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

General General

A new data augmentation method based on local image warping for medical image segmentation.

In Medical physics ; h5-index 59.0

PURPOSE : The segmentation accuracy of medical images was improved by increasing the number of training samples using a local image warping technique. The performance of the proposed method was evaluated in the segmentation of breast masses, prostate and brain tumors, and lung nodules.

METHODS : We propose a simple data augmentation method which is called stochastic evolution (SE). Specifically, the idea of stochastic evolution stems from our thinking about the deterioration of the diseased tissue and the healing process. In order to simulate this natural process, we implement it according to the local distortion algorithm in image warping. In other words, the irregular deterioration and healing processes of the diseased tissue is simulated according to the direction of the local distortion, thereby producing a natural sample that is indistinguishable by humans.

RESULTS : The proposed method is evaluated on four segmentation tasks of breast masses, prostate, brain tumors and lung nodules. Comparing the experimental results of four segmentation methods based on the UNet segmentation architecture without adding any expanded data during training, the accuracy and the Hausdorff distance obtained in our approach remain almost the same as other methods. However, the dice similarity coefficient (DSC) and sensitivity (SEN) have both improved to some extent. Among them, DSC is increased by 5.2%, 2.8%, 1.0% and 3.2%, respectively; SEN is increased by 6.9%, 4.3%, 1.2% and 4.5%, respectively.

CONCLUSIONS : Experimental results show that the proposed SE data augmentation method could improve the segmentation accuracy of breast masses, prostate, brain tumors and lung nodules. The method also shows the robustness with different image datasets and imaging modalities.

Liu Hong, Cao Haichao, Song Enmin, Ma Guangzhi, Xu Xiangyang, Jin Renchao, Liu Tengying, Liu Lei, Liu Daiyang, Hung Chih-Cheng


Data augmentation, Deep learning, Image warping, Medical image segmentation

Radiology Radiology

Deep convolutional neural networks for multi-planar lung nodule detection: improvement in small nodule identification.

In Medical physics ; h5-index 59.0

PURPOSE : Early detection of lung cancer is of importance since it can increase patients' chances of survival. To detect nodules accurately during screening, radiologists would commonly take the axial, coronal and sagittal planes into account, rather than solely the axial plane in clinical evaluation. Inspired by clinical work, the paper aims to develop an accurate deep learning framework for nodule detection by a combination of multiple planes.

METHODS : The nodule detection system is designed in two stages, multi-planar nodule candidate detection, multi-scale false positive reduction. At the first stage, a deeply-supervised encoder-decoder network is trained by axial, coronal, and sagittal slices for the candidate detection task. All possible nodule candidates from the three different planes are merged. To further refine results, a 3-D multi-scale dense convolutional neural network that extracts multi-scale contextual information is applied to remove non-nodules. In the public LIDC-IDRI dataset, 888 CT scans with 1186 nodules accepted by at least three out of four radiologists are selected to train and evaluate our proposed system via a ten-fold cross-validation scheme. The free-response receiver operating characteristic curve is used for performance assessment.

RESULTS : The proposed system achieves a sensitivity of 94.2% with 1.0 false positive/scan and a sensitivity of 96.0% with 2.0 false positives/scan. Although it is difficult to detect small nodules (i.e. < 6 mm), our designed CAD system reaches a sensitivity of 93.4% (95.0%) of these small nodules at an overall false positive rate of 1.0 (2.0) false positives/scan. At the nodule candidate detection stage, results show that the system with a multi-planar method is capable to detect more nodules compared to using a single plane.

CONCLUSION : Our approach achieves good performance not only for small nodules, but also for large lesions on this dataset. This demonstrates the effectiveness of our developed CAD system for lung nodule detection.

Zheng Sunyi, Cornelissen Ludo J, Cui Xiaonan, Jing Xueping, Veldhuis Raymond N J, Oudkerk Matthijs, van Ooijen Peter M A


Computer-aided detection, computed tomography, convolutional neural network, deep learning, pulmonary nodule detection

Public Health Public Health

Predicting regulatory variants using a dense epigenomic mapped CNN model elucidated the molecular basis of trait-tissue associations.

