Advancing maxillofacial prosthodontics by using pre-trained convolutional neural networks: Image-based classification of the maxilla

PURPOSE

The study aimed to compare the performance of four pre-trained convolutional neural networks in recognizing seven distinct prosthodontic scenarios involving the maxilla, as a preliminary step in developing an artificial intelligence (AI)-powered prosthesis design system.

MATERIALS AND METHODS

Seven distinct classes, including cleft palate, dentulous maxillectomy, edentulous maxillectomy, reconstructed maxillectomy, completely dentulous, partially edentulous, and completely edentulous, were considered for recognition. Utilizing transfer learning and fine-tuned hyperparameters, four AI models (VGG16, Inception-ResNet-V2, DenseNet-201, and Xception) were employed. The dataset, consisting of 3541 preprocessed intraoral occlusal images, was divided into training, validation, and test sets. Model performance metrics encompassed accuracy, precision, recall, F1 score, area under the receiver operating characteristic curve (AUC), and confusion matrix.

RESULTS

VGG16, Inception-ResNet-V2, DenseNet-201, and Xception demonstrated comparable performance, with maximum test accuracies of 0.92, 0.90, 0.94, and 0.95, respectively. Xception and DenseNet-201 slightly outperformed the other models, particularly compared with InceptionResNet-V2. Precision, recall, and F1 scores exceeded 90% for most classes in Xception and DenseNet-201 and the average AUC values for all models ranged between 0.98 and 1.00.

CONCLUSIONS

While DenseNet-201 and Xception demonstrated superior performance, all models consistently achieved diagnostic accuracy exceeding 90%, highlighting their potential in dental image analysis. This AI application could help work assignments based on difficulty levels and enable the development of an automated diagnosis system at patient admission. It also facilitates prosthesis designing by integrating necessary prosthesis morphology, oral function, and treatment difficulty. Furthermore, it tackles dataset size challenges in model optimization, providing valuable insights for future research.

Deep Learning-Based Automated Magnetic Resonance Image Segmentation of the Lumbar Structure and Its Adjacent Structures at the L4/5 Level

(1) Background: This study aims to develop a deep learning model based on a 3D Deeplab V3+ network to automatically segment multiple structures from magnetic resonance (MR) images at the L4/5 level. (2) Methods: After data preprocessing, the modified 3D Deeplab V3+ network of the deep learning model was used for the automatic segmentation of multiple structures from MR images at the L4/5 level. We performed five-fold cross-validation to evaluate the performance of the deep learning model. Subsequently, the Dice Similarity Coefficient (DSC), precision, and recall were also used to assess the deep learning model's performance. Pearson's correlation coefficient analysis and the Wilcoxon signed-rank test were employed to compare the morphometric measurements of 3D reconstruction models generated by manual and automatic segmentation. (3) Results: The deep learning model obtained an overall average DSC of 0.886, an average precision of 0.899, and an average recall of 0.881 on the test sets. Furthermore, all morphometry-related measurements of 3D reconstruction models revealed no significant difference between ground truth and automatic segmentation. Strong linear relationships and correlations were also obtained in the morphometry-related measurements of 3D reconstruction models between ground truth and automated segmentation. (4) Conclusions: We found it feasible to perform automated segmentation of multiple structures from MR images, which would facilitate lumbar surgical evaluation by establishing 3D reconstruction models at the L4/5 level.

A deep learning model using hyperspectral image for EUS-FNA cytology diagnosis in pancreatic ductal adenocarcinoma

BACKGROUND AND AIMS

Endoscopic ultrasonography-guided fine-needle aspiration/biopsy (EUS-FNA/B) is considered to be a first-line procedure for the pathological diagnosis of pancreatic cancer owing to its high accuracy and low complication rate. The number of new cases of pancreatic ductal adenocarcinoma (PDAC) is increasing, and its accurate pathological diagnosis poses a challenge for cytopathologists. Our aim was to develop a hyperspectral imaging (HSI)-based convolution neural network (CNN) algorithm to aid in the diagnosis of pancreatic EUS-FNA cytology specimens.

METHODS

HSI images were captured of pancreatic EUS-FNA cytological specimens from benign pancreatic tissues (n = 33) and PDAC (n = 39) prepared using a liquid-based cytology method. A CNN was established to test the diagnostic performance, and Attribution Guided Factorization Visualization (AGF-Visualization) was used to visualize the regions of important classification features identified by the model.

