False Positive Indications Research Articles

An updated concept of the polarization index enables the integration of three intensity zones in one number and identifies polarized training in an extended definition range

Introduction Polarized endurance training is an important and frequently discussed training intensity distribution (TID). The polarized TID is described as the largest fraction of training time or sessions spent with low-intensity exercise in intensity zone (z)1, followed by a considerable fraction of high intensity exercise (z3), and a relatively small fraction of threshold intensity (z2), so that z1 > z3 > z2. The polarization index concept (PI; Figure 1, equation 1) provides a clear, numerical cut-off for this rather vague, verbal description of a polarized TID by relating the fractions of z1, z2, and z3 and defining a given TID as polarized if the result is > 2.00 (Treff et al., 2019). The concept has successfully been applied in more than 36 peer-reviewed articles. Since the publication of the high-performance sports centered PI, a variety of issues related to transforming context conditions appeared, revealing some limitations of the original PI: (i) The increasing use of automatic database entries can result in TIDs with intensity zones < 1%. These can cause rounding problems. (ii) TIDs are attracting increasing attention in the public health sector (Festa et al., 2023). Here, due to lower total volume and session frequencies, TIDs may occur that are virtually never reported in high-performance sports as, e.g., z2 < z1 < z3. These TIDs can result in negative log10-arguments of equation 1. It has been proposed to limit the calculation of the PI to TIDs z1 > z3 >z2 (Arjona et al., 2023). This would, however, limit the PI’s applicability. Here, we propose an update to adapt the PI without limiting its definition space. Methods Methods: Given that z1, z2, z3 ∈ [0, 1] and z1 + z2 + z3 = 1, where z1, z2 denote the measured fractions rounded to two decimals and z3 = 1 - z1 - z2, new PI is calculated according to equation 2 (Figure 1). Results A TID (given as z1 - z2 - z3) partially consisting of fractions below 0.01 (0.991 - 0.001 - 0.008) will be quantized to (0.99 - 0.00 - 0.01) and the respective PI results in 0, thereby indicating a non-polarized TID in contrast to 2.90 with equation 1 (Figure 1). Thus, the false-positive indication of a polarized TID is avoided. Applying equation 2 (Figure 1), the PI of the TID (0.15 - 0.05 - 0.80) will result in -2.38 (inversely polarized) instead of 2.38 (polarized) with equation 1 (Figure 1). Discussion & Conclusion Particularly the exclusive consideration of intensity zones with at least 1% of the total proportion avoids negative arguments in the log10 algorithm and the indication of a polarized TID if the fraction of z3 is not higher than 0.01. Furthermore, equation 2 (Figure 1) allows for the identification of inversely polarized TIDs while avoiding false positive indication of a polarized TID when z2 < z1 < z3. For polarized distributions (z1 > z3 > z2) with z2 = 0, subtracting the 0.01 fraction from z1 accounts for the underestimation of the z3 percentages especially in heart rate-based quantifications. Thus, our proposal allows a broad application of the PI concept in health and performance-focused databases to accurately detect polarized TIDs.

Vibration based dual-criteria damage detection method using deep neural networks in highway bridges with steel girders

This article introduces a new method for assessing the damage in steel girders of highway bridges using deep neural networks. The method uses modified forms of modal strain energy damage index and modal flexibility damage index to train the neural network and prevent unsafe decisions. The neural network is trained using damage indexes achieved from numerical simulations of the bridge with different damage scenarios. Three types of neural networks, including convolutional neural network (CNN), Long short-term memory (LSTM) neural network, and conventional artificial neural network (ANN) are compared to achieve the best performance. The proposed method overcomes previous detection problems such as false-positive indications, limited damage locations, and scenarios, and insufficient accuracy in cases with multiple damage scenarios. The results show that the proposed method and CNN can accurately identify unspecified damage locations and severity in multi-span highway bridges. The new training method of the CNN deep neural network systems overcomes some shortcomings in ANN, such as being time-consuming, having a low convergence rate, and experiencing data overfitting. Additionally, the CNN training scheme can limit the need for massive amounts of input data and increase the speed and accuracy of network training, especially in multiple damage scenarios.

Use of High-Frequency Ultrasound for the Correlation of Epithelial Rete Peg Variation

ObjectiveRete pegs are projections of the oral epithelium into connective tissue. Their dimensions change during pathological conditions and may correlate with wound-healing status. Non-invasive, high-frequency ultrasound (US) may be able to capture these changes and aid in early detection of histopathological changes. The aim of this preclinical study is to correlate US images with histology and quantify epithelial layers at different tooth sites. MethodsSagittal B-mode images of mid-facial and interproximal oral soft tissue sites were recorded in a preclinical minipig model using a linear array in second harmonic mode (12/24 MHz). Histology samples from the same locations were stained (hematoxylin and eosin), digitized and registered with US images. Manual annotations were used to measure distances D1 (thickness of epithelium on histology vs. hyperechoic zone on US) and D2 (sum of epithelial thickness and length of rete pegs on histology vs. sum of hyperechoic and hypoechoic zone on US) to statistically analyze them. ResultsUltrasonic-derived dimensions yielded a mean bias of -0.64 (55% coefficient of variance [COV]: -180 to +180 µm) and -12 µm (39% COV: -260 to +240 µm) for D1 and D2, respectively. Individualized analysis of D1 and D2 by tooth type showed similar tends in the ability to differentiate between epithelium at different tooth locations, on both histology and US. ConclusionAssessing soft tissue dimensions on a sub-millimeter scale using clinical imaging hardware is still a developing area. Future research might open doors for diagnosis of oral pathologies and abnormal wound healing, and may limit false-positive indications for biopsies.

Persistence of sewage-associated genetic markers in advanced and conventional treated recycled water: implications for microbial source tracking in surface waters.

Genes in sewage-associated microorganisms are widely accepted indicators of sewage pollution in environmental waters. However, DNA persists through wastewater treatment and can reach surface waters when recycled water is discharged, potentially causing false-positive indications of sewage contamination. Previous studies have found that bacterial and viral sewage-associated genes persist through wastewater treatment; however, these studies did not compare different facilities or identify a solution to distinguish sewage from recycled water. In this study, we demonstrated the persistence of bacterial marker genes and the greater persistence of a viral marker gene (CPQ_056 of crAssphage) through varying wastewater treatment facilities. We also aim to provide a tool to confirm sewage contamination in surface waters with recycled water inputs. This work showed that the level of wastewater treatment affects the removal of microorganisms, particularly viruses, and expands our ability to identify sewage in surface waters.

Local Detection of Ground Coupled Acoustic Waves with Seismic Arrays and Their Potential Role in the Discrimination of Explosions and Earthquakes

Abstract Acoustic waves are widely used to characterize explosive sources such as volcanoes, meteorites, and controlled explosions. This study examines the potential role of ground coupled airwaves (GCA), which effectively propagate at acoustic speeds (∼0.34 km/s) before coupling to the ground near seismometers, in aiding local discrimination between low-yield explosions in shallow boreholes and earthquakes. GCA generated by shallow borehole explosions from the 2014 imaging magma under St. Helens experiment (ML 0.9–2.3) and earthquakes (ML 2–3.4) from 2014 to 2016, were recorded by various seismometers at <150 km source–receiver distance. Potential GCA are analyzed using arrays of broadband seismometers (number of seismometers, n = 85), nodal seismometers with 10-Hz geophones atop the surface (n = 904), and Texan dataloggers with shallowly buried 4.5-Hz geophones (n = 2535). Array-based detections are defined using the distributions of short-time average over long-time average functions in time windows during and adjacent to the predicted GCA arrival for direct source–receiver transmission. GCA are detected for 14 of 23 borehole explosions and 0 of 34 earthquakes. All detections occurred during times of low-mean wind speed (<0.5 m/s) at ground-based weather stations. GCA amplitudes exhibit strong spatial variability, and the number of spatially distributed receivers appears more important for GCA detection than the type of seismometer installation. GCA detections were compared with seismic P/S amplitude ratios, which are a common source discriminant, and field logs of whether the borehole explosions ejected any mass or deformed the surface. No clear correlation was found with either type of source information, suggesting that heterogeneous propagation and near-receiver effects like wind noise are more influential than variations in source processes among the 23 explosions. Our results indicate that local seismic detection of GCA may valuably complement discrimination metrics like P/S ratios, with a low tendency for false-positive indications of explosions but a high tendency for false negatives.

Use of the "Sepsis Risk Calculator" in Belgian Newborns: A Retrospective Cohort Study.

Prevention of early-onset neonatal sepsis (EONS) is a frequent reason why many newborns receive unnecessary antibiotics. The Sepsis Risk Calculator (SRC) was developed by the Kaiser Permanente Institute as a multivariate risk assessment of EONS, aiming to reduce laboratory testing and empiric neonatal antibiotic therapy. Our objective was to assess the potential of the SRC in reducing antibiotic use in our setting. Late preterm and term newborns who received antibiotics from 2019 to 2020 in a tertiary Belgian hospital were included. Newborn-specific data were collected and entered into the online SRC, retrospectively calculating a sepsis risk score and providing recommendations for antibiotic administration. False-positive indications for treatment by the SRC were estimated based on previously published data. Antibiotic therapy rates according to the SRC recommendations were compared to the actual rate of antibiotic therapy. Of 5891 births, 414 newborns received antibiotics and were eligible for this study, representing a rate of 7.6% of newborns receiving antibiotics following our current guidelines. The SRC would have recommended antibiotic administration for 2.7%, reducing antibiotic therapy by 64.5%. Of 5 possible cases of EONS, 3 would have received antibiotics in the first 24 hours according to the SRC. In this Belgian cohort, use of the SRC has the potential to significantly decrease by 64.5% the newborns that receive antibiotics. This reduction would primarily concern asymptomatic newborns. If use of the SRC was to be implemented in Belgian maternities, strict clinical surveillance practices should be ensured.

Operational data for the risk management of victim operated explosive devices in humanitarian mine action: A Practitioner's perspective

Since Mine Action's inception at the end of the 1980s, operators have collected limited data on the Victim Operated Explosive Devices (VOEDs) they clear. This includes not only data on the explosive ordnance itself but data on how they were found, where they were found and how they were processed and ultimately destroyed. In a context where detection of mines, boobytraps and certain Victim Operated Improvised Explosive Devices (VOIEDs) is an ongoing humanitarian and military challenge, significantly expanded operational data collection provides an achievable way to facilitate enhanced operational risk management. Risk decisions inherent in the clearance of VOEDs are better if made on the basis of extensive operational data. In the absence of a technological solution to detect and positively discriminate VOEDs from false positive indications, the collection of operational data offers the best prospect for “managing” if not “solving” the problem.

Umbilical cord arterial blood pH an indicator of rise in cesarean section rate due to false-positive indications by CTG

Background: Cesarean section rate is rising globally, now accounting for more than 1 in 5 (21%) of all childbirths. This number is set to continue increasing over the coming decade unless the factors which increase the rate of cesarean section are kept in check. Intrapartum cardiotocography (CTG) has shown a false-positive rate of cesarean section. This has been correlated with neonatal umbilical cord blood pH analysis. Aims and Objectives: (1) Correlation of non-reassuring CTG, resulting in cesarean section with umbilical cord arterial blood pH. (2) To find the false-positive indications of cesarean section due to CTG. Materials and Methods: This observational study was conducted in the Department of Obstetrics and Gynecology of SKIMS, Soura, from September 2020 to July 2022 over a period of 22 months. Women with a gestational age of more than 34 weeks with singleton pregnancy were included in the study. Those with CTG-documented fetal distress were subjected to umbilical cord arterial blood pH monitoring. Results: A total of 85 patients underwent cesarean sections in view of fetal distress documented by CTG, but only 45 neonates had actual distress as documented by umbilical cord blood sampling. Conclusion: Social and non-medical factors no doubt have caused an alarming rise in cesarean sections, but at the same time, there are some iatrogenic causes, namely CTG, which lead to the rise in the cesarean section rate due to false-positive indications. This has been further supported by the fact that the pH of the umbilical cord blood of neonates with intrapartum non-reassuring CTG does not correspond to fetal acidemia in all cases. Hence, a significant number of cesarean sections is being done for false-positive indications.

