Assessing the survival time of women with breast cancer in Northwestern Ethiopia: using the Bayesian approach

BACKGROUND

Despite the significant weight of difficulty, Ethiopia's survival rate and mortality predictors have not yet been identified. Finding out what influences outpatient breast cancer patients' survival time was the major goal of this study.

METHODS

A retrospective study was conducted on outpatients with breast cancer. In order to accomplish the goal, 382 outpatients with breast cancer were included in the study using information obtained from the medical records of patients registered at the University of Gondar referral hospital in Gondar, Ethiopia, between May 15, 2016, and May 15, 2020. In order to compare survival functions, Kaplan-Meier plots and the log-rank test were used. The Cox-PH model and Bayesian parametric survival models were then used to examine the survival time of breast cancer outpatients. The use of integrated layered Laplace approximation techniques has been made.

RESULTS

The study included 382 outpatients with breast cancer in total, and 148 (38.7%) patients died. 42 months was the estimated median patient survival time. The Bayesian Weibull accelerated failure time model was determined to be suitable using model selection criteria. Stage, grade 2, 3, and 4, co-morbid, histological type, FIGO stage, chemotherapy, metastatic number 1, 2, and >=3, and tumour size all have a sizable impact on the survival time of outpatients with breast cancer, according to the results of this model. The breast cancer outpatient survival time was correctly predicted by the Bayesian Weibull accelerated failure time model.

CONCLUSIONS

Compared to high- and middle-income countries, the overall survival rate was lower. Notable variables influencing the length of survival following a breast cancer diagnosis were weight loss, invasive medullar histology, comorbid disease, a large tumour size, an increase in metastases, an increase in the International Federation of Gynaecologists and Obstetricians stage, an increase in grade, lymphatic vascular space invasion, positive regional nodes, and late stages of cancer. The authors advise that it is preferable to increase the number of early screening programmes and treatment centres for breast cancer and to work with the public media to raise knowledge of the disease's prevention, screening, and treatment choices.

Quo Vadis Forensic Neuropsychological Malingering Determinations? Reply to Drs. Bush, Faust, and Jewsbury

The thoughtful commentaries in this volume of Drs. Bush, Jewsbury, and Faust add to the impact of the two reviews in this volume of statistical and methodological issues in the forensic neuropsychological determination of malingering based on performance and symptom validity tests (PVTs and SVTs). In his commentary, Dr. Bush raises, among others, the important question of whether such malingering determinations can still be considered as meeting the legal Daubert standard which is the basis for neuropsychological expert testimony. Dr. Jewsbury focuses mostly on statistical issues and agrees with two key points of the statistical review: Positive likelihood chaining is not a mathematically tenable method to combine findings of multiple PVTs and SVTs, and the Simple Bayes method is not applicable to malingering determinations. Dr. Faust adds important narrative texture to the implications for forensic neuropsychological practice and points to a need for research into factors other than malingering that may explain PVT and SVT failures. These commentaries put into even sharper focus the serious questions raised in the reviews about the scientific basis of present practices in the forensic neuropsychological determination of malingering.

Review of Statistical and Methodological Issues in the Forensic Prediction of Malingering from Validity Tests: Part I: Statistical Issues

Forensic neuropsychological examinations with determination of malingering have tremendous social, legal, and economic consequences. Thousands of studies have been published aimed at developing and validating methods to diagnose malingering in forensic settings, based largely on approximately 50 validity tests, including embedded and stand-alone performance validity tests. This is the first part of a two-part review. Part I explores three statistical issues related to the validation of validity tests as predictors of malingering, including (a) the need to report a complete set of classification accuracy statistics, (b) how to detect and handle collinearity among validity tests, and (c) how to assess the classification accuracy of algorithms for aggregating information from multiple validity tests. In the Part II companion paper, three closely related research methodological issues will be examined. Statistical issues are explored through conceptual analysis, statistical simulations, and through reanalysis of findings from prior validation studies. Findings suggest extant neuropsychological validity tests are collinear and contribute redundant information to the prediction of malingering among forensic examinees. Findings further suggest that existing diagnostic algorithms may miss diagnostic accuracy targets under most realistic conditions. The review makes several recommendations to address these concerns, including (a) reporting of full confusion table statistics with 95% confidence intervals in diagnostic trials, (b) the use of logistic regression, and (c) adoption of the consensus model on the "transparent reporting of multivariate prediction models for individual prognosis or diagnosis" (TRIPOD) in the malingering literature.