RESULTS

A total of 1913 HSI images were obtained. Our ResNet18-SimSiam model achieved an accuracy of 0.9204, sensitivity of 0.9310 and specificity of 0.9123 (area under the curve of 0.9625) when trained on HSI images for the differentiation of PDAC cytological specimens from benign pancreatic cells. AGF-Visualization confirmed that the diagnoses were based on the features of tumor cell nuclei.

CONCLUSIONS

An HSI-based model was developed to diagnose cytological PDAC specimens obtained using EUS-guided sampling. Under the supervision of experienced cytopathologists, we performed multi-staged consecutive in-depth learning of the model. Its superior diagnostic performance could be of value for cytologists when diagnosing PDAC.

Detection of Imperceptible Intervertebral Disc Fissures in Conventional MRI-An AI Strategy for Improved Diagnostics

Annular fissures in the intervertebral discs are believed to be closely related to back pain. However, no sensitive non-invasive method exists to detect annular fissures. This study aimed to propose and test a method capable of detecting the presence and position of annular fissures in conventional magnetic resonance (MR) images non-invasively. The method utilizes textural features calculated from conventional MR images combined with attention mapping and artificial intelligence (AI)-based classification models. As ground truth, reference standard computed tomography (CT) discography was used. One hundred twenty-three intervertebral discs in 43 patients were examined with MR imaging followed by discography and CT. The fissure classification model determined the presence of fissures with 100% sensitivity and 97% specificity. Moreover, the true position of the fissures was correctly determined in 90 (87%) of the analyzed discs. Additionally, the proposed method was significantly more accurate at identifying fissures than the conventional radiological high-intensity zone marker. In conclusion, the findings suggest that the proposed method is a promising diagnostic tool to detect annular fissures of importance for back pain and might aid in clinical practice and allow for new non-invasive research related to the presence and position of individual fissures.

Fuzzy Control of Pressure in a Water Supply Network Based on Neural Network System Modeling and IoT Measurements

As hydroenergetic losses are inherent to water supply systems, they are a frequent issue which water utilities deal with every day. The control of network pressure is essential to reducing these losses, providing a quality supply to consumers, saving electricity and preserving piping from excess pressure. However, to obtain these benefits, it is necessary to overcome some difficulties such as sensing the pressure of geographically distant consumer units and developing a control logic that is capable of making use of the data from these sensors and, at the same time, a good solution in terms of cost benefit. Therefore, this work has the purpose of developing a pressure monitoring and control system for water supply networks, using the ESP8266 microcontroller to collect data from pressure sensors for the integrated ScadaLTS supervisory system via the REST API. The modeling of the plant was developed using artificial neural networks together with fuzzy pressure control, both designed using the Python language. The proposed method was tested by considering a pumping station and two reference units located in the city of João Pessoa, Brazil, in which there was an excess of pressure in the supply network and low performance from the old controls, during the night period from 12:00 a.m. to 6:00 a.m. The field results estimated 2.9% energy saving in relation to the previous form of control and a guarantee that the pressure in the network was at a healthy level.

Development, Validation and Comparison of Artificial Neural Network and Logistic Regression Models Predicting Eosinophilic Chronic Rhinosinusitis With Nasal Polyps

PURPOSE

Chronic rhinosinusitis with nasal polyps (CRSwNP) can be classified into eosinophilic CRSwNP (eCRSwNP) and non-eosinophilic CRSwNP (non-eCRSwNP) by tissue biopsy, which is difficult to perform preoperatively. Clinical biomarkers have predictive value for the classification of CRSwNP. We aimed to evaluate the application of artificial neural network (ANN) modeling in distinguishing different endotypes of CRSwNP based on clinical biomarkers.

METHODS

Clinical parameters were collected from 109 CRSwNP patients, and their predictive ability was analyzed. ANN and logistic regression (LR) models were developed in the training group (72 patients) and further tested in the test group (37 patients). The output variable was the diagnosis of eCRSwNP, defined as tissue eosinophil count > 10 per high-power field. The receiver operating characteristics curve was used to assess model performance.

RESULTS

A total of 15 clinical features from 60 healthy controls, 60 eCRSwNP and 49 non-eCRSwNP were selected as candidate predictors. Nasal nitric oxide levels, peripheral eosinophil absolute count, total immunoglobulin E, and ratio of bilateral computed tomography scores for the ethmoid sinus and maxillary sinus were identified as important features for modeling. Two ANN models based on 4 and 15 clinical features were developed to predict eCRSwNP, which showed better performance, with the area under the receiver operator characteristics significantly higher than those from the respective LR models (0.976 vs. 0.902, P = 0.048; 0.970 vs. 0.845, P = 0.011). All ANN models had better fits than single variable prediction models (all P < 0.05), and ANN model 1 had the best predictive performance among all models.

