Upstream Probabilistic Meta-Imputation for Multimodal Pediatric Pancreatitis Classification

Nelson M. A., Keles E., Tasci E. S., YAZOL M., Aktas H. E., Hong Z., ...Daha Fazla

23rd IEEE International Symposium on Biomedical Imaging, ISBI 2026, London, İngiltere, 8 - 11 Nisan 2026, cilt.2026-April, (Tam Metin Bildiri) identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 2026-April
  • Doi Numarası: 10.1109/isbi61048.2026.11515465
  • Basıldığı Şehir: London
  • Basıldığı Ülke: İngiltere
  • Anahtar Kelimeler: modality fusion, Pediatric pancreatitis, small sample learning, stacking, synthetic data augmentation
  • Gazi Üniversitesi Adresli: Evet

Özet

Pediatric pancreatitis is a progressive and debilitating inflammatory condition, including acute pancreatitis and chronic pancreatitis, that presents significant clinical diagnostic challenges. Machine learning-based methods also face diagnostic challenges due to limited sample availability and multimodal imaging complexity. To address these challenges, this paper proposes Upstream Probabilistic Meta-Imputation (UPMI), a light-weight augmentation strategy that operates upstream of a stacking classifier in a low-dimensional meta-feature space rather than in image space. Modality-specific logistic regressions (T1W and T2W MRI radiomics) produce probability outputs that are transformed into a 7-dimensional meta-feature vector. Class-conditional Gaussian mixture models (GMMs) are then fit within each cross-validation fold to sample synthetic meta-features that, combined with real meta-features, train a Random Forest (RF) meta-classifier. On 67 pediatric subjects with paired T1W/T2W MRIs, UPMI achieves a mean AUC of 0.908 ± 0.072, a ∼ 5% relative gain over a real-only baseline (AUC 0.864 ± 0.061).