Lifestyle Risk Prediction Model for Prostate Cancer in a Korean Population

Artificial intelligence to predict cancer risk, are we there yet? A comprehensive review across cancer types

Cancer remains the second leading cause of death worldwide, representing a substantial challenge to global health. Although traditional risk prediction models have played a crucial role in epidemiology of several cancer types, they have limitations especially in the ability to process complex and multidimensional data. In contrast, artificial intelligence (AI) approaches represent a promising solution to overcome this limitation. AI techniques have the potential to identify complex patterns and relationships in data that traditional methods might overlook, making them especially useful for handling large and heterogeneous datasets analysed in cancer research. This review first examines the current state of the art of AI techniques, highlighting their differences and suitability for various data types. Then, offers a comprehensive analysis of the literature, focusing on the application of AI approaches in nineteen cancer types (bladder cancer, breast cancer, cervical cancer, colorectal cancer, endometrial cancer, esophageal cancer, gastric cancer, gynaecological cancers, head and neck cancer, haematological cancers, kidney cancer, liver cancer, lung cancer, melanoma, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer and overall cancer), evaluating the models, metrics, and exposure variables used. Finally, the review discusses the application of AI in the clinical practice, along with an assessment of its potential limitations and future directions.

Machine learning algorithms that predict the risk of prostate cancer based on metabolic syndrome and sociodemographic characteristics: a prospective cohort study

BACKGROUND

Given the rapid increase in the prevalence of prostate cancer (PCa), identifying its risk factors and developing suitable risk prediction models has important implications for public health. We used machine learning (ML) approach to screen participants with high risk of PCa and, specifically, investigated whether participants with metabolic syndrome (MetS) exhibited an elevated PCa risk.

METHODS

A prospective cohort study was performed with 41,837 participants in South Korea. We predicted PCa based on MetS, its components, and sociodemographic factors using Cox proportional hazards and five ML models. Integrated Brier score (IBS) and C-index were used to assess model performance.

RESULTS

A total of 210 incident PCa cases were identified. We found good calibration and discrimination for all models (C-index ≥ 0.800 and IBS = 0.01). Importantly, performance increased after excluding MetS and its components from the models; the highest C-index was 0.862 for survival support vector machine. In contrast, first-degree family history of PCa, alcohol consumption, age, and income were valuable for PCa prediction.

CONCLUSION

ML models are an effective approach to develop prediction models for survival analysis. Furthermore, MetS and its components do not seem to influence PCa susceptibility, in contrast to first-degree family history of PCa, age, alcohol consumption, and income.

To Drink or Not to Drink? Investigating Alcohol's Impact on Prostate Cancer Risk

BACKGROUND/OBJECTIVES

Prostate cancer (PCa) is a significant global health issue. The relationship between alcohol consumption and PCa risk has been the subject of extensive research, yet findings remain inconsistent. This review aims to clarify the association between alcohol intake and PCa risk, its aggressiveness, and the potential metabolic pathways involved in PCa onset.

METHODS

A comprehensive literature search was conducted across multiple databases, including PubMed and MEDLINE, focusing on epidemiological studies, meta-analyses, cohort studies, and case-control studies. Studies evaluating alcohol consumption, prostate-specific antigen (PSA) levels, and PCa risk were included. The review also explored the roles of alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH) in alcohol metabolism.

RESULTS

The analysis reveals a complex relationship between alcohol consumption and PCa. Heavy alcohol intake is associated with an increased risk of PCa, particularly more aggressive forms, and higher mortality rates. However, studies also show weak or no association between moderate alcohol consumption and PCa. The variability in findings may be attributed to differences in alcohol types, regional factors, and study methodologies.

CONCLUSIONS

The link between alcohol consumption and PCa risk is multifaceted. While heavy drinking appears to increase the risk of aggressive PCa, the overall relationship remains unclear. Further research is needed to better understand these associations and inform public health recommendations and cancer prevention strategies.