CONCLUSIONS

Machine learning models assist clinicians in predicting endotypes of nasal polyps before invasive detection. The ANN model has the potential to predict eCRSwNP with high sensitivity and specificity, and is superior to the LR model. ANNs are valuable for optimizing personalized patient management.

Performance of deep convolutional neural network for classification and detection of oral potentially malignant disorders in photographic images

Oral potentially malignant disorders (OPMDs) are a group of conditions that can transform into oral cancer. The purpose of this study was to evaluate convolutional neural network (CNN) algorithms to classify and detect OPMDs in oral photographs. In this study, 600 oral photograph images were collected retrospectively and grouped into 300 images of OPMDs and 300 images of normal oral mucosa. CNN-based classification models were created using DenseNet-121 and ResNet-50. The detection models were created using Faster R-CNN and YOLOv4. The image data were randomly selected and assigned as training, validating, and testing data. The testing data were evaluated to compare the performance of the CNN models with the diagnosis results produced by oral and maxillofacial surgeons. DenseNet-121 and ResNet-50 were found to produce high efficiency in diagnosis of OPMDs, with an area under the receiver operating characteristic curve (AUC) of 95%. Faster R-CNN yielded the highest detection performance, with an AUC of 74.34%. For the CNN-based classification model, the sensitivity and specificity were 100% and 90%, respectively. For the oral and maxillofacial surgeons, these values were 91.73% and 92.27%, respectively. In conclusion, the DenseNet-121, ResNet-50 and Faster R-CNN models have potential for the classification and detection of OPMDs in oral photographs.

Effective Connectivity in the Human Brain for Sour Taste, Retronasal Smell, and Combined Flavour

The anterior insula and rolandic operculum are key regions for flavour perception in the human brain; however, it is unclear how taste and congruent retronasal smell are perceived as flavours. The multisensory integration required for sour flavour perception has rarely been studied; therefore, we investigated the brain responses to taste and smell in the sour flavour-processing network in 35 young healthy adults. We aimed to characterise the brain response to three stimulations applied in the oral cavity—sour taste, retronasal smell of mango, and combined flavour of both—using functional magnetic resonance imaging. Effective connectivity of the flavour-processing network and modulatory effect from taste and smell were analysed. Flavour stimulation activated middle insula and olfactory tubercle (primary taste and olfactory cortices, respectively); anterior insula and rolandic operculum, which are associated with multisensory integration; and ventrolateral prefrontal cortex, a secondary cortex for flavour perception. Dynamic causal modelling demonstrated that neural taste and smell signals were integrated at anterior insula and rolandic operculum. These findings elucidated how neural signals triggered by sour taste and smell presented in liquid form interact in the brain, which may underpin the neurobiology of food appreciation. Our study thus demonstrated the integration and synergy of taste and smell.

Attractor competition enriches cortical dynamics during awakening from anesthesia

Slow oscillations (≲ 1 Hz), a hallmark of slow-wave sleep and deep anesthesia across species, arise from spatiotemporal patterns of activity whose complexity increases as wakefulness is approached and cognitive functions emerge. The arousal process constitutes an open window to the unknown mechanisms underlying the emergence of such dynamical richness in awake cortical networks. Here, we investigate the changes in network dynamics as anesthesia fades out in the rat visual cortex. Starting from deep anesthesia, slow oscillations gradually increase their frequency, eventually expressing maximum regularity. This stage is followed by the abrupt onset of an infra-slow (~0.2 Hz) alternation between sleep-like oscillations and activated states. A population rate model reproduces this transition driven by an increased excitability that brings it to periodically cross a critical point. Based on our model, dynamical richness emerges as a competition between two metastable attractor states, a conclusion strongly supported by the data.

