Biomarker Platform (Virtual Nodule Clinic) for the Management of Indeterminate Pulmonary Nodules

PRIMARY OBJECTIVES:

I. To test the hypothesis that usual care plus a radiomic prediction score impacts patient management compared to usual care alone.

II. To conduct a multicenter pragmatic randomized controlled platform trial using a validated biomarker, the radiomic prediction score.

III. To conduct a biomarker study that will evaluate the first necessary (but not sufficient) step to show clinical utility.

IV. To assess the magnitude of change in patient management with use of the radiomic prediction score.

V. To develop a platform that can be used as framework for future larger biomarker studies.

OUTLINE: Patients are randomized to 1 of 2 arms.

ARM I: Patients undergo standard of care (SOC) computed tomography (CT) evaluation and receive a Virtual Nodule Clinic radiomic prediction score on study. Patients then receive SOC lung nodule management on study.

ARM II: Patients undergo SOC CT evaluation on study. Patients then receive SOC lung nodule management on study.

Inclusion criteria

• Adults 35-year-old and older with undiagnosed IPN(s) 8-30mm referred for evaluation

• Referral includes direct in-basket messages in the electronic healthcare record (EHR) to study providers, telehealth visits or clinic visit
• For multiple nodules, we will obtain the score from the dominant or most suspicious nodule based on providers or radiologist impression
• Available CT scan with slice thickness of 3 mm or less with the nodule of interest present. Nodules identified during screening low dose computed tomography of the chest (LDCT) that have had a conventional, follow-up CT performed are eligible for inclusion

Exclusion criteria

• Pure ground glass nodule
• Patients known to be a prisoners
• Patients known to be pregnant
• Known active malignancy within the last 5 years at time of enrollment (excluding non-melanoma skin cancers)
• More than 5 IPNs present on imaging
• Nodules referred after initial LDCT for screening with only one LDCT available. The Lung Cancer Prediction Convolutional Neural Network (LCP CNN) algorithm is not currently validated for screening studies
• Thoracic implants that impact the image appearance of the nodule
• Clinician determines that use of the LCP CNN model is required or contraindicated for the optimal care of the patient