Intelligence-based medicine最新文献

Breast cancer is one of the most aggressive tumors that claims the lives of women each year. Radiologists recommend mammography to detect cancer at the early stages. Masses, micro-calcifications, and distortion in mammography indicate breast cancer. This paper proposes FuzzyDeepNets for extracting the features and the Hanman transform classifier for the classification of mammograms. In this work, mammograms are categorized based on abnormality present, type of abnormality, and the characteristics of the abnormality present. FuzzyDeepNets allows us to skip the layers thereby reducing the computational complexity of the deep learning architectures. Principal component analysis helps in reducing the dimensionality of the selected features. The results achieved using proposed method on publicly available mini-MIAS, DDSM, INbreast and private database surpasses the results of the state-of-the-art techniques used for comparison. Results of the proposed method are clinically relevant as they are validated by expert radiologists.

Detecting pharyngeal acid reflux (PAR) episodes from 24-h ambulatory hypopharyngeal multichannel intraluminal impedance-pH (HMII-pH) signals is crucial for diagnosing laryngopharyngeal reflux (LPR). Currently, a lack of effective software for PAR episode detection requires time-consuming manual interpretation, which is prone to inter-rater variability. This study introduces a deep learning-based artificial intelligence (AI) system for PAR episode detection and diagnosis using HMII-pH signals. Ninety patients with suspected LPR and 28 healthy volunteers underwent HMII-pH testing in three Taiwanese medical centers. Candidate PAR episodes were defined as esophagopharyngeal pH drops exceeding 2 units, with nadir pH below 5 within 30 seconds during esophageal acidification. A consensus review by three experts validated 84 PAR episodes in 17 subjects. Data preprocessing identified 225 candidate PAR episodes, including 84 PAR episodes and 141 swallows/artifacts, were divided into training, validation, and test datasets (6:2:2 ratio). Three cascade deep learning AI models were trained. Among them, the cascade Multivariate Long Short-Term Memory with Fully Convolutional Network (MLSTM-FCN) model performed best in the test dataset. At the episode level, this model achieved 0.936 accuracy, 0.941 precision, 0.889 recall, 0.966 specificity, 0.914 F₁ score, and 0.864 Matthew's correlation coefficient (MCC). For subject-level evaluation, the corresponding metrics were 0.917 accuracy, 1.000 precision, 0.818 recall, 1.000 specificity, 0.900 F₁ score, and 0.842 MCC. In conclusion, the cascade MLSTM-FCN model demonstrates robust accuracy in diagnosing PAR episodes from HMII-pH signals, offering a promising tool for efficient and consistent PAR episode detection in LPR diagnosis.

Diabetic foot ulcer (DFU) is a common problem among people with diabetes that can result in amputation of the affected limb. Modern DFU treatment and diagnosis methods are expensive and time-consuming. Today, the development of the computer-aided diagnosis (CAD) method makes it possible for pathologists to diagnose DFU more swiftly and accurately. This has led to a rise in interest in deep learning (DL) approaches based on CAD. In this study, we introduce a novel framework called "DFU_MultiNet," which focuses on the transfer learning approach to classify healthy and ulcer skin images using publicly available repositories. The proposed framework is developed to offer an efficient and robust method for DFU classification that determines the distinction between healthy and ulcerated skin. The proposed approach extracts features from foot samples using three well-known pre-trained CNN models: VGG19, DenseNet201, and NasNetMobile. Finally, these extracted results are merged through a summing layer to create a powerful hybrid network. Through obtaining impressive accuracy (99.06 %), precision (100.00 %), recall (98.18 %), specificity (100.00 %), F1-score (99.08 %), and AUC (99.09 %) the proposed "DFU_MultiNet" framework holds great potential as a diagnostic tool in healthcare and clinical settings.

Objective

Low-value care (i.e., costly health care treatments that provide little or no benefit) is an ongoing problem in United States hospitals. Traditional strategies for reducing low-value care are only moderately successful. Informed by behavioral science principles, we sought to use machine learning to inform a targeted prompting system that suggests preferred alternative treatments at the point of care but before clinicians have made a decision.

Methods

We used intravenous administration of albumin for fluid resuscitation in intensive care unit (ICU) patients as an exemplar of low-value care practice, identified using the electronic health record of a multi-hospital health system. We divided all ICU episodes into 4-h periods and defined a set of relevant clinical features at the period level. We then developed two machine learning models: a single-stage model that directly predicts if a patient will receive albumin in the next period; and a two-stage model that first predicts if any resuscitation fluid will be administered and then predicts albumin only among the patients with a high probability of fluid use.

Results

We examined 87,489 ICU episodes divided into approximately 1.5 million 4-h periods. The area under the receiver operating characteristic curve was 0.86 for both prediction models. The positive predictive value was 0.21 (95% confidence interval: 0.20, 0.23) for the single-stage model and 0.22 (0.20, 0.23) for the two-stage model. Applying either model in a targeted prompting system could prevent 10% of albumin administrations, with an attending physician receiving one prompt every 4.2 days of ICU service.

Conclusion

Prediction of low-value care is feasible and could enable a point-of-care, targeted prompting system that offers suggestions ahead of the moment of need before clinicians have already decided. A two-stage approach does not improve performance but does interject new levers for the calibration of such a system.

