Recent advancements in artificial intelligence (AI) and the vast data generated by modern clinical systems have driven the development of AI solutions in medical imaging, encompassing image reconstruction, segmentation, diagnosis, and treatment planning. Despite these successes and potential, many stakeholders worry about the risks and ethical implications of imaging AI, viewing it as complex, opaque, and challenging to understand, use, and trust in critical clinical applications. The FUTURE-AI guideline for trustworthy AI in healthcare was established based on six guiding principles: Fairness, Universality, Traceability, Usability, Robustness, and Explainability. Through international consensus, a set of recommendations was defined, covering the entire lifecycle of medical AI tools, from design, development, and validation to regulation, deployment, and monitoring. In this paper, we describe how these specific recommendations can be instantiated in the domain of medical imaging, providing an overview of current best practices along with guidelines and concrete metrics on how those recommendations could be met, offering a valuable resource to the international medical imaging community.

Deep learning (DL) applied to breast tissue segmentation in magnetic resonance imaging (MRI) has received increased attention in the last decade, however, the domain shift which arises from different vendors, acquisition protocols, and biological heterogeneity, remains an important but challenging obstacle on the path towards clinical implementation. In this paper, we propose a novel Multi-level Semantic-guided Contrastive Domain Adaptation (MSCDA) framework to address this issue in an unsupervised manner. Our approach incorporates self-training with contrastive learning to align feature representations between domains. In particular, we extend the contrastive loss by incorporating pixel-to-pixel, pixel-to-centroid, and centroid-to-centroid contrasts to better exploit the underlying semantic information of the image at different levels. To resolve the data imbalance problem, we utilize a category-wise cross-domain sampling strategy to sample anchors from target images and build a hybrid memory bank to store samples from source images. We have validated MSCDA with a challenging task of cross-domain breast MRI segmentation between datasets of healthy volunteers and invasive breast cancer patients. Extensive experiments show that MSCDA effectively improves the model's feature alignment capabilities between domains, outperforming state-of-the-art methods. Furthermore, the framework is shown to be label-efficient, achieving good performance with a smaller source dataset. The code is publicly available at https://github.com/ShengKuangCN/MSCDA.

Hypoxia-activated prodrugs (HAPs) present a conceptually elegant approach to not only overcome, but better yet, exploit intra-tumoural hypoxia. Despite being successful in vitro and in vivo, HAPs are yet to achieve successful results in clinical settings. It has been hypothesised that this lack of clinical success can, in part, be explained by the insufficiently stringent clinical screening selection of determining which tumours are suitable for HAP treatments. Taking a mathematical modelling approach, we investigate how tumour properties and HAP-radiation scheduling influence treatment outcomes in simulated tumours. The following key results are demonstrated in silico: (i) HAP and ionising radiation (IR) monotherapies may attack tumours in dissimilar, and complementary, ways. (ii) HAP-IR scheduling may impact treatment efficacy. (iii) HAPs may function as IR treatment intensifiers. (iv) The spatio-temporal intra-tumoural oxygen landscape may impact HAP efficacy. Our in silico framework is based on an on-lattice, hybrid, multiscale cellular automaton spanning three spatial dimensions. The mathematical model for tumour spheroid growth is parameterised by multicellular tumour spheroid (MCTS) data.

We present a multi-scale approach of tumor modeling in order to predict its evolution during radiotherapy. Within this context we focus on three different scales of tumor modeling: microscopic (individual cells in a voxel), mesoscopic (population of cells in a voxel) and macroscopic (whole tumor), with transition interfaces between these three scales. At the cellular level, the description is based on phase transfer probabilities in the cellular cycle. At the mesoscopic scale we represent populations of cells according to different stages in a cell cycle. Finally, at the macroscopic scale, the tumor description is based on the use of FDG PET image voxels. These three scales exist naturally: biological data are collected at the macroscopic scale, but the pathological behavior of the tumor is based on an abnormal cell-cycle at the microscopic scale. On the other hand, the introduction of a mesoscopic scale is essential in order to reduce the gap between the two extreme, in terms of resolution, description levels. It also reduces the computational burden of simulating a large number of individual cells. As an application of the proposed multi-scale model, we simulate the effect of oxygen on tumor evolution during radiotherapy. Two consecutive FDG PET images of 17 rectal cancer patients undergoing radiotherapy are used to simulate the tumor evolution during treatment. The simulated results are compared with those obtained on a third FDG PET image acquired two weeks after the beginning of the treatment.

