Minimal Resource Requirements Research Articles

General framework for unsteady aerodynamic prediction of airfoils based on deep transfer learning

Analyzing the unsteady aerodynamic performance of airfoils under dynamic stall using computational fluid dynamics (CFD) is computationally intensive. Although deep learning models can quickly predict aerodynamic parameters, their generalization capability on a small-scale dataset is often poor. This paper presents a novel deep transfer learning (TL) framework that combines model-based and instance-based transfer methods, termed synergistic instance-model TL. This framework facilitates rapid predictions of unsteady aerodynamic performance for various airfoils and pitch oscillations from the small-scale dataset. The framework integrates the accelerated training speed of model-based methods with the dynamic dataset expansion benefits of instance-based approaches. Initially, a pre-trained Wasserstein-deep convolutional generative adversarial network (W-DCGAN) is developed, combining a convolutional neural network with a generative adversarial network to predict aerodynamic hysteresis loops for the SC1095 airfoil in the source domain. The framework then fine-tunes the pre-trained model and incorporates weighted source domain dataset into the small-scale target domain dataset, producing the transferred model W-DCGAN-TL. This approach significantly reduces prediction inaccuracies compared to model-based and non-TL methods when applied to the small-scale dataset. The framework's flexibility allows the use of pre-trained models and datasets from related aerodynamic problems to address issues with insufficient data. Consequently, it is expected to reduce the dependency on extensive datasets, enhance design efficiency, and minimize resource requirements.

Feasibility of implementing RPA coupled with CRISPR-Cas12a (RPA-Cas12a) for Hepatozoon canis detection in dogs

Hepatozoonosis, caused by the protozoan Hepatozoon canis, is a prevalent blood disease affecting owned and stray dogs and cats. The prevalence of these parasites among companion animals in Thailand remains poorly understood. Diagnosing the old-world form of the disease is challenging due to the wide range of nonspecific clinical signs and the reliance on finding low levels of Hepatozoon gamonts in blood smears for conventional diagnosis. PCR demonstrates high specificity and sensitivity but it requires sophisticated instrumentation. Therefore, we established recombinase polymerase amplification (RPA) coupled with Cas12a for H. canis detection based on 18S rRNA. Our findings showed that RPA-Cas12a using gRNA_H was highly specific to H. canis, without yielding positives for other pathogen species including Babesia species. Even in cases of co-infection, RPA-Cas12a only detected positives in samples containing H. canis. This approach detected minimal amounts of H. canis18S rRNA-harboring plasmid at 10 copies per reaction, whereas plasmid-spiked canine blood enabled detection at a minimal amount of 100 copies per reaction. The performance of RPA-Cas12a was validated by comparing it with quantitative PCR-high resolution melting analysis (qPCR-HRM) and sequencing based on 35 canine blood samples. RPA-Cas12a demonstrated precision and accuracy values of 94 % and 90 %, respectively comparable to qPCR-HRM. Overall, these results indicate that RPA-Cas12a serves as a promising tool for H. canis detection as indicated by comparable performance to qPCR-HRM and is suitable for implementation in small animal hospitals or clinics due to its minimal resource requirements, thereby contributing to effective diagnosis and treatment for infected dogs.

Impact of Yoga Intervention on Physical Fitness Parameters in Preadolescent Children: A Randomized Controlled Trial

Purpose:This study aims to assess the impact of yoga on physical fitness among healthy primary school children using a randomized control design. Material and Methods:A total of 52 children, aged 8-13, were randomly assigned to either the Yoga Group (n = 27, mean age = 10.48 ± 1.25 years) or the Control Group (n = 25, mean age = 11.20 ± 1.50 years). The participants underwent evaluation using the EUROFIT Test Battery as the primary outcome measure.The Yoga Group engaged in 45-minute yoga sessions twice a week for a duration of 10 weeks, while the Control Group received no intervention. Outcome assessments were repeated at the conclusion of the 10-week intervention period. Results:Significantly positive enhancements were observed in the Yoga Group across various components of the EUROFIT Test Battery, including balance, speed, flexibility, muscle endurance, and cardiovascular endurance. Conclusion:These findings demonstrate that yoga interventions contribute to improvements in physical fitness among preadolescent children. Consequently, incorporating yoga exercises into the school curriculum is recommended, given their efficacy, minimal resource requirements, and positive impact on sports readiness.

A Robust Automatic Epilepsy Seizure Detection Algorithm Based on Interpretable Features and Machine Learning

Epilepsy, as a serious neurological disorder, can be detected by analyzing the brain signals produced by neurons. Electroencephalogram (EEG) signals are the most important data source for monitoring these brain signals. However, these complex, noisy, nonlinear and nonstationary signals make detecting seizures become a challenging task. Feature-based seizure detection algorithms have become a dominant approach for automatic seizure detection. This study presents an algorithm for automatic seizure detection based on novel features with clinical and statistical significance. Our algorithms achieved the best results on two benchmark datasets, outperforming traditional feature-based methods and state-of-the-art deep learning algorithms. Accuracy exceeded 99.99% on both benchmark public datasets, with the 100% correct detection of all seizures on the second one. Due to the interpretability and robustness of our algorithm, combined with its minimal computational resource requirements and time consumption, it exhibited substantial potential value in the realm of clinical application. The coefficients of variation of datasets proposed by us makes the algorithm data-specific and can give theoretical guidance on the selection of appropriate random spectral features for different datasets. This will broaden the applicability scenario of our feature-based approach.

SHARP (SHort Answer, Rationale Provision): A New Item Format to Assess Clinical Reasoning.

Many non-workplace-based assessments do not provide good evidence of a learner's problem representation or ability to provide a rationale for a clinical decision they have made. Exceptions include assessment formats that require resource-intensive administration and scoring. This article reports on research efforts toward building a scalable non-workplace-based assessment format that was specifically developed to capture evidence of a learner's ability to justify a clinical decision. The authors developed a 2-step item format called SHARP (SHort Answer, Rationale Provision), referring to the 2 tasks that comprise the item. In collaboration with physician-educators, the authors integrated short-answer questions into a patient medical record-based item starting in October 2021 and arrived at an innovative item format in December 2021. In this format, a test-taker interprets patient medical record data to make a clinical decision, types in their response, and pinpoints medical record details that justify their answers. In January 2022, a total of 177 fourth-year medical students, representing 20 U.S. medical schools, completed 35 SHARP items in a proof-of-concept study. Primary outcomes were item timing, difficulty, reliability, and scoring ease. There was substantial variability in item difficulty, with the average item answered correctly by 44% of students (range, 4%-76%). The estimated reliability (Cronbach α ) of the set of SHARP items was 0.76 (95% confidence interval, 0.70-0.80). Item scoring is fully automated, minimizing resource requirements. A larger study is planned to gather additional validity evidence about the item format. This study will allow comparisons between performance on SHARP items and other examinations, examination of group differences in performance, and possible use cases for formative assessment. Cognitive interviews are also planned to better understand the thought processes of medical students as they work through the SHARP items.

Resource-efficient quantum principal component analysis

Principal component analysis (PCA) is an important dimensionality reduction method in machine learning and data analysis. Recently, the quantum version of PCA has been established to diagonalize quantum states. Although these quantum algorithms promise quantum advantages, they require substantial resources beyond the reach of state-of-the-art quantum technologies. This work aims to reduce resource requirements and improve the efficiency of quantum PCA. Assuming that the quantum state is accessed through a purified quantum query model and a sampling model, we propose quantum algorithms that use minimal resource requirements for ancillary qubits to reveal properties of eigenvectors and eigenvalues of a state. In particular, we show that estimating eigenvalue λ with error ε and success probability larger than λ(1−η) requests a query complexity O~(ϵ−1) and a sample complexity O~(ϵ−2η−1) , respectively. To our knowledge, our result is the first quantum speedup that achieves asymptotic linear scaling in 1/ϵ for quantum PCA. As applications, we discuss estimating the minimum relative entropy of entanglement of bipartite pure-states and performing quantum state discrimination tasks. We show that quantum speedups are maintained when the pure state has a low Schmidt number and states of discrimination have a low rank. This study opens up a new quantum PCA method for high-dimensional quantum data analysis and discusses its application in quantum information processing tasks.

Quantifying pesticide residues in food matrices using statistical methods

Traditional techniques for pesticide residues detection and quantification in food using mass spectrometry often require the analysis of standards for each compound, leading to time-consuming and laborious procedures, especially considering that these methods usually involve hundreds of pesticides to be targeted. This paper presents a novel approach to compound quantitation, where multiple target compounds can be accurately quantified using few predictor compounds, significantly reducing the experimental time and minimizing resource requirements. For this purpose, data on detector response which encompassed calibration slopes of a total of 96 pesticide standards on GC-MS/MS and 66 standards on LC-MS/MS, over a period of 4 years, were collected. Two fundamental statistical techniques, Pearson correlation and linear regression, were used to create a predictive model, which accuracy was further evaluated using R-square, adjusted R-square, and Root Mean Square Error (RMSE). Four predicted compounds for the LC dataset and seven predicted compounds for GC dataset were identified, and various predictor combinations were considered. A linear regression model was developed using predictor combinations to estimate the calibration slope of each of the target compound. This model was applied for the quantitation of several pesticide residues in food samples. The results validated the model’s high accuracy.

Designing quantum annealing schedules using Bayesian optimization

We propose and analyze the use of Bayesian optimization techniques to design quantum annealing schedules with minimal user and resource requirements. We showcase our scheme with results for two paradigmatic spin models. We find that Bayesian optimization is able to identify schedules resulting in fidelities several orders of magnitude better than standard protocols for both quantum and reverse annealing, as applied to the p-spin model. We also show that our scheme can help improve the design of hybrid quantum algorithms for hard combinatorial optimization problems, such as the maximum independent set problem, and illustrate these results via experiments on a neutral-atom quantum processor available on Amazon Braket. Published by the American Physical Society 2024

A Multimodal Exertional Test for concussion: a pilot study in healthy athletes.

Exertional tests have become a promising tool to assist clinicians in the management of concussions, however require expensive equipment, extensive spaces, and specialized clinician expertise. As such, we developed a test with minimal resource requirements encompassing key elements of sport and physical activity. The purpose of this study was to pilot test the Multimodal Exertional Test (MET) protocol in a sample of healthy interuniversity athletes. The MET comprises four stages, each featuring three distinct tasks. The test begins with engaging in squats, alternating reverse lunges, and hip hinges (Stage 1). The next stage progressively evolves into executing these tasks within specified time limits (Stage 2). Following this, the test advances to a stage that incorporates cognitive tasks (Stage 3), and the final stage demands greater levels of physical exertion, cognition, and multi-directional movements (Stage 4). Heart rate (HR) was obtained during each stage of the MET and participants' symptom severity scores were recorded following each task. Fourteen healthy interuniversity athletes (n = 8 female, n = 6 male) participated in the study. HR was obtained for 10 of the 14 athletes (females: n = 6, males: n = 4). Increases in average and maximum HR were identified between pre-MET and Stage 1, and between Stages 3 and 4. Consistent with the tasks in each stage, there were no increases in average and maximum HR observed between MET Stages 1 to 3. Female athletes exhibited higher average and maximum HRs compared to male athletes during all four stages. All 14 athletes reported minimal changes in symptom severity following each task. Among healthy athletes, the MET elicits an increase in average and maximum HR throughout the protocol without symptom provocation. Female athletes exhibit higher HRs during all four stages in comparison to male athletes.

