Related Experiment Video
Updated: Mar 13, 2026
Practical Considerations in Studying Metastatic Lung Colonization in Osteosarcoma Using the Pulmonary Metastasis Assay
Published on: March 12, 2018
The Vancouver Lung Cancer Risk Prediction Model: Assessment by Using a Subset of the National Lung Screening Trial
Charles S White1, Ekta Dharaiya1, Erin Campbell1
1From the Department of Diagnostic Radiology, University of Maryland, 22 S Greene St, Baltimore, MD 21201 (C.S.W.); Philips Healthcare, Highland Heights, Ohio (E.D.); and Philips Research North America, Briarcliff Manor, NY (E.C., L.B.).
A risk calculator effectively assessed lung nodule malignancy in the National Lung Screening Trial (NLST). This tool aids in distinguishing benign from malignant nodules, improving diagnostic accuracy for lung cancer screening.
Area of Science:
- Pulmonary Medicine
- Radiology
- Oncology
Background:
- Lung nodules are common findings in lung cancer screening.
- Accurate characterization of lung nodules is crucial for patient management.
- Differentiating benign from malignant nodules can reduce unnecessary invasive procedures.
Purpose of the Study:
- To evaluate a risk calculator for assessing malignancy likelihood in lung nodules from the National Lung Screening Trial (NLST).
- To determine the effectiveness of nodule and patient characteristics in predicting malignancy.
- To compare nodule characteristics between the NLST and a Vancouver cohort.
Main Methods:
- Utilized a nodule risk calculator developed in Vancouver, applying it to a subset of NLST data.
- Compared patient populations and nodule characteristics between the NLST and Vancouver cohorts.
- Employed logistic regression and tested multiple thresholds to optimize sensitivity and specificity.
Main Results:
- Analysis included 4431 nodules (4315 benign, 116 malignant) from the NLST.
- An optimal composite risk score threshold of 10% achieved 85.3% sensitivity and 93.9% specificity.
- The logistic regression model showed a high discriminant value with an area under the receiver operating characteristic curve of 0.963.
Conclusions:
- The Vancouver risk calculator demonstrated high discriminant value when applied to NLST data.
- Risk calculator methods are valuable for distinguishing benign from malignant lung nodules.
- This approach supports improved decision-making in lung nodule management.