Application of Artificial Intelligence Techniques to Predict Risk of Recurrence of Breast Cancer: A Systematic Review

Breast cancer is the most common disease among women, with over 2.1 million new diagnoses each year worldwide. About 30% of patients initially presenting with early stage disease have a recurrence of cancer within 10 years. Predicting who will have a recurrence and who will not remains challenging, with consequent implications for associated treatment. Artificial intelligence strategies that can predict the risk of recurrence of breast cancer could help breast cancer clinicians avoid ineffective overtreatment. Despite its significance, most breast cancer recurrence datasets are insufficiently large, not publicly available, or imbalanced, making these studies more difficult. This systematic review investigates the role of artificial intelligence in the prediction of breast cancer recurrence. We summarise common techniques, features, training and testing methodologies, metrics, and discuss current challenges relating to implementation in clinical practice. We systematically reviewed works published between 1 January 2011 and 1 November 2021 using the methodology of Kitchenham and Charter. We leveraged Springer, Google Scholar, PubMed, and IEEE search engines. This review found three areas that require further work. First, there is no agreement on artificial intelligence methodologies, feature predictors, or assessment metrics. Second, issues such as sampling strategies, missing data, and class imbalance problems are rarely addressed or discussed. Third, representative datasets for breast cancer recurrence are scarce, which hinders model validation and deployment. We conclude that predicting breast cancer recurrence remains an open problem despite the use of artificial intelligence.

Artificial intelligence empowered Digital Health Technologies in Cancer Survivorship Care: a scoping review

Objective

The objectives of this systematic review are to describe features and specific application scenarios for current cancer survivorship care services of Artificial intelligence (AI)-driven digital health technologies (DHTs) and to explore the acceptance and briefly evaluate its feasibility in the application process.

Methods

Search for literatures published from 2010 to 2022 on sites MEDLINE, IEEE-Xplor, PubMed, Embase, Cochrane Central Register of Controlled Trials and Scopus systematically. The types of literatures include original research, descriptive study, randomized controlled trial, pilot study, and feasible or acceptable study. The literatures above described current status and effectiveness of digital medical technologies based on AI and used in cancer survivorship care services. Additionally, we use QuADS quality assessment tool to evaluate the quality of literatures included in this review.

Results

43 studies that met the inclusion criteria were analyzed and qualitatively synthesized. The current status and results related to the application of AI-driven DHTs in cancer survivorship care were reviewed. Most of these studies were designed specifically for breast cancer survivors’ care and focused on the areas of recurrence or secondary cancer prediction, clinical decision support, cancer survivability prediction, population or treatment stratified, anti-cancer treatment-induced adverse reaction prediction, and so on. Applying AI-based DHTs to cancer survivors actually has shown some positive outcomes, including increased motivation of patient-reported outcomes (PROs), reduce fatigue and pain levels, improved quality of life, and physical function. However, current research mostly explored the technology development and formation (testing) phases, with limited-scale population, and single-center trial. Therefore, it is not suitable to draw conclusions that the effectiveness of AI-based DHTs in supportive cancer care, as most of applications are still in the early stage of development and feasibility testing.

Conclusions

While digital therapies are promising in the care of cancer patients, more high-quality studies are still needed in the future to demonstrate the effectiveness of digital therapies in cancer care. Studies should explore how to develop uniform standards for measuring patient-related outcomes, ensure the scientific validity of research methods, and emphasize patient and health practitioner involvement in the development and use of technology.

Preoperative Non-Invasive Prediction of Breast Cancer Molecular Subtypes With a Deep Convolutional Neural Network on Ultrasound Images

Purpose

This study aimed to develop a deep convolutional neural network (DCNN) model to classify molecular subtypes of breast cancer from ultrasound (US) images together with clinical information.