Designing Trojan Detectors in Neural Networks Using Interactive Simulations

This paper addresses the problem of designing trojan detectors in neural networks (NNs) using interactive simulations. Trojans in NNs are defined as triggers in inputs that cause misclassification of such inputs into a class (or classes) unintended by the design of a NN-based model. The goal of our work is to understand encodings of a variety of trojan types in fully connected layers of neural networks. Our approach is (1) to simulate nine types of trojan embeddings into dot patterns, (2) to devise measurements of NN states, and (3) to design trojan detectors in NN-based classification models. The interactive simulations are built on top of TensorFlow Playground with in-memory storage of data and NN coefficients. The simulations provide analytical, visualization, and output operations performed on training datasets and NN architectures. The measurements of a NN include (a) model inefficiency using modified Kullback-Liebler (KL) divergence from uniformly distributed states and (b) model sensitivity to variables related to data and NNs. Using the KL divergence measurements at each NN layer and per each predicted class label, a trojan detector is devised to discriminate NN models with or without trojans. To document robustness of such a trojan detector with respect to NN architectures, dataset perturbations, and trojan types, several properties of the KL divergence measurement are presented. For the general use, the web-based simulations is deployed via GitHub pages at https://github.com/usnistgov/nn-calculator.

Finding Distributed Needles in Neural Haystacks

The human cortex encodes information in complex networks that can be anatomically dispersed and variable in their microstructure across individuals. Using simulations with neural network models, we show that contemporary statistical methods for functional brain imaging-including univariate contrast, searchlight multivariate pattern classification, and whole-brain decoding with L1 or L2 regularization-each have critical and complementary blind spots under these conditions. We then introduce the sparse-overlapping-sets (SOS) LASSO-a whole-brain multivariate approach that exploits structured sparsity to find network-distributed information-and show in simulation that it captures the advantages of other approaches while avoiding their limitations. When applied to fMRI data to find neural responses that discriminate visually presented faces from other visual stimuli, each method yields a different result, but existing approaches all support the canonical view that face perception engages localized areas in posterior occipital and temporal regions. In contrast, SOS LASSO uncovers a network spanning all four lobes of the brain. The result cannot reflect spurious selection of out-of-system areas because decoding accuracy remains exceedingly high even when canonical face and place systems are removed from the dataset. When used to discriminate visual scenes from other stimuli, the same approach reveals a localized signal consistent with other methods-illustrating that SOS LASSO can detect both widely distributed and localized representational structure. Thus, structured sparsity can provide an unbiased method for testing claims of functional localization. For faces and possibly other domains, such decoding may reveal representations more widely distributed than previously suspected.SIGNIFICANCE STATEMENT Brain systems represent information as patterns of activation over neural populations connected in networks that can be widely distributed anatomically, variable across individuals, and intermingled with other networks. We show that four widespread statistical approaches to functional brain imaging have critical blind spots in this scenario and use simulations with neural network models to illustrate why. We then introduce a new approach designed specifically to find radically distributed representations in neural networks. In simulation and in fMRI data collected in the well studied domain of face perception, the new approach discovers extensive signal missed by the other methods-suggesting that prior functional imaging work may have significantly underestimated the degree to which neurocognitive representations are distributed and variable across individuals.

Nonlinear Control in the Nematode C. elegans

Recent whole-brain calcium imaging recordings of the nematode C. elegans have demonstrated that the neural activity associated with behavior is dominated by dynamics on a low-dimensional manifold that can be clustered according to behavioral states. Previous models of C. elegans dynamics have either been linear models, which cannot support the existence of multiple fixed points in the system, or Markov-switching models, which do not describe how control signals in C. elegans neural dynamics can produce switches between stable states. It remains unclear how a network of neurons can produce fast and slow timescale dynamics that control transitions between stable states in a single model. We propose a global, nonlinear control model which is minimally parameterized and captures the state transitions described by Markov-switching models with a single dynamical system. The model is fit by reproducing the timeseries of the dominant PCA mode in the calcium imaging data. Long and short time-scale changes in transition statistics can be characterized via changes in a single parameter in the control model. Some of these macro-scale transitions have experimental correlates to single neuro-modulators that seem to act as biological controls, allowing this model to generate testable hypotheses about the effect of these neuro-modulators on the global dynamics. The theory provides an elegant characterization of control in the neuron population dynamics in C. elegans. Moreover, the mathematical structure of the nonlinear control framework provides a paradigm that can be generalized to more complex systems with an arbitrary number of behavioral states.

