Search

Background Using reports of randomised trials of smoking cessation interventions as a test case, this study aimed to develop and evaluate machine learning (ML) algorithms for extracting information from study reports and predicting outcomes as part of the Human Behaviour-Change Project. It is the first of two linked papers, with the second paper reporting on further development of a prediction system. Methods Researchers manually annotated 70 items of information (‘entities’) in 512 reports of randomised trials of smoking cessation interventions covering intervention content and delivery, population, setting, outcome and study methodology using the Behaviour Change Intervention Ontology. These entities were used to train ML algorithms to extract the information automatically. The information extraction ML algorithm involved a named-entity recognition system using the ‘FLAIR’ framework. The manually annotated intervention, population, setting and study entities were used to develop a deep-learning algorithm using multiple layers of long-short-term-memory (LSTM) components to predict smoking cessation outcomes. Results The F1 evaluation score, derived from the false positive and false negative rates (range 0–1), for the information extraction algorithm averaged 0.42 across different types of entity (SD=0.22, range 0.05–0.88) compared with an average human annotator’s score of 0.75 (SD=0.15, range 0.38–1.00). The algorithm for assigning entities to study arms (e.g., intervention or control) was not successful. This initial ML outcome prediction algorithm did not outperform prediction based just on the mean outcome value or a linear regression model. Conclusions While some success was achieved in using ML to extract information from reports of randomised trials of smoking cessation interventions, we identified major challenges that could be addressed by greater standardisation in the way that studies are reported. Outcome prediction from smoking cessation studies may benefit from development of novel algorithms, e.g., using ontological information to inform ML (as reported in the linked paper 1 ).

Background Using reports of randomised trials of smoking cessation interventions as a test case, this study aimed to develop and evaluate machine learning (ML) algorithms for extracting information from study reports and predicting outcomes as part of the Human Behaviour-Change Project. It is the first of two linked papers, with the second paper reporting on further development of a prediction system. Methods Researchers manually annotated 70 items of information (‘entities’) in 512 reports of randomised trials of smoking cessation interventions covering intervention content and delivery, population, setting, outcome and study methodology using the Behaviour Change Intervention Ontology. These entities were used to train ML algorithms to extract the information automatically. The information extraction ML algorithm involved a named-entity recognition system using the ‘FLAIR’ framework. The manually annotated intervention, population, setting and study entities were used to develop a deep-learning algorithm using multiple layers of long-short-term-memory (LSTM) components to predict smoking cessation outcomes. Results The F1 evaluation score, derived from the false positive and false negative rates (range 0-1), for the information extraction algorithm averaged 0.42 across different types of entity (SD=0.22, range 0.05-0.88) compared with an average human annotator’s score of 0.75 (SD=0.15, range 0.38-1.00). The algorithm for assigning entities to study arms (e.g., intervention or control) was not successful. This initial ML outcome prediction algorithm did not outperform prediction based just on the mean outcome value or a linear regression model. Conclusions While some success was achieved in using ML to extract information from reports of randomised trials of smoking cessation interventions, we identified major challenges that could be addressed by greater standardisation in the way that studies are reported. Outcome prediction from smoking cessation studies may benefit from development of novel algorithms, e.g., using ontological information to inform ML (as reported in the linked paper (1)).

Introduction: Unilateral primary aldosteronism (UPA) is the most frequent surgically curable form of endocrine hypertension. Adrenalectomy is the cornerstone of treatment for UPA, but outcomes after surgery are variable. The cause of resistant hypertension after surgery is still a matter of debate. Our aim was to investigate cure rates after surgery and to evaluate preoperative factors that might influence the surgical outcome. Methods: Between 2000 and 2021, the charts of 71 Tunisian patients who underwent laparoscopic adrenalectomy for UPA were retrospectively reviewed. Preoperative medical records were collected and follow-up data (1-158 months) were registered. Antihypertensive medication doses were calculated using defined daily doses (DDD) and postoperative outcomes were assessed using the Primary Aldosteronism Surgical Outcome (PASO) criterion. Results: Of 91 enrolled patients, 71 (59% women, mean age 46 years, median length of follow-up 21 months) were suitable for evaluation. Thirty-four patients (48%) had complete clinical success according to the PASO criteria. The most relevant factors associated with complete clinical success on univariate analysis were: absence of diabetes (p= 0.007), low body mass index (BMI) (p= 0.001), lower preoperative DDD (p= 0.01), preoperatively controlled blood pressure (p= 0.024), higher plasma aldosterone to renin ratio (ARR) (p= 0.001), adenoma subtyping (p <0.001) and aldosteronoma resolution score (ARS) (p= 0.002). Multivariate regression analysis showed that the major predictors of complete clinical success were absence of diabetes (OR: 5.205), a BMI < 30 (OR: 4.930), a plasma ARR > 332 (OR: 4.554) and an ARS = 3 (OR: 2.056). Conclusion: Complete and partial clinical response rates were achieved in respectively 48 and 43% of cases. The main predictors of complete resolution of hypertension were absence of diabetes, low BMI, high plasma ARR and high ARS. Taking these factors into account may help identify patients at risk of persistent postoperative hypertension who may require long-term surveillance and medication. [ABSTRACT FROM AUTHOR]

