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

General General

Machine Learning Predicts Degree of Aromaticity from Structural Fingerprints.

In Journal of chemical information and modeling

Prediction of whether a compound is 'aromatic' is at first glance a relatively simple task - does it obey Hückel's rule (planar cyclic π-system with 4n+2 electrons) or not? However, aromaticity is far from a binary property, and there are distinct variations in chemical and biological behaviour between different systems which obey Hückel's rule and are thus classified as aromatic. To that end, the aromaticity of each molecule in a large public dataset [1] [2] was quantified by an extension of the work of Raczyńska et al. [3]. Building on this data, a method is proposed for machine-learning the degree of aromaticity for each aromatic ring in a molecule. Categories are derived from the numeric results, allowing the differentiation of structural patterns between them and thus better representation of the underlying chemical and biological behaviour in expert and (Q)SAR systems.

Ponting David John, van Deursen Ruud, Ott Martin A


Radiology Radiology

Assessing and mitigating the effects of class imbalance in machine learning with application to X-ray imaging.

In International journal of computer assisted radiology and surgery

PURPOSE : Machine learning (ML) algorithms are well known to exhibit variations in prediction accuracy when provided with imbalanced training sets typically seen in medical imaging (MI) due to the imbalanced ratio of pathological and normal cases. This paper presents a thorough investigation of the effects of class imbalance and methods for mitigating class imbalance in ML algorithms applied to MI.

METHODS : We first selected five classes from the Image Retrieval in Medical Applications (IRMA) dataset, performed multiclass classification using the random forest model (RFM), and then performed binary classification using convolutional neural network (CNN) on a chest X-ray dataset. An imbalanced class was created in the training set by varying the number of images in that class. Methods tested to mitigate class imbalance included oversampling, undersampling, and changing class weights of the RFM. Model performance was assessed by overall classification accuracy, overall F1 score, and specificity, recall, and precision of the imbalanced class.

RESULTS : A close-to-balanced training set resulted in the best model performance, and a large imbalance with overrepresentation was more detrimental to model performance than underrepresentation. Oversampling and undersampling methods were both effective in mitigating class imbalance, and efficacy of oversampling techniques was class specific.

CONCLUSION : This study systematically demonstrates the effect of class imbalance on two public X-ray datasets on RFM and CNN, making these findings widely applicable as a reference. Furthermore, the methods employed here can guide researchers in assessing and addressing the effects of class imbalance, while considering the data-specific characteristics to optimize imbalance mitigating methods.

Qu Wendi, Balki Indranil, Mendez Mauro, Valen John, Levman Jacob, Tyrrell Pascal N


Class imbalance, Machine learning, Medical imaging, Radiology, X-ray

General General

Ethical dilemmas in COVID-19 times: how to decide who lives and who dies?

In Revista da Associacao Medica Brasileira (1992)

The respiratory disease caused by the coronavirus SARS-CoV-2 (COVID-19) is a pandemic that produces a large number of simultaneous patients with severe symptoms and in need of special hospital care, overloading the infrastructure of health services. All of these demands generate the need to ration equipment and interventions. Faced with this imbalance, how, when, and who decides, there is the impact of the stressful systems of professionals who are at the front line of care and, in the background, issues inherent to human subjectivity. Along this path, the idea of using artificial intelligence algorithms to replace health professionals in the decision-making process also arises. In this context, there is the ethical question of how to manage the demands produced by the pandemic. The objective of this work is to reflect, from the point of view of medical ethics, on the basic principles of the choices made by the health teams, during the COVID-19 pandemic, whose resources are scarce and decisions cause anguish and restlessness. The ethical values for the rationing of health resources in an epidemic must converge to some proposals based on fundamental values such as maximizing the benefits produced by scarce resources, treating people equally, promoting and recommending instrumental values, giving priority to critical situations. Naturally, different judgments will occur in different circumstances, but transparency is essential to ensure public trust. In this way, it is possible to develop prioritization guidelines using well-defined values and ethical recommendations to achieve fair resource allocation.

Neves Nedy M B C, Bitencourt Flávia B C S N, Bitencourt Almir G V


General General

Artificial intelligence technology for diagnosing COVID-19 cases: a review of substantial issues.

