Events

Detecting lung cancer in ambulatory care – how and when do patients present?

Matthew Thompson,
MBChB, MPH, DPhil

Helen D. Cohen Professor,
Department of Family Medicine

Lesleigh Kowalski Frank,
PhD, MOT, OTR/L

Research Scientist,
Department of Family Medicine

Monica Zigman Suchsland, MPH

Research Scientist,

Department of Family Medicine

Hannah Burkhardt, BS

Graduate Student,

Department of Family Medicine

DESCRIPTION

Lung cancer is the most common cancer-related cause of death in the US. Most patients present at a later stage of their illness where curative options are more limited. While screening for lung cancer has been available for several years, the overwhelming majority of patients are diagnosed following clinical presentation in ambulatory care settings.  However, there is very little data in the US on how patients with lung cancer present clinically in ambulatory care, how clinicians could differentiate patients with lung cancer from those with other more common illness, and the duration of the pre-diagnosis period. We present results from a case-control and cohort study of patients diagnosed with lung cancer from UW Medicine which used Natural Language Processing (NLP) to identify symptoms/signs from free text clinical notes from patients with lung cancer and matched controls from ambulatory care. We present potentially ground-breaking new evidence on how and when patients with lung cancer present in ambulatory care, and the value of NLP for extracting granular information from electronic medical records.

learning objectives

1. Recognize the role of primary care clinicians in early detection of lung cancer among patients who present symptomatically
2. Understand the use of NLP to extract detailed information on clinical features from routine electronic medical record data.
3. Identify potential for future opportunities to advance detection of lung (and other) cancers by understanding the key role of primary care in early detection

2022 RESEARCH SEMINARS: