Probabilistic Graphical Models Research Articles

Enhanced multi-model multi-scan data association and tracking algorithm via convex variational inference

Tracking multiple targets with unknown measurement-to-target association and uncertain target dynamics is a significant problem that arises in various applications such as surveillance monitoring and intelligent transportation systems. In this paper, we propose an enhanced multi-model multi-scan data association algorithm to address the problem of tracking multiple maneuvering targets. First, we use a probabilistic graphical model to represent the joint distribution of the dynamic model indices, target state, and multi-scan data association variables. This formulation transforms the inference of marginal distributions into a Bethe free energy (BFE) problem. Next, to transform the BFE problem into a convex one, we demonstrate that the BFE function can be made convex through re-weighting. Additionally, we decompose the re-weighted BFE function into a block-wise sum form. We prove that under certain regularization conditions, each block of the re-weighted BFE is convex, ensuring convergence of the primal–dual coordinate ascent algorithm to the minimum of the overall re-weighted BFE. Finally, we provide a particle implementation of the proposed algorithm, accompanied by an analysis of its complexity. Simulation results indicate that the proposed algorithm exhibits favorable performance when compared to both the single-model multi-scan algorithm and the multi-model single-scan algorithm.

A Spatiotemporal Probabilistic Graphical Model Based on Adaptive Expectation-Maximization Attention for Individual Trajectory Reconstruction Considering Incomplete Observations

Spatiotemporal information on individual trajectories in urban rail transit is important for operational strategy adjustment, personalized recommendation, and emergency command decision-making. However, due to the lack of journey observations, it is difficult to accurately infer unknown information from trajectories based only on AFC and AVL data. To address the problem, this paper proposes a spatiotemporal probabilistic graphical model based on adaptive expectation maximization attention (STPGM-AEMA) to achieve the reconstruction of individual trajectories. The approach consists of three steps: first, the potential train alternative set and the egress time alternative set of individuals are obtained through data mining and combinatorial enumeration. Then, global and local potential variables are introduced to construct a spatiotemporal probabilistic graphical model, provide the inference process for unknown events, and state information about individual trajectories. Further, considering the effect of missing data, an attention mechanism-enhanced expectation-maximization algorithm is proposed to achieve maximum likelihood estimation of individual trajectories. Finally, typical datasets of origin-destination pairs and actual individual trajectory tracking data are used to validate the effectiveness of the proposed method. The results show that the STPGM-AEMA method is more than 95% accurate in recovering missing information in the observed data, which is at least 15% more accurate than the traditional methods (i.e., PTAM-MLE and MPTAM-EM).

Exploring the Relationship Between Key Factors and Victory and Defeat in Volleyball Matches using Bayesian Network Modeling

Identifying key factors contributing to victory and defeat in volleyball matches using Bayesian network modeling involves probabilistic graphical modeling to analyze the interrelationships among various factors and their influence on match outcomes. By integrating data on player performance, team strategies, match conditions, and opponent characteristics, Bayesian networks can uncover complex dependencies and predict the likelihood of winning or losing based on different scenarios. The research conducted on exploring the relationship between key factors and victory and defeat in volleyball matches using Bayesian network modeling. Through the utilization of Bayesian network modeling, this study aims to elucidate the intricate interplay between different variables and their impact on match outcomes in the context of volleyball. The research employs a Hidden Markov Process Bayesian Network (HMPBN) to model the temporal dependencies and uncertainties inherent in volleyball matches. The dataset comprises a comprehensive collection of volleyball match data, encompassing various observable variables such as points scored, successful serves, defensive plays, and more. The simulation setup involves training the HMPBN model on a dataset consisting of 1000 instances and evaluating its performance on a separate test dataset of 500 instances. The model parameters, including the transition matrix, emission matrix, and initial state distribution, are determined through rigorous statistical estimation techniques. The results of the simulation experiments provide valuable insights into the probabilities associated with different actions and their influence on match outcomes. The study reveals that successful serves have a high probability of contributing to match victories, with an average success rate of 80% in the dataset. Conversely, defensive plays exhibit a lower success rate, with an average probability of success of 60%. The Hidden Markov Process Bayesian Network (HMPBN) achieves an overall classification accuracy of 85% on the test dataset, demonstrating its effectiveness in predicting match outcomes.