Methods

A total of 1,012 breast cancer patients with 2,284 US images (center 1) were collected as the main cohort for training and internal testing. Another cohort of 117 breast cancer cases with 153 US images (center 2) was used as the external testing cohort. Patients were grouped according to thresholds of nodule sizes of 20 mm and age of 50 years. The DCNN models were constructed based on US images and the clinical information to predict the molecular subtypes of breast cancer. A Breast Imaging-Reporting and Data System (BI-RADS) lexicon model was built on the same data based on morphological and clinical description parameters for diagnostic performance comparison. The diagnostic performance was assessed through the accuracy, sensitivity, specificity, Youden’s index (YI), and area under the receiver operating characteristic curve (AUC).

Results

Our DCNN model achieved better diagnostic performance than the BI-RADS lexicon model in differentiating molecular subtypes of breast cancer in both the main cohort and external testing cohort (all p < 0.001). In the main cohort, when classifying luminal A from non-luminal A subtypes, our model obtained an AUC of 0.776 (95% CI, 0.649–0.885) for patients older than 50 years and 0.818 (95% CI, 0.726–0.902) for those with tumor sizes ≤20 mm. For young patients ≤50 years, the AUC value of our model for detecting triple-negative breast cancer was 0.712 (95% CI, 0.538–0.874). In the external testing cohort, when classifying luminal A from non-luminal A subtypes for patients older than 50 years, our DCNN model achieved an AUC of 0.686 (95% CI, 0.567–0.806).

Conclusions

We employed a DCNN model to predict the molecular subtypes of breast cancer based on US images. Our model can be valuable depending on the patient’s age and nodule sizes.

Can artificial intelligence improve cancer treatments?

Artificial intelligence (AI) powered by the accumulating clinical and molecular data about cancer has fueled the expectation that a transformation in cancer treatments towards significant improvement of patient outcomes is at hand. However, such transformation has been so far elusive. The opacity of AI algorithms and the lack of quality annotated data being available at population scale are among the challenges to the application of AI in oncology. Fundamentally however, the heterogeneity of cancer and its evolutionary dynamics make every tumor response to therapy sufficiently different from the population, machine-learned statistical models, challenging hence the capacity of these models to yield reliable inferences about treatment recommendations that can improve patient outcomes. This article reviews the nominal elements of clinical decision-making for precision oncology and frames the utility of AI to cancer treatment improvements in light of cancer unique challenges.

What are the leading causes of fatal and severe injury crashes involving older pedestrian? Evidence from Bayesian network model

INTRODUCTION

Identifying factors contributing to the risk of older pedestrian fatal/severe injuries, along with their possible interdependency, is the first step towards improving safety. Several previous studies focused on identifying the influence of individual factors while ignoring their interdependencies. This study investigated the leading risk factors associated with older pedestrian fatalities/severe injuries by identifying the interdependency relationship among variables.

METHOD

A Bayesian Logistic Regression (BLR) model was developed to identify significant factors influencing pedestrian fatalities and severe injuries, followed by a Bayesian Network (BN) model to reveal the interdependency relationship among the statistically significant variables and crash severity. Furthermore, the probabilistic inference was conducted to identify the leading cause of fatal and severe injuries involving older pedestrians. The models were developed with data from 913 pedestrian crashes involving older pedestrians at signalized intersections in Florida from 2016 through 2018.

RESULTS

Vehicle maneuver, lighting condition, road type, and shoulder type were directly associated with older pedestrian fatality/severe injury. Vehicle maneuver (going straight ahead) was the most significant factor in influencing the severity of crashes involving older pedestrians. The interdependency of vehicle moving straight, nighttime condition, and two-way divided roadway with curbed shoulders was associated with the highest likelihood of fatal and severe-injury crashes involving older pedestrians.

CONCLUSIONS

The Bayesian Network revealed the interdependency between variables associated with fatal and severe injury-crashes involving older pedestrians. The interdependency relationship with the highest likelihood to cause fatalities/severe-injuries comprised factors with the significant individual contribution to the severity of crashes involving older pedestrians. Practical applications: The interdependencies among variables identified in this research could help devise targeted engineering, education, and enforcement strategies that could potentially have a greater effect on improving the safety of older pedestrians.

