Artificial Intelligence Deep Learning Research Articles

Deep Learning for Breast Cancer Risk Prediction: Application to a Large Representative UK Screening Cohort.

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To develop an artificial intelligence (AI) deep learning tool capable of predicting future breast cancer risk from a current negative screening mammographic examination and to evaluate the model on data from the UK National Health Service Breast Screening Program. Materials and Methods The OPTIMAM Mammography Imaging Database contains screening data, including mammograms and information on interval cancers, for > 300,000 women who attended screening at three different sites in the UK from 2012 onward. Cancer-free screening examinations from women aged 50-70 years were obtained and classified as risk-positive or risk-negative based on the occurrence of cancer within 3 years of the original examination. Examinations with confirmed cancer and images containing implants were excluded. From the resulting 5264 risk-positive and 191488 risk-negative examinations, training (n = 89285) validation (n = 2106) and test (n = 39351) datasets were produced for model development and evaluation. The AI model was trained to predict future cancer occurrence based on screening mammograms and patient age. Performance was evaluated on the test dataset using the area under the receiver operating characteristic curve (AUC) and compared across subpopulations to assess potential biases. Interpretability of the model was explored, including with saliency maps. Results On the hold-out test set, the AI model achieved an overall AUC of 0.70 (95% CI: 0.69, 0.72). There was no evidence of a difference in performance across the three sites, between patient ethnicities or across age-groups Visualization of saliency maps and sample images provided insights into the mammographic features associated with AI-predicted cancer risk. Conclusion The developed AI tool showed good performance on a multisite, UK-specific dataset. ©RSNA, 2024.

Product Design Interaction and Experience Based on Virtual Reality Technology

Artificial intelligence involves imitating human thought, consciousness, and other aspects. This is similar to having machine with human brain that can think, produce independently. However, unlike human brain, it has a speed, memory advantage. A virtual reality interactive glove is built using a nine-axis inertial sensor, an artificial intelligence deep learning algorithm. In this research work, Product Design Interaction and Experience Based on Virtual Reality Technology (PDIE-VRT-SCGAN) were proposed. Initially, input gestures data are gathered from the Virtual Reality Experiences (VRE) dataset. The input gestures data is then pre-processed using Multi-Window Savitzky-Golay Filter (MWSGF) to reduce noises, increase overall quality of the gestures data. In order to improving overall user engagement in product design interactions on virtual reality (VR) technology, the pre-processed gestures data are then fed into an adversarial network called a Semi-Cycled Generative Adversarial Network (SCGAN). In general, SCGAN does not express some adaption of optimization strategies for determining optimal parameters to promise exact to improving overall user engagement in product design interactions using VR technology. Therefore, FOX-inspired Optimization (FIO) is proposed to enhance weight parameter of SCGAN method, which precisely improving the user experience in product design interaction. The efficacy of PDIE-VRT-SCGAN method is assessed using a number of performance criteria, including tracking accuracy, frame rate, latency, rendering time, error rate, and user error. The proposed PDIE-VRT-SCGAN method attains 22.36%, 25.42% and 18.17% higher tracking accuracy, 21.26%, 15.42% and 19.27% higher latency, 28.36%, 25.32% and 28.27% higher frame rate compared with existing methods, such as design and implementation of virtual reality interactive product software depend on artificial intelligence deep learning algorithm(DVRI-PS-AI-DL), virtual evaluation system for product designing utilizing virtual reality (VES-PD-VR), and analysis of unsatisfying user experiences with unmet psychological needs for virtual reality exergames utilizing deep learning approach (AUUE-UP-VRE-DLA) respectively.

A reliable deep-learning-based method for alveolar bone quantification using a murine model of periodontitis and micro-computed tomography imaging

Objectives: This study focuses on artificial intelligence (AI)-assisted analysis of alveolar bone for periodontitis in a mouse model with the aim to create an automatic deep-learning segmentation model that enables researchers to easily examine alveolar bone from micro-computed tomography (µCT) data without needing prior machine learning knowledge.Methods: Ligature-induced experimental periodontitis was produced by placing a small-diameter silk sling ligature around the left maxillary second molar. At 4, 7, 9, or 14 days, the maxillary bone was harvested and processed with a µCT scanner (µCT-45, Scanco). Using Dragonfly (v2021.3), we developed a 3D deep learning model based on the U-Net AI deep learning engine for segmenting materials in complex images to measure alveolar bone volume (BV) and bone mineral density (BMD) while excluding the teeth from the measurements.Results: This model generates 3D segmentation output for a selected region of interest with over 98% accuracy on different formats of µCT data. BV on the ligature side gradually decreased from 0.87 mm3 to 0.50 mm3 on day 9 and then increased to 0.63 mm3 on day 14. The ligature side lost 4.6% of BMD on day 4, 9.6% on day 7, 17.7% on day 9, and 21.1% on day 14.Conclusions: This study developed an AI model that can be downloaded and easily applied, allowing researchers to assess metrics including BV, BMD, and trabecular bone thickness, while excluding teeth from the measurements of mouse alveolar bone.Clinical significance: This work offers an innovative, user-friendly automatic segmentation model that is fast, accurate, and reliable, demonstrating new potential uses of artificial intelligence (AI) in dentistry with great potential in diagnosing, treating, and prognosis of oral diseases.

