Improving risk prediction for pancreatic cancer in symptomatic patients: a Saudi Arabian study

Machine Learning Models for Pancreatic Cancer Risk Prediction Using Electronic Health Record Data-A Systematic Review and Assessment

INTRODUCTION

Accurate risk prediction can facilitate screening and early detection of pancreatic cancer (PC). We conducted a systematic review to critically evaluate effectiveness of machine learning (ML) and artificial intelligence (AI) techniques applied to electronic health records (EHR) for PC risk prediction.

METHODS

Ovid MEDLINE(R), Ovid EMBASE, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, Scopus, and Web of Science were searched for articles that utilized ML/AI techniques to predict PC, published between January 1, 2012, and February 1, 2024. Study selection and data extraction were conducted by 2 independent reviewers. Critical appraisal and data extraction were performed using the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies checklist. Risk of bias and applicability were examined using prediction model risk of bias assessment tool.

RESULTS

Thirty studies including 169,149 PC cases were identified. Logistic regression was the most frequent modeling method. Twenty studies utilized a curated set of known PC risk predictors or those identified by clinical experts. ML model discrimination performance (C-index) ranged from 0.57 to 1.0. Missing data were underreported, and most studies did not implement explainable-AI techniques or report exclusion time intervals.

DISCUSSION

AI/ML models for PC risk prediction using known risk factors perform reasonably well and may have near-term applications in identifying cohorts for targeted PC screening if validated in real-world data sets. The combined use of structured and unstructured EHR data using emerging AI models while incorporating explainable-AI techniques has the potential to identify novel PC risk factors, and this approach merits further study.

Traditional Chinese Medicine Strategy for Treating Major Depressive Disorder Based on a Famous Formulation-Baweixiaoyaosan

In this study, systematic pharmacological methods were used to reveal the potential pharmacological targets of baweixiaoyaosan in the treatment of major depressive disorder (MDD). We identified 133 potential active compounds through data mining and absorption, distribution, metabolism, and excretion evaluation systems. Then, the target of potential active compounds is predicted by a system model based on random forest and support vector machine methods. Next, construct herbal ingredient-target networks and target-disease networks for further analysis of multi-directional treatment methods. At the same time, we also performed gene ontology enrichment analysis, tissue location analysis, and pathway analysis on 76 potential targets. Finally, we conducted the Jun-Chen-Zuo-Shi compatibility analysis of the formula and scientifically explained the different functions of different herbs in the formula. In short, we found that the formula mainly exerts the effect of treating MDD through the four functional modules of inflammation inhibition, neuroprotection, monoamine neurotransmitter and liver. This research not only explores the mechanism of Traditional Chinese Medicine treatment of MDD from a multi-scale perspective, but also provides a reference for future research on BWXYS. It plays a role in promoting the widespread use of BWXYS.

Early screening and diagnosis strategies of pancreatic cancer: a comprehensive review

Pancreatic cancer is a highly malignant digestive system tumor with a poor prognosis. Most pancreatic cancer patients are diagnosed at an advanced stage or even metastasis due to its highly aggressive characteristics and lack of typical early symptoms. Thus, an early diagnosis of pancreatic cancer is crucial for improving its prognosis. Currently, screening is often applied in high‐risk individuals to achieve the early diagnosis of pancreatic cancer. Fully understanding the risk factors of pancreatic cancer and pathogenesis could help us identify the high‐risk population and achieve early diagnosis and timely treatment of pancreatic cancer. Notably, accumulating studies have been undertaken to improve the detection rate of different imaging methods and the diagnostic accuracy of endoscopic ultrasound‐guided fine‐needle aspiration (EUS‐FNA) which is the golden standard for pancreatic cancer diagnosis. In addition, there are currently no biomarkers with sufficient sensitivity and specificity for the diagnosis of pancreatic cancer to be applied in the clinic. As the only serum biomarker approved by the United States Food and Drug Administration, carbohydrate antigen 19‐9 (CA19‐9) is not recommended to be used in the early screening of pancreatic cancer because of its limited specificity. Recently, increasing numbers of studies focused on the discovering of novel serum biomarkers and exploring their combination with CA19‐9 in the detection of pancreatic cancer. Besides, the application of liquid biopsy involving circulating tumor cells (CTCs), circulating tumor DNA (ctDNA), microRNAs (miRNAs), and exosomes in blood and biomarkers in urine, and saliva in pancreatic cancer diagnosis are drawing more and more attention. Furthermore, many innovative technologies such as artificial intelligence, computer‐aided diagnosis system, metabolomics technology, ion mobility spectrometry (IMS) associated technologies, and novel nanomaterials have been tested for the early diagnosis of pancreatic cancer and have shown promising prospects. Hence, this review aims to summarize the recent progress in the development of early screening and diagnostic methods, including imaging, pathological examination, serological examination, liquid biopsy, as well as other potential diagnostic strategies for pancreatic cancer.

Survival features, prognostic factors, and determinants of diagnosis and treatment among Iranian patients with pancreatic cancer, a prospective study

OBJECTIVES

Investigating the survival features, and determinants of treatment and stage at presentation in Iran.