In Nucleic acids research ; h5-index 217.0

Assessing the causal tissues of human complex diseases is important for the prioritization of trait-associated genetic variants. Yet, the biological underpinnings of trait-associated variants are extremely difficult to infer due to statistical noise in genome-wide association studies (GWAS), and because >90% of genetic variants from GWAS are located in non-coding regions. Here, we collected the largest human epigenomic map from ENCODE and Roadmap consortia and implemented a deep-learning-based convolutional neural network (CNN) model to predict the regulatory roles of genetic variants across a comprehensive list of epigenomic modifications. Our model, called DeepFun, was built on DNA accessibility maps, histone modification marks, and transcription factors. DeepFun can systematically assess the impact of non-coding variants in the most functional elements with tissue or cell-type specificity, even for rare variants or de novo mutations. By applying this model, we prioritized trait-associated loci for 51 publicly-available GWAS studies. We demonstrated that CNN-based analyses on dense and high-resolution epigenomic annotations can refine important GWAS associations in order to identify regulatory loci from background signals, which yield novel insights for better understanding the molecular basis of human complex disease. We anticipate our approaches will become routine in GWAS downstream analysis and non-coding variant evaluation.

Pei Guangsheng, Hu Ruifeng, Dai Yulin, Manuel Astrid Marilyn, Zhao Zhongming, Jia Peilin


General General

A Process Evaluation Examining the Performance, Adherence, and Acceptability of a Physical Activity and Diet Artificial Intelligence Virtual Health Assistant.

In International journal of environmental research and public health ; h5-index 73.0

Artificial intelligence virtual health assistants are a promising emerging technology. This study is a process evaluation of a 12-week pilot physical activity and diet program delivered by virtual assistant "Paola". This single-arm repeated measures study (n = 28, aged 45-75 years) was evaluated on technical performance (accuracy of conversational exchanges), engagement (number of weekly check-ins completed), adherence (percentage of step goal and recommended food servings), and user feedback. Paola correctly asked scripted questions and responded to participants during the check-ins 97% and 96% of the time, respectively, but correctly responded to spontaneous exchanges only 21% of the time. Participants completed 63% of weekly check-ins and conducted a total of 3648 exchanges. Mean dietary adherence was 91% and was lowest for discretionary foods, grains, red meat, and vegetables. Participants met their step goal 59% of the time. Participants enjoyed the program and found Paola useful during check-ins but not for spontaneous exchanges. More in-depth knowledge, personalized advice and spontaneity were identified as important improvements. Virtual health assistants should ensure an adequate knowledge base and ability to recognize intents and entities, include personality and spontaneity, and provide ongoing technical troubleshooting of the virtual assistant to ensure the assistant remains effective.

Davis Courtney R, Murphy Karen J, Curtis Rachel G, Maher Carol A


Mediterranean diet, chatbot, conversational agent, intervention, lifestyle, physical activity, process evaluation, virtual health assistant

General General

Analyses of Risk, Racial Disparity, and Outcomes Among US Patients With Cancer and COVID-19 Infection.

In JAMA oncology ; h5-index 85.0

Importance : Patients with specific cancers may be at higher risk than those without cancer for coronavirus disease 2019 (COVID-19) and its severe outcomes. At present, limited data are available on the risk, racial disparity, and outcomes for COVID-19 illness in patients with cancer.

Objectives : To investigate how patients with specific types of cancer are at risk for COVID-19 infection and its adverse outcomes and whether there are cancer-specific race disparities for COVID-19 infection.

Design, Setting, and Participants : This retrospective case-control analysis of patient electronic health records included 73.4 million patients from 360 hospitals and 317 000 clinicians across 50 US states to August 14, 2020. The odds of COVID-19 infections for 13 common cancer types and adverse outcomes were assessed.