IntroductionUnilateral primary aldosteronism (UPA) is the most frequent surgically curable form of endocrine hypertension. Adrenalectomy is the cornerstone of treatment for UPA, but outcomes after surgery are variable. The cause of resistant hypertension after surgery is still a matter of debate. Our aim was to investigate cure rates after surgery and to evaluate preoperative factors that might influence the surgical outcome.MethodsBetween 2000 and 2021, the charts of 71 Tunisian patients who underwent laparoscopic adrenalectomy for UPA were retrospectively reviewed. Preoperative medical records were collected and follow-up data (1-158 months) were registered. Antihypertensive medication doses were calculated using defined daily doses (DDD) and postoperative outcomes were assessed using the Primary Aldosteronism Surgical Outcome (PASO) criterion.ResultsOf 91 enrolled patients, 71 (59% women, mean age 46 years, median length of follow-up 21 months) were suitable for evaluation. Thirty-four patients (48%) had complete clinical success according to the PASO criteria. The most relevant factors associated with complete clinical success on univariate analysis were: absence of diabetes (p= 0.007), low body mass index (BMI) (p= 0.001), lower preoperative DDD (p= 0.01), preoperatively controlled blood pressure (p= 0.024), higher plasma aldosterone to renin ratio (ARR) (p= 0.001), adenoma subtyping (p 332 (OR: 4.554) and an ARS ≥ 3 (OR: 2.056).ConclusionComplete and partial clinical response rates were achieved in respectively 48 and 43% of cases. The main predictors of complete resolution of hypertension were absence of diabetes, low BMI, high plasma ARR and high ARS. Taking these factors into account may help identify patients at risk of persistent postoperative hypertension who may require long-term surveillance and medication.

The present investigation consists of a bibliometric review related to the prediction systems for the production of renewable energies through artificial intelligence. In this regard, a systematic bibliographic search of the topic was carried out in the Scopus database in November 2022 using the search equation: ( TITLEABS- KEY ( "artificial intelligence") OR TITLE-ABS-KEY ( " machine learning" ) OR TITLE-ABS-KEY ( "Deep learning") AND TITLE-ABSKEY ( "renewable energies " ) OR TITLE-ABS-KEY ( "alternative energies" ) OR TITLE-ABS-KEY ( "clean energies" " ) AND TITLE-ABSKEY ( "predictive model*" ) OR TITLE-ABS-KEY ( "forecast* systems" ) OR TITLE-ABS-KEY ( "predictive method" ) OR TITLE-ABS-KEY ( "predictive analytics " ) ). The results obtained show that between the years 2021 and 2022, 58% of all the investigations carried out are concentrated. In the geographical section, China (36) turned out to be the country that contributed the most papers to the subject studied, closely followed by the United States (34). In turn, the four journals that publish the most on the subject were ENERGIES (14), IEEE ACCESS (11), JOURNAL OF CLEANER PRODUCTION (6) and APPLIED ENERGY (5), while the 3 authors with the most published articles were Deo RC (12), Ghimire S (6) and Raj N (4). Finally, the three publications with the highest number of citations were Hu Q, 2016, Renew Energy (301), Ahmad MW, 2018, J Clean Prod (151) and Ahmad T, 2020, Sustainable Cities SOC (149). [ABSTRACT FROM AUTHOR]

