Childhood Interstitial Lung Disease: A Pattern-Based Imaging Approach

A practical radiology review of Childhood Interstitial Lung Disease: A Pattern-Based Imaging Approach, focused on imaging findings, differential diagnosis, reporting points, and high-yield teaching pearls.

• What is it?
• Heterogeneous diffuse lung diseases unique to childhood

Not same entities or patterns as adult ILD

• Rare but high morbidity — esp. surfactant disorders (100% neonatal mortality w/o transplant)

Classification logic (ATS-2013)

1. First exclude common causes (infection, aspiration, CF, BPD)

Then group by age at onset

• Infancy-specific (developmental, lung growth, undefined, surfactant dysfunction)
• Not-infancy-specific (normal host, systemic disease, immunocompromise, mimics)

Imaging pillars

• CT = reference standard (confirms ILD, maps extent, narrows DDx, biopsy targeting)
• Secondary pulmonary lobule anatomy is central to pattern reading
• CXR low-specificity but OK for follow-up / screening

Infancy-predominant buckets & Critical CT clues

• Diffuse developmental arrest → GGO + septal thickening ± lymphangiectasia
• Lung growth abnormalities → overinflation, cysts, hypoplasia patterns
• NEHI → middle-lobe/lingula GGO + air-trapping (diagnosable on CT alone in many)
• Surfactant dysfunction (SFTPB/C, ABCA3, TTF-1) → PAP / CPI / NSIP / DIP blends; GGO + cysts common
• Mutations in surfactant machinery genes (SFTPB, SFTPC, ABCA3, TTF-1/NKX2-1)
• Leads to mis-processing / mis-clearing / mis-packaging of surfactant
• Result = injury + inflammation + abnormal filling or fibrosis → ILD patterns on CT
• Why multiple patterns (PAP / CPI / NSIP / DIP)?

Surfactant homeostasis failure can

• Fill alveoli with protein/lipid (→ PAP phenotype)
• Inflame & injure alveolar walls (→ CPI / NSIP phenotype)
• Accumulate macrophages in alveoli (→ DIP phenotype)
• Hence one gene defect can phenocopy several classic ILD patterns
• Dominant CT phenotypes — what you actually see
• PAP-like → crazy paving (GGO + septal thickening)
• CPI/NSIP-like → diffuse/basal GGO + mild septal thickening + subtle fibrosis
• DIP-like → diffuse GGO with fine reticulation + macrophage filling inferred
• Cysts (esp. SFTPC) → rounded thin-walled or irregular cysts within GGO background
• Gene-pattern pearls (high-yield)
• SFTPB → neonatal, severe, PAP-dominant, lethal without transplant
• ABCA3 → later than SFTPB, PAP ± CPI/NSIP mix, somewhat milder
• SFTPC → infancy/childhood, GGO + cysts common, variable severity
• TTF-1 (NKX2-1) → lung + brain + thyroid triad; CT like SFTPC/ABCA3 group

Why radiologists must recognize this group

• CT pattern + age strongly suggests genetic testing, not open lung biopsy
• Syndromic implications (brain/thyroid in TTF-1; transplant in SFTPB)

Not-infancy buckets (older kids)

• Post-infectious COB → mosaic + air-trapping ± bronchiectasis
• HP → centrilobular nodules + GGO (acute), fibrosis (chronic)
• Aspiration → dependent consolidation
• Systemic disease (EGPA / SLE / JDM / storage) → NSIP/OP/GGO/cysts variable
• Immunocompromised → LIP/follicular bronchiolitis nodules ± GGO
• Mimics (PVOD/PCH/lymphatic) → septal thickening + PH signs + effusions

How to read chILD CT like an expert

• Always anchor to age of onset
• Use secondary lobule map (centrilobular vs septal vs subpleural vs diffuse)
• Look for signature constellations (NEHI pattern / PAP crazy-paving / COB mosaicism)
• Decide if imaging is specific enough vs triggers biopsy / genetics
• Infant with GGO + cysts → surfactant genes until proven otherwise
• Middle-lobe/lingula GGO + air-trapping → NEHI until proven otherwise
• Mosaic + chronic cough post-viral → call COB, do not biopsy first
• Systemic disease on labs → assume lung involvement and map pattern, do not start from ILD list