Combinations of lifestyle behaviors and cancer risk among Korean adults

Combinations of lifestyle behaviors may lead to different cancer risks. This study aimed to identify the latent classes based on lifestyle behavior trajectories and to investigate the association between these latent classes and cancer risk. Participants in the 2002-2003 National Health Insurance Service general health examination were included. Data on smoking, alcohol drinking, body mass index (BMI), and physical activity measured four times between 2002 and 2009 were analyzed. Incident cancer cases were tracked from 2010 to 2018. Patterns of alcohol drinking, smoking, BMI, and physical activity and latent classes based on trajectories of smoking, alcohol drinking, BMI, and physical activity were identified. Among 2,735,110 adults (1,787,486 men and 947,624 women), 111,218 (69,089 men and 42,129 women) developed incident cancer. Six latent classes of lifestyle behavior were identified, with Class 1 (healthy class) involving only 0.2% of men and 0.5% of women. The highest risk class in males tended to be steady light drinkers and steady moderate smokers, have steady low frequency of physical activity, and be obese. This class showed a 1.47 times higher (95% CI = 1.29-1.69) risk of all cancers than did the healthy class. Among women, there was only an association between the highest risk class (tendency to be non-drinkers, light smokers) and colorectal cancer (HR = 1.70, 95% CI = 1.02-2.83). Only a small percentage of participants maintained a long-term healthy lifestyle. Identifying classes of behavior combinations and their links to cancer development is therefore critical for cancer prevention.

Prostate cancer risk prediction based on clinical factors and prostate-specific antigen

INTRODUCTION

The incidence rate of prostate cancer (PCa) has continued to rise in Korea. This study aimed to construct and evaluate a 5-year PCa risk prediction model using a cohort with PSA < 10 ng/mL by incorporating PSA levels and individual factors.

METHODS

The PCa risk prediction model including PSA levels and individual risk factors was constructed using a cohort of 69,319 participants from the Kangbuk Samsung Health Study. 201 registered PCa incidences were observed. A Cox proportional hazards regression model was used to generate the 5-year risk of PCa. The performance of the model was assessed using standards of discrimination and calibration.

RESULTS

The risk prediction model included age, smoking status, alcohol consumption, family history of PCa, past medical history of dyslipidemia, cholesterol levels, and PSA level. Especially, an elevated PSA level was a significant risk factor of PCa (hazard ratio [HR]: 1.77, 95% confidence interval [CI]: [1.67-1.88]). This model performed well with sufficient discrimination ability and satisfactory calibration (C-statistic: 0.911, 0.874; Nam-D'Agostino test statistic:19.76, 4.21 in the development and validation cohort, respectively).

CONCLUSIONS

Our risk prediction model was effective in predicting PCa in a population according to PSA levels. When PSA levels are inconclusive, an assessment of both PSA and specific individual risk factors (e.g., age, total cholesterol, and family history of PCa) could provide further information in predicting PCa.

Cigarette smoking and prostate cancer: A systematic review and meta-analysis of prospective cohort studies

INTRODUCTION

Cigarette smoking is a well-known cancer-causing behavior and a leading cause of death from cancer. However, according to previously published research and meta-analyses, cigarette smoking has a significant inverse association with prostate cancer incidence. Therefore, this study aims to investigate this association based on updated evidence by conducting a systematic review and meta-analysis.

METHODS

A search for relevant articles was performed in PubMed and Scopus databases to obtain the pooled relative risk (RR) and the corresponding 95% confidence intervals (CIs) for the risk of prostate cancer incidence among smokers compared to non-smokers. Our search was limited to prospective cohort studies.

RESULTS

A total of 17 cohort studies were included in the systematic review. Fifteen studies were included in the meta-analysis and showed that cigarette smoking has an inverse association with prostate cancer incidence with a relative risk of 0.84 (95% CI: 0.78-0.91). From all cohorts included in this systematic review, five studies examined the association between current smokers and the risk of death from prostate cancer. Therefore, a meta-analysis of these cohort studies was performed and showed that current smokers had a 42% higher risk of death from prostate cancer when compared to non-smokers with a relative risk of 1.42 (95% CI: 1.20-1.68).

CONCLUSIONS

Data from observational studies suggest that cigarette smoking has an inverse association with prostate cancer incidence. However, smokers have an increased risk of death from prostate cancer. Important to realize that this lower risk for smokers might be attributed to low prostate cancer screening uptake among smokers, misclassification bias, or selection bias from the included original studies. In summary, our results indicate that the incidence of prostate cancer is lower among smokers. Nevertheless, smokers who develop the disease have a significantly worse prognosis.

Diagnostic accuracy of prostate-specific antigen below 4 ng/mL as a cutoff for diagnosing prostate cancer in a hospital setting: A systematic review and meta-analysis

PURPOSE

A prostate-specific antigen (PSA) cutoff of 4 ng/mL has been widely used for prostate cancer screening in population-based settings. However, the accuracy of PSA below 4 ng/mL as a cutoff for diagnosing prostate cancer in a hospital setting is inconclusive. We systematically reviewed the accuracy of PSA below 4 ng/mL cutoff in a hospital setting.

MATERIALS AND METHODS

We systematically reviewed the literature by searching major databases until March 2020, and a meta-analysis and quality assessment were performed.

