Development and validation of an explainable artificial intelligence-based decision-supporting tool for prostate biopsy

OBJECTIVES

To develop and validate a risk calculator for prostate cancer (PCa) and clinically significant PCa (csPCa) using explainable artificial intelligence (XAI).

PATIENTS AND METHODS

We used data of 3791 patients to develop and validate the risk calculator. We initially divided the data into development and validation sets. An extreme gradient-boosting algorithm was applied to the development calculator using five-fold cross-validation with hyperparameter tuning following feature selection in the development set. The model feature importance was determined based on the Shapley value. The area under the curve (AUC) of the receiver operating characteristic curve was analysed for each validation set of the calculator.

RESULTS

Approximately 1216 (32.7%) and 562 (14.8%) patients were diagnosed with PCa and csPCa. The data of 2843 patients were used for development, whereas the data of 948 patients were used as a test set. We selected the variables for each PCa and csPCa risk calculation according to the least absolute shrinkage and selection operator regression. The AUC of the final PCa model was 0.869 (95% confidence interval [CI] 0.844-0.893), whereas that of the csPCa model was 0.945 (95% CI 0.927-0.963). The prostate-specific antigen (PSA) level, free PSA level, age, prostate volume (both the transitional zone and total), hypoechoic lesions on ultrasonography, and testosterone level were found to be important parameters in the PCa model. The number of previous biopsies was not associated with the risk of csPCa, but was negatively associated with the risk of PCa.

CONCLUSION

We successfully developed and validated a decision-supporting tool using XAI for calculating the probability of PCa and csPCa prior to prostate biopsy.

Supercomputer-Based Ensemble Docking Drug Discovery Pipeline with Application to Covid-19

We present a supercomputer-driven pipeline for in-silico drug discovery using enhanced sampling molecular dynamics (MD) and ensemble docking. We also describe preliminary results obtained for 23 systems involving eight protein targets of the proteome of SARS CoV-2. THe MD performed is temperature replica-exchange enhanced sampling, making use of the massively parallel supercomputing on the SUMMIT supercomputer at Oak Ridge National Laboratory, with which more than 1ms of enhanced sampling MD can be generated per day. We have ensemble docked repurposing databases to ten configurations of each of the 23 SARS CoV-2 systems using AutoDock Vina. We also demonstrate that using Autodock-GPU on SUMMIT, it is possible to perform exhaustive docking of one billion compounds in under 24 hours. Finally, we discuss preliminary results and planned improvements to the pipeline, including the use of quantum mechanical (QM), machine learning, and AI methods to cluster MD trajectories and rescore docking poses.

Bladder wall thickness and detrusor wall thickness can help to predict the bladder outlet obstruction in men over the age of 70 years with symptomatic benign prostatic hyperplasia

Purpose

We investigated the possible association between preoperative bladder wall thickness (BWT) or detrusor wall thickness (DWT) and bladder outlet obstruction (BOO) based on urodynamic studies in men with symptomatic benign prostatic hyperplasia (BPH).

Materials and Methods

Data were prospectively collected from a BPH surgery database. A total of 196 men who underwent prostate vaporization for symptomatic BPH were included in this study. BWT and DWT were measured in the suprapubic area after uroflowmetry.

Results

No significant difference was noted in BWT and DWT in any patient according to the presence of BOO; however, subgroup analysis showed that BWT and DWT were significantly thicker in the obstruction group in men aged 70 years or older than in those under age 70 (BWT: 3.6+0.9 mm vs. 3.1+0.9 mm, p=0.022, DWT: 2.8±0.8 mm vs. 2.3±0.8 mm, p=0.007). In this older age group, the classification based on a BWT ≥4.0 mm showed 31% sensitivity, 87% specificity, and 65% diagnostic accuracy for the diagnosis of BOO, whereas DWT ≥3.0 mm showed 49% sensitivity, 82% specificity, and 69% diagnostic accuracy.

Conclusions

BWT and DWT were associated with BOO in men aged 70 years or older. Therefore, BWT and DWT will be a useful non-invasive parameter for deciding the management strategy for elderly men with symptomatic BPH. An appropriate measurement method should be established as soon as possible for further application of the relationship among BWT, DWT and BOO.

Differences in risk factors for biochemical recurrence after radical prostatectomy stratified by the degree of obesity: Focused on surgical methods

This study aims to evaluate differences in the risk factors for biochemical recurrence (BCR) for radical prostatectomy stratified by degree of obesity, focusing on the surgical method used. All 3099 patients who underwent radical prostatectomy in two medical centres from January 2008 to March 2018 were retrospectively reviewed. Patients were divided into three groups based on body mass index: ‘normal’, ‘overweight’, and ‘obese’. Risk factors associated with biochemical recurrence were compared between different degrees of obesity. We analysed the contributing risk factors of BCR-free survival using univariate and multivariable Cox hazard models. There were 378 (12.2%) and 1815 (58.6%) patients in the ‘obese’ and ‘overweight’ groups, respectively. Accordingly, 1324 patients underwent retropubic radical prostatectomy (RRP), and 1775 underwent robotic-assisted laparoscopic prostatectomy (RALP). Multivariable analysis showed that patients who underwent RALP had significantly lower BCR free survival in the ‘overweight’ and ‘obese’ groups than those who underwent RRP, with adjusted hazard ratios of 0.75 (95% CI 0.59–0.95, p-value = 0.01) and 0.55 (95% CI 0.33–0.90, p-value = 0.02), respectively. The degree of obesity was not directly associated with BCR-free survival. Robotic prostatectomy showed greater benefits in BCR-free survival than open prostatectomy in the ‘overweight’ and ‘obese’ groups.

Optimal high-density lipoprotein cholesterol level for decreasing benign prostatic hyperplasia in men not taking statin medication: A historical cohort study

BACKGROUND

We evaluated the optimal high-density lipoprotein cholesterol level for benign prostatic hyperplasia (BPH) prevention in men not taking statin medication using a large historical cohort.

METHODS

We initially selected 130 454 men who underwent health checkups in 2009 from the National Health Information Database of the National Health Insurance Service. After excluding 36 854 men with BPH in 2009, and 45 061 men for statin use, 48 539 men were ultimately included in the analysis. A Kaplan-Meier analysis and multivariable Cox regression analysis was performed to assess the optimal high-density lipoprotein cholesterol level for preventing BPH.

RESULTS

High-density lipoprotein cholesterol levels were less than 40 mg/dL in 7431 (15.3%) men, 40 to 49 in 15 861 (32.7%), 50 to 59 in 15 328 (27.5%), and greater than or equal to 60 in 11 919 (24.6%). The overall cumulative incidence of BPH was 4.4%, 8.7%, 13.0%, and 17.8% at the 1-, 2-, 3-, and 4-year follow-up periods, respectively. In multivariable analysis, high-density lipoprotein greater than or equal to 60 mg/dL were significantly associated with a decreased incidence of BPH, as were age, residence, income, body mass index, diabetes, hypertension, triglyceride, and increased annual clinic visits, especially in men in their 40s.

CONCLUSION

Elevated serum high-density lipoprotein cholesterol levels were negatively associated with BPH incidence. In addition, maintaining high-density lipoprotein greater than or equal to 60 mg/dL was associated with a decreased BPH incidence compared with high-density lipoprotein less than 40 mg/dL, especially in men in their 40s.

A risk prediction model for the development of subsequent primary melanoma in a population-based cohort

BACKGROUND

Guidelines for follow-up of patients with melanoma are based on limited evidence.

OBJECTIVES

To guide skin surveillance, we developed a risk prediction model for subsequent primary melanomas, using demographic, phenotypical, histopathological, sun exposure and genomic risk factors.

METHODS

Using Cox regression frailty models, we analysed data for 2613 primary melanomas from 1266 patients recruited to the population-based Genes, Environment and Melanoma study in New South Wales, Australia, with a median of 14 years' follow-up via the cancer registry. Discrimination and calibration were assessed.

RESULTS

The median time to diagnosis of a subsequent primary melanoma decreased with each new primary melanoma. The final model included 12 risk factors. Harrell's C-statistic was 0·73 [95% confidence interval (CI) 0·68-0·77], 0·65 (95% CI 0·62-0·68) and 0·65 (95% CI 0·61-0·69) for predicting second, third and fourth primary melanomas, respectively. The risk of a subsequent primary melanoma was 4·75 times higher (95% CI 3·87-5·82) for the highest vs. the lowest quintile of the risk score. The mean absolute risk of a subsequent primary melanoma within 5 years was 8·0 ± SD 4.1% after the first primary melanoma, and 46·8 ± 15·0% after the second, but varied substantially by risk score.

CONCLUSIONS

The risk of developing a subsequent primary melanoma varies considerably between individuals and is particularly high for those with two or more primary melanomas. The risk prediction model and its associated nomograms enable estimation of the absolute risk of subsequent primary melanoma, on the basis of on an individual's risk factors, and can be used to tailor surveillance intensity, communicate risk and provide patient education. What's already known about this topic? Current guidelines for the frequency and length of follow-up to detect new primary melanomas in patients with one or more previous primary melanomas are based on limited evidence. People with one or more primary melanomas have, on average, a higher risk of developing another primary invasive melanoma, compared with the general population, but an accurate way of estimating individual risk is needed. What does this study add? We provide a comprehensive risk prediction model for subsequent primary melanomas, using data from 1266 participants with melanoma (2613 primary melanomas), over a median 14 years' follow-up. The model includes 12 risk factors comprising demographic, phenotypical, histopathological and genomic factors, and sun exposure. It enables estimation of the absolute risk of subsequent primary melanomas, and can be used to tailor surveillance intensity, communicate individual risk and provide patient education.

Selective multi-nanosoldering for fabrication of advanced solution-processed micro/nanoscale metal grid structures

Solution-processed metal grid transparent conductors with low sheet resistance, high optical transmittance and good mechanical flexibility have great potential for use in flexible optoelectronic devices. However, there are still remaining challenges to improve optoelectrical properties and electromechanical stability of the metallic structures due to random loose packings of nanoparticles and the existence of many pores. Here we introduce a selective multi-nanosoldering method to generate robust metallic layers on the thin metal grid structures (< a thickness of 200 nm), which are generated via self-pining assisted direct inking of silver ions. The selective multi-nanosoldering leads to lowering the sheet resistance of the metal grid transparent conductors, while keeping the optical transmittance constant. Also, it reinforces the electromechanical stability of flexible metal grid transparent conductors against a small bending radius or a repeated loading. Finally, organic light-emitting diodes based on the flexible metal grid transparent conductors are demonstrated. Our approach can open a new route to enhance the functionality of metallic structures fabricated using a variety of solution-processed metal patterning methods for next-generation optoelectronic and micro/nanoelectronic applications.

Can we improve the detection rate of prostate cancer using standard 12-core TRUS-guided prostate biopsy? Focused on the location of prostate biopsy

Background

We assessed the effect of biopsy location on the prostate cancer detection and clinically significant prostate cancer.

Methods

A total of 2774 patients with 12‐core prostate transrectal ultrasound‐guided prostate biopsy were included for per core analysis. Multivariate Cox regression analysis was performed to evaluate the effect of the location of biopsy on the prostate cancer and clinically significant prostate cancer detection.

Results

Prostate cancer was found in 775 patients (27.9%) and 576 prostate cancer patients (20.8%) were found to be clinically significant. The core length (P = .043), tumor length (P < .001), and % tumor length (P < .001) were significantly different according to the biopsy location. The detection rates for prostate cancer and clinically significant prostate cancer differed significantly according to the location of biopsy. Multivariate analysis revealed that the apical core was significantly related with increased detection of prostate cancer and clinically significant prostate cancer. The lateral core, in addition to apical core, was found to be significantly related with increased detection rates of prostate cancer and clinically significant prostate cancer in men with prostate‐specific antigen <10 ng/mL.

Conclusions

More in‐depth discussions on the location of standard 12‐core prostate biopsy are considered necessary. Apical core and lateral core biopsies may be helpful, especially in patients with prostate‐specific antigen ˂10 ng/mL if additional biopsies are planned following findings of no target lesions on imaging studies.

Postoperative renal functional changes assessed by 99mTc-DTPA scintigraphy and predictive factors after miniaturized percutaneous nephrolithotomy and retrograde intrarenal surgery: An observational 1-year follow-up study

Purpose

We evaluated the comparative effect of miniaturized percutaneous nephrolithotomy (mini-PCNL) and retrograde intrarenal surgery (RIRS) on perioperative kidney function by use of diethylenetriamine penta-acetic acid (^99mTc-DTPA) scintigraphy and identified significant predictors associated with deterioration or amelioration of renal function after surgery.

Materials and Methods

All 70 patients who underwent mini-PCNL or RIRS between 2012 and 2016 were monitored by ^99mTc-DTPA scintigraphy preoperatively. Patients with abnormal renal function were monitored from 3 to 12 months postoperatively. Logistic regression analyses were conducted to estimate the predictors of aggravated renal dysfunction and improvement.

Results

The difference in preoperative renal function between the contralateral and the operative side was >10% in 57 patients (81.4%). Among those in the group with abnormal renal function, 40 (70.2%), 10 (17.5%), and 7 (12.3%) patients showed stability, deterioration, and improvement in renal function at postoperative year 1, respectively. Functional changes did not differ according to the type of surgery. A high level of serum creatinine preoperatively (p=0.060) and a history of previous stone procedures (p=0.051) showed borderline significance for prediction of deterioration in renal function.

Conclusions

RIRS and mini-PCNL had similar effects and favorable outcomes on renal function during a 1-year follow-up period. High baseline serum creatinine levels and a history of procedures warrant careful attention.

Patients with Biopsy Gleason Score 3 + 4 Are Not Appropriate Candidates for Active Surveillance

OBJECTIVE

The aim of this study was to evaluate the feasibility of including patients with biopsy Gleason score (bGS) 3 + 4 prostate cancer in an active surveillance (AS) protocol.

METHODS

A total of 615 patients underwent a radical prostatectomy and satisfied the following requirements: prostate-specific antigen ≤10 ng/dL, clinical stage T1c or T2a, 2 or fewer positive biopsy cores, and bGS 6 or 3 + 4 prostate cancer. The patients were divided into two groups according to their bGS (bGS 6 group, n =534; bGS 3 + 4 group, n = 81).

RESULTS

The adverse pathological features were significantly higher in the bGS 3 + 4 group (16.7 vs. 49.4%, p< 0.001). Biochemical recurrence (BCR)-free survival was also significantly lower in this group (p < 0.001). In a multivariate analysis, clinical stage (odds ratio [OR] 2.026, p =0.007), maximum percentage of biopsy core involvement (OR 1.015, p = 0.014), and bGS (OR 1.913, p = 0.030) were independent risk factors for adverse pathological features. However, the bGS was the only variable to forecast BCR (hazard ratio 3.567, p < 0.001).

CONCLUSIONS

A bGS 3 + 4 was the leading risk factor for a worse postoperative prognosis. Therefore, patients with a bGS 3 + 4 are not appropriate candidates for AS.

Quality of life after photo-selective vaporization and holmium-laser enucleation of the prostate: 5-year outcomes

This study was aimed to compare serial long-term postoperative changes in quality-of-life (QoL) between photoselective-vaporization (PVP) using 120W-High-Performance-System and holmium-laser-enucleation (HoLEP) in benign-prostatic-hyperplasia (BPH) patients and to identify factors influencing the QoL improvement at the short-term, mid-term and long-term follow-up visits after surgery. We analyzed 1,193 patients with a baseline QoL-index ≥2 who underwent PVP (n = 439) or HoLEP (n = 754). Surgical outcomes were serially compared between the two groups at up to 60-months using the International-Prostatic-Symptom-Score (I-PSS), uroflowmetry, and serum PSA. We used logistic regression analysis to identify predictors of QoL improvement (a reduction in the QoL-index ≥50% compared with baseline) at the short-term (12-months), mid-term (36-months), and long term (60-months) follow-up after surgery. In both groups, the QoL-index was decreased throughout the entire follow-up period compared with that at baseline. There were no significant differences in postoperative changes from the baseline QoL-index between the two groups during the 48-month follow-up, except at 60-months. The degree of improvement in QoL at 60-months after HoLEP was greater than that after PVP. A lower baseline storage-symptom-subscore and a higher bladder-outlet-obstruction-index (BOOI) were independent factors influencing QoL improvement at the short-term. No independent factor influences QoL improvement at the mid- or long-term.