Evaluation of Teaching Effectiveness Based on Bayesian Network Algorithm in Teaching and Learning Process in Higher Education Institutions

The Bayesian network algorithm plays a pivotal role in enhancing the teaching and learning process in higher education institutions by facilitating personalized and adaptive learning experiences. This algorithm leverages probabilistic graphical models to represent relationships among various educational variables, such as student performance, learning resources, and teaching methods. Additionally, Bayesian networks aid instructors in optimizing course design and instructional strategies by identifying areas for improvement and evaluating the effectiveness of different teaching approaches. This paper explores the application of Hidden Markov Chain Bayesian Networks (HMCBN) in analyzing teaching effectiveness within higher education institutions. Teaching effectiveness is a multifaceted concept influenced by various factors, including teaching strategies, student characteristics, and learning outcomes. Traditional statistical methods often struggle to capture the dynamic and interdependent nature of these factors. Through the construction of HMCBN models, this study investigates the probabilistic relationships between teaching strategies, student characteristics, learning outcomes, and teaching effectiveness. For teaching strategies, the probabilities [0.3, 0.5, 0.2] indicate the likelihood of each strategy being employed, suggesting that Strategy 2 is the most frequently utilized, followed by Strategy 1 and Strategy 3. Similarly, the probabilities [0.4, 0.3, 0.3] for student characteristics suggest a relatively balanced distribution among the three characteristic categories. In terms of learning outcomes, the conditional probabilities reflect the influence of both teaching strategies and student characteristics on the outcomes achieved. For each combination of teaching strategy and student characteristic, the probabilities [0.1, 0.3, 0.6], [0.4, 0.5, 0.1], [0.3, 0.4, 0.3], [0.2, 0.5, 0.3], [0.3, 0.4, 0.3], [0.4, 0.3, 0.3], [0.1, 0.7, 0.2], [0.4, 0.5, 0.1], and [0.5, 0.4, 0.1] represent the probabilities of achieving different learning outcomes given specific combinations of teaching strategies and student characteristics.

A Blended Teaching Model of College English Based on Deep Learning

Blending traditional face-to-face instruction with online learning components, this model aims to enhance student engagement and improve learning outcomes. This paper introduces a novel blended teaching model for college English courses, integrating Conditional Random Field (CRF) techniques with deep learning frameworks. Blended learning, combining traditional face-to-face instruction with online learning components, has gained traction in higher education for its potential to enhance student engagement and learning outcomes. CRF blended teaching integrates the principles of conditional random fields, a powerful probabilistic graphical model, with traditional teaching methodologies to enhance learning outcomes. Unlike conventional teaching approaches that often rely on linear or static instructional methods, CRF blended teaching embraces a more dynamic and adaptive framework. By leveraging CRF, which models the dependencies between different instructional variables, educators can design personalized learning experiences tailored to individual student needs and learning styles. This approach enables the seamless integration of diverse instructional modalities, including face-to-face instruction, online resources, and interactive learning activities. The proposed model CRF algorithms to effectively capture the sequential dependencies inherent in language learning tasks, such as sentence parsing and part-of-speech tagging. By integrating deep learning architectures, including recurrent neural networks (RNNs) and transformers, model with large-scale linguistic corpora to improve accuracy and generalization. The model achieves a 15% increase in students' English proficiency scores and a 25% improvement in comprehension compared to traditional teaching methods, underscoring its effectiveness in enhancing learning outcomes.

3D Multiresolution Velocity Model Fusion with Probability Graphical Models

ABSTRACT The variability in spatial resolution of seismic velocity models obtained via tomographic methodologies is attributed to many factors, including inversion strategies, ray-path coverage, and data integrity. Integration of such models, with distinct resolutions, is crucial during the refinement of community models, thereby enhancing the precision of ground-motion simulations. Toward this goal, we introduce the probability graphical model (PGM), combining velocity models with heterogeneous resolutions and nonuniform data point distributions. The PGM integrates data relations across varying resolution subdomains, enhancing detail within low-resolution (LR) domains by utilizing information and prior knowledge from high-resolution (HR) subdomains through a maximum posterior problem. Assessment of efficacy, utilizing both 2D and 3D velocity models—consisting of synthetic checkerboard models and a fault-zone model from Ridgecrest, California—demonstrates noteworthy improvements in accuracy, compared to state-of-the-art fusion techniques. Specifically, we find reductions of 30% and 44% in computed travel-time residuals for 2D and 3D models, respectively, as compared to conventional smoothing techniques. Unlike conventional methods, the PGM’s adaptive weight selection facilitates preserving and learning details from complex, nonuniform HR models and applies the enhancements to the LR background domain.

