Background/Objectives: Breast cancer diagnosis based on histopathological examination remains the most reliable and widely accepted approach in clinical practice, despite being time-consuming and prone to inter-observer variability. While deep learning methods have achieved high accuracy in medical image classification, their cross-site generalization remains limited due to differences in staining protocols and image acquisition. This study aims to evaluate and compare three clinically relevant adaptation strategies to improve model robustness under domain shift. Methods: The ResNet50V2 model, pretrained on ImageNet and further fine-tuned on the Kaggle Breast Histopathology Images dataset, was subsequently adapted to the BreaKHis dataset under three clinically relevant transfer strategies: (i) threshold calibration without retraining (site calibration), (ii) head-only fine-tuning (light FT), and (iii) full fine-tuning (full FT). Experiments were performed on an internal balanced dataset and on the public BreaKHis dataset using strict patient-level splitting to avoid data leakage. Evaluation metrics included accuracy, precision, recall, F1-score, ROC-AUC, and PR-AUC, computed per magnification level (40×, 100×, 200×, 400×). Results: Full fine-tuning consistently yielded the highest performance across all magnifications, reaching up to 0.983 ROC-AUC and 0.980 sensitivity at 400×. At 40× and 100×, the model correctly identified over 90% of malignant cases, with ROC-AUC values of 0.9500 and 0.9332, respectively. Head-only fine-tuning led to moderate gains (e.g., sensitivity up to 0.859 at 200×), while threshold calibration showed limited improvements (ROC-AUC ranging between 0.60–0.73). Grad-CAM analysis revealed more stable and focused attention maps after full fine-tuning, though they did not always align with diagnostically relevant regions. Conclusions: Our findings confirm that full fine-tuning is essential for robust cross-site deployment of histopathology AI systems, particularly at high magnifications. Lighter strategies such as threshold calibration or head-only fine-tuning may serve as practical alternatives in resource-constrained environments where retraining is not feasible.

Background/Objectives: Accurate nuclear segmentation is a fundamental step in computational pathology, enabling reliable estimation of cellularity and nuclear morphology. However, segmentation models are typically evaluated under ideal imaging conditions, while real-world microscopy data are affected by staining variability, noise, and image degradation. This study aims to comparatively evaluate three representative convolutional architectures for nuclei segmentation, with emphasis on robustness and clinical relevance under perturbed imaging conditions. Methods: U-Net, Attention U-Net, and U-Net++ were trained and evaluated on the BBBC038 nuclei microscopy dataset using fixed train–validation–test splits. Robustness was assessed under three types of synthetic perturbations: Gaussian blur, additive noise, and color jitter. Segmentation performance was quantified using the Dice coefficient and Intersection-over-Union (IoU). Paired Wilcoxon signed-rank tests with Holm correction and Cliff’s delta were used for statistical comparison. In addition, clinically relevant nuclear descriptors—nuclear count, median nuclear area, area interquartile range (IQR), and nuclear density—were extracted from predicted masks, and descriptor stability was analyzed as relative deviation from clean conditions. Results: Under clean imaging conditions, Attention U-Net achieved the highest mean Dice score, while paired statistical analysis indicated that U-Net++ exhibited the most consistent performance across test samples. Under image perturbations, Attention U-Net demonstrated greater robustness to blur and noise, whereas U-Net++ showed superior stability under color variations. Descriptor-based analysis further indicated that U-Net++ preserved nuclear count and density most reliably under chromatic perturbations, while U-Net exhibited larger instability in nuclear count and density, particularly under noise. Conclusions: Architectural design choices strongly influence not only pixel-level segmentation accuracy but also the stability of clinically relevant nuclear morphology descriptors. Robustness evaluation under multiple perturbation types reveals important trade-offs between architectures that are not captured by clean-image benchmarks alone. These findings highlight the necessity of multi-level evaluation strategies combining overlap metrics, statistical testing, robustness analysis, and descriptor stability assessment for future benchmarking and clinically reliable deployment of nuclei segmentation systems.

Background/Objectives: Deep learning approaches have demonstrated strong performance in breast histopathology image classification; however, reliable generalization across heterogeneous acquisition environments remains challenging due to domain shift. In clinical practice, missed malignant cases are particularly critical because they may directly affect diagnostic decisions and patient outcomes. This study systematically investigates the behavior of modern deep learning architectures and adaptation strategies under realistic cross-domain conditions, with particular emphasis on malignant case detection and false-negative reduction. Methods: Three modern architectures—ConvNeXt-Tiny, Swin-Tiny, and MaxViT-Tiny—were initially trained on a large-scale breast histopathology dataset and subsequently evaluated on the BreaKHis dataset using strict patient-level separation to avoid information leakage. Three transfer settings were investigated: direct zero-shot transfer, head-only adaptation, and full fine-tuning. Performance was evaluated independently across four magnification levels (40×, 100×, 200×, and 400×) using accuracy, precision, sensitivity, F1-score, ROC–AUC, PR–AUC, and false-negative rates. Results: Direct zero-shot transfer produced substantial performance degradation across all architectures, with mean false-negative rates ranging from 75.85% to 90.11%, highlighting the limited transferability of source-domain representations under heterogeneous acquisition conditions. Both adaptation strategies substantially improved performance and reduced missed malignant cases to below 10%. Swin-Tiny under head-only adaptation achieved the most favorable malignant detection profile, reaching a mean sensitivity of 97.36% while reducing the average false-negative rate to 2.64%. In contrast, MaxViT-Tiny achieved the highest mean ROC–AUC value (0.849) after full fine-tuning, although this did not correspond to the lowest false-negative burden. Conclusions: The findings demonstrate that maximizing global discrimination performance does not necessarily correspond to optimal malignant detection under cross-domain conditions. Sensitivity and missed-case analysis provide complementary information beyond conventional discrimination metrics and may support more informed model assessment. Furthermore, the proposed methodology provides a reproducible framework for investigating adaptation performance in AI-assisted breast histopathology systems.