What Interval of Daily Pain Assessment Is Required to Reliably Diagnose Chronic Pain in Sickle Cell Disease? the Pisces Study

Background: In Sickle Cell Disease (SCD), chronic pain, i.e., pain on most days over a 6 month time period, is a major complication associated with substantial morbidity and poor outcomes. In the landmark Pain in Sickle Cell Epidemiology Study (PiSCES), the largest study of pain in SCD, one-third of individuals age 16 years or older had daily or near-daily pain, and >50% had chronic pain. However, PiSCES required patients to have 5-6 months of completed diaries in order to qualify for chronic pain analysis. Practically, the International Association for the Study of Pain (IASP) has used three months as the interval upon which to base a definition of chronic pain. Shorter daily assessment intervals upon which to reliably base a definition of chronic pain would reduce respondent burden, and would improve the frequency and accuracy of pain phenotyping for clinical and research purposes. Method: We chose as the gold-standard for chronic pain definition subjects enrolled in PiSCES that had at least 5 months of daily diaries (n=127). We excluded 10 patients who had at least 182 diaries (essentially 6 months), but had at least 50% of diaries missing in any of months 1-4. We simulated daily diary assessment over less burdensome, shorter intervals, always beginning on day one of assessment, to mimic what would be done clinically. We chose the time frames of 2 weeks (2w), 4 weeks (4w), 1 month (1m), 2 months (2m), 3 months (3m), and 4 months (4m). To operationalize the definition of chronic pain, we used a definition of pain severity rated as >0 on > 50% of diary days, regardless of interval. To determine if using diaries collected over a shorter period of time would still classify patients correctly, we calculated the sensitivity and the number of false negatives, as well as the specificity and number of false positives of each simulated diary interval. Results: Within the gold standard sample, 51.3% of patients had chronic pain based on all diaries submitted. The Table shows the percent of patients diagnosed with chronic pain, the sensitivity as well as the number of false negatives, and the specificity as well as the number of false positives, for each simulated diary interval. Simulated intervals of two months or more of daily diaries yielded chronic pain percentage estimates with both high sensitivity and high specificity, compared to the gold standard estimate. Simulated intervals of two week and four weeks yielded increasingly lower specificity, i.e., increasingly more false-positive indications of chronic pain, as the interval shortened. However, each interval still yielded high sensitivity-few patients with gold-standard chronic pain were missed, even at two weeks. Results were very similar even if data from the entire 127 participants was used. Conclusion : For determining the presence of chronic SCD pain, defined as pain on more than 50% of days enrolled in PiSCES, simulated intervals of two months or more of daily diary collection yield both very good sensitivity and very good specificity, compared to intervals of 5-6 months. Clinicians and researchers may reliably choose to adopt two months of dairy collection, as a matter of practicality, to reduce respondent burden when phenotyping SCD pain as acute versus chronic.

An investigation of the potential clinical utility of critical slowing down as an early warning sign for recurrence of depression

Background and objectivesMuch of the burden of depressive illness is due to relapses that occur after treatment into remission. Prediction of an individual's imminent depressive relapse could lead to just-in-time interventions to prevent relapse, reducing depression's substantial burden of disability, costs, and suicide risk. Increasingly strong relationships in the form of autocorrelations between depressive symptoms, a signal of a phenomenon described as critical slowing down (CSD), have been proposed as a means of predicting relapse. MethodsIn the current study, four participants in remission from depression, one of whom relapsed, responded to daily smartphone surveys with depression symptoms. We used p-technique factor analysis to identify depression factors from over 100 survey responses. We then tested for the presence of CSD using time-varying vector autoregression and detrended fluctuation analysis. ResultsWe found evidence that CSD provided an early warning sign for depression in the participant who relapsed, but we also detected false positive indications of CSD in participants who did not relapse. Results from time-varying vector autoregression and detrended fluctuation analysis were not in agreement. LimitationsLimitations include use of secondary data and a small number of participants with daily responding to a subset of depression symptoms. ConclusionsCSD provides a compelling framework for predicting depressive relapse and future research should focus on improving detection of early warning signs reliably. Improving early detection methods for depression is clinically significant, as it would allow for the development of just-in-time interventions.

Abstract P390: E-value Over P-value And Its Adjustments: Improving Dna Methylation Analysis With An Application To Hypertension Studies

