), lindsay Silas (3), Rogerio lilenbaum (2), (1) Jupiter Medical Center, Palm Beach Gardens, FL, (2) Jupiter Medical Center, jupiter, FL, (3) Jupiter Medical Center, Jupiter, FL

Moishe Liberman

Invited Discussant

Dr. Liberman is a minimally invasive thoracic surgeon interested in interventional thoracic endoscopy, thoracic oncology and airway surgery. He is the Director of the CHUM Endoscopic Tracheobronchial and Oesophageal Center (C.E.T.O.C.) and was appointed to be the inaugural Medical Director of the Multidisciplinary CHUM Robotic, Endoscopic and Minimally Invasive Surgery Program. He runs one of the busiest thoracic clinical trials programs in the country. He has run over 60 clinical trials at the CHUM over the last 13 years in Thoracic Surgical Oncology in the areas of neoadjuvant therapy, adjuvant therapy, surgical technology and novel devices.

 

His research interests include new minimally invasive techniques for staging and treatment of lung cancer and ultrasonographic technology in thoracic surgery. He also is very interested in trans-oral, non-surgical techniques for lung cancer diagnosis, staging and ablation as well as novel methods for performing VATS and Robotic lobectomy. His research laboratory focuses on technology assessment and technology development in minimally invasive thoracic oncology, endoscopy, robotics and natural orifice surgical techniques as well as clinical trials in thoracic cancer ablation and technology.

 

Dr. Liberman’s Laboratory is funded through the Canadian Foundation for Innovation, the Fonds de Recherche en Santé du Quebec (FRSQ), the Society of University Surgeons (SUS), the Thoracic Surgery Research Foundation of Montreal, Axelys and multiple industry partners. He has published well over 160 peer reviewed manuscripts, 50 book chapters and has been an invited lecturer over 160 times.

K Lee

Abstract Presenter

A pioneer and leading practitioner of thoracic surgery using minimally invasive techniques and was the first in the nation to perform a da Vinci robotic four arm lobectomy for lung cancer.  I have lectured and trained internationally other surgeons in robotic thoracic surgery, as well as on electromagnetic navigational bronchoscopy, and development of cancer centers in the regional and community settings.

 

Implemented community and regional Thoracic institutes on principles of Screening/Diagnostics/Treatment/Clinical Research, utilizing and marketing Multidisciplinary tumor boards, multidisciplinary clinics, and most recently a Women’s Lung program.