MultiTrans: Multi-branch transformer network for medical image segmentation.

BACKGROUND AND OBJECTIVE: Transformer, which is notable for its ability of global context modeling, has been used to remedy the shortcomings of Convolutional neural networks (CNN) and break its dominance in medical image segmentation. However, the self-attention module is both memory and computational inefficient, so many methods have to build their Transformer branch upon largely downsampled feature maps or adopt the tokenized image patches to fit their model into accessible GPUs. This patch-wise operation restricts the network in extracting pixel-level intrinsic structural or dependencies inside each patch, hurting the performance of pixel-level classification tasks. METHODS: To tackle these issues, we propose a memory- and computation-efficient self-attention module to enable reasoning on relatively high-resolution features, promoting the efficiency of learning global information while effective grasping fine spatial details. Furthermore, we design a novel Multi-Branch Transformer (MultiTrans) architecture to provide hierarchical features for handling objects with variable shapes and sizes in medical images. By building four parallel Transformer branches on different levels of CNN, our hybrid network aggregates both multi-scale global contexts and multi-scale local features. RESULTS: MultiTrans achieves the highest segmentation accuracy on three medical image datasets with different modalities: Synapse, ACDC and M&Ms. Compared to the Standard Self-Attention (SSA), the proposed Efficient Self-Attention (ESA) can largely reduce the training memory and computational complexity while even slightly improve the accuracy. Specifically, the training memory cost, FLOPs and Params of our ESA are 18.77%, 20.68% and 74.07% of the SSA. CONCLUSIONS: Experiments on three medical image datasets demonstrate the generality and robustness of the designed network. The ablation study shows the efficiency and effectiveness of our proposed ESA. Code is available at: https://github.com/Yanhua-Zhang/MultiTrans-extension.

Personalized Detection of Motion Artifacts for Telemonitoring Applications.

Among its main benefits, telemonitoring enables personalized management of chronic diseases by means of biomarkers extracted from signals. In these applications, a thorough quality assessment is required to ensure the reliability of the monitored parameters. Motion artifacts are a common problem in recordings with wearable devices. In this work, we propose a fully automated and personalized method to detect motion artifacts in multimodal recordings devoted to the monitoring of the Cardiac Time Intervals (CTIs). The detection of motion artifacts was carried out by using template matching with a personalized template. The method yielded a balanced accuracy of 86%. Moreover, it proved effective to decrease the variability of the estimated CTIs by at least 17%. Our preliminary results show that personalized detection of motion artifacts improves the robustness of the assessment CTIs and opens to the use in wearable systems.

Acquisition Devices for Fetal Phonocardiography: A Scoping Review.

Timely and reliable fetal monitoring is crucial to prevent adverse events during pregnancy and delivery. Fetal phonocardiography, i.e., the recording of fetal heart sounds, is emerging as a novel possibility to monitor fetal health status. Indeed, due to its passive nature and its noninvasiveness, the technique is suitable for long-term monitoring and for telemonitoring applications. Despite the high share of literature focusing on signal processing, no previous work has reviewed the technological hardware solutions devoted to the recording of fetal heart sounds. Thus, the aim of this scoping review is to collect information regarding the acquisition devices for fetal phonocardiography (FPCG), focusing on technical specifications and clinical use. Overall, PRISMA-guidelines-based analysis selected 57 studies that described 26 research prototypes and eight commercial devices for FPCG acquisition. Results of our review study reveal that no commercial devices were designed for fetal-specific purposes, that the latest advances involve the use of multiple microphones and sensors, and that no quantitative validation was usually performed. By highlighting the past and future trends and the most relevant innovations from both a technical and clinical perspective, this review will represent a useful reference for the evaluation of different acquisition devices and for the development of new FPCG-based systems for fetal monitoring.

Medical Device Software: From Requirements to Certification.

The role of software in healthcare is getting more and more pervasive. Nevertheless, manufacturers sometimes forget that these software are medical devices and must be certified according to the EU Medical Device Regulation 2017/745. In this work we propose a pipeline for developing a Medical Device Software (MDS) compliant with the regulations and certifiable. The pipeline includes the phase of requirements elicitation, risk assessment and analysis of effectiveness as key elements. The preparation of the technical file should be carried out in parallel with the MDS development. In the overall, it can be stated that the certification process starts with the conceptualization of the MDS and proceeds all along its design and implementation.

Design of a Service for the Management of Heart Failure Patients Using Telemedicine.