Development of a Nomogram-Based Tool to Predict Neurocognitive Impairment Among HIV-positive Charter Participants

Background:

Despite the widespread use of combination antiretroviral therapy (cART), HIV-associated neurocognitive impairment (NCI) persists in people living with HIV (PLWH). Studies have generated inconsistent results regarding etiological factors for NCI in PLWH. Furthermore, a user-friendly and readily available predictive tool is desirable in clinical practice to screen PLWH for NCI.

Objective:

This study aimed to identify factors associated with NCI using a large and diverse sample of PLWH and build a nomogram based on demographic, clinical, and behavioral variables.

Methods:

We performed Bayesian network analysis using a supervised learning technique with the Markov Blanket (MB) algorithm. Logistic regression was also conducted to obtain the adjusted regression coefficients to construct the nomogram.

Results:

Among 1,307 participants, 21.6% were neurocognitively impaired. During the MB analysis, age provided the highest amount of mutual information (0.0333). Logistic regression also showed that old age (>50 vs. ≤50 years) had the strongest association (OR=2.77, 95% CI=1.99-3.85) with NCI. The highest possible points on the nomogram were 626, translated to a nomogram-predicted probability of NCI to be approximately 0.95. The receiver operating characteristic (ROC) curve's concordance index was 0.75, and the nomogram's calibration plot exhibited an excellent agreement between observed and predicted probabilities.

Conclusion:

The nomogram used variables that can be easily measured in clinical settings and, thus, easy to implement within a clinic or web-interface platform. The nomogram may help clinicians screen for patients with a high probability of having NCI and thus needing a comprehensive neurocognitive assessment for early diagnosis and appropriate management.

Benchmarking prognosis methods for survivability - A case study for patients with contingent primary cancers

BACKGROUND

There is an increasing number of patients with a first primary cancer who are diagnosed with a second primary cancer, but prognosis methods to predict the survivability of a patient with multiple primary cancers have not been fully benchmarked.

METHODS

This study investigated the five-year survivability prognosis performances of six machine learning approaches. These approaches are: artificial neural network, decision tree (DT), logistic regression, support vector machine, naïve Bayes (NB), and Bayesian network (BN). A synthetic minority over-sampling technique (SMOTE) was used to solve the imbalanced problem, and a nationwide cancer patient database containing 7,845 subjects in Taiwan was used as a sample source. Ten primary and secondary cancers and their key variables affecting the survivability of the patients were identified.

RESULTS

All the models using SMOTE improved sensitivity and specificity significantly. NB has the highest performance in terms of accuracy and specificity, whereas BN has the highest performance in terms of sensitivity. Further, the computational time and the power of knowledge representation of NB, BN, and DT outperformed the others.

CONCLUSIONS

Selecting the appropriate prognosis models to predict survivability of patients with two contingent primary cancers can aid precise prediction and can support appropriate treatment advice.

Validation of the Fracture Risk Evaluation Model (FREM) in predicting major osteoporotic fractures and hip fractures using administrative health data

BACKGROUND

Prevention of osteoporotic fractures remains largely insufficient, and effective means to identify patients at high, short-term fracture risk are needed. The FREM tool is available for automated case finding of men and women aged 45 years or older at high imminent (1-year) risk of osteoporotic fractures, based on administrative health data with a 15-year look-back. The aim of this study was to validate the performance of FREM, and the effect of applying a shorter look-back period. We also evaluated FREM for 5-year fracture risk prediction.

METHODS

Using Danish national health registers we generated consecutive general population cohorts for the years 2014 through 2018. Within each year and across the full time period we estimated the individual fracture risk scores and determined the actual occurrence of major osteoporotic fractures (MOF) and hip fractures. Risk scores were calculated with 15- and 5-year look-back periods. The discriminative ability was evaluated by area under the receiver operating curve (AUC), and negative predictive value (NPV) and positive predictive value (PPV) were estimated applying a calculated risk cut-off of 2% for MOF and 0.3% for hip fractures.

RESULTS

Applying a 15-year look-back, AUC was around 0.75-0.76 for MOF and 0.84-0.87 for hip fractures in 2014, with minor decreases in the subsequent fracture cohorts (2015 to 2018). Applying a 5-year look-back generated similar results, with only marginally lower AUC. In the 5-year risk prediction setting, AUC-values were 0.70-0.72 for MOF and 0.81-0.84 for hip fractures. Generally, PPVs were low, while NPVs were very high.