Hypertrophic cardiomyopathy detection with artificial intelligence electrocardiography in international cohorts: an external validation study

Abstract Aims Recently, deep learning artificial intelligence (AI) models have been trained to detect cardiovascular conditions, including hypertrophic cardiomyopathy (HCM), from the 12-lead electrocardiogram (ECG). In this external validation study, we sought to assess the performance of an AI-ECG algorithm for detecting HCM in diverse international cohorts. Methods and results A convolutional neural network-based AI-ECG algorithm was developed previously in a single-centre North American HCM cohort (Mayo Clinic). This algorithm was applied to the raw 12-lead ECG data of patients with HCM and non-HCM controls from three external cohorts (Bern, Switzerland; Oxford, UK; and Seoul, South Korea). The algorithm’s ability to distinguish HCM vs. non-HCM status from the ECG alone was examined. A total of 773 patients with HCM and 3867 non-HCM controls were included across three sites in the merged external validation cohort. The HCM study sample comprised 54.6% East Asian, 43.2% White, and 2.2% Black patients. Median AI-ECG probabilities of HCM were 85% for patients with HCM and 0.3% for controls (P < 0.001). Overall, the AI-ECG algorithm had an area under the receiver operating characteristic curve (AUC) of 0.922 [95% confidence interval (CI) 0.910–0.934], with diagnostic accuracy 86.9%, sensitivity 82.8%, and specificity 87.7% for HCM detection. In age- and sex-matched analysis (case–control ratio 1:2), the AUC was 0.921 (95% CI 0.909–0.934) with accuracy 88.5%, sensitivity 82.8%, and specificity 90.4%. Conclusion The AI-ECG algorithm determined HCM status from the 12-lead ECG with high accuracy in diverse international cohorts, providing evidence for external validity. The value of this algorithm in improving HCM detection in clinical practice and screening settings requires prospective evaluation.

Prospective validation of a seizure diary forecasting falls short.

Recently, a deep learning artificial intelligence (AI) model forecasted seizure risk using retrospective seizure diaries with higher accuracy than random forecasts. The present study sought to prospectively evaluate the same algorithm. We recruited a prospective cohort of 46 people with epilepsy; 25 completed sufficient data entry for analysis (median = 5 months). We used the same AI method as in our prior study. Group-level and individual-level Brier Skill Scores (BSSs) compared random forecasts and simple moving average forecasts to the AI. The AI had an area under the receiver operating characteristic curve of .82. At the group level, the AI outperformed random forecasting (BSS = .53). At the individual level, AI outperformed random in 28% of cases. At the group and individual level, the moving average outperformed the AI. If pre-enrollment (nonverified) diaries (with presumed underreporting) were included, the AI significantly outperformed both comparators. Surveys showed most did not mind poor-quality LOW-RISK or HIGH-RISK forecasts, yet 91% wanted access to these forecasts. The previously developed AI forecasting tool did not outperform a very simple moving average forecasting in this prospective cohort, suggesting that the AI model should be replaced.

AI-driven remote sensing enhances Mediterranean seagrass monitoring and conservation to combat climate change and anthropogenic impacts

Seagrasses are undergoing widespread loss due to anthropogenic pressure and climate change. Since 1960, the Mediterranean seascape lost 13–50% of the areal extent of its dominant and endemic seagrass-Posidonia oceanica, which regulates its ecosystem. Many conservation and restoration projects failed due to poor site selection and lack of long-term monitoring. Here, we present a fast and efficient operational approach based on a deep-learning artificial intelligence model using Sentinel-2 data to map the spatial extent of the meadows, enabling short and long-term monitoring, and identifying the impacts of natural and human-induced stressors and changes at different timescales. We apply ACOLITE atmospheric correction to the satellite data and use the output to train the model along with the ancillary data and therefore, map the extent of the meadows. We apply noise-removing filters to enhance the map quality. We obtain 74–92% of overall accuracy, 72–91% of user’s accuracy, and 81–92% of producer’s accuracy, where high accuracies are observed at 0–25 m depth. Our model is easily adaptable to other regions and can produce maps in in-situ data-scarce regions, providing a first-hand overview. Our approach can be a support to the Mediterranean Posidonia Network, which brings together different stakeholders such as authorities, scientists, international environmental organizations, professionals including yachting agents and marinas from the Mediterranean countries to protect all P. oceanica meadows in the Mediterranean Sea by 2030 and increase each country’s capability to protect these meadows by providing accurate and up-to-date maps to prevent its future degradation.

