Surveillance Algorithms Research Articles

Infectious Complications in Pediatric, Adolescent and Young Adult Patients Undergoing CD19-CAR T Cell Therapy.

CD19-specific chimeric antigen receptor (CAR) T cell therapy has changed the treatment paradigm for pediatric, adolescent and young adult (AYA) patients with relapsed/refractory B-cell acute lymphoblastic leukemia (B-ALL). However, data on the associated infectious disease challenges in this patient population are scarce. Knowledge of infections presenting during treatment, and associated risk factors, is critical for pediatric cellular therapy and infectious disease specialists as we seek to formulate effective anti-infective prophylaxis, infection monitoring schemas, and empiric therapy regimens. In this work we describe our institutional experience in a cohort of 38 pediatric and AYA patients with CD19-positive malignancy treated with lymphodepleting chemotherapy (fludarabine/cyclophosphamide) followed by a single infusion of CD19-CAR T cells (total infusions, n=39), including tisagenlecleucel (n=19; CD19/4-1BB) or on an institutional clinical trial (n=20; CD19/4-1BB; NCT03573700). We demonstrate that infections were common in the 90 days post CAR T cells, with 19 (50%) patients experiencing a total of 35 infections. Most of these (73.7%) occurred early post infusion (day 0 to 28; infection density of 2.36 per 100 patient days-at-risk) compared to late post infusion (day 29 to 90; infection density 0.98 per 100 patient days-at-risk), respectively. Bacterial infections were more frequent early after CAR T cell therapy, with a predominance of bacterial blood stream infections. Viral infections occurred throughout the post infusion period and included primarily systemic reactivations and gastrointestinal pathogens. Fungal infections were rare. Pre-infusion disease burden, intensity of bridging chemotherapy, lymphopenia post lymphodepleting chemotherapy/CAR T cell infusion and development of CAR-associated hemophagocytic lymphohistiocytosis (carHLH) were all significantly associated with either infection density or time to first infection post CAR T cell infusion. A subset of patients (n=6) had subsequent CAR T cell reinfusion and did not appear to have increased risk of infectious complications. Our experience highlights the risk of infections after CD19-CAR T cell therapy, and the need for continued investigation of infectious outcomes as we seek to improve surveillance, prophylaxis and treatment algorithms.

Intracranial Metastatic Disease: Present Challenges, Future Opportunities.

Intracranial metastatic disease (IMD) is a prevalent complication of cancer that significantly limits patient survival and quality of life. Over the past half-century, our understanding of the epidemiology and pathogenesis of IMD has improved and enabled the development of surveillance and treatment algorithms based on prognostic factors and tumor biomolecular characteristics. In addition to advances in surgical resection and radiation therapy, the treatment of IMD has evolved to include monoclonal antibodies and small molecule antagonists of tumor-promoting proteins or endogenous immune checkpoint inhibitors. Moreover, improvements in the sensitivity and specificity of imaging as well as the development of new serological assays to detect brain metastases promise to revolutionize IMD diagnosis. In this review, we will explore current treatment principles in patients with IMD, including the emerging role of targeted and immunotherapy in select primary cancers, and discuss potential areas for further investigation.

Addressing the Competitive Harms of Opaque Online Surveillance and Recommendation Algorithms

Facebook and Alphabet operate free internet services that are widely used. They provide these services for free because users are online ad targets. Together Facebook and Alphabet have a large share of the market for online advertising in the U.S. Their dominance delivers monopolistic returns, reflected in the persistently high valuations financial markets place on each company. Online ad sales depend on the ability of these platforms to individually target ads and messages to huge numbers of people. Targeting is made possible by surveillance which is large in scale, scope, and effectiveness. User engagement, which helps determine target numbers, is stimulated and directed by “recommendation” algorithms on Facebook and Alphabet’s YouTube platform. These algorithms can affect what users read and view, and can influence their attitudes, emotions, and behavior. While surveillance has negative effects on user privacy, and algorithms have had powerful effects on user attitudes and behavior, platform users have limited knowledge about how these practices operate or their impacts. These information asymmetries between platforms and users have important competitive effects. They divert users from competing platforms that do not engage in these business practices, and inhibit entry and the innovation it would stimulate, thereby helping sustain the monopoly power of dominant incumbents. Section 5 of the Federal Trade Commission Act, which prohibits “unfair methods of competition” and includes rulemaking authority, may be the most effective way to address anticompetitive practices that are technically complex, can evolve rapidly, and are difficult for industry outsiders to observe.