CONCLUSION

FREM predicts the 1- and 5-year risk of MOF and hip fractures with acceptable vs excellent discriminative power, respectively, when applying both a 15- and a 5-year look-back. Hence, the FREM tool may be applied to improve identification of individuals at high imminent risk of fractures using administrative health data.

Making Sense of Genetic Information: The Promising Evolution of Clinical Stratification and Precision Oncology Using Machine Learning

Precision medicine is a medical approach to administer patients with a tailored dose of treatment by taking into consideration a person's variability in genes, environment, and lifestyles. The accumulation of omics big sequence data led to the development of various genetic databases on which clinical stratification of high-risk populations may be conducted. In addition, because cancers are generally caused by tumor-specific mutations, large-scale systematic identification of single nucleotide polymorphisms (SNPs) in various tumors has propelled significant progress of tailored treatments of tumors (i.e., precision oncology). Machine learning (ML), a subfield of artificial intelligence in which computers learn through experience, has a great potential to be used in precision oncology chiefly to help physicians make diagnostic decisions based on tumor images. A promising venue of ML in precision oncology is the integration of all available data from images to multi-omics big data for the holistic care of patients and high-risk healthy subjects. In this review, we provide a focused overview of precision oncology and ML with attention to breast cancer and glioma as well as the Bayesian networks that have the flexibility and the ability to work with incomplete information. We also introduce some state-of-the-art attempts to use and incorporate ML and genetic information in precision oncology.

The prediction of asymptomatic carotid atherosclerosis with electronic health records: a comparative study of six machine learning models

BACKGROUND

Screening carotid B-mode ultrasonography is a frequently used method to detect subjects with carotid atherosclerosis (CAS). Due to the asymptomatic progression of most CAS patients, early identification is challenging for clinicians, and it may trigger ischemic stroke. Recently, machine learning has shown a strong ability to classify data and a potential for prediction in the medical field. The combined use of machine learning and the electronic health records of patients could provide clinicians with a more convenient and precise method to identify asymptomatic CAS.

METHODS

Retrospective cohort study using routine clinical data of medical check-up subjects from April 19, 2010 to November 15, 2019. Six machine learning models (logistic regression [LR], random forest [RF], decision tree [DT], eXtreme Gradient Boosting [XGB], Gaussian Naïve Bayes [GNB], and K-Nearest Neighbour [KNN]) were used to predict asymptomatic CAS and compared their predictability in terms of the area under the receiver operating characteristic curve (AUCROC), accuracy (ACC), and F1 score (F1).

RESULTS

Of the 18,441 subjects, 6553 were diagnosed with asymptomatic CAS. Compared to DT (AUCROC 0.628, ACC 65.4%, and F1 52.5%), the other five models improved prediction: KNN + 7.6% (0.704, 68.8%, and 50.9%, respectively), GNB + 12.5% (0.753, 67.0%, and 46.8%, respectively), XGB + 16.0% (0.788, 73.4%, and 55.7%, respectively), RF + 16.6% (0.794, 74.5%, and 56.8%, respectively) and LR + 18.1% (0.809, 74.7%, and 59.9%, respectively). The highest achieving model, LR predicted 1045/1966 cases (sensitivity 53.2%) and 3088/3566 non-cases (specificity 86.6%). A tenfold cross-validation scheme further verified the predictive ability of the LR.

CONCLUSIONS

Among machine learning models, LR showed optimal performance in predicting asymptomatic CAS. Our findings set the stage for an early automatic alarming system, allowing a more precise allocation of CAS prevention measures to individuals probably to benefit most.

Risk Prediction Using Bayesian Networks: An Immunotherapy Case Study in Patients With Metastatic Renal Cell Carcinoma

To address the need for more accurate risk stratification models for cancer immuno-oncology, this study aimed to develop a machine-learned Bayesian network model (BNM) for predicting outcomes in patients with metastatic renal cell carcinoma (mRCC) being treated with immunotherapy.