A scoping review of literature on deep learning and symbolic AI-based framework for detecting Covid-19 using computerized tomography scans

This scoping review aims to explore various Deep Learning and Symbolic Artificial Intelligence (AI) models that can be integrated into explainable hybrid AI for the purpose of detecting COVID-19 based on Computerized Tomography (CT) scans. We followed the PRISMA-ScR framework as the foundation for our scoping review protocol. Our approach included a thorough search across 13 databases, complemented by an additional random internet search for relevant articles. Due to the voluminous number of articles returned, the search was further narrowed using the keywords: Deep Learning, Symbolic AI and Hybrid AI. These keywords were used because they are more visible in the earmarked literature. A screening of all articles by title was performed to remove duplicates. The final screening process centered on the publication year, ensuring that all considered articles fell within the range of 2019 to 2023, inclusive. Subsequently, abstract or text synthesis was conducted. Our search query retrieved a total of 3,312 potential articles from the thirteen databases, and an additional 12 articles from a random internet search, resulting in a cumulative count of 3,324 identified articles. After the deduplication and screening steps, 260 articles met our inclusion criteria. These articles were categorized based on the year of publication, the type of aim, and the type of AI used. An analysis of the year of publication revealed a linear trend, indicating growth in the field of Hybrid AI. Out of the five aim categories identified, we deliberately excluded articles that lacked a specified aim. It's noteworthy that 3% of the articles focused on the integration of AI models. The low percentage value suggests that the integration aspect is overlooked, thereby transcripting the integration of Deep Learning and symbolic AI into hybrid AI as an area worth exploring. This scoping review gives an overview of how a Deep Learning and Symbolic AI-based framework has been used in the detection of COVID-19 based on CT scans.

Impact of AI for Digital Breast Tomosynthesis on Breast Cancer Detection and Interpretation Time.

Purpose To develop an artificial intelligence (AI) model for the diagnosis of breast cancer on digital breast tomosynthesis (DBT) images and to investigate whether it could improve diagnostic accuracy and reduce radiologist reading time. Materials and Methods A deep learning AI algorithm was developed and validated for DBT with retrospectively collected examinations (January 2010 to December 2021) from 14 institutions in the United States and South Korea. A multicenter reader study was performed to compare the performance of 15 radiologists (seven breast specialists, eight general radiologists) in interpreting DBT examinations in 258 women (mean age, 56 years ± 13.41 [SD]), including 65 cancer cases, with and without the use of AI. Area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and reading time were evaluated. Results The AUC for stand-alone AI performance was 0.93 (95% CI: 0.92, 0.94). With AI, radiologists' AUC improved from 0.90 (95% CI: 0.86, 0.93) to 0.92 (95% CI: 0.88, 0.96) (P = .003) in the reader study. AI showed higher specificity (89.64% [95% CI: 85.34%, 93.94%]) than radiologists (77.34% [95% CI: 75.82%, 78.87%]) (P < .001). When reading with AI, radiologists' sensitivity increased from 85.44% (95% CI: 83.22%, 87.65%) to 87.69% (95% CI: 85.63%, 89.75%) (P = .04), with no evidence of a difference in specificity. Reading time decreased from 54.41 seconds (95% CI: 52.56, 56.27) without AI to 48.52 seconds (95% CI: 46.79, 50.25) with AI (P < .001). Interreader agreement measured by Fleiss κ increased from 0.59 to 0.62. Conclusion The AI model showed better diagnostic accuracy than radiologists in breast cancer detection, as well as reduced reading times. The concurrent use of AI in DBT interpretation could improve both accuracy and efficiency. Keywords: Breast, Computer-Aided Diagnosis (CAD), Tomosynthesis, Artificial Intelligence, Digital Breast Tomosynthesis, Breast Cancer, Computer-Aided Detection, Screening Supplemental material is available for this article. © RSNA, 2024 See also the commentary by Bae in this issue.

Deep learning and capsule endoscopy: Automatic multi-brand and multi-device panendoscopic detection of vascular lesions.