In European review for medical and pharmacological sciences

Today, the world suffers from the rapid spread of COVID-19, which has claimed thousands of lives. Unfortunately, its treatment is yet to be developed. Nevertheless, this phenomenon can be decelerated by diagnosing and quarantining patients with COVID-19 at early stages, thereby saving numerous lives. In this study, the early diagnosis of this disease through artificial intelligence (AI) technology is explored. AI is a revolutionizing technology that drives new research opportunities in various fields. Although this study does not provide a final solution, it highlights the most promising lines of research on AI technology for the diagnosis of COVID-19. The major contribution of this work is a discussion on the following substantial issues of AI technology for preventing the severe effects of COVID-19: (1) rapid diagnosis and detection, (2) outbreak and prediction of virus spread, and (3) potential treatments. This study profoundly investigates these controversial research topics to achieve a precise, concrete, and concise conclusion. Thus, this study provides significant recommendations on future research directions related to COVID-19.

Alsharif M H, Alsharif Y H, Chaudhry S A, Albreem M A, Jahid A, Hwang E


Internal Medicine Internal Medicine

Large scale assessment of consistency in sleep stage scoring rules among multiple sleep centers using an interpretable machine learning algorithm.

In Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine

STUDY OBJECTIVES : Polysomnography is the gold standard in identifying sleep stages; however, there are discrepancies in how technicians use the standards. Because organizing meetings to evaluate this discrepancy and/or reach a consensus among multiple sleep centers is time consuming, we developed an artificial intelligence (AI) system to efficiently evaluate the reliability and consistency of sleep scoring, and hence the sleep center quality.

METHODS : An interpretable machine learning algorithm was used to evaluate interrater reliability (IRR) of sleep stage annotation among sleep centers. The AI system was trained to learn raters from one hospital, and applied to subjects from the same or other hospitals. The results were compared with the experts' annotation to determine IRR. Intra-center and inter-center assessments were conducted on 679 subjects without sleep apnea from six sleep centers in Taiwan. Centers with potential quality issues were identified by the estimated IRR.

RESULTS : In the intra-center assessment, the median accuracy ranged from 80·3% to 83·3% with the exception of one hospital (designated E) with an accuracy of 72·3%. In the inter-center assessment, the median accuracy ranged from 75·7% to 83·3% when hospital E was excluded from testing and training. The performance of the proposed method was higher for N2, awake, and REM, compared to N1 and N3. The significant IRR discrepancy of hospital E suggested a quality issue. This quality issue is confirmed by the physicians in charge of hospital E.

CONCLUSIONS : The proposed AI system proved effective in assessing IRR and hence the sleep center quality.

Liu Gi-Ren, Lin Ting-Yu, Wu Hau-Tieng, Sheu Yuan-Chung, Liu Ching-Lung, Liu Wen-Te, Yang Mei-Chen, Ni Yung-Lun, Chou Kun-Ta, Chen Chao-Hsien, Wu Dean, Lan Chou-Chin, Chiu Kuo-Liang, Chiu Hwa-Yen, Lo Yu-Lun


inter-center assessments, interrater reliability, intra-center assessments, machine learning, sleep stage scoring

Public Health Public Health

Cross-trial prediction in psychotherapy: External validation of the Personalized Advantage Index using machine learning in two Dutch randomized trials comparing CBT versus IPT for depression.

In Psychotherapy research : journal of the Society for Psychotherapy Research

AbstractObjective: Optimizing treatment selection may improve treatment outcomes in depression. A promising approach is the Personalized Advantage Index (PAI), which predicts the optimal treatment for a given individual. To determine the generalizability of the PAI, models needs to be externally validated, which has rarely been done. Method: PAI models were developed within each of two independent trials, with substantial between-study differences, that both compared CBT and IPT for depression (STEPd: n = 151 and FreqMech: n = 200). Subsequently, both PAI models were tested in the other dataset. Results: In the STEPd study, post-treatment depression was significantly different between individuals assigned to their PAI-indicated treatment versus those assigned to their non-indicated treatment (d = .57). In the FreqMech study, post-treatment depression was not significantly different between patients receiving their indicated treatment versus those receiving their non-indicated treatment (d = .20). Cross-trial predictions indicated that post-treatment depression was not significantly different between those receiving their indicated treatment and those receiving their non-indicated treatment (d = .16 and d = .27). Sensitivity analyses indicated that cross-trial prediction based on only overlapping variables didn't improve the results. Conclusion: External validation of the PAI has modest results and emphasizes between-study differences and many other challenges.

Van Bronswijk Suzanne C, Bruijniks Sanne J E, Lorenzo-Luaces Lorenzo, Derubeis Robert J, Lemmens Lotte H J M, Peeters Frenk P M L, Huibers Marcus J H


cognitive behavioural therapy, depression, external validation, interpersonal psychotherapy, precision medicine, prediction