The tremendous prevalence and mortality of heart failure (HF), along with the social and economic impact of its consequences, make an appropriate disease management utmost important. In this context, telemedicine offers promising possibilities. Current clinical guidelines and technological solutions do not address the problem of monitoring at-risk patients and patients affected by mild HF for prevention purposes. The goal of this work is to design a service based on a telemedicine framework for the management of heart failure patients. The proposed service grounds the monitoring of the patient on a custom multi-sensor array that we designed and developed for the purpose. The description of the processes involved in the service was carried out by means of Process Modelling tools, and in particular through Swim Lane Activity Diagrams. The results look promising for the implementation of the service in a real-life scenario. The main strength of the service resides in a) the use of noninvasive monitoring technologies to include patients with a mild HF or at-risk patients; and b) the integration of hospital and territory services to grant continuity and coherence in the treatment.

A Wearable Multi-Sensor Array Enables the Recording of Heart Sounds in Homecare.

The home monitoring of patients affected by chronic heart failure (CHF) is of key importance in preventing acute episodes. Nevertheless, no wearable technological solution exists to date. A possibility could be offered by Cardiac Time Intervals extracted from simultaneous recordings of electrocardiographic (ECG) and phonocardiographic (PCG) signals. Nevertheless, the recording of a good-quality PCG signal requires accurate positioning of the stethoscope over the chest, which is unfeasible for a naïve user as the patient. In this work, we propose a solution based on multi-source PCG. We designed a flexible multi-sensor array to enable the recording of heart sounds by inexperienced users. The multi-sensor array is based on a flexible Printed Circuit Board mounting 48 microphones with a high spatial resolution, three electrodes to record an ECG and a Magneto-Inertial Measurement Unit. We validated the usability over a sample population of 42 inexperienced volunteers and found that all subjects could record signals of good to excellent quality. Moreover, we found that the multi-sensor array is suitable for use on a wide population of at-risk patients regardless of their body characteristics. Based on the promising findings of this study, we believe that the described device could enable the home monitoring of CHF patients soon.

Machine-Learning-Based Tool to Predict Target Prostate Biopsy Outcomes: An Internal Validation Study.

The aim of this study is to present a personalized predictive model (PPM) with a machine learning (ML) system that is able to identify and classify patients with suspected prostate cancer (PCa) following mpMRI. We extracted all the patients who underwent fusion biopsy (FB) from March 2014 to December 2019, while patients from August 2020 to April 2021 were included as a validation set. The proposed system was based on the following four ML methods: a fuzzy inference system (FIS), the support vector machine (SVM), k-nearest neighbors (KNN), and self-organizing maps (SOMs). Then, a system based on fuzzy logic (FL) + SVM was compared with logistic regression (LR) and standard diagnostic tools. A total of 1448 patients were included in the training set, while 181 patients were included in the validation set. The area under the curve (AUC) of the proposed FIS + SVM model was comparable with the LR model but outperformed the other diagnostic tools. The FIS + SVM model demonstrated the best performance, in terms of negative predictive value (NPV), on the training set (78.5%); moreover, it outperformed the LR in terms of specificity (92.1% vs. 83%). Considering the validation set, our model outperformed the other methods in terms of NPV (60.7%), sensitivity (90.8%), and accuracy (69.1%). In conclusion, we successfully developed and validated a PPM tool using the FIS + SVM model to calculate the probability of PCa prior to a prostate FB, avoiding useless ones in 15% of the cases.

Could normalization improve robustness of abdominal MRI radiomic features?

Radiomics-based systems could improve the management of oncological patients by supporting cancer diagnosis, treatment planning, and response assessment. However, one of the main limitations of these systems is the generalizability and reproducibility of results when they are applied to images acquired in different hospitals by different scanners. Normalization has been introduced to mitigate this issue, and two main approaches have been proposed: one rescales the image intensities (image normalization), the other the feature distributions for each center (feature normalization). The aim of this study is to evaluate how different image and feature normalization methods impact the robustness of 93 radiomics features acquired using a multicenter and multi-scanner abdominal Magnetic Resonance Imaging (MRI) dataset. To this scope, 88 rectal MRIs were retrospectively collected from 3 different institutions (4 scanners), and for each patient, six 3D regions of interest on the obturator muscle were considered. The methods applied were min-max, 1st-99th percentiles and 3-Sigma normalization, z-score standardization, mean centering, histogram normalization, Nyul-Udupa and ComBat harmonization. The Mann-Whitney U-test was applied to assess features repeatability between scanners, by comparing the feature values obtained for each normalization method, including the case in which no normalization was applied. Most image normalization methods allowed to reduce the overall variability in terms of intensity distributions, while worsening or showing unpredictable results in terms of feature robustness, except for thez-score, which provided a slight improvement by increasing the number of statistically similar features from 9/93 to 10/93. Conversely, feature normalization methods positively reduced the overall variability across the scanners, in particular, 3sigma,z_scoreandComBatthat increased the number of similar features (79/93). According to our results, it emerged that none of the image normalization methods was able to strongly increase the number of statistically similar features.

