Gradual Self-Training via Confidence and Volume Based Domain Adaptation for Multi Dataset Deep Learning-Based Brain Metastases Detection Using Nonlocal Networks on MRI Images.

BACKGROUND: Research suggests that treatment of multiple brain metastases (BMs) with stereotactic radiosurgery shows improvement when metastases are detected early, providing a case for BM detection capabilities on small lesions. PURPOSE: To demonstrate automatic detection of BM on three MRI datasets using a deep learning-based approach. To improve the performance of the network is iteratively co-trained with datasets from different domains. A systematic approach is proposed to prevent catastrophic forgetting during co-training. STUDY TYPE: Retrospective. POPULATION: A total of 156 patients (105 ground truth and 51 pseudo labels) with 1502 BM (BrainMetShare); 121 patients with 722 BM (local); 400 patients with 447 primary gliomas (BrATS). Training/pseudo labels/validation data were distributed 84/51/21 (BrainMetShare). Training/validation data were split: 121/23 (local) and 375/25 (BrATS). FIELD STRENGTH/SEQUENCE: A 5 T and 3 T/T1 spin-echo postcontrast (T1-gradient echo) (BrainMetShare), 3 T/T1 magnetization prepared rapid acquisition gradient echo postcontrast (T1-MPRAGE) (local), 0.5 T, 1 T, and 1.16 T/T1-weighted-fluid-attenuated inversion recovery (T1-FLAIR) (BrATS). ASSESSMENT: The ground truth was manually segmented by two (BrainMetShare) and four (BrATS) radiologists and manually annotated by one (local) radiologist. Confidence and volume based domain adaptation (CAVEAT) method of co-training the three datasets on a 3D nonlocal convolutional neural network (CNN) architecture was implemented to detect BM. STATISTICAL TESTS: The performance was evaluated using sensitivity and false positive rates per patient (FP/patient) and free receiver operating characteristic (FROC) analysis at seven predefined (1/8, 1/4, 1/2, 1, 2, 4, and 8) FPs per scan. RESULTS: The sensitivity and FP/patient from a held-out set registered 0.811 at 2.952 FP/patient (BrainMetShare), 0.74 at 3.130 (local), and 0.723 at 2.240 (BrATS) using the CAVEAT approach with lesions as small as 1 mm being detected. DATA CONCLUSION: Improved sensitivities at lower FP can be achieved by co-training datasets via the CAVEAT paradigm to address the problem of data sparsity. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY STAGE: 2.

Not out of the woods yet: Signatures of the prolonged negative genetic consequences of a population bottleneck in a rapidly re-expanding wader, the black-faced spoonbill Platalea minor.

The long-term persistence of a population which has suffered a bottleneck partly depends on how historical demographic dynamics impacted its genetic diversity and the accumulation of deleterious mutations. Here we provide genomic evidence for the genetic effect of a recent population bottleneck in the endangered black-faced spoonbill (Platalea minor) after its rapid population recovery. Our data suggest that the bird's effective population size, Ne , had been relatively stable (7500-9000) since 22,000 years ago; however, a recent brief yet severe bottleneck (Ne  = 20) which we here estimated to occur around the 1940s wiped out >99% of its historical Ne in roughly three generations. Despite a >15-fold population recovery since 1988, we found that black-faced spoonbill population has higher levels of inbreeding (7.4 times more runs of homozygosity) than its sister species, the royal spoonbill (P. regia), which is not thought to have undergone a marked population contraction. Although the two spoonbills have similar levels of genome-wide genetic diversity, our results suggest that selection on more genes was relaxed in the black-faced spoonbill; moreover individual black-faced spoonbills carry more putatively deleterious mutations (Grantham's score > 50), and may therefore express more deleterious phenotypic effects than royal spoonbills. Here we demonstrate the value of using genomic indices to monitor levels of genetic erosion, inbreeding and mutation load in species with conservation concerns. To mitigate the prolonged negative genetic effect of a population bottleneck, we recommend that all possible measures should be employed to maintain population growth of a threatened species.