Background: Using reports of randomised trials of smoking cessation interventions as a test case, this study aimed to develop and evaluate machine learning (ML) algorithms for extracting information from study reports and predicting outcomes as part of the Human Behaviour-Change Project. It is the first of two linked papers, with the second paper reporting on further development of a prediction system., Methods: Researchers manually annotated 70 items of information ('entities') in 512 reports of randomised trials of smoking cessation interventions covering intervention content and delivery, population, setting, outcome and study methodology using the Behaviour Change Intervention Ontology. These entities were used to train ML algorithms to extract the information automatically. The information extraction ML algorithm involved a named-entity recognition system using the 'FLAIR' framework. The manually annotated intervention, population, setting and study entities were used to develop a deep-learning algorithm using multiple layers of long-short-term-memory (LSTM) components to predict smoking cessation outcomes., Results: The F1 evaluation score, derived from the false positive and false negative rates (range 0-1), for the information extraction algorithm averaged 0.42 across different types of entity (SD=0.22, range 0.05-0.88) compared with an average human annotator's score of 0.75 (SD=0.15, range 0.38-1.00). The algorithm for assigning entities to study arms ( e.g. , intervention or control) was not successful. This initial ML outcome prediction algorithm did not outperform prediction based just on the mean outcome value or a linear regression model., Conclusions: While some success was achieved in using ML to extract information from reports of randomised trials of smoking cessation interventions, we identified major challenges that could be addressed by greater standardisation in the way that studies are reported. Outcome prediction from smoking cessation studies may benefit from development of novel algorithms, e.g. , using ontological information to inform ML (as reported in the linked paper 1 )., Competing Interests: Competing interests: RW and SM are unpaid directors of the Unlocking Behaviour Change Community Interest Company., (Copyright: © 2024 West R et al.)

Seasonal predictions in the Mediterranean region have relevant socio-economic implications, especially in the context of a changing climate. To date, sources of predictability have not been sufficiently investigated at the seasonal scale in this region. To fill this gap, we explore sources of predictability using a weather regimes (WRs) framework. The role of WRs in influencing regional weather patterns in the climate state has generated interest in assessing the ability of climate models to reproduce them. We identify four Mediterranean WRs for the winter (DJF) season and explore their sources of predictability looking at teleconnections with sea surface temperature (SST). In particular, we assess how SST anomalies affect the WRs frequencies during winter focussing on the two WRs that are associated with the teleconnections in which the signal is more intense: the Meridional and the Anticyclonic regimes. These sources of predictability are sought in five state-of-the-art seasonal forecasting systems included in the Copernicus Climate Change Services (C3S) suite finding a weaker signal but an overall good agreement with reanalysis data. Finally, we assess the ability of the C3S models in reproducing the reanalysis data WRs frequencies finding that their moderate skill increases during ENSO intense years, indicating that this teleconnection is well reproduced by the models and yields improved predictability in the Mediterranean region. [ABSTRACT FROM AUTHOR]

Monsoons affect the economy, agriculture, and human health of two thirds of the world’s population. Therefore, predicting variations in monsoon precipitation is societally important. We explore the ability of climate models from the sixth phase of the Climate Model Intercomparison Project to predict summer monsoon precipitation variability by using hindcasts from the Decadal Climate Prediction Project (Component A). The multi-model ensemble-mean shows significant skill at predicting summer monsoon precipitation from one year to 6–9 years ahead. However, this skill is dependent on the model, monsoon domain, and lead-time. In general, the skill of the multi-model ensemble-mean prediction is low in year 1 but increases for longer-lead times and is largely consistent with externally forced changes. The best captured region is northern Africa for the 2–5 and 6–9 year forecast lead times. In contrast, there is no significant skill using the ensemble-mean over East and South Asia and, furthermore, there is significant spread in skill among models for these domains. By sub-sampling the ensemble we show that the difference in skill between models is tied to the simulation of the externally forced response over East and South Asia, with models with a more skilful forced response capable of better predictions. A further contribution is from skilful prediction of Pacific Ocean temperatures for the South Asian summer monsoon at longer lead-times. Therefore, these results indicate that predictions of the East and South Asian monsoons could be significantly improved.