Normalization Strategies in Multi-Center Radiomics Abdominal MRI: Systematic Review and Meta-Analyses.

Goal: Artificial intelligence applied to medical image analysis has been extensively used to develop non-invasive diagnostic and prognostic signatures. However, these imaging biomarkers should be largely validated on multi-center datasets to prove their robustness before they can be introduced into clinical practice. The main challenge is represented by the great and unavoidable image variability which is usually addressed using different pre-processing techniques including spatial, intensity and feature normalization. The purpose of this study is to systematically summarize normalization methods and to evaluate their correlation with the radiomics model performances through meta-analyses. This review is carried out according to the PRISMA statement: 4777 papers were collected, but only 74 were included. Two meta-analyses were carried out according to two clinical aims: characterization and prediction of response. Findings of this review demonstrated that there are some commonly used normalization approaches, but not a commonly agreed pipeline that can allow to improve performance and to bridge the gap between bench and bedside.

Feature Selection Based on a Genetic Algorithm for Optimizing Weaning Success.

Finding the right time for weaning from ventilator is a difficult clinical decision. Several systems based on machine or deep learning are reported in literature. However, the results of these applications are not completely satisfactory and may be improved. An important aspect is represented by the features used as input of these systems. In this paper we present the results of the application of genetic algorithms to perform feature selection on a dataset containing 13688 patients under mechanical ventilation characterizing by 58 variables, extracted from the MIMIC III database. The results show that all features are important, but four of them are essential: 'Sedation_days', 'Mean_Airway_Pressure', 'PaO2', and 'Chloride'. This is only the initial step to obtain a tool to be added to the other clinical indices for minimize the risk of extubation failure.

Key Aspects to Teach Medical Device Software Certification.

Certification of Medical Device Software (MDS) according to the EU Medical Device Regulation 2017/745 requires demonstrating safety and effectiveness. Thus, the syllabus of a course on MDS development must provide tools for addressing these issues. To assure safety, risk analysis has to be performed using a four-step procedure. Effectiveness could be demonstrated by literature systematic review combined with meta-analysis, to compare the MDS performances with those of similar tools.

A fully automatic deep learning algorithm to segment rectal Cancer on MR images: a multi-center study.

The aim of the study is to present and tune a fully automatic deep learning algorithm to segment colorectal cancers (CRC) on MR images, based on a U-Net structure. It is a multicenter study, including 3 different Italian institutions, that used 4 different MRI scanners. Two of them were used for training and tuning the systems, while the other two for the validation. The implemented algorithm consists of a pre-processing step to normalize and to highlight the tumoral area, followed by the CRC segmentation using different U-net structures. Automatic masks were compared with manual segmentations performed by three experienced radiologists, one at each center. The two best performing systems (called mdl2 and mdl3), obtained a median Dice Similarity Coefficient of 0.68(mdl2) - 0.69(mdl3), precision of 0.75(md/2) - 0.71(md/3), and recall of 0.69(mdl2) - 0.73(mdl3) on the validation set. Both systems reached high detection rates, 0.98 and 0.95, respectively, on the validation set. These encouraging results, if confirmed on larger dataset, might improve the management of patients with CRC, since it can be used as a fast and precise tool for further radiomics analyses. Clinical Relevance - To provide a reliable tool able to automatically segment CRC tumors that can be used as first step in future radiomics studies aimed at predicting response to chemotherapy and personalizing treatment.

Artificial intelligence for target prostate biopsy outcomes prediction the potential application of fuzzy logic.