CASPIANET++: A multidimensional Channel-Spatial Asymmetric attention network with Noisy Student Curriculum Learning paradigm for brain tumor segmentation.

Convolutional neural networks (CNNs) have been used quite successfully for semantic segmentation of brain tumors. However, current CNNs and attention mechanisms are stochastic in nature and neglect the morphological indicators used by radiologists to manually annotate regions of interest. In this paper, we introduce a channel and spatial wise asymmetric attention (CASPIAN) by leveraging the inherent structure of tumors to detect regions of saliency. To demonstrate the efficacy of our proposed layer, we integrate this into a well-established convolutional neural network (CNN) architecture to achieve higher Dice scores, with less GPU resources. Also, we investigate the inclusion of auxiliary multiscale and multiplanar attention branches to increase the spatial context crucial in semantic segmentation tasks. The resulting architecture is the new CASPIANET++, which achieves Dice Scores of 91.19%, 87.6% and 81.03% for whole tumor, tumor core and enhancing tumor respectively. Furthermore, driven by the scarcity of brain tumor data, we investigate the Noisy Student method for segmentation tasks. Our new Noisy Student Curriculum Learning paradigm, which infuses noise incrementally to increase the complexity of the training images exposed to the network, further boosts the enhancing tumor region to 81.53%. Additional validation performed on the BraTS2020 data shows that the Noisy Student Curriculum Learning method works well without any additional training or finetuning.

Morphological variations in a widely distributed Eastern Asian passerine cannot be consistently explained by ecogeographic rules.

Ecogeographic rules that describe quantitative relationships between morphologies and climate might help us predict how morphometrics of animals was shaped by local temperature or humidity. Although the ecogeographic rules had been widely tested in animals of Europe and North America, they had not been fully validated for species in regions that are less studied. Here, we investigate the morphometric variation of a widely distributed East Asian passerine, the vinous-throated parrotbill (Sinosuthora webbiana), to test whether its morphological variation conforms to the prediction of Bergmann's rule, Allen's rules, and Gloger's rule. We at first described the climatic niche of S. webbiana from occurrence records (n = 7838) and specimen records (n = 290). The results of analysis of covariance (ANCOVA) suggested that the plumage coloration of these parrotbills was darker in wetter/warmer environments following Gloger's rule. However, their appendage size (culmen length, beak volume, tarsi length) was larger in colder environments, the opposite of the predictions of Allen's rule. Similarly, their body size (wing length) was larger in warmer environments, the opposite of the predictions of Bergmann's rule. Such disconformity to both Bergmann's rule and Allen's rule suggests that the evolution of morphological variations is likely governed by multiple selection forces rather than dominated by thermoregulation. Our results suggest that these ecogeographic rules should be validated prior to forecasting biological responses to climate change especially for species in less-studied regions.

Evaluation of the poly(ADP-ribose) polymerase-1 gene variants in Alzheimer's disease.

Amyloid peptide is thought to play a critical role in neuronal death in Alzheimer's disease (AD), most likely through oxidative stress. Free radical-related injury leads to DNA breaks, which subsequently activates the repair enzyme poly(ADP-ribose) polymerase-1 (PARP-1). In this study, the relationship between genetic variants situated at the PARP-1 gene and AD development was investigated. We performed a case and control study from a Taiwanese population enrolled 120 AD patients and 111 healthy controls by using a polymerase chain reaction restriction fragment length polymorphism approach for two PARP-1 exonic polymorphisms, 414C/T (rs1805404) and 2456T/C (rs1136410), corresponding to protein residues at positions 81Asp/Asp and 762Val/Ala. There were no significant differences in allele or genotype frequencies for either PARP-1 gene variant between the case and control groups; however, upon analysis of the haplotype distribution, four haplotypes (Hts) were identified. We found that the distributions of Ht3-TT and Ht4-CC were significantly associated with an increased risk of AD (P<0.0001), whereas the Ht1-TC haplotype showed a protective effect for cases compared with the control group (P<0.05). These results reveal that the PARP-1 gene is highly associated with AD susceptibility and might contribute to a critical mechanism that mediates cell survival or death as a response to cytotoxic stress.