Objective: Hypertension cure following adrenalectomy in unilateral primary aldosteronism is not guaranteed. Its likelihood is associated with pre‐operative parameters, which have been variably combined in six different predictive scoring systems. The relative performance of these systems is currently unknown. The objective of this work was to identify the best performing scoring system for predicting hypertension cure following adrenalectomy for primary aldosteronism. Design: Retrospective analysis in a single tertiary referral centre. Patients: Eighty‐seven adult patients with unilateral primary aldosteronism who had undergone adrenalectomy between 2004 and 2018 for whom complete data sets were available to calculate all scoring systems. Measurements: Prediction of hypertension cure by each of the six scoring systems. Results: Hypertension cure was achieved in 36/87 (41.4%) patients within the first post‐operative year, which fell to 18/71 (25.4%) patients at final follow‐up (median 53 months, P =.002). Analysis of receiver operating characteristic area under the curves for the different scoring systems identified a difference in performance at early, but not late, follow‐up. For all systems, the area under the curve was lower at early compared with late follow‐up and compared to performance in the cohorts in which they were originally defined. Conclusions: No single scoring system performed significantly better than all others when applied in our cohort, although two did display particular advantages. It remains to be determined how best such scoring systems can be incorporated into the routine clinical care of patients with PA. [ABSTRACT FROM AUTHOR]

Extreme Wind Speed events (EWS) are responsible for the worst damages caused by wind in wind farms. An accurate estimation of the frequency and intensity of EWS is essential to avoid wind turbine damage and to minimize cut-out events in these facilities. In this paper we discuss how generative Data Augmentation (DA) techniques improve the performance of Machine Learning (ML) and Deep Learning (DL) algorithms in EWS prediction problems. These problems are usually tackled as classification tasks, which are highly unbalanced due to the small number of EWS events in wind farms. Different versions of Variational AutoEncoders (VAE) are proposed and analysed in this work (VAEs, Conditional VAEs (CVAEs) and Class-Informed VAEs (CI-VAE)) as generative DA techniques to balance the data in EWS problems, leading to better performance of the prediction systems. The proposed generative DA techniques have been compared against traditional DA algorithms in a real problem of EWS prediction in Spain, considering ERA5 reanalysis data as predictive variables. The results showed that the CI-VAE with a Convolutional Neural Network approach obtained the best results, with values of Precision 0.62, Recall 0.74 and F1 score 0.67, improving up to 4% the results of the method without data augmentation techniques. [Display omitted] • Generative data augmentation techniques can improve extreme wind prediction. • Variational Autoencoders modifications are used to generate synthetic wind extreme samples. • Experiments on real wind speed data from a Spanish wind farm are presented. • Comparisons with traditional data augmentation techniques show prediction improvements. [ABSTRACT FROM AUTHOR]

The paper presents the application of recurrent LSTM neural networks for short-time load forecasting in the Polish Power System (PPS) and a small region of a power system in Central Poland. The objective of the present work was to develop an efficient and accurate method of forecasting the 24-h pattern of power load with a 1-h and 24-h horizon. LSTM showed effectiveness in predicting the irregular trends in time series. The final forecast is estimated using an ensemble consisted of five independent predictions. Numerical experiments proved the superiority of the ensemble above single predictor resulting in a reduction of the MAPE the RMSE error by more than 6% in both forecasting tasks.

Quasi-optical beam forming structures are very interesting and involve different kinds of methods for their analysis. In order to optimize the cost function of our target solution, it is important to develop a predictive system to determine where the receiving antenna should be located for more efficient propagation of electromagnetic waves. Solving this antenna problem with the mathematical theory of diffraction is a complicated task due to the involved integrals with hyper-singular singularities in their kernels. This has raised the objective to develop a discrete mathematical model which could solve those difficulties. [ABSTRACT FROM AUTHOR]

We construct universal prediction systems in the spirit of Popper's falsifiability and Kolmogorov complexity and randomness. These prediction systems do not depend on any statistical assumptions (but under the IID assumption they dominate, to within the usual accuracy, conformal prediction). Our constructions give rise to a theory of algorithmic complexity and randomness of time containing analogues of several notions and results of the classical theory of Kolmogorov complexity and randomness. [ABSTRACT FROM AUTHOR]

We study the problem of predicting likely places of visit of users using their past tweets. What people write on their microblogs reflects their intent and desire relating to most of their common day interests. Taking this as a strong evidence, we hypothesize that tweets of the person can also be treated as source of strong indicator signals for predicting their places of visits. In this paper, we propose a novel approach for predicting place of visit within a given geospatial range considering the past tweets and the time of visit. These predictions can be used for generating places recommendation or for promotions. In this approach, we analyze use of various features that can be extracted from the historical tweets-for example, personality traits estimated from the past tweets and the actual words mentioned in the tweets. We performed extensive empirical experiments involving, real data derived from twitter timelines of 4600 persons with multi-label classification as predictive model. The performances of proposed approach outperform the four baselines with accuracy reaching 90 % for top five predictions. Based on our experimental study, we come up with general guidelines on building the prediction model in terms of the type of features extracted from historical tweets, window size of historical tweets and on the optimal radius of query around the place of visit at a given time. [ABSTRACT FROM AUTHOR]