Automated surveillance systems for healthcare-associated infections: results from a European survey and experiences from real-life utilization

As most automated surveillance (AS) methods to detect healthcare-associated infections (HAIs) have been developed and implemented in research settings, information about the feasibility of large-scale implementation is scarce. To describe key aspects of the design of AS systems and implementation in European institutions and hospitals. An online survey was distributed via e-mail in February/March 2019 among (i) PRAISE (Providing a Roadmap for Automated Infection Surveillance in Europe) network members; (ii) corresponding authors of peer-reviewed European publications on existing AS systems; and (iii) the mailing list of national infection prevention and control focal points of the European Centre for Disease Prevention and Control. Three AS systems from the survey were selected, based on quintessential features, for in-depth review focusing on implementation in practice. Through the survey and the review of three selected AS systems, notable differences regarding the methods, algorithms, data sources, and targeted HAIs were identified. The majority of AS systems used a classification algorithm for semi-automated surveillance and targeted HAIs were mostly surgical site infections, urinary tract infections, sepsis, or other bloodstream infections. AS systems yielded a reduction of workload for hospital staff. Principal barriers of implementation were strict data security regulations as well as creating and maintaining an information technology infrastructure. AS in Europe is characterized by heterogeneity in methods and surveillance targets. To allow for comparisons and encourage homogenization, future publications on AS systems should provide detailed information on source data, methods, and the state of implementation.

Real-Time Traffic and Road Surveillance With Parallel Edge Intelligence

Edge computing, which is an emerging concept to complement cloud computing, processes data closer to where the data is generated at the network edge. Transportation system is a critical component of our cities and is generating more and more data. To leverage the power of the increasing amount of data from traffic sensors, researchers have started exploring the feasibility of edge computing for traffic big data analytics. Given the resource constraints on edge devices, it is very challenging to achieve high intelligence and real-time capability at the same time. There are few edge transportation systems that function in a multi-tasking mode with high efficiency and intelligence on the edge. This paper proposes a multi-thread parallel edge system architecture and several lightweight algorithms for traffic and road surveillance. The designs ensure robustness and real-time capability for the key modules of the system.

The accuracy of fully automated algorithms for surveillance of healthcare-onset Clostridioides difficile infections in hospitalized patients.

We developed and validated a set of fully automated surveillance algorithms for healthcare-onset CDI using electronic health records. In a validation data set of 750 manually annotated admissions, the algorithm based on International Classification of Disease, Tenth Revision (ICD-10) code A04.7 had insufficient sensitivity. Algorithms based on microbiological test results with or without addition of symptoms performed well.

Foreground Object Detection in Visual Surveillance With Spatio-Temporal Fusion Network

Object detection generally shows promising results only using spatial information, but foreground object detection in visual surveillance requires proper use of temporal information in addition to spatial information. Recently, deep learning-based visual surveillance algorithms have shown improved results, in an environment similar to training one, compared to traditional background subtraction (BGS) algorithms. However, in unseen environments, they show poor performance compared to BGS algorithms. This paper proposes an algorithm that improves performance in unseen environments by integrating spatial and temporal information. We propose a spatio-temporal fusion network (STFN) that extracts temporal and spatial information from 3D and 2D networks. Also, we propose a method for stable training of the proposed STFN using a semi-foreground map. STFN can generate a compliant background model image and operate in real-time on a desktop with GPU. The proposed algorithm performs well in an environment different from training and is demonstrated by experiments using various public datasets.

