Longitudinal smoothing techniques resolve inconsistencies in smoking history across multiple time points.
Risk model-based surveillance incorporating enhanced smoking data outperformed NCCN Guidelines in identifying second malignancies.
Guideline-Based Recommendations
Diagnosis
Use detailed smoking history including pack-years, duration, and quit-years for lung cancer risk assessment.
Incorporate longitudinal smoking data rather than single time-point status for accurate risk stratification.
Management
Apply automated LLM-based extraction methods to improve smoking documentation quality in EHRs.
Use enhanced smoking history data to guide lung cancer screening eligibility and surveillance strategies.
Monitoring & Follow-up
Implement longitudinal data smoothing to identify and correct inconsistencies in smoking history over time.
Regularly update smoking status and related variables to inform ongoing risk assessment.
Risks
Be aware of potential inaccuracies and hallucinations in LLM-extracted data; apply rule-based corrections.
Recognize that incomplete or inconsistent smoking documentation can impair risk assessment and surveillance.
Patient & Prescribing Data
Lung cancer patients with documented smoking histories across multiple healthcare systems
Enhanced smoking history extraction enables improved identification of patients eligible for lung cancer screening and surveillance, potentially reducing missed second malignancies.
Clinical Best Practices
Combine generative LLMs with rule-based longitudinal smoothing for robust smoking history extraction.
Validate extraction models across diverse healthcare settings to ensure generalizability.
Use comprehensive smoking variables beyond status (e.g., pack-years, quit-years) for clinical decision-making.
Incorporate smoking history data into risk models to optimize lung cancer monitoring and follow-up.
by Ingrid Luo, Anna Graber-Naidich, Mengrui Zhang, Rakshit Kaushik, Grant M. Nieda, Tony Chen, Bo Gu, Eunji Choi, Victoria Y. Ding, Fatma Gunturkun, Mina Satoyoshi, Archana Bhat, Tae Yoon Lee, Chloe C. Su, Timothy John Ellis-Caleo, A. Solomon Henry, Manisha Desai, Leah M. Backhus, Natalie S. Lui, Ann Leung, Joel W. Neal, Allison W. Kurian, Curtis P. Langlotz, Heather A. Wakelee, Su-Ying Liang, Aparajita Khan, Summer S. Han