Estudou-se o efeito da infiltração de magnésio e cálcio no desenvolvimento de sintomas de bitter pit de maçãs 'Gala'. Examinou-se, também, a eficiência da infiltração de magnésio e da aceleração da maturação de maçãs 'Royal Gala' como métodos de previsão da ocorrência de bitter pit após o armazenamento. No primeiro experimento, maçãs 'Gala' foram infiltradas com MgCl2; CaCl2 ou MgCl2 mais CaCl2. No segundo experimento, maçãs 'Royal Gala' foram infiltradas com MgCl2; ou imersas em solução contendo ethephon por cinco minutos. Como controle, frutos de cada cultivar foram armazenados em atmosfera controlada por seis meses mais 18 dias a 20°C. Frutos infiltrados com MgCl2 apresentaram significativo acréscimo na incidência e severidade de bitter pit, proporcional à concentração de MgCl2 na solução. Frutos infiltrados com CaCl2 não apresentaram sintomas de bitter pit. A aceleração da maturação de maçãs 'Royal Gala' na colheita foi efetiva na previsão de bitter pit após o armazenamento.The effect of magnesium and calcium infiltration on bitter pit symptoms development in Gala apples was studied. It was also examined the efficiency of magnesium infiltration and the ripening hastening of Royal Gala apples as a predictive methods of bitter pit occurrence after storage. In the first experiment, Gala apples were infiltrated with MgCl2; CaCl2; or MgCl2 plus CaCl2. In the second experiment Royal Gala apples were infiltrated with MgCl2; or immersed into solution containing ethephon during five minutes. As a control a sample of each cultivar was stored in controlled atmosphere during 6 months plus eighteen days at 20°C. Fruits infiltrated with MgCl2 showed increased incidence and severity of bitter pit proportional to the MgCl2 concentration in the solution. Infiltration with CaCl2 prevents bitter pit-like symptoms. The ripening hastening of Royal Gala apples at harvest was effective on the bitter pit prediction after storage.

Objectives The EuResist expert system is a novel data-driven online system for computing the probability of 8-week success for any given pair of HIV-1 genotype and combination antiretroviral therapy regimen plus optional patient information. The objective of this study was to compare the EuResist system vs. human experts (EVE) for the ability to predict response to treatment. Methods The EuResist system was compared with 10 HIV-1 drug resistance experts for the ability to predict 8-week response to 25 treatment cases derived from the EuResist database validation data set. All current and past patient data were made available to simulate clinical practice. The experts were asked to provide a qualitative and quantitative estimate of the probability of treatment success. Results There were 15 treatment successes and 10 treatment failures. In the classification task, the number of mislabelled cases was six for EuResist and 6-13 for the human experts [mean ± standard deviation (SD) 9.1 ± 1.9]. The accuracy of EuResist was higher than the average for the experts (0.76 vs. 0.64, respectively). The quantitative estimates computed by EuResist were significantly correlated (Pearson r = 0.695, P = 0.0001) with the mean quantitative estimates provided by the experts. However, the agreement among experts was only moderate (for the classification task, inter-rater k = 0.355; for the quantitative estimation, mean ± SD coefficient of variation = 55.9 ± 22.4%). Conclusions With this limited data set, the EuResist engine performed comparably to or better than human experts. The system warrants further investigation as a treatment-decision support tool in clinical practice. [ABSTRACT FROM AUTHOR]

The effect of magnesium and calcium infiltration on bitter pit symptoms development in Gala apples was studied. It was also examined the efficiency of magnesium infiltration and the ripening hastening of Royal Gala apples as a predictive methods of bitter pit occurrence after storage. In the first experiment, Gala apples were infiltrated with MgCl2; CaCl2; or MgCl2 plus CaCl2. In the second experiment Royal Gala apples were infiltrated with MgCl2; or immersed into solution containing ethephon during five minutes. As a control a sample of each cultivar was stored in controlled atmosphere during 6 months plus eighteen days at 20°C. Fruits infiltrated with MgCl2 showed increased incidence and severity of bitter pit proportional to the MgCl2 concentration in the solution. Infiltration with CaCl2 prevents bitter pit-like symptoms. The ripening hastening of Royal Gala apples at harvest was effective on the bitter pit prediction after storage. [ABSTRACT FROM AUTHOR]

The article focuses on the application of adaptive network-based fuzzy inference system (ANFIS) algorithm to lessen the delay in real-time gaze control. It describes the fuzzy inference system and learning algorithm for ANFIS architecture and reveals the use of colour segmentation algorithm and of proportional-integral-derivative (PID) controller to explain the whole system implementation. It reveals that the neurofuzzy system allows fast and precise motion control of the target object.

