Exponentially weighted ensemble deep learning for histologic growth pattern classification in lung adenocarcinoma

Accurate classification of lung adenocarcinoma growth patterns from hematoxylin and eosin (H&E)-stained whole-slide images is essential for treatment planning and prognosis. We develop and evaluate an exponentially weighted ensemble deep learning approach that combines five architectures (EfficientNet-B3, DeiT3, Swin Transformer, Vision Transformer, ConvNeXt) with weights determined by a softmax transformation of per-model Quadratic Weighted Kappa (QWK) scores. Using 25,545 image patches at 40× magnification from the Dartmouth-Hitchcock dataset, we evaluated performance with stratified 5-fold cross-validation. Individual model performance ranged from Cohen’s kappa κ = 0.5563 (Con-vNeXt) to κ = 0.9561 (EfficientNet-B3). The ensemble achieved 96.52% accuracy and κ = 0.9648, exceeding the best individual model by Δ κ = +0.0087 with statistical significance (McNemar’s test, χ ² = 50.84, p < 10 ^{− 12} ; bootstrap 95% CI on Δ κ [+0.0056, +0.0117]). Pairwise disagreement rates of 16.6–43.2% between architectures indicate that the models capture different aspects of tissue morphology, and the framework automatically down-weights unreliable components, as illustrated by ConvNeXt receiving only 5.7% ensemble weight despite a single-fold training failure. The approach offers a clinically grounded method for ensemble aggregation in digital pathology.