This paper investigates the adaptability of four state-of-the-art Artificial Intelligence (AI) models to the Australian mammographic context through transfer learning, explores the impact of image enhancement on model performance and analyses the relationship between AI outputs and histopathological features for clinical relevance and accuracy assessment. A total of 1712 screening mammograms (n=856 cancer cases and n=856 matched normal cases) were used in this study. The 856 cases with cancer lesions were annotated by two expert radiologists and the level of concordance between their annotations was used to establish image subsets. The area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of Globally aware Multiple Instance Classifier (GMIC), Global-Local Activation Maps (GLAM), I&H and End2End AI models, both in the pre-trained and transfer learning modes, with and without applying the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm. The four AI models with and without transfer learning in the high-concordance subset outperformed those in the entire dataset. Applying the CLAHE algorithm to mammograms improved the performance of the AI models. In the high-concordance subset with transfer learning and CLAHE algorithm applied, the AUC of the GMIC model was highest (0.912), followed by GLAM model (0.909), I&H (0.893) and End2End (0.875). There were significant differences (P<0.05) in the performances of the four AI models between high-concordance subset and entire dataset. The AI models demonstrated significant differences in malignancy probability concerning different tumour size categories in mammograms. The performance of AI models was affected by several factors such as concordance classification, image enhancement and transfer learning. Mammograms with strong concordance of radiologists’ annotations, applying image enhancement and transfer learning could enhance the accuracy of AI models.

Background: Various vitamins and minerals have been implicated in the aetiology of depression. Objective: To estimate the effects of micronutrient exposures on major depressive disorder (MDD) and recurrent depression (rMDD) using Mendelian randomization (MR). Methods: We undertook a comprehensive bidirectional MR study of multiple micronutrient exposures on MDD and rMDD. Summary statistics were obtained from the Psychiatric Genomics Consortium (PGC) genome wide association studies (GWASs) of MDD (cases = 116,209; controls = 314,566), and rMDD (cases=17,451; controls=62,482). Results: None of the micronutrients with available genetic instruments were strongly associated with MDD or rMDD using traditional MR methods. However, using methods to increase analytical power by accounting for genetically correlated variants (e.g. cIVW), highlighted five micronutrients with possible causal effects. Point estimates for rMDD were the largest magnitude, with three micronutrients suggestive of a protective effect: serum iron (OR_cIVW 0.90 per SD increase; 95% CI 0.85- 0.95; p=0.0003); erythrocyte copper (OR_cIVW 0.97; 95% CI 0.95- 0.99; p=0.0004); and 25(OH) vitamin D (OR_cIVW 0.81; 0.66- 0.99; p=0.04). Apparent adverse effects of increased selenium on the risk of MDD (OR_cIVW 1.03; 95% CI 1.02-1.05; p=0.0003) and rMDD (OR_cIVW 1.08; 95% CI 1.00-1.08; p=0.06), and serum magnesium on rMDD (OR_cIVW 1.21; 1.01- 1.44; p=0.04); were less consistent between methods and may be driven by pleiotropy. Conclusions: Our results suggest weak evidence for a protective effect of iron, copper and 25(OH)D on major depressive outcomes, with mixed evidence for selenium and magnesium. There was no evidence to support a causal effect of any other micronutrients on MDD or rMDD, although genetic instruments were lacking, with insufficient power to detect small but important effects. Future micronutrient supplementation trials should ensure ample statistical power given modest causal effect estimates, and consider potential risks of supplementation, as some micronutrient effect estimates suggested potential harm in excess.