BACKGROUND: In current precision prostate cancer (PCa) surgery era the identification of the best patients candidate for prostate biopsy still remains an open issue. The aim of this study was to evaluate if the prostate target biopsy (TB) outcomes could be predicted by using artificial intelligence approach based on a set of clinical pre-biopsy. METHODS: Pre-biopsy characteristics in terms of PSA, PSA density, digital rectal examination (DRE), previous prostate biopsies, number of suspicious lesions at mp-MRI, lesion volume, lesion location, and Pi-Rads score were extracted from our prospectively maintained TB database from March 2014 to December 2019. Our approach is based on Fuzzy logic and associative rules mining, with the aim to predict TB outcomes. RESULTS: A total of 1448 patients were included. Using the Frequent-Pattern growth algorithm we extracted 875 rules and used to build the fuzzy classifier. 963 subjects were classified whereas for the remaining 484 subjects were not classified since no rules matched with their input variables. Analyzing the classified subjects we obtained a specificity of 59.2% and sensitivity of 90.8% with a negative and the positive predictive values of 81.3% and 76.6%, respectively. In particular, focusing on ISUP ≥ 3 PCa, our model is able to correctly predict the biopsy outcomes in 98.1% of the cases. CONCLUSIONS: In this study we demonstrated that the possibility to look at several pre-biopsy variables simultaneously with artificial intelligence algorithms can improve the prediction of TB outcomes, outclassing the performance of PSA, its derivates and MRI alone.

Evaluation of Muscle Function by Means of a Muscle-Specific and a Global Index.

Gait analysis applications in clinics are still uncommon, for three main reasons: (1) the considerable time needed to prepare the subject for the examination; (2) the lack of user-independent tools; (3) the large variability of muscle activation patterns observed in healthy and pathological subjects. Numerical indices quantifying the muscle coordination of a subject could enable clinicians to identify patterns that deviate from those of a reference population and to follow the progress of the subject after surgery or completing a rehabilitation program. In this work, we present two user-independent indices. First, a muscle-specific index (MFI) that quantifies the similarity of the activation pattern of a muscle of a specific subject with that of a reference population. Second, a global index (GFI) that provides a score of the overall activation of a muscle set. These two indices were tested on two groups of healthy and pathological children with encouraging results. Hence, the two indices will allow clinicians to assess the muscle activation, identifying muscles showing an abnormal activation pattern, and associate a functional score to every single muscle as well as to the entire muscle set. These opportunities could contribute to facilitating the diffusion of surface EMG analysis in clinics.

Automated Assessment of the Quality of Phonocardographic Recordings through Signal-to-Noise Ratio for Home Monitoring Applications.

The signal quality limits the applicability of phonocardiography at the patients' domicile. This work proposes the signal-to-noise ratio of the recorded signal as its main quality metrics. Moreover, we define the minimum acceptable values of the signal-to-noise ratio that warrantee an accuracy of the derived parameters acceptable in clinics. We considered 25 original heart sounds recordings, which we corrupted by adding noise to decrease their signal-to-noise ratio. We found that a signal-to-noise ratio equal to or higher than 14 dB warrants an uncertainty of the estimate of the valve closure latencies below 1 ms. This accuracy is higher than that required by most clinical applications. We validated the proposed method against a public database, obtaining results comparable to those obtained on our sample population. In conclusion, we defined (a) the signal-to-noise ratio of the phonocardiographic signal as the preferred metric to evaluate its quality and (b) the minimum values of the signal-to-noise ratio required to obtain an uncertainty of the latency of heart sound components compatible with clinical applications. We believe these results are crucial for the development of home monitoring systems aimed at preventing acute episodes of heart failure and that can be safely operated by naïve users.

Simulation of the Impact on the Workload of the Enlargement of the Clinical Staff of a Specialistic Reference Center.

Quality of care and patient satisfaction are important aspects of high standard care. If clinical staff is subject to an elevated workload there is a possible decrease of both. This justifies the development of tools to quantify the workload and to find organizational changes that will normalize it. We have previously developed a simulation system to quantify the workload of the staff working in a regional reference center for the treatment of bleeding and hemorrhagic disorders. The goal of this new work is to simulate, through an agent-based model, the impact of adding a physician to the staff. Ten sets of initial parameters were defined to simulate ten typical weeks. Results show that the introduction of the new physician together with a second ambulatory room can reduce the workload of all the staff to the expected 8-hour. In this situation, in which the staff workload does not exceed the daily capacity, we may suppose that an increase in the quality of care and patient satisfaction will be possible.

An innovative radiomics approach to predict response to chemotherapy of liver metastases based on CT images.