Abstract: Scorpion toxins are common experimental tools for studies of biochemical and pharmacological properties of ion channels. The number of functionally annotated scorpion toxins is steadily growing, but the number of identified toxin sequences is increasing at much faster pace. With an estimated 100,000 different variants, bioinformatic analysis of scorpion toxins is becoming a necessary tool for their systematic functional analysis. Here, we report a bioinformatics-driven system involving scorpion toxin structural classification, functional annotation, database technology, sequence comparison, nearest neighbour analysis, and decision rules which produces highly accurate predictions of scorpion toxin functional properties. [Copyright &y& Elsevier]

Both qualitative and quantitative modeling methods relating chemical structure to biological activity, called structure-activity relationship analyses or SAR, are applied to the prediction and characterization of chemical toxicity. This minireview will discuss some generic issues and modeling approaches that are tailored to problems in toxicology. Different approaches to, and some facets and limitations of the practice and science of, SAR as they pertain to current toxicology analyses, and the basic elements of SAR and SAR-model development and prediction systems are discussed. Other topics include application of 3-D SAR to understanding of the propensity of chemicals to cause endocrine disruption, and the use of models to analyze biological activity of metal ions in toxicology. An example of integration of knowledge pertaining to mechanisms into an expert system for prediction of skin sensitization to chemicals is also discussed. This minireview will consider the utility of modeling approaches as one component for better integration of physicochemical and biological properties into risk assessment, and also consider the potential for both environmental and human health effects of chemicals and their interactions. [ABSTRACT FROM PUBLISHER]

The paper presents the application of recurrent LSTM neural networks for short-time load forecasting in the Polish Power System (PPS) and a small region of a power system in Central Poland. The objective of the present work was to develop an efficient and accurate method of forecasting the 24-h pattern of power load with a 1-h and 24-h horizon. LSTM showed effectiveness in predicting the irregular trends in time series. The final forecast is estimated using an ensemble consisted of five independent predictions. Numerical experiments proved the superiority of the ensemble above single predictor resulting in a reduction of the MAPE the RMSE error by more than 6% in both forecasting tasks. [ABSTRACT FROM AUTHOR]

Different software reliability models can produce very different answers when called upon to predict future reliability in a reliability growth context. Users need to know which, if any, of the competing predictions are trustworthy. Some techniques are presented which form the basis of a partial solution to this problem. Rather than attempting to decide which model is generally best, the approach adopted here allows a user to decide upon the most appropriate model for each application. [ABSTRACT FROM AUTHOR]

There are two distinct notions of validation for software measures. On the one hand there is the informal notion that a measure is only valid if it is useful and practical. Although this view of validation is important, it is almost impossible to test. On the other hand there is a format notion of validation of measures and it is this view which is addressed here. By considering software measures in the context of measurement theory, and by considering a classification of software entities into products, processes and resources the paper describes what the authors believe are the necessary (but not sufficient) activities required for validation, and show how far this deviates from the `accepted' view of validation. [ABSTRACT FROM AUTHOR]

Pheochromocytoma and paraganglioma (PPGL) are rare neuroendocrine tumors that arising from the adrenal medulla or extra-adrenal autonomic ganglia. Traditionally, PPGL was classified as benign or malignant based on the presence of distant metastasis at the time of initial surgery. However, according to WHO 2017 Classification of Tumors of Endocrine Organs, all PPGL has metastatic potential. The term "metastatic" is used, replacing "malignant" in this group of tumors. The prediction of PPGL's metastatic potential is a clinical concern, although many relevant indicators such as genetics, histology, pathology and molecular biology markers have been proved to be related to the metastasis of PPGL, but none of them is 100% predictive; various types of prediction systems had been created, but previous studies had demonstrated that they still need to be validated in multicenter studies. There is no unified clinical standard to differentiate metastatic from non-metastatic and a highly effective prediction system is of urgent need. In this review, we summarized all reported prediction systems, including the PASS system, the GAPP system, the COPPs system and the ASES system. Additional potential indicators that related to metastatic PPGL were also introduced., Competing Interests: None., (AJCR Copyright © 2020.)