Lead Distribution in Urban Soil in a Medium-Sized City: Household-Scale Analysis

This study characterizes potential soil lead (Pb) exposure risk at the household scale in Greensboro, North Carolina, using an innovative combination of field sampling, statistical analysis, and machine-learning techniques. Soil samples were collected at the dripline, yard, and street side at 462 households (total sample size = 2310). Samples were analyzed for Pb and then combined with publicly available data on potential historic Pb sources, soil properties, and household and neighborhood demographic characteristics. This curated data set was then analyzed with statistical and machine-learning techniques to identify the drivers of potential soil Pb exposure risks and to build predictive models. Among all samples, 43% exceeded current guidelines for Pb in residential gardens. There were significant racial disparities in potential soil Pb exposure risk; soil Pb at the dripline increased by 19% for every 25% increase in the neighborhood population identifying as Black. A machine-learned Bayesian network model was able to classify residential parcels by risk of exceeding residential gardening standards with excellent reproducibility in cross validation. These findings underscore the need for targeted outreach programs to prevent Pb exposure in residential areas and demonstrate an approach for prioritizing outreach locations.

Probabilistic Graphical Modeling for Estimating Risk of Coronary Artery Disease: Applications of a Flexible Machine-Learning Method

Objectives. Coronary artery disease (CAD) is the leading cause of death and disease burden worldwide, causing 1 in 7 deaths in the United States alone. Risk prediction models that can learn the complex causal relationships that give rise to CAD from data, instead of merely predicting the risk of disease, have the potential to improve transparency and efficacy of personalized CAD diagnosis and therapy selection for physicians, patients, and other decision makers. Methods. We use Bayesian networks (BNs) to model the risk of CAD using the Z-Alizadehsani data set-a published real-world observational data set of 303 Iranian patients at risk for CAD. We also describe how BNs can be used for incorporation of background knowledge, individual risk prediction, handling missing observations, and adaptive decision making under uncertainty. Results. BNs performed on par with machine-learning classifiers at predicting CAD and showed better probability calibration. They achieved a mean 10-fold area under the receiver-operating characteristic curve (AUC) of 0.93 ± 0.04, which was comparable with the performance of logistic regression with L1 or L2 regularization (AUC: 0.92 ± 0.06), support vector machine (AUC: 0.92 ± 0.06), and artificial neural network (AUC: 0.91 ± 0.05). We describe the use of BNs to predict with missing data and to adaptively calculate prognostic values of individual variables under uncertainty. Conclusion. BNs are powerful and versatile tools for risk prediction and health outcomes research that can complement traditional statistical techniques and are particularly useful in domains in which information is uncertain or incomplete and in which interpretability is important, such as medicine.

Prediction of drivers and pedestrians' behaviors at signalized mid-block Danish offset crosswalks using Bayesian networks

INTRODUCTION

This study presents the prediction of driver yielding compliance and pedestrian tendencies to press pushbuttons at signalized mid-block Danish offset crosswalks.

METHOD

It applies Bayesian Networks (BNs) analysis, which is basically a graphical non-functional form model, on observational survey data collected from five signalized crosswalks in Las Vegas, Nevada. The BNs structures were learnt from the data by the application of several score functions. By considering prediction accuracy and the Area under the Receiver Operating Characteristic (ROC) curves, the BN learnt using the Bayesian Information Criterion (BIC) score resulted as the best network structure, compared to the ones learnt using K2 and the Akaike Information Criterion (AIC). The BIC score-based structure was then used for parameter learning and probabilistic inference.

RESULTS

Results show that, when considering an individual scenario, the highest predicted yielding compliance (81%) is attained when pedestrians arrive at the crosswalk while the flashes are active, whereas the lowest predicted yielding compliance (23.4%) is observed when the pedestrians cross between the yield line and advanced pedestrian crosswalk sign. On the other hand, crossing within marked stripes, approaching the crosswalk from the near side of the pushbutton pole, inactive flashing lights, and being the first to arrive at the crosswalk result in relatively high-predicted probabilities of pedestrians pressing pushbutton. Furthermore, with a combination of scenarios, the maximum achievable predicted yielding probability is 87.5%, while that of pressing the button was 96.3%. Practical applications: Traffic engineers and planners may use these findings to improve the safety of crosswalk users.