Liver metastases (mts) from colorectal cancer (CRC) can have different responses to chemotherapy in the same patient. The aim of this study is to develop and validate a machine learning algorithm to predict response of individual liver mts. 22 radiomic features (RF) were computed on pretreatment portal CT scans following a manual segmentation of mts. RFs were extracted from 7x7 Region of Interests (ROIs) that moved across the image by step of 2 pixels. Liver mts were classified as non-responder (R-) if their largest diameter increased more than 3 mm after 3 months of treatment and responder (R+), otherwise. Features selection (FS) was performed by a genetic algorithm and classification by a Support Vector Machine (SVM) classifier. Sensitivity, specificity, negative (NPV) and positive (PPV) predictive values were evaluated for all lesions in the training and validation sets, separately. On the training set, we obtained sensitivity of 86%, specificity of 67%, PPV of 89% and NPV of 61%, while, on the validation set, we reached a sensitivity of 73%, specificity of 47%, PPV of 64% and NPV of 57%. Specificity was biased by the low number of R- lesions on the validation set. The promising results obtained in the validation dataset should be extended to a larger cohort of patient to further validate our method.Clinical Relevance- to personalize treatment of patients with metastastic colorectal cancer, based on the likelihood of response to chemotherapy of each liver metastasis.

A Convolutional Neural Network based system for Colorectal cancer segmentation on MRI images.

The aim of the study is to present a new Convolutional Neural Network (CNN) based system for the automatic segmentation of the colorectal cancer. The algorithm implemented consists of several steps: a pre-processing to normalize and highlights the tumoral area, the classification based on CNNs, and a post-processing aimed at reducing false positive elements. The classification is performed using three CNNs: each of them classifies the same regions of interest acquired from three different MR sequences. The final segmentation mask is obtained by a majority voting. Performances were evaluated using a semi-automatic segmentation revised by an experienced radiologist as reference standard. The system obtained Dice Similarity Coefficient (DSC) of 0.60, Precision (Pr) of 0.76 and Recall (Re) of 0.55 on the testing set. After applying the leave-one-out validation, we obtained a median DSC=0.58, Pr=0.74, Re=0.54. The promising results obtained by this system, if validated on a larger dataset, could strongly improve personalized medicine.

Surface Electromyography Applied to Gait Analysis: How to Improve Its Impact in Clinics?

Surface electromyography (sEMG) is the main non-invasive tool used to record the electrical activity of muscles during dynamic tasks. In clinical gait analysis, a number of techniques have been developed to obtain and interpret the muscle activation patterns of patients showing altered locomotion. However, the body of knowledge described in these studies is very seldom translated into routine clinical practice. The aim of this work is to analyze critically the key factors limiting the extensive use of these powerful techniques among clinicians. A thorough understanding of these limiting factors will provide an important opportunity to overcome limitations through specific actions, and advance toward an evidence-based approach to rehabilitation based on objective findings and measurements.

Radiomics predicts response of individual HER2-amplified colorectal cancer liver metastases in patients treated with HER2-targeted therapy.

The aim of our study was to develop and validate a machine learning algorithm to predict response of individual HER2-amplified colorectal cancer liver metastases (lmCRC) undergoing dual HER2-targeted therapy. Twenty-four radiomics features were extracted after 3D manual segmentation of 141 lmCRC on pretreatment portal CT scans of a cohort including 38 HER2-amplified patients; feature selection was then performed using genetic algorithms. lmCRC were classified as nonresponders (R-), if their largest diameter increased more than 10% at a CT scan performed after 3 months of treatment, responders (R+) otherwise. Sensitivity, specificity, negative (NPV) and positive (PPV) predictive values in correctly classifying individual lesion and overall patient response were assessed on a training dataset and then validated on a second dataset using a Gaussian naïve Bayesian classifier. Per-lesion sensitivity, specificity, NPV and PPV were 89%, 85%, 93%, 78% and 90%, 42%, 73%, 71% respectively in the testing and validation datasets. Per-patient sensitivity and specificity were 92% and 86%. Heterogeneous response was observed in 9 of 38 patients (24%). Five of nine patients were carriers of nonresponder lesions correctly classified as such by our radiomics signature, including four of seven harboring only one nonresponder lesion. The developed method has been proven effective in predicting behavior of individual metastases to targeted treatment in a cohort of HER2 amplified patients. The model accurately detects responder lesions and identifies nonresponder lesions in patients with heterogeneous response, potentially paving the way to multimodal treatment in selected patients. Further validation will be needed to confirm our findings.