RESULTS

A total of 11 studies were included at the completion of the screening process. The meta-analysis showed a sensitivity of 0.92 and a specificity of 0.16 for a PSA cutoff below 4 ng/mL. The area under the hierarchical summary receiver operating characteristic curve was 0.87, the positive likelihood ratio was 1.23, the negative likelihood ratio was 0.46, and the diagnostic odds ratio was 2.64. PSA sensitivities and specificities varied according to the cutoff range: 0.94 and 0.17 for 2 to 2.99 ng/mL, and 0.92 and 0.16 for 3 to 3.99 ng/mL, respectively. No significant differences in the sensitivity and specificity of PSA cutoffs in the range of 2 to 2.99 ng/mL and 3 to 3.99 ng/mL were found.

CONCLUSIONS

Although a PSA cutoff <3 ng/mL is relatively more sensitive and specific than PSA ≥3 ng/mL, no significant differences in sensitivity and specificity were found in the diagnosis of prostate cancer. Therefore, clinicians should choose an appropriate PSA cutoff on the basis of clinical circumstances and patients' characteristics.

Alcohol and Prostate Cancer: Time to Draw Conclusions

It has been a long-standing debate in the research and medical societies whether alcohol consumption is linked to the risk of prostate cancer (PCa). Many comprehensive studies from different geographical areas and nationalities have shown that moderate and heavy drinking is positively correlated with the development of PCa. Nevertheless, some observations could not confirm that such a correlation exists; some even suggest that wine consumption could prevent or slow prostate tumor growth. Here, we have rigorously analyzed the evidence both for and against the role of alcohol in PCa development. We found that many of the epidemiological studies did not consider other, potentially critical, factors, including diet (especially, low intake of fish, vegetables and linoleic acid, and excessive use of red meat), smoking, family history of PCa, low physical activity, history of high sexual activities especially with early age of first intercourse, and sexually transmitted infections. In addition, discrepancies between observations come from selectivity criteria for control groups, questionnaires about the type and dosage of alcohol, and misreported alcohol consumption. The lifetime history of alcohol consumption is critical given that a prostate tumor is typically slow-growing; however, many epidemiological observations that show no association monitored only current or relatively recent drinking status. Nevertheless, the overall conclusion is that high alcohol intake, especially binge drinking, is associated with increased risk for PCa, and this effect is not limited to any type of beverage. Alcohol consumption is also directly linked to PCa lethality as it may accelerate the growth of prostate tumors and significantly shorten the time for the progression to metastatic PCa. Thus, we recommend immediately quitting alcohol for patients diagnosed with PCa. We discuss the features of alcohol metabolism in the prostate tissue and the damaging effect of ethanol metabolites on intracellular organization and trafficking. In addition, we review the impact of alcohol consumption on prostate-specific antigen level and the risk for benign prostatic hyperplasia. Lastly, we highlight the known mechanisms of alcohol interference in prostate carcinogenesis and the possible side effects of alcohol during androgen deprivation therapy.

Association Between Red and Processed Meat Consumption and Risk of Prostate Cancer: A Systematic Review and Meta-Analysis

Background

Debate on the potential carcinogenic effects of meat intake is open and the relationship between meat consumption and risk of prostate cancer remains uncertain. This meta-analysis was conducted to summarize earlier prospective studies on the association of meat consumption with risk of prostate cancer.

Methods

Relevant studies were identified by exploring PubMed/Medline, Scopus, Web of Science, EMBASE, and Google Scholar databases up to December 2020. Fixed-effects and random-effects meta-analyses were used for pooling the relative risks (RRs). Heterogeneity across studies was evaluated using the Q-statistic and I-square (I²). A funnel plot and Egger's test was used to detect publication bias. Linear and non-linear dose-response analyses were performed to estimate the dose-response relations between meat intake and risk of prostate cancer.

Results

Twenty-five prospective studies were included in this meta-analysis. Totally, 1,900,910 participants with 35,326 incident cases of prostate cancer were investigated. Pooling the eligible effect sizes, we observed that high consumption of processed meat might be associated with an increased risk of “total prostate cancer” (RR: 1.06; 95% CI: 1.01, 1.10; I² = 1.5%, P = 0.43) and “advanced prostate cancer” (1.17; 1.09, 1.26; I² = 58.8%, P = 0.01). However, the association between processed meat and “advanced prostate cancer” was not significant in the random-effects model: 1.12 (95% CI: 0.98, 1.29). A linear dose-response analysis indicated that an increment of 50 grams per day of processed meat intake might be related to a 4% greater risk of “total prostate cancer” (1.04; 1.00, 1.08; I² = 0.0%, P = 0.51). “Total meat intake” was marginally associated with all outcomes of prostate cancer risk (1.04; 1.01, 1.07; I² = 58.4%, P < 0.001).

