Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Radiology Radiology

Automatic fetal biometry prediction using a novel deep convolutional network architecture.

In Physica medica : PM : an international journal devoted to the applications of physics to medicine and biology : official journal of the Italian Association of Biomedical Physics (AIFB)

PURPOSE : Fetal biometric measurements face a number of challenges, including the presence of speckle, limited soft-tissue contrast and difficulties in the presence of low amniotic fluid. This work proposes a convolutional neural network for automatic segmentation and measurement of fetal biometric parameters, including biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL) from ultrasound images that relies on the attention gates incorporated into the multi-feature pyramid Unet (MFP-Unet) network.

METHODS : The proposed approach, referred to as Attention MFP-Unet, learns to extract/detect salient regions automatically to be treated as the object of interest via the attention gates. After determining the type of anatomical structure in the image using a convolutional neural network, Niblack's thresholding technique was applied as pre-processing algorithm for head and abdomen identification, whereas a novel algorithm was used for femur extraction. A publicly-available dataset (HC18 grand-challenge) and clinical data of 1334 subjects were utilized for training and evaluation of the Attention MFP-Unet algorithm.

RESULTS : Dice similarity coefficient (DSC), hausdorff distance (HD), percentage of good contours, the conformity coefficient, and average perpendicular distance (APD) were employed for quantitative evaluation of fetal anatomy segmentation. In addition, correlation analysis, good contours, and conformity were employed to evaluate the accuracy of the biometry predictions. Attention MFP-Unet achieved 0.98, 1.14 mm, 100%, 0.95, and 0.2 mm for DSC, HD, good contours, conformity, and APD, respectively.

CONCLUSIONS : Quantitative evaluation demonstrated the superior performance of the Attention MFP-Unet compared to state-of-the-art approaches commonly employed for automatic measurement of fetal biometric parameters.

Ghelich Oghli Mostafa, Shabanzadeh Ali, Moradi Shakiba, Sirjani Nasim, Gerami Reza, Ghaderi Payam, Sanei Taheri Morteza, Shiri Isaac, Arabi Hossein, Zaidi Habib


Convolutional neural network, Deep learning, Fetal biometry, Image classification, Ultrasound imaging

Pathology Pathology

Spatially Constrained Context-Aware Hierarchical Deep Correlation Filters for Nucleus Detection in Histology Images.

In Medical image analysis

Nucleus detection in histology images is a fundamental step for cellular-level analysis in computational pathology. In clinical practice, quantitative nuclear morphology can be used for diagnostic decision making, prognostic stratification, and treatment outcome prediction. Nucleus detection is a challenging task because of large variations in the shape of different types of nucleus such as nuclear clutter, heterogeneous chromatin distribution, and irregular and fuzzy boundaries. To address these challenges, we aim to accurately detect nuclei using spatially constrained context-aware correlation filters using hierarchical deep features extracted from multiple layers of a pre-trained network. During training, we extract contextual patches around each nucleus which are used as negative examples while the actual nucleus patch is used as a positive example. In order to spatially constrain the correlation filters, we propose to construct a spatial structural graph across different nucleus components encoding pairwise similarities. The correlation filters are constrained to act as eigenvectors of the Laplacian of the spatial graphs enforcing these to capture the nucleus structure. A novel objective function is proposed by embedding graph-based structural information as well as the contextual information within the discriminative correlation filter framework. The learned filters are constrained to be orthogonal to both the contextual patches and the spatial graph-Laplacian basis to improve the localization and discriminative performance. The proposed objective function trains a hierarchy of correlation filters on different deep feature layers to capture the heterogeneity in nuclear shape and texture. The proposed algorithm is evaluated on three publicly available datasets and compared with 15 current state-of-the-art methods demonstrating competitive performance in terms of accuracy, speed, and generalization.

Javed Sajid, Mahmood Arif, Dias Jorge, Werghi Naoufel, Rajpoot Nasir


Computational pathology, Correlation filters, Deep learning, Nucleus detection

General General

Application of computer tongue image analysis technology in the diagnosis of NAFLD.

In Computers in biology and medicine

Nonalcoholic fatty liver disease (NAFLD), a leading cause of chronic hepatic disease, can progress to liver fibrosis, cirrhosis, and hepatocellular carcinoma. Therefore, it is extremely important to explore early diagnosis and screening methods. In this study, we developed models based on computer tongue image analysis technology to observe the tongue characteristics of 1778 participants (831 cases of NAFLD and 947 cases of non-NAFLD). Combining quantitative tongue image features, basic information, and serological indexes, including the hepatic steatosis index (HSI) and fatty liver index (FLI), we utilized machine learning methods, including Logistic Regression, Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Adaptive Boosting Algorithm (AdaBoost), Naïve Bayes, and Neural Network for NAFLD diagnosis. The best fusion model for diagnosing NAFLD by Logistic Regression, which contained the tongue image parameters, waist circumference, BMI, GGT, TG, and ALT/AST, achieved an AUC of 0.897 (95% CI, 0.882-0.911), an accuracy of 81.70% with a sensitivity of 77.62% and a specificity of 85.22%; in addition, the positive likelihood ratio and negative likelihood ratio were 5.25 and 0.26, respectively. The application of computer intelligent tongue diagnosis technology can improve the accuracy of NAFLD diagnosis and may provide a convenient technical reference for the establishment of early screening methods for NAFLD, which is worth further research and verification.

Jiang Tao, Guo Xiao-Jing, Tu Li-Ping, Lu Zhou, Cui Ji, Ma Xu-Xiang, Hu Xiao-Juan, Yao Xing-Hua, Cui Long-Tao, Li Yong-Zhi, Huang Jing-Bin, Xu Jia-Tuo


Computer intelligent technology, Deep learning, Diagnostic model, NAFLD, Tongue image

oncology Oncology

CT synthesis from MRI using multi-cycle GAN for head-and-neck radiation therapy.

In Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society

Magnetic Resonance Imaging (MRI) guided Radiation Therapy is a hot topic in the current studies of radiotherapy planning, which requires using MRI to generate synthetic Computed Tomography (sCT). Despite recent progress in image-to-image translation, it remains challenging to apply such techniques to generate high-quality medical images. This paper proposes a novel framework named Multi-Cycle GAN, which uses the Pseudo-Cycle Consistent module to control the consistency of generation and the domain control module to provide additional identical constraints. Besides, we design a new generator named Z-Net to improve the accuracy of anatomy details. Extensive experiments show that Multi-Cycle GAN outperforms state-of-the-art CT synthesis methods such as Cycle GAN, which improves MAE to 0.0416, ME to 0.0340, PSNR to 39.1053.

Liu Yanxia, Chen Anni, Shi Hongyu, Huang Sijuan, Zheng Wanjia, Liu Zhiqiang, Zhang Qin, Yang Xin


CT synthesis, Cycle GAN, Deep learning, MRI

General General

The protein inputs of an ultra-predictive aging clock represent viable anti-aging drug targets.

In Ageing research reviews ; h5-index 66.0

Machine learning models capable of predicting age given a set of inputs are referred to as aging clocks. We recently developed an aging clock that utilizes 491 plasma protein inputs, has an exceptional accuracy, and is capable of measuring biological age. Here, we demonstrate that this clock is extremely predictive (r = 0.95) when used to measure age in a novel plasma proteomic dataset derived from 370 human subjects aged 18-69 years. Over-representation analyses of the proteins that make up this clock in the Gene Ontology and Reactome databases predominantly implicated innate and adaptive immune system processes. Immunological drugs and various age-related diseases were enriched in the DrugBank and GLAD4U databases. By performing an extensive literature review, we find that at least 269 (54.8 %) of these inputs regulate lifespan and/or induce changes relevant to age-related disease when manipulated in an animal model. We also show that, in a large plasma proteomic dataset, the majority (57.2 %) of measurable clock proteins significantly change their expression level with human age. Different subsets of proteins were overlapped with distinct epigenetic, transcriptomic, and proteomic aging clocks. These findings indicate that the inputs of this age predictor likely represent a rich source of anti-aging drug targets.

Johnson Adiv A, Shokhirev Maxim N, Lehallier Benoit


Age prediction, Aging clock, Bioinformatics, Biomarker, Healthspan, Machine learning

Radiology Radiology

Application of a deep learning image reconstruction (DLIR) algorithm in head CT imaging for children to improve image quality and lesion detection.

In BMC medical imaging

BACKGROUND : To evaluate the performance of a Deep Learning Image Reconstruction (DLIR) algorithm in pediatric head CT for improving image quality and lesion detection with 0.625 mm thin-slice images.

METHODS : Low-dose axial head CT scans of 50 children with 120 kV, 0.8 s rotation and age-dependent 150-220 mA tube current were selected. Images were reconstructed at 5 mm and 0.625 mm slice thickness using Filtered back projection (FBP), Adaptive statistical iterative reconstruction-v at 50% strength (50%ASIR-V) (as reference standard), 100%ASIR-V and DLIR-high (DL-H). The CT attenuation and standard deviation values of the gray and white matters in the basal ganglia were measured. The clarity of sulci/cisterns, boundary between white and gray matters, and overall image quality was subjectively evaluated. The number of lesions in each reconstruction group was counted.

RESULTS : The 5 mm FBP, 50%ASIR-V, 100%ASIR-V and DL-H images had a subjective score of 2.25 ± 0.44, 3.05 ± 0.23, 2.87 ± 0.39 and 3.64 ± 0.49 in a 5-point scale, respectively with DL-H having the lowest image noise of white matter at 2.00 ± 0.34 HU; For the 0.625 mm images, only DL-H images met the diagnostic requirement. The 0.625 mm DL-H images had similar image noise (3.11 ± 0.58 HU) of the white matter and overall image quality score (3.04 ± 0.33) as the 5 mm 50% ASIR-V images (3.16 ± 0.60 HU and 3.05 ± 0.23). Sixty-five lesions were recognized in 5 mm 50%ASIR-V images and 69 were detected in 0.625 mm DL-H images.

CONCLUSION : DL-H improves the head CT image quality for children compared with ASIR-V images. The 0.625 mm DL-H images improve lesion detection and produce similar image noise as the 5 mm 50%ASIR-V images, indicating a potential 85% dose reduction if current image quality and slice thickness are desired.

Sun Jihang, Li Haoyan, Wang Bei, Li Jianying, Li Michelle, Zhou Zuofu, Peng Yun


CT, Children, Deep learning, Head, IR, Low-dose