Background and study aims Capsule endoscopy (CE) is commonly used as the initial exam for suspected mid-gastrointestinal bleeding after normal upper and lower endoscopy. Although the assessment of the small bowel is the primary focus of CE, detecting upstream or downstream vascular lesions may also be clinically significant. This study aimed to develop and test a convolutional neural network (CNN)-based model for panendoscopic automatic detection of vascular lesions during CE. Patients and methods A multicentric AI model development study was based on 1022 CE exams. Our group used 34655 frames from seven types of CE devices, of which 11091 were considered to have vascular lesions (angiectasia or varices) after triple validation. We divided data into a training and a validation set, and the latter was used to evaluate the model's performance. At the time of division, all frames from a given patient were assigned to the same dataset. Our primary outcome measures were sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), and an area under the precision-recall curve (AUC-PR). Results Sensitivity and specificity were 86.4% and 98.3%, respectively. PPV was 95.2%, while the NPV was 95.0%. Overall accuracy was 95.0%. The AUC-PR value was 0.96. The CNN processed 115 frames per second. Conclusions This is the first proof-of-concept artificial intelligence deep learning model developed for pan-endoscopic automatic detection of vascular lesions during CE. The diagnostic performance of this CNN in multi-brand devices addresses an essential issue of technological interoperability, allowing it to be replicated in multiple technological settings.

Abstract 4915: Identification and validation of novel prognostic genetic markers in HER2-negative advanced gastric cancer (AGC) by artificial intelligence (AI) deep learning and machine learning algorithm

Abstract Introduction: Despite advances in anticancer treatments, there has been urgent need to identify novel prognostic markers for HER2-negative AGC. We propose AI based deep learning and machine learning pipeline to explore genetic markers that predict patient response to chemotherapy in HER2-negative AGC. Methods: We retrospectively assessed 179 stage IV HER2-negative AGC patients who have been treated with first-line chemotherapy in Yonsei Cancer Center, Korea between 2015 and 2021. In-house targeted sequencing panel data (CancerSCANTM and CancerMaster) were used to identify candidate prognostic markers. DeepSurv was mainly used to analyze the progression-free survival (PFS), which is a deep learning algorithm that can investigate prognostic roles. The SHapley Additive exPlanations (SHAP) method was applied to open the deep learning models’ black-box and calculate the importance ranking of gene features. Machine learning models-random survival forest, elastic-net, and lasso-were also investigated to identify potentially meaningful gene variants. To rank the gene features, we calculated variable importance (VIMP) for random survival forest, and penalized regression coefficients for elastic-net and lasso. All deep learning and machine learning models were trained using 5-folds cross-validation in 1,000 iterations of re-sampling bootstrapping with hyperparameter tuning by grid search. To classify responder or non-responder related genetic markers to chemotherapy, we used the signs of averaged penalize coefficients obtained from elastic-net and lasso. Based on the median linear predictor with candidate gene features, patients were classified into high-risk and low-risk group. Results: A total of 7,824 common variants were analyzed and 20 prognostic genetic markers were identified based on the top average values of SHAP, VIMP, and penalized regression coefficients. The top 10 non-responder genetic markers are included as follows: PARP4(p.S873N), CHUK(p.V268I), BARD1(359_365del), CREBBP(p.Y1125F), BARD1(p.P24S), PHLPP2(p.R1312Q), FAT3(p.Q3375R), CASP5(p.T106A), PHLPP2(p.I544V), and PTCH2(p.T988M). Otherwise, the top 10 responder genetic markers are defined as follows: IL7R(p.I66T), NOTCH4(L16delinsLL), FGFR3, BCL2A1(p.G82D), ZNF217(p.T548I), BRCA1(p.K1183R), FANCA(p.A412V), ADGRA2, CDKN2B, and FANCM(p.S175F). The high-risk group had worse PFS compared with low-risk group (median PFS 4.1 vs. 9.1 months; P-value &amp;lt;0.0001). Conclusion: This study showed the potential of AI deep learning and machine learning pipeline that employs an innovate prognostic genetic markers for HER2-negative AGC. Current ongoing analyses of larger cohort with more comprehensive clinical data will be presented. Citation Format: Sejung Park, Seok-Jae Heo, Choong-Kun Lee, Yaeji Lee, Woo Sun Kwon, Jingmin Che, Minseok Kim, Hyun Cheol Chung, Sun Young Rha. Identification and validation of novel prognostic genetic markers in HER2-negative advanced gastric cancer (AGC) by artificial intelligence (AI) deep learning and machine learning algorithm [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4915.