Health Care Analytics With Time-Invariant and Time-Variant Feature Importance to Predict Hospital-Acquired Acute Kidney Injury: Observational Longitudinal Study

BackgroundAcute kidney injury (AKI) develops in 4% of hospitalized patients and is a marker of clinical deterioration and nephrotoxicity. AKI onset is highly variable in hospitals, which makes it difficult to time biomarker assessment in all patients for preemptive care.ObjectiveThe study sought to apply machine learning techniques to electronic health records and predict hospital-acquired AKI by a 48-hour lead time, with the aim to create an AKI surveillance algorithm that is deployable in real time.MethodsThe data were sourced from 20,732 case admissions in 16,288 patients over 1 year in our institution. We enhanced the bidirectional recurrent neural network model with a novel time-invariant and time-variant aggregated module to capture important clinical features temporal to AKI in every patient. Time-series features included laboratory parameters that preceded a 48-hour prediction window before AKI onset; the latter’s corresponding reference was the final in-hospital serum creatinine performed in case admissions without AKI episodes.ResultsThe cohort was of mean age 53 (SD 25) years, of whom 29%, 12%, 12%, and 53% had diabetes, ischemic heart disease, cancers, and baseline eGFR <90 mL/min/1.73 m2, respectively. There were 911 AKI episodes in 869 patients. We derived and validated an algorithm in the testing dataset with an AUROC of 0.81 (0.78-0.85) for predicting AKI. At a 15% prediction threshold, our model generated 699 AKI alerts with 2 false positives for every true AKI and predicted 26% of AKIs. A lowered 5% prediction threshold improved the recall to 60% but generated 3746 AKI alerts with 6 false positives for every true AKI. Representative interpretation results produced by our model alluded to the top-ranked features that predicted AKI that could be categorized in association with sepsis, acute coronary syndrome, nephrotoxicity, or multiorgan injury, specific to every case at risk.ConclusionsWe generated an accurate algorithm from electronic health records through machine learning that predicted AKI by a lead time of at least 48 hours. The prediction threshold could be adjusted during deployment to optimize recall and minimize alert fatigue, while its precision could potentially be augmented by targeted AKI biomarker assessment in the high-risk cohort identified.

Real-time performance evaluation of BGSLibrary algorithms for intelligent surveillance

&lt;p&gt;Background subtraction is the first and basic stage in video analysis and smart surveillance to extract moving objects. In fact, the background subtraction library (BGSLibrary) was created by Andrews Sobral in 2012, which currently combines 43 background subtraction algorithms from the most popular and widely used in the field of video analysis. Each algorithm has its own characteristics, strengths and weaknesses in extracting moving objects. The evaluation allows the identification of these characteristics and helps researchers to design the best methods. Unfortunately, the literature lacks a comprehensive evaluation of the algorithms included in the library. Accordingly, the present work will evaluate these algorithms in the BGSLibrary through the segmentation performance, execution time and processor, so as to, achieve a perfect, comprehensive, real-time evaluation of the system. Indeed, a background modeling challenge (BMC) dataset was selected using the synthetic video with the presence of noise. Results are presented in tables, columns and foreground masks.&lt;/p&gt;

Surveillance Strategies for Hepatocellular Carcinoma: Recent Advances and the Shifting Paradigm

Currently, international liver societies recommend screening at-risk individuals for HCC (patients with cirrhosis regardless of etiology, and/or chronic hepatitis B virus, and/or advanced liver fibrosis) with biannual abdominal ultrasound (USG) with or without alpha-fetoprotein (AFP). The global acceptance of USG in surveillance relies on the absence of risks, non-invasiveness, and lower costs. However, the suboptimal performance of USG ± AFP in reaching direct and indirect goals of HCC surveillance highlights the need for alternative surveillance strategies. Several studies targeted contrast-enhanced magnetic resonance imaging techniques, but the main barriers for their entrance to surveillance programs have been concerns about cost-effectivity and long scan times. Overall, the HCC risk stratification appears at hand by several validated multiple score systems, but their optimal performance is obtained only in populations who show highly homogenous clinical, pathological, epidemiologic, etiologic, and therapeutic characteristics, and this limitation poses a major drawback to their sustainable use in clinical practice. We need globally validated and molecular integrated risk stratification tools to shape the future tailored HCC surveillance decision algorithms. A dynamic process for HCC surveillance algorithms awaits us owing to the expected further prospective studies focusing on risk-stratified screening strategy.