VertiBayes: learning Bayesian network parameters from vertically partitioned data with missing values

AbstractFederated learning makes it possible to train a machine learning model on decentralized data. Bayesian networks are widely used probabilistic graphical models. While some research has been published on the federated learning of Bayesian networks, publications on Bayesian networks in a vertically partitioned data setting are limited, with important omissions, such as handling missing data. We propose a novel method called VertiBayes to train Bayesian networks (structure and parameters) on vertically partitioned data, which can handle missing values as well as an arbitrary number of parties. For structure learning we adapted the K2 algorithm with a privacy-preserving scalar product protocol. For parameter learning, we use a two-step approach: first, we learn an intermediate model using maximum likelihood, treating missing values as a special value, then we train a model on synthetic data generated by the intermediate model using the EM algorithm. The privacy guarantees of VertiBayes are equivalent to those provided by the privacy preserving scalar product protocol used. We experimentally show VertiBayes produces models comparable to those learnt using traditional algorithms. Finally, we propose two alternative approaches to estimate the performance of the model using vertically partitioned data and we show in experiments that these give accurate estimates.

Causal chain event graphs for remedial maintenance.

The analysis of system reliability has often benefited from graphical tools such as fault trees and Bayesian networks. In this article, instead of conventional graphical tools, we apply a probabilistic graphical model called the chain event graph (CEG) to represent the failures and processes of deterioration of a system. The CEG is derived from an event tree and can flexibly represent the unfolding of asymmetric processes. For this application, we need to define a new class of formal intervention we call remedial to model the causal effects of remedial maintenance. This fixes the root causes of a failure and returns the status of the system to as good as new. We demonstrate that the semantics of the CEG are rich enough to express this novel type of intervention. Furthermore, through the bespoke causal algebras, the CEG provides a transparent framework with which to guide and express the rationale behind predictive inferences about the effects of various types of remedial intervention. A backdoor theorem is adapted to apply to these interventions to help discover when a system is only partiallyobserved.

Enhancing Breast Cancer Detection with Ensemble Machine Learning Models

Breast cancer remains a significant public health issue worldwide, underlining the need for accurate and efficient diagnostic methods. In this paper, we propose a new technique to enhance breast cancer diagnosis through the integration of multiple machine-learning models. Our strategy employs a combination of the Naive Bayes classifier, Stochastic Gradient Descent (SGD), Bagging, and the ZeroR classifier, alongside Bayes Network learning. The cornerstone of our approach is Bayes Network learning, a probabilistic graphical model designed to map out the intricate interconnections among various diagnostic factors. This is significant in the way that it can help to uncover complex relationships in the data for the sake of leading to more accurate predictions. Added to the above, we use the Naïve Bayes classifier, a classifier showing good validity in classification tasks and based on probabilistic reasoning, for the screening of breast cancer. Further, a refined model's parameter is included using the SGD and leads to enhancement of the generalization and overall performance of the model. In addition, as part of controlling overfitting, one can also use Bagging. It is an ensemble method in the sense that it considers several models. ZeroR classifier is a very basic classifier and is just used to compare its performance with our composite approach. We are comparing complex ensemble results to its simplicity. We will validate the ability of our proposed methodology to compare the performance of our integrated models against ZeroR.

An IoT‐based multiple teaching quality evaluation method for English translation with improved deep learning