Medical information processing is a staple of modern medicine with its increasing focus on the collection of numeric medical data such as questionnaires, biophysiological signals, and medical images. Although these modalities have long existed and guided medical practice, the movement towards using algorithms to transform, curate, summarise, and otherwise interact with this data is relatively new. Novel algorithms now form the interface between clinical users and data, extracting information that would otherwise be inaccessible or cumbersome. Recently, machine learning has expanded the capacities of these algorithms, using a priori acquired (and often annotated) datasets to learn a complex computational task. Validation of these techniques is inherently important for determining their safety and efficacy in a particular clinical context. However, methodological considerations such as the definition of reference data and validation procedures can obscure validation issues such as inaccurate reporting, a lack of standardisation, and a variety of biases. The purpose of this paper is to develop a framework for understanding medical information processing algorithms with a focus on validation that is adapted for machine learning approaches as well as traditional ones. This framework is instantiated in two example literature reviews which serve as the starting point for a discussion on how validation can be improved cognisant of machine learning.

Nowadays, one of the most important illnesses is a heart disease which causes most patients dead. The medical diagnosis of heart disease is quite difficult. This diagnosis is a challenging process that requires accuracy and efficiency. The chance of death will be decreased with early heart disease detection. Because cardiac problems are now a fairly frequent ailment, predicting heart disease has become one of the most difficult medical jobs in recent years. Researchers looked at a variety of closely related traits to discover the most reliable predictors of these conditions. In this study, Machine Learning (ML) techniques are used to identify the presence of cardiac abnormalities. The proposed method predicts the chances of heart disease and classifies patient's risk level by using different ML algorithm techniques such as Decision Tree (DT), Ada-Boost Classifier (AB), Extra trees Classifier (ET), Support vector Machine (SVM), Gradient boost, MLP, extreme gradient boost (XGB), Random Forest (RF), KNN, and LR. Three different datasets are combined to train and test the proposed system. The experimental results show that, when compared to other ML algorithms, the Decision Tree algorithm has the highest accuracy, at 99.16%.

Objectives

The primary goal of this study is to examine if a convolutional neural network (CNN) can be applied as a diagnostic tool for predicting Alzheimer's Disease (AD) from magnetic resonance imaging (MRI) using the MIRIAD-dataset (Minimal Interval Resonance Imaging in Alzheimer's Disease) from one single central slice of the brain.

Methods

The MIRIAD dataset contains patients' health records represented by a set of MRI scans of the brain and further diagnostic data. Hyperparameters and configurations of CNNs were optimized to determine the best-performing model. The CNN was implemented in Python with the deep learning library ‘Keras’ using Linux/Ubuntu as the operating system.

Results

This study obtained the following best performance metrics for predicting Alzheimer's Disease from MRI with Matthew's Correlation Coefficient (MCC) of 0.77; accuracy of 0.89; F1-score of 0.89; AUC of 0.92. The computational time for the training of a CNN takes less than 30 sec. s with a GPU (graphics processing unit). The prediction takes less than 1 sec. on a standard PC.

Conclusions

The study suggests that an axial MRI scan can be used to diagnose if a patient has Alzheimer's Disease with an AUC score of 0.92.

In fulfilling their duty to regulate the practice of medicine, state medical boards face complex regulatory challenges and patient safety concerns in adapting regulations and standards for the provision of medical care where the use of artificial intelligence becomes more prevalent. This article raises preliminary, yet foundational, questions of how artificial intelligence will continue to change the patient experience and the duties of a physician, and calls for increased regulatory attention from state and federal regulators. This article introduces the important role of state medical boards and why those interested in deploying artificial intelligence in clinical settings should be aware of how medical boards approach issues of standard of care and ethics. It also offers suggestions on how regulators may be able to improve collaboration to promote an innovation-friendly regulatory strategy.

Background

Patient-reported outcomes (PRO) allow clinicians to measure health-related quality of life (HRQOL) and understand patients’ treatment priorities, but obtaining PRO requires surveys which are not part of routine care. We aimed to develop a preliminary natural language processing (NLP) pipeline to extract HRQOL trajectory based on deep learning models using patient language.

Materials and methods

Our data consisted of transcribed interviews of 100 patients undergoing surgical intervention for low-risk thyroid cancer, paired with HRQOL assessments completed during the same visits. Our outcome measure was HRQOL trajectory measured by the SF-12 physical and mental component scores (PCS and MCS), and average THYCA-QoL score.

We constructed an NLP pipeline based on BERT, a modern deep language model that captures context semantics, to predict HRQOL trajectory as measured by the above endpoints. We compared this to baseline models using logistic regression and support vector machines trained on bag-of-words representations of transcripts obtained using Linguistic Inquiry and Word Count (LIWC). Finally, given the modest dataset size, we implemented two data augmentation methods to improve performance: first by generating synthetic samples via GPT-2, and second by changing the representation of available data via sequence-by-sequence pairing, which is a novel approach.

Results

A BERT-based deep learning model, with GPT-2 synthetic sample augmentation, demonstrated an area-under-curve of 76.3% in the classification of HRQOL accuracy as measured by PCS, compared to the baseline logistic regression and bag-of-words model, which had an AUC of 59.9%. The sequence-by-sequence pairing method for augmentation had an AUC of 71.2% when used with the BERT model.

Conclusions

NLP methods show promise in extracting PRO from unstructured narrative data, and in the future may aid in assessing and forecasting patients’ HRQOL in response to medical treatments. Our experiments with optimization methods suggest larger amounts of novel data would further improve performance of the classification model.