New Binary Particle Swarm Optimization Algorithm for Surveillance and Camera Situation Assessments

New Binary Particle Swarm Optimization Algorithm for Surveillance and Camera Situation Assessments

A new proposal for secondary surveillance following potentially curative therapy of HCC: alternating MRI and CEUS

PurposeA high recurrence rate following ablative therapy of hepatocellular carcinoma (HCC) necessitates routine follow-up imaging (secondary surveillance) to facilitate early re-treatment. We evaluate our unique secondary surveillance algorithm (with use of alternating MRI and CEUS) by assessment of the relative diagnostic accuracy of MRI and CEUS in detection of residual/recurrent tumor. Potential benefits of alternating surveillance are compared to the use of MRI alone.Materials and methodsThis prospective observational IRB approved study included 231 patients with 354 treated tumors between January 2017 and June 2020. Treated lesions underwent secondary surveillance for a minimum of 7 months and up to 3 years, median follow-up 14 months. Secondary surveillance involved MRI performed at 1 month after treatment, followed by CEUS and MRI at alternate 3-month intervals (i.e., CEUS at month 4, MRI at month 7, etc.), for a total of 2 years. An equivocal finding on one imaging modality triggered expeditious evaluation with the alternate modality. Arterial phase hyperenhancement and washout comprise the classic features of recurrent tumor on both modalities.ResultsA total of 746 MRI and 712 CEUS examinations were performed, and a total of 184 tumor recurrences detected, MRI (n = 82) and CEUS (n = 102) (p = 0.19). There was no difference in the sensitivity (71.0–85.0% and 80.9–92.0%), specificity (97.4–99.2% and 98.5–99.9%), and area under the ROC curve (0.85–0.92 and 0.91–0.96) between MRI and CEUS, respectively. 23 of 82 recurrent tumors identified on MRI were equivocal and confirmed with expedited CEUS. 9 equivocal cases on MRI were disproved by expedited CEUS. On CEUS, 1 of the 102 recurrent tumors was equivocal and confirmed on MRI, and 2 equivocal CEUS cases were disproved by MRI.ConclusionMRI and CEUS performed similarly in our secondary surveillance algorithm for HCC in their ability to detect tumor recurrence, and showed no significant difference in their relative diagnostic test accuracy measures. Of greater interest, equivocal results on MRI (typically due to difficulty in distinguishing tumor recurrence from post-treatment change/shunting) were either confirmed or disproven by CEUS in all cases. Secondary surveillance of treated HCC with alternating MRI and CEUS shows equivalent performance of each modality. CEUS resolves equivocal MRI and optimally demonstrates APHE and washout in tumor recurrence.Graphic abstract

NIMG-57. MANAGEMENT OF TECTAL GLIOMAS IN ADULTS

Abstract BACKGROUND Tectal gliomas (TG) are rare tumors occurring primarily in children but also found in adults during workup of various neurological symptoms. Surgery is not required in asymptomatic cases, so histopathological information is sparse. No consensus on timing of imaging surveillance or management has been established. OBJECTIVE We seek to standardize neuroimaging, including MRI protocol and surveillance time intervals, and clinical management of symptoms and disease progression, including surgery, radiotherapy, and chemotherapy. METHODS At our institution, patients with TG were identified through a search of radiology reports and clinic notes between 1989 and 2020. Initial and serial MRI exams were evaluated for tumor size, enhancement, edema, hydrocephalus, and other radiographic features. When tissue was available, cellularity, mitotic activity, and morphology were described. We documented neurological symptoms and signs potentially related to the tumor. RESULTS 37 cases were identified: 22 female, 15 male; 5 children, 32 adults. Age of diagnosis ranged from 7 to 69 years. Presenting symptoms included headache (59%), visual symptoms (35%), and imbalance (14%), less commonly: seizure, weakness, nausea/vomiting, and dizziness. Surgical procedures included biopsy (9), resection (7), endoscopic third ventriculostomy (15), and shunt placement (11). Eight patients received radiotherapy, including IMRT, CyberKnife, GammaKnife, and Zap-X (all adults; 4 at diagnosis, 3 at progression, 1 at diagnosis and again at progression). Four patients received chemotherapy (all adults; 1 at diagnosis, 3 at progression), all with temozolomide. One additionally received bevacizumab for radionecrosis. Three patients died with progressive disease, two following treatment and one without. Of interest, 5 adult patients developed signs of parkinsonism during their follow-up period. CONCLUSION Management of TG encompasses both neoplastic progression and symptom control, either from local compression or infiltrative disease. We have developed an algorithm for imaging surveillance and treatment, including MRI protocol, definition of progressive disease, and indications for antineoplastic therapies.