AbstractTo enhance the teaching quality (TQ) of English translation (ET) courses, multivariate evaluation methods have garnered significant attention for their ability to identify deficiencies in the teaching process and subsequently improve instructional standards. Traditional evaluation techniques have shown limited effectiveness. Therefore, this paper proposes an ET in multiple teaching quality (MTQ) evaluation method with improved deep learning in internet of things (IoT). First, the paper establishes evaluation indicators for ET in MTQ, which mainly encompass ET teaching effectiveness, pedagogic competency in ET, teaching methods, content, as well as attitudes towards ET teaching. It then utilizes IoT technology to preliminarily collect data on these indicators, and using clustering algorithm based on weighted attributes and density (CABWAD) algorithm to mine data on English teaching MTQ evaluation indicators for ET. The extracted evaluation indicator data is then denoised using a probabilistic undirected graph model. Ultimately, the multilayer perceptron in deep learning is improved through Wolfe line search optimization, and this enhanced multilayer perceptron is employed to construct an ET in MTQ evaluation model. The denoised indicator data is inputted into the model, which then outputs precise MTQ evaluation results. The results show that the absolute value of the average Pearson correlation coefficient of this method is the highest, the Spearman correlation coefficient is the lowest, the mean average precision value is 0.965, and the positive category imbalance degree and the negative category imbalance degree are the lowest, indicating that the proposed method has outstanding performance in all aspects, and has certain application value in the field of TQ evaluation.

Dynamic Bayesian network structure learning based on an improved bacterial foraging optimization algorithm

With the rapid development of artificial intelligence and data science, Dynamic Bayesian Network (DBN), as an effective probabilistic graphical model, has been widely used in many engineering fields. And swarm intelligence algorithm is an optimization algorithm based on natural selection with the characteristics of distributed, self-organization and robustness. By applying the high-performance swarm intelligence algorithm to DBN structure learning, we can fully utilize the algorithm's global search capability to effectively process time-based data, improve the efficiency of network generation and the accuracy of network structure. This study proposes an improved bacterial foraging optimization algorithm (IBFO-A) to solve the problems of random step size, limited group communication, and the inability to maintain a balance between global and local searching. The IBFO-A algorithm framework comprises four layers. First, population initialization is achieved using a logistics-sine chaotic mapping strategy as the basis for global optimization. Second, the activity strategy of a colony foraging trend is constructed by combining the exploration phase of the Osprey optimization algorithm. Subsequently, the strategy of bacterial colony propagation is improved using a "genetic" approach and the Multi-point crossover operator. Finally, the elimination-dispersal activity strategy is employed to escape the local optimal solution. To solve the problem of complex DBN learning structures due to the introduction of time information, a DBN structure learning method called IBFO-D, which is based on the IBFO-A algorithm framework, is proposed. IBFO-D determines the edge direction of the structure by combining the dynamic K2 scoring function, the designed V-structure orientation rule, and the trend activity strategy. Then, according to the improved reproductive activity strategy, the concept of "survival of the fittest" is applied to the network candidate solution while maintaining species diversity. Finally, the global optimal network structure with the highest score is obtained based on the elimination-dispersal activity strategy. Multiple tests and comparison experiments were conducted on 10 sets of benchmark test functions, two non-temporal and temporal data types, and six data samples of two benchmark 2T-BN networks to evaluate and analyze the optimization performance and structure learning ability of the proposed algorithm under various data types. The experimental results demonstrated that IBFO-A exhibits good convergence, stability, and accuracy, whereas IBFO-D is an effective approach for learning DBN structures from data and has practical value for engineering applications.

Probabilistic Graph Model Based Recommendation Algorithm for Material Selection in Self-directed Learning

Faced the vast amount of information, choosing the appropriate materials is a prerequisite for effective self-directed learning. The recommendation algorithm is a kind of intelligent technology that can accurately locate the required information which the users care about most. However, many recommendation techniques experience can not be trained adequately in scenarios with small sample data and extremely sparse ratings. Moreover, DLRAs (Deep learning based Recommendation Algorithms) require high hardware support. The probabilistic graph (PG) can effectively represent the implicit complex relations among nodes, but it still has the problem of sparse data sensitivity. Therefore, we propose a Matrix-Factorization-based Probabilistic Graph Model for Recommendation Algorithm (MF-PGMRA): By matrix-factorizing the sparse rating matrix, the users and items are mapped to the user/item spaces, respectively; We employ the inner product to data-enhance and overcome the problems of sparse data and cold start; Then, we build Probabilistic Graph to construct the “user-item” latent spaces and estimate the probability distribution based on expectation maximization (EM), so as to predict the ratings; Finally, we built a library management system with the recommendation module to highlight the benefits of MF-PGMRA for students’ subject learning. According to a questionnaire, we confirmed that the students are satisfied with the system from four aspects of speed, accuracy, usability and convenience, which can confirm that the library management system based on MF-PGMRA can efficiently and accurately recommend suitable materials for students from the huge amount of learning materials to improve students’ self-directed learning efficiency.