DNA methylation plays a crucial role in transcriptional regulation. Reduced representation bisulfite sequencing (RRBS) is a technique of increasing use for analyzing genome-wide methylation profiles. Many computational tools such as Metilene, MethylKit, BiSeq and DMRfinder have been developed to use RRBS data to detect differentially methylated regions (DMRs) that may be involved in epigenetic regulations of gene expression. For DMR detection tools, as for countless other medical literature applications, P-values and their adjustments are among the most widely reported statistics used to assess the statistical significance of biological findings. However, P-values are coming under increasing criticism related to their questionable accuracy and relatively high levels of false positive or negative indications. Here, we proposed a method to calculate E-value for DMR detection in RRBS data, which is defined as likelihood ratios falling into the null hypothesis over the entire parameter space. We also provided the corresponding R package ‘ metevalue ’ as a user-friendly interface to implement E-value calculations into various DMR detection tools. To evaluate the performance of E-value, we generated various RRBS benchmarking datasets using our simulator ‘ RRBSsim ’ with 8 samples in each experimental group. Our comprehensive benchmarking analyses showed that using E-value not only significantly improved accuracy, AUC and power, over that of P-value or adjusted P-value, but also reduced false discovery rates and Type I errors. To illustrate the utility of E-value, we applied it to identify DMRs in two real RRBS datasets. One was genome-wide DNA methylome and transcriptome of renal T lymphocytes from Dahl Salt Sensitive rats treated with high- or low-salt diets. The other was genome-wide DNA methylation and gene expression data of human arterioles from our clinical trial in which 10 subjects were placed on 2-week low-salt diet (1200-mg sodium/day). Compared to the classical adjusted P-value, the use of E-value detected biologically more relevant DMRs and improved the negative association between DNA methylation and gene expression.

Evaluating scent detection dogs as a tool to detect pathogenic Phytophthora species

AbstractThe fungal genus Phytophthora includes an array of destructive plant pathogens that have had severe impacts on native, agricultural, and horticultural systems worldwide. Preventing the spread of Phytophthora species is critical for protecting vulnerable plants and ecosystems; yet detection remains a challenge due to their microscopic size, broad host range, and latent and cryptic expression in host plants. We tested the effectiveness of trained detection dogs to discriminate the odor of a target Phytophthora species, from among other non‐target odors, by conducting a multi‐choice experiment. We tasked two dogs with discriminating the scent of P. agathidicida—the causal agent of the lethal root rot disease “kauri dieback”—from two non‐target Phytophthora scents (P. cinnamomi and P. multivora), and four non‐target control treatments. We assessed the dogs' scent detection abilities by measuring the sensitivity and precision of their indications toward the target scent over 120 randomized trials. The dogs had a combined sensitivity of 68.6% (CI: 64.1–72.9) and precision of 52.2% (CI: 48.1–56.4), meaning they often identified P. agathidicida when it was present but also signaled on the other non‐target Phytophthora species. Moreover, we characterized the nature of false positive indications made for non‐target scents, which has important implications for how future multi‐choice experiments should be conducted. Our study shows that detection dogs are likely to be an adequate first‐pass detection tool for Phytophthora within a wider biosecurity framework. However, further research is warranted before dogs can be deployed for this purpose in the field.

E-value: a superior alternative to P-value and its adjustments in DNA methylation studies.

DNA methylation plays a crucial role in transcriptional regulation. Reduced representation bisulfite sequencing (RRBS) is a technique of increasing use for analyzing genome-wide methylation profiles. Many computational tools such as Metilene, MethylKit, BiSeq and DMRfinder have been developed to use RRBS data for the detection of the differentially methylated regions (DMRs) potentially involved in epigenetic regulations of gene expression. For DMR detection tools, as for countless other medical applications, P-values and their adjustments are among the most standard reporting statistics used to assess the statistical significance of biological findings. However, P-values are coming under increasing criticism relating to their questionable accuracy and relatively high levels of false positive or negative indications. Here, we propose a method to calculate E-values, as likelihood ratios falling into the null hypothesis over the entire parameter space, for DMR detection in RRBS data. We also provide the R package 'metevalue' as a user-friendly interface to implement E-value calculations into various DMR detection tools. To evaluate the performance of E-values, we generated various RRBS benchmarking datasets using our simulator 'RRBSsim' with eight samples in each experimental group. Our comprehensive benchmarking analyses showed that using E-values not only significantly improved accuracy, area under ROC curve and power, over that of P-values or adjusted P-values, but also reduced false discovery rates and type I errors. In applications using real RRBS data of CRL rats and a clinical trial on low-salt diet, the use of E-values detected biologically more relevant DMRs and also improved the negative association between DNA methylation and gene expression.