Abstract 299: FAP/αSMA association with tumor visibility and progression in prostate cancer

Abstract Introduction Prostate cancer (PCa) ranks as the second most frequently diagnosed cancer in the male population worldwide. The challenge in effectively managing PCa is the absence of well-established biological markers, which often leads to both over and under treatment. To address this issue, a deeper understanding of the molecular alterations within PCa is essential for more accurate patient stratification. In this research, we embark on a comprehensive examination of how the presence of fibroblast activation protein (FAP) and alpha smooth muscle actin (αSMA) in PCa tissue associates with the visibility of the disease in Magnetic Resonance Imaging (MRI) and their impact on cancer progression. The objective of this study is to analyze cell compartmental expression of FAP and aSMA using an AI based image analysis method. We also study the role of FAP and αSMA expression in predicting disease progression and mortality. Materials and Methods We applied FAP (ab207178, Abcam)/αSMA (M0851, DAKO) double chromogen immunohistochemical (IHC) staining to three independent cohorts of formalin-fixed paraffin-embedded (FFPE) prostate cancer tissue microarrays. All together, we analyzed 835 patients’ samples. First cohort consists of 387 patients with preoperative multiparametric MRI and robot-assisted laparoscopic prostatectomy (RALP) as primary treatment. Second cohort was case vs control cohort of Grade group 2-4 localized PCa at radical prostatectomy (RP) with 168 patients. Third cohort consists PCa patents from continuous, population-based collection of radical prostatectomies of 319 patients. Image analysis was performed with Aiforia Create, which is cloud-based deep-learning artificial intelligence software that provides platform for training and validating automated image analysis pipelines using convolutional neural networks. Analysis algorithm in this project was based on semantic, pixel-level area segmentation. Results FAP expression was higher and αSMA lower in MRI-visible tumors compared with MRI-invisible tumors. Stromal and epithelial FAP expression associate with tumor progression. We could also see statistically significant differences in FAP positivity between clinically significant (GS ≥7) and insignificant (GS ≤6) tumors. Hazard ratios imply that FAP and αSMA have significant impact on risk of biochemical recurrence in patients with MRI-visible tumors. Conclusions FAP is closely linked to tumor MRI visibility, cancer progression, and Gleason scores, marking it as a valuable biomarker in PCa assessment. High stromal αSMA is primarily associated with MRI visibility. Our automated AI-based image analysis method has demonstrated reliability, aligning with our previous findings. This advances the potential for more accurate and efficient prostate cancer diagnosis and prognosis, offering a potential benefit to future patient care. Citation Format: Jenni Säilä, Timo-Pekka Lehto, Annabrita Schoonenberg, Katja Välimäki, Andrew Erickson, Antti Rannikko, Olli Kallioniemi, Tuomas Mirtti, Teijo Pellinen. FAP/αSMA association with tumor visibility and progression in prostate cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 299.

A characterization of soft-error sensitivity in data-parallel and model-parallel distributed deep learning

The latest advances in artificial intelligence deep learning models are unprecedented. A wide spectrum of application areas is now thriving thanks to available massive training datasets and gigantic complex neural network models. Those two characteristics demand outstanding computing power that only advanced computing platforms can provide. Therefore, distributed deep learning has become a necessity in capitalizing on the potential of cutting-edge artificial intelligence. Two basic schemes have emerged in distributed learning. First, the data-parallel approach, which aims at dividing the training dataset into multiple computing nodes. Second, the model-parallel approach, which splits layers of a model into several computing nodes. Each scheme has its upsides and downsides, particularly when running on large machines that are susceptible to soft errors. Those errors occur as a consequence of several factors involved in the manufacturing process of current electronic components of supercomputers. On many occasions, those errors are expressed as bit flips that do not cause the whole system to crash, but generate wrong numerical results in computations. To study the effect of soft error on different approaches for distributed learning, we leverage checkpoint alteration, a technique that injects bit flips on checkpoint files. It allows researchers to understand the effect of soft errors on applications that produce checkpoint files in HDF5 format. This paper uses the popular deep learning PyTorch tool on two distributed-learning platforms: one for data-parallel training and one for model-parallel training. We use well-known deep learning models with popular training datasets to provide a picture of how soft errors challenge the training phase of a deep learning model.