Conclusions

This systematic review and meta-analysis of prospective studies indicated that increased consumption of “total meat” and “processed meat” might be associated with a higher risk of prostate cancer.

Systematic Review Registration

https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=230824, identifier: CRD42021230824.

Personalized 5-Year Prostate Cancer Risk Prediction Model in Korea Based on Nationwide Representative Data

Prostate cancer is the fourth most common cause of cancer in men in Korea, and there has been a rapid increase in cases. In the present study, we constructed a risk prediction model for prostate cancer using representative data from Korea. Participants who completed health examinations in 2009, based on the Korean National Health Insurance database, were eligible for the present study. The crude and adjusted risks were explored with backward selection using the Cox proportional hazards model to identify possible risk variables. Risk scores were assigned based on the adjusted hazard ratios, and the standardized points for each risk factor were proportional to the β-coefficient. Model discrimination was assessed using the concordance statistic (c-statistic), and calibration ability was assessed by plotting the mean predicted probability against the mean observed probability of prostate cancer. Among the candidate predictors, age, smoking intensity, body mass index, regular exercise, presence of type 2 diabetes mellitus, and hypertension were included. Our risk prediction model showed good discrimination (c-statistic: 0.826, 95% confidence interval: 0.821-0.832). The relationship between model predictions and actual prostate cancer development showed good correlation in the calibration plot. Our prediction model for individualized prostate cancer risk in Korean men showed good performance. Using easily accessible and modifiable risk factors, this model can help individuals make decisions regarding prostate cancer screening.

Prostate Cancer Risk Calculators for Healthy Populations: Systematic Review

Background

Screening for prostate cancer has long been a debated, complex topic. The use of risk calculators for prostate cancer is recommended for determining patients’ individual risk of cancer and the subsequent need for a prostate biopsy. These tools could lead to better discrimination of patients in need of invasive diagnostic procedures and optimized allocation of health care resources

Objective

The goal of the research was to systematically review available literature on the performance of current prostate cancer risk calculators in healthy populations by comparing the relative impact of individual items on different cohorts and on the models’ overall performance.

Methods

We performed a systematic review of available prostate cancer risk calculators targeted at healthy populations. We included studies published from January 2000 to March 2021 in English, Spanish, French, Portuguese, or German. Two reviewers independently decided for or against inclusion based on abstracts. A third reviewer intervened in case of disagreements. From the selected titles, we extracted information regarding the purpose of the manuscript, analyzed calculators, population for which it was calibrated, included risk factors, and the model’s overall accuracy.

Results

We included a total of 18 calculators from 53 different manuscripts. The most commonly analyzed ones were the Prostate Cancer Prevention Trial (PCPT) and European Randomized Study on Prostate Cancer (ERSPC) risk calculators developed from North American and European cohorts, respectively. Both calculators provided high diagnostic ability of aggressive prostate cancer (AUC as high as 0.798 for PCPT and 0.91 for ERSPC). We found 9 calculators developed from scratch for specific populations that reached a diagnostic ability as high as 0.938. The most commonly included risk factors in the calculators were age, prostate specific antigen levels, and digital rectal examination findings. Additional calculators included race and detailed personal and family history.

Conclusions

Both the PCPR and ERSPC risk calculators have been successfully adapted for cohorts other than the ones they were originally created for with no loss of diagnostic ability. Furthermore, designing calculators from scratch considering each population’s sociocultural differences has resulted in risk tools that can be well adapted to be valid in more patients. The best risk calculator for prostate cancer will be that which has been calibrated for its intended population and can be easily reproduced and implemented.

Trial Registration

PROSPERO CRD42021242110; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=242110

Combination possibility and deep learning model as clinical decision-aided approach for prostate cancer

This study aims to introduce as proof of concept a combination model for classification of prostate cancer using deep learning approaches. We utilized patients with prostate cancer who underwent surgical treatment representing the various conditions of disease progression. All possible combinations of significant variables from logistic regression and correlation analyses were determined from study data sets. The combination possibility and deep learning model was developed to predict these combinations that represented clinically meaningful patient's subgroups. The observed relative frequencies of different tumor stages and Gleason score Gls changes from biopsy to prostatectomy were available for each group. Deep learning models and seven machine learning approaches were compared for the classification performance of Gleason score changes and pT2 stage. Deep models achieved the highest F1 scores by pT2 tumors (0.849) and Gls change (0.574). Combination possibility and deep learning model is a useful decision-aided tool for prostate cancer and to group patients with prostate cancer into clinically meaningful groups.