A Method of Damage Detection Efficiency Enhancement of PZT Sensor Networks under Influence of Environmental and Operational Conditions

Two performance parameters are particularly important for the assessment of structural health monitoring (SHM) systems, i.e., their damage detection capabilities and risk of false positive indications due to varying environmental and operational conditions (EOCs). A reduced ratio of false-positive indications can be of significant importance for particular applications, for example, in aerospace, where the costs of unplanned maintenance procedures can be very high. In such cases, the reduction of the false calls ratio can be critical for the possibility of the practical application of the system, apart from damage detection efficiency and system costs. Among various sensor technologies, PZT networks are proven to be one of the most universal approaches to SHM, and they were successfully applied in different scenarios. Moreover, many EOCs which may have an impact on the risk of false positive indications have been identified. Over the years, different approaches to the influence of EOCs compensation have been proposed. Compensation methods can be tailored to the particular way in which a given measurement condition, for example, ambient temperature, alters signals acquired by the PZT network or can be formulated to be also applied in the more general case. In the paper, a method for enhancement of damage detection efficiency under influence of EOCs of general nature is proposed. The particular measurement condition affecting signals acquired by PZT sensors neither needs to be measured, which could be hard in some cases, but also nor even have to be identified. The efficiency of the proposed compensation algorithms is verified based on the example of experimental results obtained under varying temperatures.

False Negative Awareness in Indicator-Based Caching Systems

Distributed caching systems such as content distribution networks often advertise their content via lightweight approximate indicators (e.g., Bloom filters) to efficiently inform clients where each datum is likely cached. While false-positive indications are necessary and well understood, most existing works assume no false-negative indications. Our work illustrates practical scenarios where false-negatives are unavoidable and ignoring them significantly impacts system performance. Specifically, we focus on false-negatives induced by indicator staleness, which arises whenever the system advertises the indicator only periodically, rather than immediately reporting every change in the cache. Such scenarios naturally occur, e.g., in bandwidth-constraint environments or when latency impedes each client’s ability to obtain an updated indicator. Our work introduces novel false-negative aware access policies that continuously estimate the false-negative ratio and sometimes access caches despite negative indications. We present optimal policies for homogeneous settings and provide approximation guarantees for our algorithms in heterogeneous environments. We further perform an extensive simulation study with multiple real system traces. We show that our false-negative aware algorithms incur a significantly lower service cost than existing approaches or match the cost of these approaches while requiring an order of magnitude fewer resources (e.g., caching capacity or bandwidth).

A Bayesian hypothesis testing-based statistical decision philosophy for structural damage detection

In this article, a new statistical decision philosophy is developed to tackle structural damage detection problems which are defined in the context of novelty detection. In line with this philosophy, structural damage detection is achieved by deriving the posterior probability of damage presence from the Bayesian testing of two competing hypotheses, with the null and alternative hypotheses representing the undamaged and damaged states of a structure of concern, respectively. To resolve the tricky problem of prior appropriateness in Bayesian hypothesis testing, a general prior specification criterion is devised based on the notion of risk management, including the risk of false positive indication (an undamaged structure is incorrectly identified as damaged) and the risk of false negative indication (a damaged structure is incorrectly identified as undamaged). To determine an optimal risk level, two principles, namely the principle of posterior probability difference minimization (PPDM) and the principle of posterior probability product maximization (PPPM), are defined. The PPDM principle is to minimize the difference between the ability of a novelty detector to avoid a false positive and its ability to avoid a false negative, and the PPPM principle is to maximize the product of the two capabilities. Both principles essentially act as a means of achieving an optimal trade-off between the false positive and false negative risks stipulated in pursuing damage detection. To demonstrate the effectiveness of the proposed statistical decision philosophy for structural damage detection, experimental data obtained from a 5-story steel frame model and a 38-story concrete building model have been investigated.

Depth image conversion model based on CycleGAN for growing tomato truss identification

BackgroundOn tomato plants, the flowering truss is a group or cluster of smaller stems where flowers and fruit develop, while the growing truss is the most extended part of the stem. Because the state of the growing truss reacts sensitively to the surrounding environment, it is essential to control its growth in the early stages. With the recent development of information and artificial intelligence technology in agriculture, a previous study developed a real-time acquisition and evaluation method for images using robots. Furthermore, we used image processing to locate the growing truss to extract growth information. Among the different vision algorithms, the CycleGAN algorithm was used to generate and transform unpaired images using generated learning images. In this study, we developed a robot-based system for simultaneously acquiring RGB and depth images of the growing truss of the tomato plant.ResultsThe segmentation performance for approximately 35 samples was compared via false negative (FN) and false positive (FP) indicators. For the depth camera image, we obtained FN and FP values of 17.55 ± 3.01% and 17.76 ± 3.55%, respectively. For the CycleGAN algorithm, we obtained FN and FP values of 19.24 ± 1.45% and 18.24 ± 1.54%, respectively. When segmentation was performed via image processing through depth image and CycleGAN, the mean intersection over union (mIoU) was 63.56 ± 8.44% and 69.25 ± 4.42%, respectively, indicating that the CycleGAN algorithm can identify the desired growing truss of the tomato plant with high precision.ConclusionsThe on-site possibility of the image extraction technique using CycleGAN was confirmed when the image scanning robot drove in a straight line through a tomato greenhouse. In the future, the proposed approach is expected to be used in vision technology to scan tomato growth indicators in greenhouses using an unmanned robot platform.