METHODS

461 patients with pancreatic ductal adenocarcinoma (PC) were prospectively enrolled from Shariati hospital, Tehran, Iran, between 2011-2018. All patients underwent endoscopic ultrasonography, computed tomography scanning, and physical examination. Validated questionnaire was completed for the participants and all were actively followed on monthly basis.

RESULTS

Median survival time was 6.5 months, and 1-, and 5-year survival rates were 26.2%, and 1.5%. Patients who were older (p<0.001), illiterate (p = 0.004), unmarried (p = 0.003), rural inhabitant (p = 0.013), opium user (p = 0.039), and had lower body mass index (BMI) (p = 0.002) had lower overall survival. Tumors located in the head of pancreas were more commonly diagnosed at lower stages (p<0.001). Only 10.4% of patients underwent surgery who were more commonly educated (p<0.001), married (p = 0.005), had a tumor located in the head of pancreas (p = 0.016), and were diagnosed at lower stages (p<0.001). After adjustment for potential confounders and risk factors, rural inhabitance (HR: 1.33 (95% CI: 1.01-1.74)), having more symptoms (HR for each increasing symptom: 1.06 (1.02-1.11)), using opium (HR: 1.51 (1.04-2.20)), having a tumor located in the body of pancreas (HR: 1.33 (1.02-1.75)), and having an advanced tumor stage (HR: 2.07 (1.34-3.19)) remained significantly associated with increased risk of mortality. After the adjusting for potential confounders, we did not find significant relationships between smoking, alcohol intake, and BMI with the risk of death among patients with pancreatic cancer.

CONCLUSIONS

Iranian patients with PC have very poor long-term survival. Besides tumor's stage and location, socioeconomic disparities could affect the probabilities of receiving treatment and/or survival in these patients. Opium use is an independent risk factor for mortality among PC patients in Iran.

Risk Assessment of Repeated Suicide Attempts Among Youth in Saudi Arabia

Introduction

Although the incidence of suicide attempts continues to increase among youth in Saudi Arabia, no risk assessment tool has been established for suicide attempt repetition in the country's youth population. The objective of the study was to develop risk assessment of suicide attempt repetition among youth in Saudi Arabia.

Methods

This is a retrospective study of youth (10-24 years) with intentional suicide attempt(s) who presented to the emergency departments (ED) at King Abdullah Specialist Children's Hospital (KASCH) and King Abdulaziz Medical City-Riyadh (KAMC-R), Saudi Arabia between 1 January 2015 and 31 December 2017. We excluded youth having unintentional suicide attempts. Data were retrieved for the 157 eligible as having attempted suicide.

Results

Forty-one of 157 (26.1%) had repeated suicide attempts (95% confidence limits: 19.433.7%). Four independent factors were identified that were associated with an increased risk of repeated suicide attempts: age (adjusted odds ratio [aOR] = 1.147, 95% confidence interval (CI) = 11.015-1.297, P=0.028), family problems (aOR = 4.218, 95% CI = 1.690-10.528, P=0.002), psychiatric disorders (aOR = 3.497, 95% CI = 1.519-8.051, P=0.003), and hospitalization (aOR = 5.143, 95% CI = 1.421-18.610, P=0.013). This risk model showed adequate utility with an area under the receiver operating characteristic (ROC) curve (AUC): 77.9%, 95% CI: 69.486.3% with optimism-corrected AUC = 71.8%. Youden index defined a probability of ≥0.38 to predict a high risk of repeated suicide attempts.

Conclusion

The risk of repeated suicide attempts among Saudi youth was high, compatible with what has been reported among youth in England and in France. Age, family problems, psychiatric disorders, and hospitalization are risk factors for repeated suicide attempts. A prevention program for suicide attempts in youth may take into account family problems, screening for psychiatric disorders, and suicidal behavior.

Modelling risk assessment for cervical cancer in symptomatic Saudi women

OBJECTIVES

To assess whether the utility of cervical cancer screening could be improved by combining multiple factors in addition to the pap test. Methods: A retrospective cohort study of 300 symptomatic women who were suspected to have cervical cancer and referred for biopsy examination at King Abdulaziz Medical City, Riyadh, Saudi Arabia between February 2017 and December 2017. Results: A high risk of cervical cancer in Saudi women was associated with 4 risk factors: family history (adjusted odds ratio [aOR], 4.216; 95% confidence intervals [CI], 1.433-12.400), vaginal bleeding (aOR, 3.959; 95% CI, 1.272-12.318), hypertension (aOR, 4.554; 95% CI, 1.606-12.912), and an abnormal pap smear test (aOR, 13.985; 95% CI, 5.108-38.284). The model yields an adequate utility (area under the curve, 87.5%, 95% CI, 80.9-94.0%) with acceptable goodness-of-fit (p=0.6915). Conclusion: The pap smear test alone is inadequate to assess high risk for cervical cancer in our center. Early detection of cervical cancer may require consideration of a combination of factors including the pap test. This study has shown that using a combination of abnormal family history, vaginal bleeding, hypertension, and the pap smear test improved the effectiveness of cervical cancer screening.