Interstitial lung diseases (ILDs) include a wide range of inflammatory and fibrotic conditions that vary in cause, severity and outcome. Their diagnosis has long relied on pulmonary function tests and CT imaging, but these conventional approaches often lack sensitivity for early disease and show variability between readers. Quantitative imaging has emerged as an objective, reproducible method to measure both global and regional abnormalities, improving detection and monitoring. Combined with artificial intelligence (AI) and machine learning (ML), it enables more precise classification, prognosis and treatment evaluation in key ILD subtypes such as idiopathic pulmonary fibrosis (IPF), hypersensitivity pneumonitis (HP) and connective tissue disease–related ILD (CTD-ILD).
Machine Learning and Quantitative CT Development
Quantitative imaging builds on ML and deep learning (DL) models that extract measurable features from medical images. Traditional ML depends on expert input to define parameters, while DL learns directly from large imaging datasets. Quantitative CT (QCT) methods range from simple threshold analysis, which quantifies areas of abnormal density, to texture-based models that capture spatial and morphological features. The Computer-Aided Lung Informatics for Pathology Evaluation and Rating (CALIPER) system, for example, segments the lungs and categorises tissue patterns such as ground glass, reticulation and honeycombing. These automated outputs correlate with functional indices and help assess disease extent, progression and response to therapy.
Advances in CT technology, particularly photon-counting detector systems, further refine image resolution and reduce radiation exposure. These improvements support the detection of small or early fibrotic changes and enhance the accuracy of ILD classification. Studies comparing such systems with conventional scanners show better depiction of fibrotic patterns and increased reader confidence, although further model calibration is still required. Together, these innovations allow quantitative imaging to provide consistent data for diagnosis and follow-up across patient groups.
Must Read: AI Triage Cuts Workload in Low-Dose Lung Screening
Clinical Applications Across Major ILDs
In IPF, quantitative imaging has progressed from density measurements to AI-driven analysis. Histogram and texture-based parameters help differentiate between fibrotic subtypes and predict clinical outcomes. CALIPER-derived features, including vascular-related structures, assist in surveillance and prognostication. DL tools have achieved higher accuracy than expert radiologists in identifying usual interstitial pneumonia patterns and estimating progression risk. Quantitative CT is also being incorporated into clinical trials of antifibrotic drugs, supporting objective assessment of treatment efficacy.
For HP, where both inflammation and fibrosis coexist, whole-lung averages are less reliable. Texture-based methods can distinguish prognostic groups and assess disease heterogeneity more effectively than visual scoring. Automated segmentation and volumetric tools link radiologic features with functional impairment and outcome. In CTD-ILD, quantitative imaging helps separate autoimmune-related fibrosis from idiopathic forms. Increased vascular structural volumes have been associated with poorer prognosis, while unsupervised and DL models have identified imaging markers linked to progression in systemic sclerosis. Some algorithms also integrate clinical data to refine risk prediction and mortality assessment.
Challenges and Emerging Functional Imaging
Despite rapid progress, several challenges limit routine clinical adoption. Variations in CT acquisition, reconstruction and patient cooperation can affect model accuracy and reproducibility. Preprocessing techniques and kernel conversion can reduce such inconsistencies, but validation across centres remains essential. Automated segmentation may still misclassify structures when affected by artefacts or suboptimal scans, highlighting the need for expert oversight. The development and maintenance of AI models also require large, well-curated datasets and significant computational resources, which restrict access beyond major academic settings.
Beyond structural analysis, functional imaging techniques are expanding the scope of quantitative evaluation. Dual-energy CT provides insight into pulmonary perfusion and early functional changes, while photon-counting CT combines high resolution with lower radiation dose. Magnetic resonance–based approaches, such as hyperpolarised xenon MRI and phase-resolved functional lung imaging, enable visualisation of gas exchange and ventilation-perfusion balance. These methods reveal physiological alterations that may precede irreversible fibrosis. Although currently limited to research environments, they demonstrate the potential to complement structural imaging for a fuller picture of ILD pathophysiology.
Quantitative imaging has become a transformative tool for assessing interstitial lung disease, offering objective and reproducible measures of structure and function. Supported by AI and advanced CT and MRI technologies, it enhances early detection, characterisation and monitoring, paving the way for more individualised management. While technical variability, cost and access remain barriers to widespread use, continued innovation and standardisation are likely to embed quantitative imaging into routine clinical practice. Its integration promises more accurate diagnosis, improved treatment planning and better long-term outcomes for patients with ILD.
Source: Radiology: Cardiothoracic Imaging
Image Credit: iStock
