Convolutional neural network assistance significantly improves dermatologists' diagnosis of cutaneous tumours using clinical images.

BACKGROUND: Convolutional neural networks (CNNs) have demonstrated expert-level performance in cutaneous tumour classification using clinical images, but most previous studies have focused on dermatologist-versus-CNN comparisons rather than their combination. The objective of our study was to evaluate the potential impact of CNN assistance on dermatologists for clinical image interpretation. METHODS: A multi-class CNN was trained and validated using a dataset of 25,773 clinical images comprising 10 categories of cutaneous tumours. The CNN's performance was tested on an independent dataset of 2107 images. A total of 400 images (40 per category) were randomly selected from the test dataset. A fully crossed, self-control, multi-reader multi-case (MRMC) study was conducted to compare the performance of 18 board-certified dermatologists (experience: 13/18 ≤ 10 years; 5/18＞10 years) in interpreting the 400 clinical images with or without CNN assistance. RESULTS: The CNN achieved an overall accuracy of 78.45% and kappa of 0.73 in the classification of 10 types of cutaneous tumours on 2107 images. CNN-assisted dermatologists achieved a higher accuracy (76.60% vs. 62.78%, P < 0.001) and kappa (0.74 vs. 0.59, P < 0.001) than unassisted dermatologists in interpreting the 400 clinical images. Dermatologists with less experience benefited more from CNN assistance. At the binary classification level (malignant or benign), the sensitivity (89.56% vs. 83.21%, P < 0.001) and specificity (87.90% vs. 80.92%, P < 0.001) of dermatologists with CNN assistance were also significantly improved than those without. CONCLUSIONS: CNN assistance improved dermatologist accuracy in interpreting cutaneous tumours and could further boost the acceptance of this new technique.

Dietary fat alters the response of hypothalamic neuropeptide Y to subsequent energy intake in broiler chickens.

Dietary fat affects appetite and appetite-related peptides in birds and mammals; however, the effect of dietary fat on appetite is still unclear in chickens faced with different energy statuses. Two experiments were conducted to investigate the effects of dietary fat on food intake and hypothalamic neuropeptides in chickens subjected to two feeding states or two diets. In Experiment 1, chickens were fed a high-fat (HF) or low-fat (LF) diet for 35 days, and then subjected to fed (HF-fed, LF-fed) or fasted (HF-fasted, LF-fasted) conditions for 24 h. In Experiment 2, chickens that were fed a HF or LF diet for 35 days were fasted for 24 h and then re-fed with HF (HF-RHF, LF-RHF) or LF (HF-RLF, LF-RLF) diet for 3 h. The results showed that chickens fed a HF diet for 35 days had increased body fat deposition despite decreasing food intake even when the diet was altered during the re-feeding period (P<0.05). LF diet (35 days) promoted agouti-related peptide (AgRP) expression compared with HF diet (P<0.05) under both fed and fasted conditions. LF-RHF chickens had lower neuropeptide Y (NPY) expression compared with LF-RLF chickens; conversely, HF-RHF chickens had higher NPY expression than HF-RLF chickens (P<0.05). These results demonstrate: (1) that HF diet decreases food intake even when the subsequent diet is altered; (2) the orexigenic effect of hypothalamic AgRP; and (3) that dietary fat alters the response of hypothalamic NPY to subsequent energy intake. These findings provide a novel view of the metabolic perturbations associated with long-term dietary fat over-ingestion in chickens.

Leucine alleviates dexamethasone-induced suppression of muscle protein synthesis via synergy involvement of mTOR and AMPK pathways.

Glucocorticoids (GCs) are negative muscle protein regulators that contribute to the whole-body catabolic state during stress. Mammalian target of rapamycin (mTOR)-signalling pathway, which acts as a central regulator of protein metabolism, can be activated by branched-chain amino acids (BCAA). In the present study, the effect of leucine on the suppression of protein synthesis induced by GCs and the pathway involved were investigated. In vitro experiments were conducted using cultured C2C12 myoblasts to study the effect of GCs on protein synthesis, and the involvement of mTOR pathway was investigated as well. After exposure to dexamethasone (DEX, 100 µmol/l) for 24 h, protein synthesis in muscle cells was significantly suppressed (P<0.05), the phosphorylations of mTOR, ribosomal protein S6 protein kinase 1 (p70s6k1) and eukaryotic initiation factor 4E binding protein 1 (4EBP1) were significantly reduced (P<0.05). Leucine supplementation (5 mmol/l, 10 mmol/l and 15 mmol/l) for 1 h alleviated the suppression of protein synthesis induced by DEX (P<0.05) and was accompanied with the increased phosphorylation of mTOR and decreased phosphorylation of AMPK (P<0.05). Branched-chain amino transferase 2 (BCAT2) mRNA level was not influenced by DEX (P>0.05) but was increased by leucine supplementation at a dose of 5 mmol/l (P<0.05).