Enhancing additive manufacturing precision: Intelligent inspection and optimization for defect-free continuous carbon fiber-reinforced polymer

Artificial intelligence (AI) has emerged as a pivotal tool in managing extensive datasets, enabling pattern recognition, and deriving solutions, particularly revolutionizing additive manufacturing (AM). This study intends to develop AI deep machine learning image processing techniques for real-time defects detection in additively manufactured continuous carbon fiber-reinforced polymer(cCFRP) specimens. Leveraging YOLOv8- a state-of-the-art, single-stage object detection algorithm, this study focuses on the relationship between printing parameters and defect occurrences, specifically misalignment errors. The research delineates a methodological advancement by correlating detected defects with parameter optimization, leading to significant quality improvements in cCFRP specimens. An impressive 94 % accuracy in detecting misalignments was achieved through fine-tuning the nozzle temperature adjustment, resulting in significant reductions in misalignment errors, while minimal impact is observed from print bed temperature, feed amount, and feed rate/sec on refining the proposed model for identifying optimal parameters.

A Review of AI in Breast Cancer Detection

Cancer stands out as one of the most pressing global health challenges, and over the past decade, significant advancements have been made in diagnostic tests and methodologies. These tests fall into categories such as imaging tests, and endoscopic procedures, generating substantial volumes of data. This data needs expert evaluation to distinguish between benign and malignant tumors. Enter artificial intelligence (AI), which offers improved accuracy in analyzing large quantities of diagnostic imaging, thereby enhancing the efficiency of healthcare systems. The integration of new AI algorithms, technical advances, and enhanced computer hardware enables the training of diagnostic neural networks. This allows machines to learn from a diverse range of scans, leading to a comprehensive understanding of cancer scanning data. AI software has been evaluated against conventional diagnostic tools used by cancer specialists, and the results show significantly increased precision, making it highly effective in early diagnosis and extended forecasting for various types of cancers. In the realm of breast cancer prognosis, AI systems have demonstrated the potential to surpass human specialists, enabling much earlier diagnosis. Similarly, informatics has developed AI algorithms and deep learning techniques capable of predicting individuals' likelihood of developing lung cancer through low-dose CT analysis. The use of convolutional neural networks (CNNs) has been instrumental in diagnosing the invasion depth of gastric cancer based on gastric endoscopy.

Cost-Effectiveness of Artificial Intelligence–Based Opportunistic Compression Fracture Screening of Existing Radiographs

PurposeOsteoporotic vertebral compression fractures (OVCFs) are a highly prevalent source of morbidity and mortality, and preventive treatment has been demonstrated to be both effective and cost effective. To take advantage of the information available on existing chest and abdominal radiographs, the authors’ study group has developed software to access these radiographs for OVCFs with high sensitivity and specificity using an established artificial intelligence deep learning algorithm. The aim of this analysis was to assess the potential cost-effectiveness of implementing this software. MethodsA deterministic expected-value cost-utility model was created, combining a tree model and a Markov model, to compare the strategies of opportunistic screening for OVCFs against usual care. Total costs and total quality-adjusted life-years were calculated for each strategy. Screening and treatment costs were considered from a limited societal perspective, at 2022 prices. ResultsIn the base case, assuming a cost of software implantation of $10 per patient screened, the screening strategy dominated the nonscreening strategy: it resulted in lower cost and increased quality-adjusted life-years. The lower cost was due primarily to the decreased costs associated with fracture treatment and decreased probability of requiring long-term care in patients who received preventive treatment. The screening strategy was dominant up to a cost of $46 per patient screened. ConclusionsArtificial intelligence–based opportunistic screening for OVCFs on existing radiographs can be cost effective from a societal perspective.

YOLOv5 based object detection in reel package X-ray images of semiconductor component

The industrial manufacturing landscape is currently shifting toward the incorporation of technologies based on artificial intelligence (AI). This transition includes an evolution toward smart factory infrastructure, with a specific focus on AI-driven strategies in production and quality control. Specifically, AI-empowered computer vision has emerged as a potent tool that offers a departure from extant rule-based systems and provides enhanced operational efficiency at manufacturing sites. As the manufacturing sector embraces this new paradigm, the impetus to integrate AI-integrated manufacturing is evident. Within this framework, one salient application is AI deep learning–facilitated small-object detection, which is poised to have extensive implications for diverse industrial applications. This study describes an optimized iteration of the YOLOv5 model, which is known for its efficacious single-stage object-detection abilities underpinned by PyTorch. Our proposed “improved model” incorporates an additional layer to the model's canonical three-layer architecture, augmenting accuracy and computational expediency. Empirical evaluations using semiconductor X-ray imagery reveal the model's superior performance metrics. Given the intricate specifications of surface-mount technologies, which are characterized by a plethora of micro-scale components, our model makes a seminal contribution to real-time, in-line production assessments. Quantitative analyses show that our improved model attained a mean average precision of 0.622, surpassing YOLOv5's 0.349, and a marked accuracy enhancement of 0.865, which is a significant improvement on YOLOv5's 0.552. These findings bolster the model's robustness and potential applicability, particularly in discerning objects at reel granularities during real-time inferencing.