Algorithm for Processing Measurements of a Hybrid Aviation Surveillance System

It is shown that a significant disadvantage of the broadcast-type automatic dependent surveillance system (ADS-B) used to solve the problem of aviation surveillance is its vulnerability to spoofing attempts. To eliminate this disadvantage, it is currently proposed to use the monitoring of ADS-B data using multilateration aviation surveillance systems (MLAT). The work shows the necessity of MLAT modernization to ensure a reliable solution to the problem of aviation surveillance in the event of failure of one or more reception points. To do this, it is proposed to use hybrid methods for assessing the coordinates of an aircraft. It can reduce the number of minimum required receiving positions due to the structural and informational redundancy of the aviation surveillance system. Structures (scenarios) of hybrid multi-position aviation surveillance systems and algorithms for processing their measurements have been developed. The algorithms ensure an increase in the reliability of the formation of estimates of aircraft coordinates.

A survey on horizon detection algorithms for maritime video surveillance: advances and future techniques

A survey on horizon detection algorithms for maritime video surveillance: advances and future techniques

Impact of including second and later cancers in cause-specific survival estimates using population-based registry data.

Second or later primary cancers account for approximately 20% of incident cases in the United States. Currently, cause-specific survival (CSS) analyses exclude these cancers because the cause of death (COD) classification algorithm was available only for first cancers. The authors added rules for later cancers to the Surveillance, Epidemiology, and End Results cause-specific death classification algorithm and evaluated CSS to include individuals with prior tumors. The authors constructed 2 cohorts: 1) the first ever primary cohort, including patients whose first cancer was diagnosed during 2000 through 2016) and 2) the earliest matching primary cohort, including patients with any cancer who matched the selection criteria irrespective of whether it was the first or a later cancer diagnosed during 2000 through 2016. The cohorts' CSS estimates were compared using follow-up through December 31, 2017. The new rules were used in the second cohort for patients whose first cancers during 2000 through 2016 were their second or later cancers. Overall, there were no statistically significant differences in CSS estimates between the 2 cohorts. Estimates were similar by age, stage, race, and time since diagnosis, except for patients with leukemia and those aged 65 to 74years (3.4 percentage point absolute difference). The absolute difference in CSS estimates for the first cancer ever cohort versus earliest of any cancers cohort in the study period was small for most cancer types. As the number of newly diagnosed patients with prior cancers increases, the algorithm will make CSS more inclusive and enable estimating survival for a group of patients with cancer for whom life tables are not available or life tables are available but do not capture other-cause mortality appropriately.

Occlusion-Aware UAV Path Planning for Reconnaissance and Surveillance

Unmanned Aerial Vehicles (UAVs) have become necessary tools for a wide range of activities including but not limited to real-time monitoring, surveillance, reconnaissance, border patrol, search and rescue, civilian, scientific and military missions, etc. Their advantage is unprecedented and irreplaceable, especially in environments dangerous to humans, for example, in radiation or pollution-exposed areas. Two path-planning algorithms for reconnaissance and surveillance are proposed in this paper, which ensures every point on the target ground area can be seen at least once in a complete surveillance circle. Moreover, the geometrically complex environments with occlusions are considered in our research. Compared with many existing methods, we decompose this problem into a waypoint-determination problem and an instance of the traveling-salesman problem.