Population coupling predicts the plasticity of stimulus responses in cortical circuits

Some neurons have stimulus responses that are stable over days, whereas other neurons have highly plastic stimulus responses. Using a recurrent network model, we explore whether this could be due to an underlying diversity in their synaptic plasticity. We find that, in a network with diverse learning rates, neurons with fast rates are more coupled to population activity than neurons with slow rates. This plasticity-coupling link predicts that neurons with high population coupling exhibit more long-term stimulus response variability than neurons with low population coupling. We substantiate this prediction using recordings from the Allen Brain Observatory, finding that a neuron's population coupling is correlated with the plasticity of its orientation preference. Simulations of a simple perceptual learning task suggest a particular functional architecture: a stable 'backbone' of stimulus representation formed by neurons with low population coupling, on top of which lies a flexible substrate of neurons with high population coupling.

Differentiation of Small (≤ 4 cm) Renal Masses on Multiphase Contrast-Enhanced CT by Deep Learning

OBJECTIVE. This study evaluated the utility of a deep learning method for determining whether a small (≤ 4 cm) solid renal mass was benign or malignant on multiphase contrast-enhanced CT. MATERIALS AND METHODS. This retrospective study included 1807 image sets from 168 pathologically diagnosed small (≤ 4 cm) solid renal masses with four CT phases (unenhanced, corticomedullary, nephrogenic, and excretory) in 159 patients between 2012 and 2016. Masses were classified as malignant (n = 136) or benign (n = 32). The dataset was randomly divided into five subsets: four were used for augmentation and supervised training (48,832 images), and one was used for testing (281 images). The Inception-v3 architecture convolutional neural network (CNN) model was used. The AUC for malignancy and accuracy at optimal cutoff values of output data were evaluated in six different CNN models. Multivariate logistic regression analysis was also performed. RESULTS. Malignant and benign lesions showed no significant difference of size. The AUC value of corticomedullary phase was higher than that of other phases (corticomedullary vs excretory, p = 0.022). The highest accuracy (88%) was achieved in corticomedullary phase images. Multivariate analysis revealed that the CNN model of corticomedullary phase was a significant predictor for malignancy compared with other CNN models, age, sex, and lesion size. CONCLUSION. A deep learning method with a CNN allowed acceptable differentiation of small (≤ 4 cm) solid renal masses in dynamic CT images, especially in the corticomedullary image model.

Bridging Single Neuron Dynamics to Global Brain States

Biological neural networks produce information backgrounds of multi-scale spontaneous activity that become more complex in brain states displaying higher capacities for cognition, for instance, attentive awake versus asleep or anesthetized states. Here, we review brain state-dependent mechanisms spanning ion channel currents (microscale) to the dynamics of brain-wide, distributed, transient functional assemblies (macroscale). Not unlike how microscopic interactions between molecules underlie structures formed in macroscopic states of matter, using statistical physics, the dynamics of microscopic neural phenomena can be linked to macroscopic brain dynamics through mesoscopic scales. Beyond spontaneous dynamics, it is observed that stimuli evoke collapses of complexity, most remarkable over high dimensional, asynchronous, irregular background dynamics during consciousness. In contrast, complexity may not be further collapsed beyond synchrony and regularity characteristic of unconscious spontaneous activity. We propose that increased dimensionality of spontaneous dynamics during conscious states supports responsiveness, enhancing neural networks' emergent capacity to robustly encode information over multiple scales.

Predicting the Functional Impact of KCNQ1 Variants of Unknown Significance

BACKGROUND

An emerging standard-of-care for long-QT syndrome uses clinical genetic testing to identify genetic variants of the KCNQ1 potassium channel. However, interpreting results from genetic testing is confounded by the presence of variants of unknown significance for which there is inadequate evidence of pathogenicity.

METHODS AND RESULTS

In this study, we curated from the literature a high-quality set of 107 functionally characterized KCNQ1 variants. Based on this data set, we completed a detailed quantitative analysis on the sequence conservation patterns of subdomains of KCNQ1 and the distribution of pathogenic variants therein. We found that conserved subdomains generally are critical for channel function and are enriched with dysfunctional variants. Using this experimentally validated data set, we trained a neural network, designated Q1VarPred, specifically for predicting the functional impact of KCNQ1 variants of unknown significance. The estimated predictive performance of Q1VarPred in terms of Matthew's correlation coefficient and area under the receiver operating characteristic curve were 0.581 and 0.884, respectively, superior to the performance of 8 previous methods tested in parallel. Q1VarPred is publicly available as a web server at http://meilerlab.org/q1varpred.