Coronary physiology instantaneous wave-free ratio (iFR) derived from x-ray angiography using artificial intelligence deep learning models: a pilot study.

Coronary angiography (CAG)-derived physiology methods have been developed in an attempt to simplify and increase the usage of coronary physiology, based mostly on dynamic fluid computational algorithms. We aimed to develop a different approach based on artificial intelligence methods, which has seldom been explored. Consecutive patients undergoing invasive instantaneous free-wave ratio (iFR) measurements were included. We developed artificial intelligence (AI) models capable of classifying target lesions as positive (iFR ≤ 0.89) or negative (iFR > 0.89). The predictions were then compared to the true measurements. Two hundred-fifty measurements were included, and 3 models were developed. Model 3 had the best overall performance: accuracy, negative predictive value (NPV), positive predictive value (PPV), sensitivity, and specificity were 69%, 88%, 44%, 74%, and 67%, respectively. Performance differed per target vessel. For the left anterior descending artery (LAD), model 3 had the highest accuracy (66%), while model 2 the highest NPV (86%) and sensitivity (91%). PPV was always low/modest. Model 1 had the highest specificity (68%). For the right coronary artery, model 1's accuracy was 86%, NPV was 97%, and specificity was 87%, but all models had low PPV (maximum 25%) and low/modest sensitivity (maximum 60%). For the circumflex, model 1 performed best: accuracy, NPV, PPV, sensitivity, and specificity were 69%, 96%, 24%, 80%, and 68%, respectively. We developed 3 AI models capable of binary iFR estimation from CAG images. Despite modest accuracy, the consistently high NPV is of potential clinical significance, as it would enable avoiding further invasive maneuvers after CAG. This pivotal study offers proof of concept for further development.

Dynamic changes in AI-based analysis of endometrial cellular composition: Analysis of PCOS and RIF endometrium

BackgroundThe human endometrium undergoes a monthly cycle of tissue growth and degeneration. During the mid-secretory phase, the endometrium establishes an optimal niche for embryo implantation by regulating cellular composition (e.g., epithelial and stromal cells) and differentiation. Impaired endometrial development observed in conditions such as polycystic ovary syndrome (PCOS) and recurrent implantation failure (RIF) contributes to infertility. Surprisingly, despite the importance of the endometrial lining properly developing prior to pregnancy, precise measures of endometrial cellular composition in these two infertility-associated conditions are entirely lacking. Additionally, current methods for measuring the epithelial and stromal area have limitations, including intra- and inter-observer variability and efficiency. MethodsWe utilized a deep-learning artificial intelligence (AI) model, created on a cloud-based platform and developed in our previous study. The AI model underwent training to segment both areas populated by epithelial and stromal endometrial cells. During the training step, a total of 28.36 mm2 areas were annotated, comprising 2.56 mm2 of epithelium and 24.87 mm2 of stroma. Two experienced pathologists validated the performance of the AI model. 73 endometrial samples from healthy control women were included in the sample set to establish cycle phase-dependent dynamics of the endometrial epithelial-to-stroma ratio from the proliferative (PE) to secretory (SE) phases. In addition, 91 samples from PCOS cases, accounting for the presence or absence of ovulation and representing all menstrual cycle phases, and 29 samples from RIF patients on day 5 after progesterone administration in the hormone replacement treatment cycle were also included and analyzed in terms of cellular composition. ResultsOur AI model exhibited reliable and reproducible performance in delineating epithelial and stromal compartments, achieving an accuracy of 92.40% and 99.23%, respectively. Moreover, the performance of the AI model was comparable to the pathologists’ assessment, with F1 scores exceeding 82% for the epithelium and >96% for the stroma. Next, we compared the endometrial epithelial-to-stromal ratio during the menstrual cycle in women with PCOS and in relation to endometrial receptivity status in RIF patients. The ovulatory PCOS endometrium exhibited epithelial cell proportions similar to those of control and healthy women’s samples in every cycle phase, from the PE to the late SE, correlating with progesterone levels (control SE, r2 = 0.64, FDR < 0.001; PCOS SE, r2 = 0.52, FDR < 0.001). The mid-SE endometrium showed the highest epithelial percentage compared to both the early and late SE endometrium in both healthy women and PCOS patients. Anovulatory PCOS cases showed epithelial cellular fractions comparable to those of PCOS cases in the PE (Anovulatory, 14.54%; PCOS PE, 15.56%, p = 1.00). We did not observe significant differences in the epithelial-to-stroma ratio in the hormone-induced endometrium in RIF patients with different receptivity statuses. ConclusionThe AI model rapidly and accurately identifies endometrial histology features by calculating areas occupied by epithelial and stromal cells. The AI model demonstrates changes in epithelial cellular proportions according to the menstrual cycle phase and reveals no changes in epithelial cellular proportions based on PCOS and RIF conditions. In conclusion, the AI model can potentially improve endometrial histology assessment by accelerating the analysis of the cellular composition of the tissue and by ensuring maximal objectivity for research and clinical purposes.

