The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): a completed reference...

Area of Science:

• Medical Imaging
• Radiology
• Computer-Aided Diagnosis (CAD)

Background:

• Development of computer-aided diagnostic (CAD) methods for lung nodule detection, classification, and quantitative assessment requires well-characterized computed tomography (CT) scan repositories.
• The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI) established a publicly available reference database to facilitate medical imaging research.
• This initiative was a public-private partnership involving the National Cancer Institute (NCI), Foundation for the National Institutes of Health (FNIH), and Food and Drug Administration (FDA).

Purpose of the Study:

• To create a comprehensive and well-characterized database of thoracic CT scans to support the development and validation of computer-aided diagnostic (CAD) systems for lung nodule analysis.
• To provide a publicly accessible resource for the medical imaging research community, fostering advancements in lung nodule detection, classification, and quantitative assessment.
• To demonstrate the success of a consensus-based, public-private partnership in building a robust medical imaging database.

Main Methods:

• Collaboration between seven academic centers and eight medical imaging companies to address organizational, technical, and clinical challenges.
• Compilation of the LIDC/IDRI Database containing 1018 thoracic CT cases, each with associated XML files detailing a two-phase image annotation process.
• A two-phase annotation process involving four experienced thoracic radiologists: an initial blinded review for lesion marking and a subsequent unblinded review for refining opinions, aiming for comprehensive nodule identification without forced consensus.

Main Results:

• The database includes 7371 lesions marked as 'nodule' by at least one radiologist.
• Out of 2669 lesions marked as 'nodule >= 3 mm' by at least one radiologist, 928 (34.7%) were consistently marked as such by all four radiologists.
• The dataset comprises nodule outlines and subjective characteristic ratings for these identified lesions.

Conclusions:

• The LIDC/IDRI Database serves as an essential resource for medical imaging research.
• It is expected to accelerate the development, validation, and clinical implementation of CAD systems for lung nodule analysis.
• The database will spur innovation and dissemination of advanced diagnostic tools in clinical practice.