Multimodal data-driven multitask learning for enhanced identification and classification of chronic obstructive pulmonary disease: a retrospective study

Key Points

• This research aims to assess the effectiveness of multitask learning models in identifying and classifying chronic obstructive pulmonary disease using integrated data.
• Utilized a multimodal approach combining chest CT images and clinical variables
• Compared performance of multitask learning models with traditional single-task methods
• Conducted a retrospective analysis of existing clinical data
• Multitask learning models surpassed single-task models in accuracy for COPD identification
• Enhanced classification capabilities were noted when integrating CT and clinical data
• Supports more informed clinical decision-making in the treatment of COPD