Development and validation of a deep learning model for detection of breast cancers in mammography from multi-institutional datasets.

ObjectivesThe objective of this study was to develop and validate a state-of-the-art, deep learning (DL)-based model for detecting breast cancers on mammography.MethodsMammograms in a hospital development dataset, a hospital test dataset, and a clinic test dataset were retrospectively collected from January 2006 through December 2017 in Osaka City University Hospital and Medcity21 Clinic. The hospital development dataset and a publicly available digital database for screening mammography (DDSM) dataset were used to train and to validate the RetinaNet, one type of DL-based model, with five-fold cross-validation. The model’s sensitivity and mean false positive indications per image (mFPI) and partial area under the curve (AUC) with 1.0 mFPI for both test datasets were externally assessed with the test datasets.ResultsThe hospital development dataset, hospital test dataset, clinic test dataset, and DDSM development dataset included a total of 3179 images (1448 malignant images), 491 images (225 malignant images), 2821 images (37 malignant images), and 1457 malignant images, respectively. The proposed model detected all cancers with a 0.45–0.47 mFPI and had partial AUCs of 0.93 in both test datasets.ConclusionsThe DL-based model developed for this study was able to detect all breast cancers with a very low mFPI. Our DL-based model achieved the highest performance to date, which might lead to improved diagnosis for breast cancer.

Deep learning-based detection of parathyroid adenoma by 99mTc-MIBI scintigraphy in patients with primary hyperparathyroidism.

It is important to detect parathyroid adenomas by parathyroid scintigraphy with 99m-technetium sestamibi (99mTc-MIBI) before surgery. This study aimed to develop and validate deep learning (DL)-based models to detect parathyroid adenoma in patients with primary hyperparathyroidism, from parathyroid scintigrams with 99mTc-MIBI. DL-based models for detecting parathyroid adenoma in early- and late-phase parathyroid scintigrams were, respectively, developed and evaluated. The training dataset used to train the models was collected from 192 patients (165 adenoma cases, mean age: 64years ± 13, 145 women) and the validation dataset used to tune the models was collected from 45 patients (30 adenoma cases, mean age: 67years ± 12, 37 women). The images were collected from patients who were pathologically diagnosed with parathyroid adenomas or in whom no lesions could be detected by either parathyroid scintigraphy or ultrasonography at our institution from June 2010 to March 2019. The models were tested on a dataset collected from 44 patients (30 adenoma cases, mean age: 67years ± 12, 38 women) who took scintigraphy from April 2019 to March 2020. The models' lesion-based sensitivity and mean false positive indications per image (mFPI) were assessed with the test dataset. The sensitivity was 82% [95% confidence interval 72-92%] with mFPI of 0.44 for the scintigrams of the early-phase model and 83% [73-92%] with mFPI of 0.31 for the scintigrams of the delayed-phase model in the test dataset, respectively. The DL-based models were able to detect parathyroid adenomas with a high sensitivity using parathyroid scintigraphy with 99m-technetium sestamibi.

Artificial intelligence–assisted detection of colorectal polyps in Lynch syndrome

A young patient with a familial history of Lynch syndrome presented for colonoscopy. She had undergone colonic polypectomy at another hospital 1 year previously. A computer-aided detection system (EX-1, Fujifilm Holdings Corp, Tokyo, Japan) was applied during colonoscopy in our hospital. When polyps that arouse suspicion are detected by this system, the lesions are indicated by cyan squares (A). On observation of the highlighted areas, the false positive indicators disappeared (B). Blue-light imaging revealed a regular vessel pattern, and computer-aided detection indicated that the detected polyp was a neoplastic lesion with a yellow mark on the position map (C).