Evaluating the use of smartphones for processing computational tasks: "A case study for the annotation of prokaryotic genomes"

Currently, there is a growing dependence on smartphones in the daily lives of thousands of people, noting that the number of users continues to rise, surpassing the number of computers in use globally. In addition, the performance of smartphone hardware has become an essential factor when choosing these devices, along with price, brand and operating system. Constant innovations in smartphones have made these devices increasingly powerful, even rivaling the performance of many computers. The example of the Samsung S23 5G, which has features comparable to a low-performance notebook, illustrates the advancement of the hardware in these devices. Despite the growing demand for advanced hardware in smartphones, many users may not fully understand the technical specifications of the processors. The article proposes an evaluation of the use of this hardware in the execution of specific tasks, with a focus on the annotation of prokaryotic genomes. The methodology of the study involved selecting popular smartphones from Samsung and Motorola, choosing the Android operating system due to its efficiency in resource management, using the Prokka tool for genomic annotation due to its minimal resource requirements and using Termux, a terminal emulator that allows users to run a full Linux distribution on Android devices. To evaluate performance, the article uses the Linux 'time' command to measure execution time, memory consumption, CPU utilization and other relevant metrics

Innovative approaches for Microcystin removal: Bacterioplankton biodegradation and multi-soil-layering system performance assessment

Microcystins (MCs) pose a significant threat to human health and the environment due to their persistence and toxicity as potent cyanobacterial toxins. Effective biodegradation strategies are required for their removal and detoxification. In this study, bacterioplankton were isolated from eutrophic reservoir water used for crop irrigation and tested for their ability to degrade MCs. These bacteria are hydrologically transported from the contaminated reservoir to downstream farms where they are able to grow under the favorable conditions provided by a multi-soil-layering (MSL) system, a low-cost and locally applicable ecotechnology for the removal and detoxification of MCs. These bacteria enhance the ability of the MSL to remove MCs by forming biofilms within the MSL system. The bacterial strains were then sequenced and their partial 16S rRNA was compared to the reference strains in the NCBI database. The effectiveness of the MSL system in the removal of MCs from water was evaluated. The MSL utilizes a layered soil matrix that promotes microbial growth and physicochemical interactions that enable the removal of various contaminants. The mechanisms of MC removal in the MSL were evaluated by studying the physicochemical properties of the system substrate before and after treatment using X-ray fluorescence, zeta potential, and scanning electron microscopy (SEM) analyses. The microbial functional groups, including ammonifying and denitrifying bacteria, were also investigated in the MSL soil matrices before and after treatment. The results showed that seven bacterial strains exhibited high MC degradation capabilities, with MC degradation performance exceeding 97%. These isolated bacterial strains had a nucleotide identity score of 100% and an E value of 0.0 compared to reference strains. The MSL showed a 99% removal rate for MC-LR and MC-YR, as well as significant reductions in other pollutants, including COD (60%), BOD5 (60%), chlorophyll-a (96%), pheophytin (70%), organic nitrogen (80%), and nitrites (90%). After treatment, there was a statistically significant increase in the total bacterial count, as well as ammonifying and denitrifying bacteria, in the MSL soil matrices compared to their state before treatment. SEM analyses revealed the formation of dense biofilm microbiota, indicating additional biological processes besides adsorption for MC reduction. Overall, the MSL provides a low-cost alternative for the removal of MCs and cyanobacteria with minimal resource requirements and maintenance. It offers a promising solution for mitigating toxic cyanobacterial blooms, ensuring safe agricultural water reuse.