Out-of-Sample Tuning for Causal Discovery.

Causal discovery is continually being enriched with new algorithms for learning causal graphical probabilistic models. Each one of them requires a set of hyperparameters, creating a great number of combinations. Given that the true graph is unknown and the learning task is unsupervised, the challenge to a practitioner is how to tune these choices. We propose out-of-sample causal tuning (OCT) that aims to select an optimal combination. The method treats a causal model as a set of predictive models and uses out-of-sample protocols for supervised methods. This approach can handle general settings like latent confounders and nonlinear relationships. The method uses an information-theoretic approach to be able to generalize to mixed data types and a penalty for dense graphs to penalize for complexity. To evaluate OCT, we introduce a causal-based simulation method to create datasets that mimic the properties of real-world problems. We evaluate OCT against two other tuning approaches, based on stability and in-sample fitting. We show that OCT performs well in many experimental settings and it is an effective tuning method for causal discovery.

베이지언 네트워크를 이용한 인지진단모형의 적용

Bayesian networks(BN) are a probabilistic graphical model that represents a system of variables and
 their conditional interdependencies using directed acyclic graphs. Once the probabilistic models are
 built and embedded in a graphical structure, evidence of variables can be efficiently propagated
 through the network and inference about students can be made. This study is to explore the potential
 of BN as a new framework for cognitive diagnostic modeling by developing and applying BN-based
 cognitive diagnostic models to a real test data set and comparing the results with those from the
 conventional cognitive diagnostic modeling. For this purpose, the current study surveyed theoretical and
 pedagogical aspects of reading in English as a foreign/second language, introduced provisional models,
 and applied the models to a set of reading comprehension data from a TOEIC practice test. The
 results show that BN-based cognitive diagnostic modeling can be an efficient framework that can
 handle diverse aspects of reading comprehension. The study then addresses some issues from the
 analysis and discusses the implications in relation to application of cognitive diagnostic modeling.

Bayesian networks supporting management practices: A multifaceted perspective based on the literature

Bayesian network is a probabilistic graphical model within machine learning that supports decision-making under conditions of uncertainty in different domains. Although the scientific literature has increasingly addressed the implementation of Bayesian networks to support management practices (BNM), a thorough review is currently lacking. This bibliometric review investigates the transformative potential of Bayesian networks in reshaping decision-making paradigms across multidisciplinary domains. The knowledge set findings reveal a predominant focus on risk management within the Engineering domain; the scientific openings involve significant progress in both theoretical frameworks and practical applications across Computer Science, Engineering, Medicine, and Environmental Science; and research trends indicate a progressive BNM within Engineering and Medicine, contrasting with a decline in innovative studies related to Computer Science. This study acts as a catalyst, propelling inventive BNM applications and fostering interdisciplinary advancements. It lays a foundation for pioneering BNM strategies.

Collaborative Synthesis of Patient Records through Multi-Visit Health State Inference

Electronic health records (EHRs) have become the foundation of machine learning applications in healthcare, while the utility of real patient records is often limited by privacy and security concerns. Synthetic EHR generation provides an additional perspective to compensate for this limitation. Most existing methods synthesize new records based on real EHR data, without consideration of different types of events in EHR data, which cannot control the event combinations in line with medical common sense. In this paper, we propose MSIC, a Multi-visit health Status Inference model for Collaborative EHR synthesis to address these limitations. First, we formulate the synthetic EHR generation process as a probabilistic graphical model and tightly connect different types of events by modeling the latent health states. Then, we derive a health state inference method tailored for the multi-visit scenario to effectively utilize previous records to synthesize current and future records. Furthermore, we propose to generate medical reports to add textual descriptions for each medical event, providing broader applications for synthesized EHR data. For generating different paragraphs in each visit, we incorporate a multi-generator deliberation framework to collaborate the message passing of multiple generators and employ a two-phase decoding strategy to generate high-quality reports. Our extensive experiments on the widely used benchmarks, MIMIC-III and MIMIC-IV, demonstrate that MSIC advances state-of-the-art results on the quality of synthetic data while maintaining low privacy risks.

