Medical image analysis is crucial for the efficient diagnosis of many diseases. Typically, hospitals maintain vast repositories of images, which can be leveraged for various purposes, including research. However, access to such image collections is largely restricted to safeguard the privacy of the individuals whose images are being stored, as data protection concerns come into play. Recently, the development of solutions for Automated Medical Image Analysis has gained significant attention, with Deep Learning being one solution that has achieved remarkable results in this area. One promising approach for medical image analysis is Federated Learning (FL), which enables the use of a set of physically distributed data repositories, usually known as nodes, satisfying the restriction that the data do not leave the repository. Under these conditions, FL can build high-quality, accurate deep-learning models using a lot of available data wherever it is. Therefore, FL can help researchers and clinicians diagnose diseases and support medical decisions more efficiently and robustly. This article provides a systematic survey of FL in medical image analysis, specifically based on Magnetic Resonance Imaging, Computed Tomography, X-radiography, and histology images. Hence, it discusses applications, contributions, limitations, and challenges and is, therefore, suitable for those who want to understand how FL can contribute to the medical imaging domain.

Background: As of today, cancer is still one of the most prevalent and high-mortality diseases, summing more than 9 million deaths in 2018. This has motivated researchers to study the application of machine learning-based solutions for cancer detection to accelerate its diagnosis and help its prevention. Among several approaches, one is to automatically classify tumor samples through their gene expression analysis. Methods: In this work, we aim to distinguish five different types of cancer through RNA-Seq datasets: thyroid, skin, stomach, breast, and lung. To do so, we have adopted a previously described methodology, with which we compare the performance of 3 different autoencoders (AEs) used as a deep neural network weight initialization technique. Our experiments consist in assessing two different approaches when training the classification model — fixing the weights after pre-training the AEs, or allowing fine-tuning of the entire network — and two different strategies for embedding the AEs into the classification network, namely by only importing the encoding layers, or by inserting the complete AE. We then study how varying the number of layers in the first strategy, the AEs latent vector dimension, and the imputation technique in the data preprocessing step impacts the network’s overall classification performance. Finally, with the goal of assessing how well does this pipeline generalize, we apply the same methodology to two additional datasets that include features extracted from images of malaria thin blood smears, and breast masses cell nuclei. We also discard the possibility of overfitting by using held-out test sets in the images datasets. Results: The methodology attained good overall results for both RNA-Seq and image extracted data. We outperformed the established baseline for all the considered datasets, achieving an average F₁ score of 99.03, 89.95, and 98.84 and an MCC of 0.99, 0.84, and 0.98, for the RNA-Seq (when detecting thyroid cancer), the Malaria, and the Wisconsin Breast Cancer data, respectively. Conclusions: We observed that the approach of fine-tuning the weights of the top layers imported from the AE reached higher results, for all the presented experiences, and all the considered datasets. We outperformed all the previous reported results when comparing to the established baselines.

A systematic survey was carried out in an unbiased sample of 815 protein chains with a maximum of 20% homology selected from the Protein Data Bank, whose structures were solved at a resolution higher than 1.6 Å and with a R‐factor lower than 25%. A set of 5556 subsequences with α‐helix or 3₁₀‐helix motifs was extracted from the protein chains considered. Global and local propensities were then calculated for all possible amino acid pairs of the type (i, i + 1), (i, i + 2), (i, i + 3), and (i, i + 4), starting at the relevant helical positions N1, N2, N3, C3, C2, C1, and N‐int (interior positions), and also at the first nonhelical positions in both termini of the helices, namely, N‐cap and C‐cap. The statistical analysis of the propensity values has shown that pairing is significantly dependent on the type of the amino acids and on the position of the pair. A few sequences of three and four amino acids were selected and their high prevalence in helices is outlined in this work. The Glu‐Lys‐Tyr‐Pro sequence shows a peculiar distribution in proteins, which may suggest a relevant structural role in α‐helices when Pro is located at the C‐cap position. A bioinformatics tool was developed, which updates automatically and periodically the results and makes them available in a web site. Proteins 2008.