LLaMA-LoRA Neural Prompt Engineering: A Deep Tuning Framework for Automatically Generating Chinese Text Logical Reasoning Thinking Chains

Abstract The exption of Chinese natural language processing (NLP) has stimulated research in the broader NLP domain. However, existing large language models have limitations in comprehending and reasoning in Chinese. This paper addresses these limitations by enhancing Chinese language models comprehension and reasoning capabilities while minimizing resource requirements. We propose LLaMA-LoRA, a neural prompt engineering framework that builds upon the LLaMA-13B model and incorporates the Low-Rank Adaptation(LoRA) of Large Language Models technique for refinement. Chain-of-Thought(CoT) are crucial for generating intermediate reasoning chains in language models, but their effectiveness can be limited by isolated language patterns. Erroneous reasoning resulting from conventional prompts negatively impacts model performance. Automatic prompts are introduced to encourage reasoning chain generation and accurate answer inference. Training the model with an extensive corpus of Chinese CoT data enhances its comprehension and reasoning abilities. The LLaMA-LoRA model demonstrates exceptional performance across numerous Chinese language tasks, surpassing benchmark performance achieved by related language models such as GPT-3.5, Chat-GLM, and OpenAssistant, delivering accurate, comprehensive, and professional answers. The availability of our open-source model code facilitates further research in the field of Chinese text logical reasoning thinking chains.

Sustainable Farming and Soil Health Enhancement through Millet Cultivation: A Review

Historically integral to Indian agriculture, millets are experiencing resurgence, driven by their adaptability to harsh climatic conditions and minimal resource requirements. The article explores the significant benefits of millets in soil nutrient management, demonstrating their ability to thrive in nutrient-poor soils while contributing to soil fertility through organic matter addition and improved soil structure. The integration of millets in crop rotation and intercropping systems is highlighted as a sustainable practice that enhances soil biodiversity and reduces the reliance on chemical inputs. A key environmental benefit of millet cultivation is its low water requirement and drought resistance, crucial in water-scarce regions, making it a strategic crop for adapting to climate change. Also examines the socio-economic impacts of millet cultivation. In rural areas, millets play a vital role in livelihoods by providing a sustainable food source and income generation, particularly in marginal environments. However, challenges in market accessibility and supply chain inefficiencies pose significant hurdles. Increasing consumer awareness and acceptance of millets, once considered traditional or 'poor man’s food', is critical for reviving their cultivation. This is complemented by emerging research and technologies in millet cultivation, focused on developing improved varieties and precision agriculture techniques tailored to the needs of millet farming. The potential for scaling up millet cultivation is immense, particularly in regions facing environmental constraints. The article emphasizes the integration of millets into global sustainable farming strategies, aligning with several Sustainable Development Goals. This integration is supported by international organizations advocating for millets in agricultural policies and programs. In conclusion, the review underscores the importance of millets in the context of global food security and sustainable agriculture. The growing importance of millets backed by research, policy support, and market trends, positions them as a key crop in the quest for a sustainable and resilient agricultural future.

Seed classification with random forest models.

To improve forest conservation monitoring, we developed a protocol to automatically count and identify the seeds of plant species with minimal resource requirements, making the process more efficient and less dependent on human operators. Seeds from six North American conifer tree species were separated from leaf litter and imaged on a flatbed scanner. In the most successful species-classification approach, an ImageJ macro automatically extracted measurements for random forest classification in the software R. The method allows for good classification accuracy, and the same process can be used to train the model on other species. This protocol is an adaptable tool for efficient and consistent identification of seed species or potentially other objects. Automated seed classification is efficient and inexpensive, making it a practical solution that enhances the feasibility of large-scale monitoring projects in conservation biology.

Enhanced Trajectory Tracking of 3D Overhead Crane Using Adaptive Sliding-Mode Control and Particle Swarm Optimization

Cranes hold a prominent position as one of the most extensively employed systems across global industries. Given their critical role in various sectors, a comprehensive examination was necessary to enhance their operational efficiency, performance, and facilitate the control of transporting loads. Furthermore, due to the complexities involved in disassembling and reinstalling cranes, as well as the challenges associated with precisely determining system parameters, it became essential to implement adaptive control methods capable of efficiently managing the system with minimal resource requirements. This work proposes a trajectory tracking control using adaptive sliding-mode control (SMC) with particle swarm optimization (PSO) to control the position and rope length of a 3D overhead crane system with unknown parameters. The PSO is mainly used to identify the model and estimate the uncertain parameters. Then, sliding-mode control is adapted using the PSO algorithm to minimize the tracking error and ensure robustness against model uncertainties. A model of the systems is derived assuming changing rope length. The model is nonlinear of second order with five states, three actuated states: position x and y, and rope length l, and two unactuated states, which are the rope angles θx and θy. The system has uncertain parameters, which are the system’s masses Mx, My and Mz, and viscous damping coefficients Dx, Dy and Dy. A simulation study is established to illustrate the influence and robustness of the developed controller and it can enhance the tracking trajectory under different scenarios to test the scheme.