CoLAL: Co-learning Active Learning for Text Classification

In the machine learning field, the challenge of effectively learning with limited data has become increasingly crucial. Active Learning (AL) algorithms play a significant role in this by enhancing model performance. We introduce a novel AL algorithm, termed Co-learning (CoLAL), designed to select the most diverse and representative samples within a training dataset. This approach utilizes noisy labels and predictions made by the primary model on unlabeled data. By leveraging a probabilistic graphical model, we combine two multi-class classifiers into a binary one. This classifier determines if both the main and the peer models agree on a prediction. If they do, the unlabeled sample is assumed to be easy to classify and is thus not beneficial to increase the target model's performance. We prioritize data that represents the unlabeled set without overlapping decision boundaries. The discrepancies between these boundaries can be estimated by the probability that two models result in the same prediction. Through theoretical analysis and experimental validation, we reveal that the integration of noisy labels into the peer model effectively identifies target model's potential inaccuracies. We evaluated the CoLAL method across seven benchmark datasets: four text datasets (AGNews, DBPedia, PubMed, SST-2) and text-based state-of-the-art (SOTA) baselines, and three image datasets (CIFAR100, MNIST, OpenML-155) and computer vision SOTA baselines. The results show that our CoLAL method significantly outperforms existing SOTA in text-based AL, and is competitive with SOTA image-based AL techniques.

Abstract 792: Continuous, probabilistic variant interpretation with Bayesian graphical models

Abstract Introduction: Current variant classification (VC) frameworks rely on rules-based approaches that use heuristic weighting of various types of evidence, resulting in > 50% of variants being classified as variants of uncertain significance (VUS), and leaving many patients with uncertainty about their disease risk or diagnosis. We propose a probabilistic and scalable Bayesian approach to model the causal relationships between various types of evidence. A fully quantitative system maximizes the utility and integration of various evidence types, empowering clinicians to make more nuanced management decisions. Methods: Probabilistic graphical models (PGMs) are uniquely suited to the needs of clinical VC. Two different component PGMs were developed to model two of the evidence categories that will ultimately be used in a comprehensive VC system: population allele frequency (Population PGM) and reported phenotype observations (Reported Phenotype PGM). Results: The Population PGM treats population allele frequency observations as a binomial process. By conditioning the model on partial observations, the probabilistic relationships between pathogenicity and allele frequencies can be estimated while stochastic variational inference allows uncertainty to be efficiently propagated. The resulting model performs well across a large number of genes at inferring pathogenicity of a variant from its allele frequency, with an average precision of >99% for benign variants. In the Reported Phenotype PGM, phenotypic features characteristic of a disorder are learned from patients expected to be affected based on genotype. Patient-level predictions are in turn combined at the variant level to derive variant pathogenicity likelihoods. To date, models representing at least 102 inherited conditions and 259 genes have demonstrated high predictive performance (>0.8 AUROC) for both patient-level and variant-level predictions. Finally, as a proof-of-concept, we demonstrate how each component PGM can be combined into a probabilistic Bayesian VC framework that also includes protein structure and stability, evolutionary conservation, and sequence context. This framework has high concordance with known, well-accepted pathogenic and benign variants classified with rules-based systems, and could make high-confidence predictions for many variants currently classified as VUS. Conclusion: We present a Bayesian approach that can integrate diverse types of evidence to achieve high VC accuracy while quantifying uncertainty. Future expansion of this Bayesian framework to all evidence types relevant to VC may allow for more accurate risk management guidelines and further inform medical and genetic counseling recommendations. Citation Format: Wolfgang Michael Korn, Yuya Kobayashi, Flavia M. Facio, Arun Nampally, Keith Nykamp, Robert Nussbaum, Alexandre Colavin, Britt Johnson, Toby Manders. Continuous, probabilistic variant interpretation with Bayesian graphical models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 792.