BACKGROUND: The reproducibility of transcriptomic biomarkers across datasets remains poor, limiting clinical application. We and others have suggested that this is in-part caused by differential error-structure between datasets, and their incomplete removal by pre-processing algorithms. METHODS: To test this hypothesis, we systematically assessed the effects of pre-processing on biomarker classification using 24 different pre-processing methods and 15 distinct signatures of tumour hypoxia in 10 datasets (2,143 patients). RESULTS: We confirm strong pre-processing effects for all datasets and signatures, and find that these differ between microarray versions. Importantly, exploiting different pre-processing techniques in an ensemble technique improved classification for a majority of signatures. CONCLUSIONS: Assessing biomarkers using an ensemble of pre-processing techniques shows clear value across multiple diseases, datasets and biomarkers. Importantly, ensemble classification improves biomarkers with initially good results but does not result in spuriously improved performance for poor biomarkers. While further research is required, this approach has the potential to become a standard for transcriptomic biomarkers.

Missing data is a given in the medical domain, so machine learning models should have satisfactory performance even when missing data occurs. Our previous work has focused on support vector machines (SVM), but we hypothesize that Bayesian networks (BN) can handle missing data better. To test the hypothesis, we trained a BN and SVM model for 2 year survival on 322 lung cancer patients and compared their performance in three separate external datasets (35, 47, 33 patients), each with their own characteristics in terms of missing data. The models used tumor size, clinical T and N stage, involved lymph nodes and WHO performance as prognostic features. We found that the BN model performed better than SVM (AUC 0.77, 0.72. 0.70 vs. 0.71, 0.68, 0.69), especially if tumor size was missing. We conclude that BN models are better suited for the medical domain, as they can handle missing data better.

Privacy-preserving data mining (PPDM) is an emergent research area that addresses the incorporation of privacy preserving concerns to data mining techniques. In this paper we propose a privacy-preserving (PP) Cox model for survival analysis, and consider a real clinical setting where the data is horizontally distributed among different institutions. The proposed model is based on linearly projecting the data to a lower dimensional space through an optimal mapping obtained by solving a linear programming problem. Our approach differs from the commonly used random projection approach since it instead finds a projection that is optimal at preserving the properties of the data that are important for the specific problem at hand. Since our proposed approach produces an sparse mapping, it also generates a PP mapping that not only projects the data to a lower dimensional space but it also depends on a smaller subset of the original features (it provides explicit feature selection). Real data from several European healthcare institutions are used to test our model for survival prediction of non-small-cell lung cancer patients. These results are also confirmed using publicly available benchmark datasets. Our experimental results show that we are able to achieve a near-optimal performance without directly sharing the data across different data sources. This model makes it possible to conduct large-scale multi-centric survival analysis without violating privacy-preserving requirements.

In biology and medical sciences, highly parallel biological assays spurred a revolution leading to the emergence of the '-omics' era. Dimensionality reduction techniques are necessary to be able to analyze, interpret, validate and take advantage of the tremendous wealth of highly dimensional data they provide. This paper is based on a DNA microarray study providing gene signatures for hypoxia. These gene signatures were tested on a large breast cancer data set for assessing their prognostic power by means of Kaplan-Meier survival, univariate, and multivariate analyses. We explore the use of several mathematical programming-based techniques that aim to reduce the gene signature sizes as much as possible while maintaining the key characteristics of the original signature, more precisely: the signature prognostic and diagnostic significance. The proposed signature reduction techniques have very interesting potential uses. Indeed, by downsizing the relevant data to a manageable size, one can then patent the core set of biomarkers and also create a dedicated assay (e.g.: on a customized array) for routine applications (e.g.: in the clinical set up) leading to individualized medicine capabilities. Our experiments show that the reduced hypoxia signatures reproduced qualitatively and quantitatively in a similar way that of the original ones.

In this paper, we show that classical survival analysis involving censored data can naturally be cast as a ranking problem. The concordance index (CI), which quantifies the quality of rankings, is the standard performance measure for model \emph{assessment} in survival analysis. In contrast, the standard approach to \emph{learning} the popular proportional hazard (PH) model is based on Cox's partial likelihood. In this paper we devise two bounds on CI--one of which emerges directly from the properties of PH models--and optimize them \emph{directly}. Our experimental results suggest that both methods perform about equally well, with our new approach giving slightly better results than the Cox's method. We also explain why a method designed to maximize the Cox's partial likelihood also ends up (approximately) maximizing the CI.