Outpatient-department-based intensified tuberculosis case-finding activity in the tertiary care hospitals in West Bengal, India.: An observational study on the yields and outputs

BackgroundImproved tuberculosis (TB) case detection using intensified case finding (ICF) is one of the cornerstones of the Stop TB Strategy. A Non-Government Organization (NGO)-led project took place in two tertiary care hospitals of West Bengal for outpatient department (OPD)-based ICF activities. Methodology: A secondary analysis was conducted with the routinely collected project data to assess the yields of the model. Risk ratios have been calculated to identify associations of attrition before TB testing and diagnosis. ResultsA total of 2324 presumptive TB patients (PTBP) were identified from a screening of 105,391 OPD attendees. The screening yield of PTBP was 2.2 % (2.1%–2.3 %). Total 143 TB patients were detected with testing of 770 (out of 2324 identified) presumptive cases. The pretesting attrition rate was 66.9 % (64.9%–68.8 %). The screening yield of TB was 0.14 % (0.11%–0.16 %).Being younger than fifteen years of age, of the Islam religion, a homemaker or unemployed was independently associated with “attrition before TB tests”.Being Hindu, a homemaker or a student was independently associated with “diagnosis with TB” among the tested. ConclusionsThe OPD-based screening yield of TB is low. However, minimal resource requirements in this model advocate for its utility. Similar model can be replicated in other hospitals at low resource settings with appropriate setting-specific modifications.

A Runtime Switchable Multi-Phase Convolutional Neural Network for Resource-Constrained Systems

Convolutional Neural Networks (CNNs) are widely used in various systems, including resource-constrained embedded systems or IoT devices. In such systems, it is typical to deploy compressed or pruned CNNs, instead of original ones, at the cost of reduced accuracy. Existing CNN pruning techniques have primarily focused on minimizing resource requirements. However, today’s embedded systems are increasingly dynamic in both resource demands and availability. Thus, the previous techniques that only consider given static cases are no longer efficient. In this paper, we propose a novel multi-phase CNN that enables a multi-objective exploration of a number of pruning candidates out of a single CNN. In the proposed technique, a CNN can operate in various versions depending on which subsets of weights are used and can be transformed to the one best matches to the given constraint adaptively and efficiently. For that, a CNN is first pruned to the sparsest form; then a set of parameters (sub-network) is additionally supplemented as the phase goes by. As a result, a number of network versions for all different phases can be represented by a single network and they form a pareto solution over the accuracy and resource usage trade-off. In this work, we target CPU-based CNN inference engines as most embedded systems do not have the luxury of specialized co-processor support such as GPUs or HW accelerators. The proposed technique has been implemented in a publicly available CPU inference engine, Darknet, and its effectiveness has been validated with a popular CNN in terms of design space exploration capability and runtime switchability.

Optimizing the number of measurements for vibrational structure on quantum computers: coordinates and measurement schemes.

One of the primary challenges prohibiting demonstrations of practical quantum advantages for near-term devices amounts to excessive measurement overheads for estimating relevant physical quantities such as ground state energies. However, with major differences between the electronic and vibrational structures of molecules, the question of how the resource requirements of computing anharmonic, vibrational states can be reduced remains relatively unexplored compared to its electronic counterpart. Importantly, bosonic commutation relations, distinguishable Hilbert spaces and vibrational coordinates allow manipulations of the vibrational system that can be exploited to minimize resource requirements. In this work, we investigate the impact of different coordinate systems and measurement schemes on the number of measurements needed to estimate anharmonic, vibrational states for a variety of three-mode (six-mode) molecules. We demonstrate an average of 3-fold (1.5-fold), with up to 7-fold (2.5-fold), reduction in the number of measurements required by employing appropriate coordinate transformations, based on an automized construction of qubit Hamiltonians from a conventional vibrational structure program.