Exploring the Value of MRI Measurement of Hippocampal Volume for Predicting the Occurrence and Progression of Alzheimer's Disease Based on Artificial Intelligence Deep Learning Technology and Evidence-Based Medicine Meta-Analysis.

Alzheimer's disease (AD), a major dementia cause, lacks effective treatment. MRI-based hippocampal volume measurement using artificial intelligence offers new insights into early diagnosis and intervention in AD progression. This study, involving 483 AD patients, 756 patients with mild cognitive impairment (MCI), and 968 normal controls (NC), investigated the predictive capability of MRI-based hippocampus volume measurements for AD risk using artificial intelligence and evidence-based medicine. Utilizing data from ADNI and OASIS-brains databases, three convolutional neural networks (InceptionResNetv2, Densenet169, and SEResNet50) were employed for automated AD classification based on structural MRI imaging. A multitask deep learning model and a densely connected 3D convolutional network were utilized. Additionally, a systematic meta-analysis explored the value of MRI-based hippocampal volume measurement in predicting AD occurrence and progression, drawing on 23 eligible articles from PubMed and Embase databases. InceptionResNetv2 outperformed other networks, achieving 99.75% accuracy and 100% AUC for AD-NC classification and 99.16% accuracy and 100% AUC for MCI-NC classification. Notably, at a 512×512 size, InceptionResNetv2 demonstrated a classification accuracy of 94.29% and an AUC of 98% for AD-NC and 97.31% accuracy and 98% AUC for MCI-NC. The study concludes that MRI-based hippocampal volume changes effectively predict AD onset and progression, facilitating early intervention and prevention.

Diagnostic accuracy of an artificial intelligence algorithm versus radiologists for fracture detection on cervical spine CT.

To compare diagnostic accuracy of a deep learning artificial intelligence (AI) for cervical spine (C-spine) fracture detection on CT to attending radiologists and assess which undetected fractures were injuries in need of stabilising therapy (IST). This single-centre, retrospective diagnostic accuracy study included consecutive patients (age ≥18 years; 2007-2014) screened for C-spine fractures with CT. To validate ground truth, one radiologist and three neurosurgeons independently examined scans positive for fracture. Negative scans were followed up until 2022 through patient files and two radiologists reviewed negative scans that were flagged positive by AI. The neurosurgeons determined which fractures were ISTs. Diagnostic accuracy of AI and attending radiologists (index tests) were compared using McNemar. Of the 2368 scans (median age, 48, interquartile range 30-65; 1441 men) analysed, 221 (9.3%) scans contained C-spine fractures with 133 IST. AI detected 158/221 scans with fractures (sensitivity 71.5%, 95% CI 65.5-77.4%) and 2118/2147 scans without fractures (specificity 98.6%, 95% CI 98.2-99.1). In comparison, attending radiologists detected 195/221 scans with fractures (sensitivity 88.2%, 95% CI 84.0-92.5%, p < 0.001) and 2130/2147 scans without fracture (specificity 99.2%, 95% CI 98.8-99.6, p = 0.07). Of the fractures undetected by AI 30/63 were ISTs versus 4/26 for radiologists. AI detected 22/26 fractures undetected by the radiologists, including 3/4 undetected ISTs. Compared to attending radiologists, the artificial intelligence has a lower sensitivity and a higher miss rate of fractures in need of stabilising therapy; however, it detected most fractures undetected by the radiologists, including fractures in need of stabilising therapy. Clinical relevance statement The artificial intelligence algorithm missed more cervical spine fractures on CT than attending radiologists, but detected 84.6% of fractures undetected by radiologists, including fractures in need of stabilising therapy. The impact of artificial intelligence for cervical spine fracture detection on CT on fracture management is unknown. The algorithm detected less fractures than attending radiologists, but detected most fractures undetected by the radiologists including almost all in need of stabilising therapy. The artificial intelligence algorithm shows potential as a concurrent reader.