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

General General

Probing the clinical and brain structural boundaries of bipolar and major depressive disorder.

In Translational psychiatry ; h5-index 60.0

Bipolar disorder (BD) and major depressive disorder (MDD) have both common and distinct clinical features, that pose both conceptual challenges in terms of their diagnostic boundaries and practical difficulties in optimizing treatment. Multivariate machine learning techniques offer new avenues for exploring these boundaries based on clinical neuroanatomical features. Brain structural data were obtained at 3 T from a sample of 90 patients with BD, 189 patients with MDD, and 162 healthy individuals. We applied sparse partial least squares discriminant analysis (s-PLS-DA) to identify clinical and brain structural features that may discriminate between the two clinical groups, and heterogeneity through discriminative analysis (HYDRA) to detect patient subgroups with reference to healthy individuals. Two clinical dimensions differentiated BD from MDD (area under the curve: 0.76, P < 0.001); one dimension emphasized disease severity as well as irritability, agitation, anxiety and flight of ideas and the other emphasized mostly elevated mood. Brain structural features could not distinguish between the two disorders. HYDRA classified patients in two clusters that differed in global and regional cortical thickness, the distribution proportion of BD and MDD and positive family history of psychiatric disorders. Clinical features remain the most reliable discriminant attributed of BD and MDD depression. The brain structural findings suggests that biological partitions of patients with mood disorders are likely to lead to the identification of subgroups, that transcend current diagnostic divisions into BD and MDD and are more likely to be aligned with underlying genetic variation. These results set the foundation for future studies to enhance our understanding of brain-behavior relationships in mood disorders.

Yang Tao, Frangou Sophia, Lam Raymond W, Huang Jia, Su Yousong, Zhao Guoqing, Mao Ruizhi, Zhu Na, Zhou Rubai, Lin Xiao, Xia Weiping, Wang Xing, Wang Yun, Peng Daihui, Wang Zuowei, Yatham Lakshmi N, Chen Jun, Fang Yiru


General General

Unsupervised Phenotypic Clustering for Determining Clinical Status in Children with Cystic Fibrosis.

In The European respiratory journal

BACKGROUND : Cystic Fibrosis (CF) is a multisystem disease in which assessing disease severity based on lung function alone may not be appropriate. The aim of the study was to develop a comprehensive machine-learning algorithm to assess clinical status independent of lung function in children.

METHODS : A comprehensive prospectively collected clinical database (Toronto, Canada) was used to apply unsupervised cluster analysis. The defined clusters were then compared by current and future lung function, risk of future hospitalisation, and risk of future pulmonary exacerbation (PEx) treated with oral antibiotics. A K-Nearest Neighbours (KNN) algorithm was used to prospectively assign clusters. The methods were validated in a paediatric clinical CF dataset from Great Ormond Street Hospital (GOSH).

RESULTS : The optimal cluster model identified four (A-D) phenotypic clusters based on 12 200 encounters from 530 individuals. Two clusters (A,B) consistent with mild disease were identified with high FEV1, and low risk of both hospitalisation and PEx treated with oral antibiotics. Two clusters (C,D) consistent with severe disease were also identified with low FEV1. Cluster D had the shortest time to both hospitalisation and PEx treated with oral antibiotics. The outcomes were consistent in 3124 encounters from 171 children at GOSH. The KNN cluster allocation error rate was low, at 2.5% (Toronto), and 3.5% (GOSH).

CONCLUSION : Machine learning derived phenotypic clusters can predict disease severity independent of lung function and could be used in conjunction with functional measures to predict future disease trajectories in CF patients.

Filipow Nicole, Davies Gwyneth, Main Eleanor, Sebire Neil J, Wallis Colin, Ratjen Felix, Stanojevic Sanja


General General

Learning the language of viral evolution and escape.

In Science (New York, N.Y.)

The ability for viruses to mutate and evade the human immune system and cause infection, called viral escape, remains an obstacle to antiviral and vaccine development. Understanding the complex rules that govern escape could inform therapeutic design. We modeled viral escape with machine learning algorithms originally developed for human natural language. We identified escape mutations as those that preserve viral infectivity but cause a virus to look different to the immune system, akin to word changes that preserve a sentence's grammaticality but change its meaning. With this approach, language models of influenza hemagglutinin, HIV-1 envelope glycoprotein (HIV Env), and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Spike viral proteins can accurately predict structural escape patterns using sequence data alone. Our study represents a promising conceptual bridge between natural language and viral evolution.

Hie Brian, Zhong Ellen D, Berger Bonnie, Bryson Bryan


Surgery Surgery

The sequence of disease-modifying anti-rheumatic drugs: pathways to and predictors of tocilizumab monotherapy.

In Arthritis research & therapy ; h5-index 60.0

BACKGROUND : There are numerous non-biologic and biologic disease-modifying anti-rheumatic drugs (bDMARDs) for rheumatoid arthritis (RA). Typical sequences of bDMARDs are not clear. Future treatment policies and trials should be informed by quantitative estimates of current treatment practice.