Prediction of regrowth and biomass of perennial sorghum using unoccupied aerial systems

AbstractPerennial grain sorghum [Sorghum bicolor (L.) Moench] has potential to produce grain and forage while improving soil health, ecosystem services, and carbon soil sequestration but requires further genetic improvement. Unoccupied aerial systems (UAS, also known as drones and unmanned aerial systems) provide opportunities to quickly evaluate plant traits on a large scale with precision. Unoccupied aerial system flights were used to evaluate biomass yield and rhizome characteristics of 100 diverse sorghum hybrids, most being from an interspecific hybridization program, in the establishment year and first year of regrowth. Twenty‐one vegetation indices (VIs) with canopy height measurements (CHMs) were processed from seven UAS flights made temporally during each growing season. Regression of the temporal data (VI and CHM) and phenotypic traits, including rhizome characteristics based on plant stand count (PSC), rhizome‐derived shoots (RDS), and fresh and dry biomass yields, showed useful predictions when combining temporal VI with CHM and machine learning. Blue chromatic coordinate index (BCC) best predicted all measured traits. If predictions could be generalized, UAS would reduce field evaluation time for perennial sorghum or breeding perennial grasses in general and allow breeders to evaluate additional genotypes. In this study, we found that optimizing flights to specific dates after planting could minimize resource requirements and costs in prediction of regrowth and biomass yield of perennial sorghum.

Full automation of total metabolic tumor volume from FDG-PET/CT in DLBCL for baseline risk assessments

BackgroundCurrent radiological assessments of 18fluorodeoxyglucose-positron emission tomography (FDG-PET) imaging data in diffuse large B-cell lymphoma (DLBCL) can be time consuming, do not yield real-time information regarding disease burden and organ involvement, and hinder the use of FDG-PET to potentially limit the reliance on invasive procedures (e.g. bone marrow biopsy) for risk assessment.MethodsOur aim is to enable real-time assessment of imaging-based risk factors at a large scale and we propose a fully automatic artificial intelligence (AI)-based tool to rapidly extract FDG-PET imaging metrics in DLBCL. On availability of a scan, in combination with clinical data, our approach generates clinically informative risk scores with minimal resource requirements. Overall, 1268 patients with previously untreated DLBCL from the phase III GOYA trial (NCT01287741) were included in the analysis (training: n = 846; hold-out: n = 422).ResultsOur AI-based model comprising imaging and clinical variables yielded a tangible prognostic improvement compared to clinical models without imaging metrics. We observed a risk increase for progression-free survival (PFS) with hazard ratios [HR] of 1.87 (95% CI: 1.31–2.67) vs 1.38 (95% CI: 0.98–1.96) (C-index: 0.59 vs 0.55), and a risk increase for overall survival (OS) (HR: 2.16 (95% CI: 1.37–3.40) vs 1.40 (95% CI: 0.90–2.17); C-index: 0.59 vs 0.55). The combined model defined a high-risk population with 35% and 42% increased odds of a 4-year PFS and OS event, respectively, versus the International Prognostic Index components alone. The method also identified a subpopulation with a 2-year Central Nervous System (CNS)-relapse probability of 17.1%.ConclusionOur tool enables an enhanced risk stratification compared with IPI, and the results indicate that imaging can be used to improve the prediction of central nervous system relapse in DLBCL. These findings support integration of clinically informative AI-generated imaging metrics into clinical workflows to improve identification of high-risk DLBCL patients.Trial RegistrationRegistered clinicaltrials.gov number: NCT01287741.Graphical