CONCLUSIONS

Although a plethora of tools are available for making pathogenicity predictions over a genome-wide scale, previous tools fail to perform in a robust manner when applied to KCNQ1. The contrasting and favorable results for Q1VarPred suggest a promising approach, where a machine-learning algorithm is tailored to a specific protein target and trained with a functionally validated data set to calibrate informatics tools.

Artificial neural networks predict the incidence of portosplenomesenteric venous thrombosis in patients with acute pancreatitis

Essentials Predicting the occurrence of portosplenomesenteric vein thrombosis (PSMVT) is difficult. We studied 72 patients with acute pancreatitis. Artificial neural networks modeling was more accurate than logistic regression in predicting PSMVT. Additional predictive factors may be incorporated into artificial neural networks.

SUMMARY

Objective To construct and validate artificial neural networks (ANNs) for predicting the occurrence of portosplenomesenteric venous thrombosis (PSMVT) and compare the predictive ability of the ANNs with that of logistic regression. Methods The ANNs and logistic regression modeling were constructed using simple clinical and laboratory data of 72 acute pancreatitis (AP) patients. The ANNs and logistic modeling were first trained on 48 randomly chosen patients and validated on the remaining 24 patients. The accuracy and the performance characteristics were compared between these two approaches by SPSS17.0 software. Results The training set and validation set did not differ on any of the 11 variables. After training, the back propagation network training error converged to 1 × 10-20 , and it retained excellent pattern recognition ability. When the ANNs model was applied to the validation set, it revealed a sensitivity of 80%, specificity of 85.7%, a positive predictive value of 77.6% and negative predictive value of 90.7%. The accuracy was 83.3%. Differences could be found between ANNs modeling and logistic regression modeling in these parameters (10.0% [95% CI, -14.3 to 34.3%], 14.3% [95% CI, -8.6 to 37.2%], 15.7% [95% CI, -9.9 to 41.3%], 11.8% [95% CI, -8.2 to 31.8%], 22.6% [95% CI, -1.9 to 47.1%], respectively). When ANNs modeling was used to identify PSMVT, the area under receiver operating characteristic curve was 0.849 (95% CI, 0.807-0.901), which demonstrated better overall properties than logistic regression modeling (AUC = 0.716) (95% CI, 0.679-0.761). Conclusions ANNs modeling was a more accurate tool than logistic regression in predicting the occurrence of PSMVT following AP. More clinical factors or biomarkers may be incorporated into ANNs modeling to improve its predictive ability.

Cost minimization using an artificial neural network sleep apnea prediction tool for sleep studies

RATIONALE

More than a million polysomnograms (PSGs) are performed annually in the United States to diagnose obstructive sleep apnea (OSA). Third-party payers now advocate a home sleep test (HST), rather than an in-laboratory PSG, as the diagnostic study for OSA regardless of clinical probability, but the economic benefit of this approach is not known.

OBJECTIVES

We determined the diagnostic performance of OSA prediction tools including the newly developed OSUNet, based on an artificial neural network, and performed a cost-minimization analysis when the prediction tools are used to identify patients who should undergo HST.

METHODS

The OSUNet was trained to predict the presence of OSA in a derivation group of patients who underwent an in-laboratory PSG (n = 383). Validation group 1 consisted of in-laboratory PSG patients (n = 149). The network was trained further in 33 patients who underwent HST and then was validated in a separate group of 100 HST patients (validation group 2). Likelihood ratios (LRs) were compared with two previously published prediction tools. The total costs from the use of the three prediction tools and the third-party approach within a clinical algorithm were compared.

MEASUREMENTS AND MAIN RESULTS

The OSUNet had a higher +LR in all groups compared with the STOP-BANG and the modified neck circumference (MNC) prediction tools. The +LRs for STOP-BANG, MNC, and OSUNet in validation group 1 were 1.1 (1.0-1.2), 1.3 (1.1-1.5), and 2.1 (1.4-3.1); and in validation group 2 they were 1.4 (1.1-1.7), 1.7 (1.3-2.2), and 3.4 (1.8-6.1), respectively. With an OSA prevalence less than 52%, the use of all three clinical prediction tools resulted in cost savings compared with the third-party approach.

CONCLUSIONS

The routine requirement of an HST to diagnose OSA regardless of clinical probability is more costly compared with the use of OSA clinical prediction tools that identify patients who should undergo this procedure when OSA is expected to be present in less than half of the population. With OSA prevalence less than 40%, the OSUNet offers the greatest savings, which are substantial when the number of sleep studies done annually is considered.