Abstract 2493: Prospective evaluation of predictive biomarkers for Palbociclib (P) and endocrine therapy (ET) benefit in hormone receptor (HR)-positive [+]/human epidermal growth factor receptor 2 (HER2)-negative [-] advanced breast cancer (ABC) patients (pts) from PARSIFAL clinical trial: The transFAL sub-study

Abstract BACKGROUND Combining cyclin-dependent kinases 4 and 6 inhibitors (CDK4/6i) with first-line ET is the standard of care for pts with HR[+]/HER2[-] ABC. PARSIFAL was a randomized, open-label, phase II trial that evaluated the efficacy and safety of first line P plus fulvestrant (F) or letrozole (L) in this pts population. Both treatments had comparable efficacy and safety results (JAMAOncol. 2021;7(12):1791-1799). Although multiple CDK4/6i clinical outcomes biomarkers have been reported, the only validated predictive factor for response is HR positivity. TransFAL, a translational study of PARSIFAL, assessed a comprehensive set of biomarkers of sensitivity/resistance to P plus ET. MATERIALS AND METHODS TransFAL employed a sample biorepository from PARSIFAL comprising formalin-fixed paraffin-embedded (FFPE) tissue and blood samples obtained from pts included in either study arm. Tumor samples (N=33) derived from the primary tumor at initial diagnosis or from primary tumor or metastatic lesions at study entry. Blood samples were obtained at baseline (N=73) and disease progression (N=22). Pts were categorized as sensitive or resistant based on progression-free survival (PFS). CDK4/6 pathway activation was assessed by immunohistochemistry for CDK4, CDK6, Rb, pRb, ER, PgR, and Ki67. Proteins extracted from FFPE slides were analyzed using hierarchical clustering, significance analysis of microarrays, and probabilistic graphical models. RNA expression and circulating tumor DNA (ctDNA) analysis were assessed by next-generation sequencing (NGS) and analysis of panel of 56 genes/gene regions, respectively. RESULTS A worse median PFS was found in pts with CDK6 positive (≥ 1% positive cells; p < 0.001) and high Ki67 (≥ 10% positive cells; p = 0.013)at baseline. ctDNA density at baseline was significantly higher for resistant (PFS < 9 months; 17.1 ng/mL) pts compared to sensitive (PFS > 31 months; 11.3 ng/mL) pts (p < 0.001). Most pts (83%) presented a genomic alteration detected in ctDNA, TP53 (28%) and PIK3CA (28%) being the most frequently mutated genes. Pts (9%) who harbored ARID1A mutation had lower PFS (p = 0.052). ESR1 mutations at progression were numerically higher in the L arm (37.5%) compared to F arm (14.3%); (p=0.31), but numbers were low. Proteomic and RNA analyses did not yield confirmable distinctive expression pattern between resistant/sensitive pts… CONCLUSIONS High Ki67 levels and CDK6 expression in the tumor and high ctDNA density, but not proteomic or RNA sequencing data, were associated to poor outcome in this analysis of pts treated with CDK4/6i and ET. This data adds to the growing evidence of poor prognosis biomarkers in this setting, highlighting the need of novel strategies for these pts. Citation Format: Joan Albanell, Angelo Gámez Pozo, Carlos L. Arteaga, Meritxell Bellet, Federico Rojo, Abel González, Beatriz Bellosillo, Violeta Serra, Petra Gener, José Antonio Guerrero, Laura López-Montero, Mario Mancino, Jose Rodríguez-Morató, Leonardo Mina, José Manuel Pérez-García, Javier Cortés, Antonio Llombart-Cussac. Prospective evaluation of predictive biomarkers for Palbociclib (P) and endocrine therapy (ET) benefit in hormone receptor (HR)-positive [+]/human epidermal growth factor receptor 2 (HER2)-negative [-] advanced breast cancer (ABC) patients (pts) from PARSIFAL clinical trial: The transFAL sub-study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2493.

Staged trees and asymmetry-labeled DAGs

AbstractBayesian networks are a widely-used class of probabilistic graphical models capable of representing symmetric conditional independence between variables of interest using the topology of the underlying graph. For categorical variables, they can be seen as a special case of the much more general class of models called staged trees, which can represent any non-symmetric conditional independence. Here we formalize the relationship between these two models and introduce a minimal Bayesian network representation of a staged tree, which can be used to read conditional independences intuitively. A new labeled graph termed asymmetry-labeled directed acyclic graph is defined, with edges labeled to denote the type of dependence between any two random variables. We also present a novel algorithm to learn staged trees which only enforces a specific subset of non-symmetric independences. Various datasets illustrate the methodology, highlighting the need to construct models that more flexibly encode and represent non-symmetric structures.