Arteriovenous fistula maturation: Physical exam versus flow study

The purpose of this study was to determine the utility of routine duplex flow study 4 to 6 weeks after primary AVF creation and to compare physical exam against a duplex flow study in predicting fistula maturation. A surveillance algorithm was established to evaluate the naïve fistula after primary creation. This was a single institution retrospective review of 155 veterans with primary autogenous AVF creation from 2016 to 2018. All patients received a duplex flow study evaluation after primary creation. A comparison was made between physical exam (PE) and flow study at 4 to 6 weeks post creation. Sensitivities and specificities of physical exam and duplex flow study were compared head-to-head in predicting unassisted fistula maturation. A mature AVF was defined as a fistula that could be repetitively cannulated and provided adequate flow for dialysis. Failure of maturation was defined as an AVF that was never usable for dialysis. An abnormal duplex included thrombosis, stenosis (> 50% on gray scale imaging), inadequate vein diameter (< 4 mm), inadequate vein length or superficialization, or poor flow (< 500 ml/min). Bivariate comparisons were conducted using Pearson's χ², Fishers exact test, and Wilcoxon test depending on distribution. Significance was defined as P < 0.05. There were 53 patients with radiocephalic (RC) fistulas, 41 patients with brachiocephalic (BC) fistulas, and 6 patients with brachiobasilic (BB) fistulas. Of patients with a confirmed abnormal duplex ultrasound, 53% had an abnormal PE (sensitivity 53%; PPV 96.3%, P < 0.001). Of the patients with a confirmed normal duplex, 98% had a normal PE (specificity 98%; NPV 68.5%, P < 0.001). An abnormal duplex flow study had a 67% sensitivity for predicting AVF failure or need for reintervention while an abnormal physical examination had a 42% sensitivity in predicting AVF failure or need for reintervention (P < 0.001). In total, 48 fistulas needed reintervention, however only 20 (42%) were associated with an abnormal physical examination. Of those 48 reinterventions, 20 (42%) fistulas exhibited primary assisted maturation (P < 0.001). On duplex flow study alone, 32 patients had hemodynamically significant lesions necessitating reintervention, which went on to afford 9 (28%) primary assisted mature fistulas (P=0.69). Abnormal duplex flow studies have a better sensitivity for detecting AVF failure or the need for reintervention compared to physical exam alone. An abnormal duplex correlates more with needing a reintervention to achieve maturation than physical exam. Therefore, we advocate routine use of a postoperative duplex flow study to identify potentially correctable issues and optimize fistula maturation.

Automated healthcare-associated infection surveillance using an artificial intelligence algorithm

SummaryHealthcare-associated infections (HAIs) are among the most common adverse events in hospitals. We used artificial intelligence (AI) algorithms for infection surveillance in a cohort study. The model correctly detected 67 out of 73 patients with HAIs. The final model used a multilayer perceptron neural network achieving an area under receiver operating curve (AUROC) of 90.27%; specificity of 78.86%; sensitivity of 88.57%. Respiratory infections had the best results (AUROC ≥93.47%). The AI algorithm could identify most HAIs. AI is a feasible method for HAI surveillance, has the potential to save time, promote accurate hospital-wide surveillance, and improve infection prevention performance.

Evolutionary clustering and community detection algorithms for social media health surveillance

The prominent rise of social networks within the past decade have become a gold mine for data mining operations seeking to model the real world through these virtual worlds. One of the most important applications that has been proposed is utilizing information generated from social networks as a supplemental health surveillance system to monitor disease epidemics. At the time this research was conducted in 2020, the COVID-19 virus had evolved into a global pandemic, forcing many countries to implement preventative measures to halt its expanse. Health surveillance has been a powerful tool in placing further preventative measures, however it is not a perfect system, and slowly collected, misidentified information can prove detrimental to these efforts. This research proposes a new potential surveillance avenue through unsupervised machine learning using dynamic, evolutionary variants of clustering algorithms DBSCAN and the Louvain method to allow for community detection in temporal networks. This technique is paired with geographical data collected directly from the social media Twitter, to create an effective and accurate health surveillance system that grows as time passes. The experimental results show that the proposed system is promising and has the potential to be an advancement on current machine learning health surveillance techniques.