METHODS : We used data from Corrona, a large real-world RA registry, to develop a method for quantifying sequential patterns in treatment with bDMARDs. As a proof of concept, we study patients who eventually use tocilizumab monotherapy (TCZm), an IL-6 antagonist with similar benefits used as monotherapy or in combination. Patients starting a bDMARD were included and were followed using a discrete-state Markov model, observing changes in treatments every 6 months and determining whether they used TCZm. A supervised machine learning algorithm was then employed to determine longitudinal patient factors associated with TCZm use.

RESULTS : 7300 patients starting a bDMARD were followed for up to 5 years. Their median age was 58 years, 78% were female, median disease duration was 5 years, and 57% were seropositive. During follow-up, 287 (3.9%) reported use of TCZm with median time until use of 25.6 (11.5, 56.0) months. Eighty-two percent of TCZm use began within 3 years of starting any bDMARD. Ninety-three percent of TCZm users switched from TCZ combination, a TNF inhibitor, or another bDMARD. Very few patients are given TCZm as their first DMARD (0.6%). Variables associated with the use of TCZm included prior use of TCZ combination therapy, older age, longer disease duration, seronegative, higher disease activity, and no prior use of a TNF inhibitor.

CONCLUSIONS : Improved understanding of treatment sequences in RA may help personalize care. These methods may help optimize treatment decisions using large-scale real-world data.

Solomon Daniel H, Xu Chang, Collins Jamie, Kim Seoyoung C, Losina Elena, Yau Vincent, Johansson Fredrik D


DMARDs, Rheumatoid arthritis, Treatment

Radiology Radiology

Deep learning-based automatic delineation of the hippocampus by MRI: geometric and dosimetric evaluation.

In Radiation oncology (London, England)

BACKGROUND : Whole brain radiotherapy (WBRT) can impair patients' cognitive function. Hippocampal avoidance during WBRT can potentially prevent this side effect. However, manually delineating the target area is time-consuming and difficult. Here, we proposed a credible approach of automatic hippocampal delineation based on convolutional neural networks.

METHODS : Referring to the hippocampus contouring atlas proposed by RTOG 0933, we manually delineated (MD) the hippocampus on the MRI data sets (3-dimensional T1-weighted with slice thickness of 1 mm, n = 175), which were used to construct a three-dimensional convolutional neural network aiming for the hippocampus automatic delineation (AD). The performance of this AD tool was tested on three cohorts: (a) 3D T1 MRI with 1-mm slice thickness (n = 30); (b) non-3D T1-weighted MRI with 3-mm slice thickness (n = 19); (c) non-3D T1-weighted MRI with 1-mm slice thickness (n = 11). All MRIs confirmed with normal hippocampus has not been violated by any disease. Virtual radiation plans were created for AD and MD hippocampi in cohort c to evaluate the clinical feasibility of the artificial intelligence approach. Statistical analyses were performed using SPSS version 23. P < 0.05 was considered significant.

RESULTS : The Dice similarity coefficient (DSC) and Average Hausdorff Distance (AVD) between the AD and MD hippocampi are 0.86 ± 0.028 and 0.18 ± 0.050 cm in cohort a, 0.76 ± 0.035 and 0.31 ± 0.064 cm in cohort b, 0.80 ± 0.015 and 0.24 ± 0.021 cm in cohort c, respectively. The DSC and AVD in cohort a were better than those in cohorts b and c (P < 0.01). There is no significant difference between the radiotherapy plans generated using the AD and MD hippocampi.

CONCLUSION : The AD of the hippocampus based on a deep learning algorithm showed satisfying results, which could have a positive impact on improving delineation accuracy and reducing work load.

Pan Kaicheng, Zhao Lei, Gu Song, Tang Yi, Wang Jiahao, Yu Wen, Zhu Lucheng, Feng Qi, Su Ruipeng, Xu Zhiyong, Li Xiadong, Ding Zhongxiang, Fu Xiaolong, Ma Shenglin, Yan Jun, Kang Shigong, Zhou Tao, Xia Bing


Artificial intelligence, Hippocampus, MRI

General General

Deep learning encodes robust discriminative neuroimaging representations to outperform standard machine learning.

In Nature communications ; h5-index 260.0

Recent critical commentaries unfavorably compare deep learning (DL) with standard machine learning (SML) approaches for brain imaging data analysis. However, their conclusions are often based on pre-engineered features depriving DL of its main advantage - representation learning. We conduct a large-scale systematic comparison profiled in multiple classification and regression tasks on structural MRI images and show the importance of representation learning for DL. Results show that if trained following prevalent DL practices, DL methods have the potential to scale particularly well and substantially improve compared to SML methods, while also presenting a lower asymptotic complexity in relative computational time, despite being more complex. We also demonstrate that DL embeddings span comprehensible task-specific projection spectra and that DL consistently localizes task-discriminative brain biomarkers. Our findings highlight the presence of nonlinearities in neuroimaging data that DL can exploit to generate superior task-discriminative representations for characterizing the human brain.

Abrol Anees, Fu Zening, Salman Mustafa, Silva Rogers, Du Yuhui, Plis Sergey, Calhoun Vince