Exposures : The exposure groups were patients diagnosed with a specific cancer, whereas the unexposed groups were patients without the specific cancer.

Main Outcomes and Measures : The adjusted odds ratio (aOR) and 95% CI were estimated using the Cochran-Mantel-Haenszel test for the risk of COVID-19 infection.

Results : Among the 73.4 million patients included in the analysis (53.6% female), 2 523 920 had at least 1 of the 13 common cancers diagnosed (all cancer diagnosed within or before the last year), and 273 140 had recent cancer (cancer diagnosed within the last year). Among 16 570 patients diagnosed with COVID-19, 1200 had a cancer diagnosis and 690 had a recent cancer diagnosis of at least 1 of the 13 common cancers. Those with recent cancer diagnosis were at significantly increased risk for COVID-19 infection (aOR, 7.14 [95% CI, 6.91-7.39]; P < .001), with the strongest association for recently diagnosed leukemia (aOR, 12.16 [95% CI, 11.03-13.40]; P < .001), non-Hodgkin lymphoma (aOR, 8.54 [95% CI, 7.80-9.36]; P < .001), and lung cancer (aOR, 7.66 [95% CI, 7.07-8.29]; P < .001) and weakest for thyroid cancer (aOR, 3.10 [95% CI, 2.47-3.87]; P < .001). Among patients with recent cancer diagnosis, African Americans had a significantly higher risk for COVID-19 infection than White patients; this racial disparity was largest for breast cancer (aOR, 5.44 [95% CI, 4.69-6.31]; P < .001), followed by prostate cancer (aOR, 5.10 [95% CI, 4.34-5.98]; P < .001), colorectal cancer (aOR, 3.30 [95% CI, 2.55-4.26]; P < .001), and lung cancer (aOR, 2.53 [95% CI, 2.10-3.06]; P < .001). Patients with cancer and COVID-19 had significantly worse outcomes (hospitalization, 47.46%; death, 14.93%) than patients with COVID-19 without cancer (hospitalization, 24.26%; death, 5.26%) (P < .001) and patients with cancer without COVID-19 (hospitalization, 12.39%; death, 4.03%) (P < .001).

Conclusions and Relevance : In this case-control study, patients with cancer were at significantly increased risk for COVID-19 infection and worse outcomes, which was further exacerbated among African Americans. These findings highlight the need to protect and monitor patients with cancer as part of the strategy to control the pandemic.

Wang QuanQiu, Berger Nathan A, Xu Rong


General General

Machine learning models for synthesizing actionable care decisions on lower extremity wounds.

In Smart health (Amsterdam, Netherlands)

Lower extremity chronic wounds affect 4.5 million Americans annually. Due to inadequate access to wound experts in underserved areas, many patients receive non-uniform, non-standard wound care, resulting in increased costs and lower quality of life. We explored machine learning classifiers to generate actionable wound care decisions about four chronic wound types (diabetic foot, pressure, venous, and arterial ulcers). These decisions (target classes) were: (1) Continue current treatment, (2) Request non-urgent change in treatment from a wound specialist, (3) Refer patient to a wound specialist. We compare classification methods (single classifiers, bagged & boosted ensembles, and a deep learning network) to investigate (1) whether visual wound features are sufficient for generating a decision and (2) whether adding unstructured text from wound experts increases classifier accuracy. Using 205 wound images, the Gradient Boosted Machine (XGBoost) outperformed other methods when using both visual and textual wound features, achieving 81% accuracy. Using only visual features decreased the accuracy to 76%, achieved by a Support Vector Machine classifier. We conclude that machine learning classifiers can generate accurate wound care decisions on lower extremity chronic wounds, an important step toward objective, standardized wound care. Higher decision-making accuracy was achieved by leveraging clinical comments from wound experts.

Nguyen Holly, Agu Emmanuel, Tulu Bengisu, Strong Diane, Mombini Haadi, Pedersen Peder, Lindsay Clifford, Dunn Raymond, Loretz Lorraine


Chronic wounds, Classification, Lower extremity ulcers, Machine learning