Electrostatic Origin of Element Selectivity during Rare Earth Adsorption

Rare earths, which are fundamental components of modern technologies, are often extracted from aqueous solutions using surfactants at oil-water interfaces. Heavier lanthanides are more easily extracted, even though all lanthanides are chemically very similar. Using x-ray fluorescence measurements and theoretical arguments, we show that there is a sharp bulk-concentration-dependent transition in the interfacial adsorption of cations from aqueous solutions containing Er^{3+} or Nd^{3+} in contact with a floating monolayer. The threshold bulk concentration of erbium (Z=68) is an order of magnitude lower than that of neodymium (Z=60), and erbium is preferentially adsorbed when the solution contains both ions. This implies that elemental selectivity during separation originates at the surfactant interface. Electrostatic effects arising from the interface dielectric mismatch, ionic correlations, and sizes of the ions explain the sharp adsorption curve and selectivity.

Effect of preoperative detrusor underactivity on long-term surgical outcomes of photovaporization and holmium laser enucleation in men with benign prostatic hyperplasia: a lesson from 5-year serial follow-up data

OBJECTIVES

To investigate the impact of preoperative detrusor underactivity (DU) on serial treatment outcomes over the course of 5 years after photovaporization (PV) or holmium laser enucleation (HoLEP) in patients with benign prostatic hyperplasia (BPH), to compare its impact after PV vs HoLEP, and to identify predictors of long-term lower urinary tract symptoms (LUTS) improvement.

MATERIALS AND METHODS

This study involved 245 patients with BPH who had complete 5-year follow-up data (PV using 120W-HPS, n = 143, HoLEP, n = 102), grouped as follows: PV-HPS-DU(+), n = 114; PV-HPS-DU(-), n = 29; HoLEP-DU(+), n = 56; and HoLEP-DU(-), n = 46. Bladder contractility index (BCI) < 100 was regarded as DU. Serial treatment outcomes for the International Prostate Symptom Score (IPSS) questionnaire, uroflowmetry and serum PSA level at 6 months, and at 1, 2, 3, 4 and 5 years after surgery, were compared among the groups. LUTS improvement was defined as a reduction in total IPSS of ≥50% relative to baseline.

RESULTS

Improvement in total IPSS, quality of life (QoL) index and post-void residual urine volume (PVR) in the PV-HPS-DU(+) and PV-HPS-DU(-) groups were maintained up to 5 years after PV, except for maximum urinary flow rate (Q_max ) and bladder voiding efficiency. In the HoLEP-DU(+) and HoLEP-DU(-) groups, improvements in all outcome variables were maintained up to 5 years after HoLEP. Deteriorations in subtotal voiding symptom score, total IPSS and Q_max with time during the long-term period after surgery were more pronounced in the PV-HPS-DU(+) and HoLEP-DU(+) groups than in the PV-HPS-DU(-) and HoLEP-DU(-) groups. Reductions in subtotal voiding symptom score, total IPSS, QoL index, and serum PSA were greater in the HoLEP-DU(+) group than in the PV-HPS-DU(+) group throughout follow-up. The type of surgery (HoLEP vs PV) and higher baseline BCI were independent predictors of LUTS improvement at 5 years after surgery.

CONCLUSION

Generally, improvement of micturition symptoms, QoL and PVR in patients with DU appears to be maintained up to 5 years after PV or HoLEP. Deterioration of voiding symptoms and urinary flow rate at long-term follow-up visits after PV or HoLEP was more pronounced in patients with LUTS/BPH with DU than in those without DU. Patients with BPH with DU may benefit from more complete removal of prostatic adenoma by HoLEP and greater baseline bladder contractility in terms of micturition symptoms and QoL.

Effects of metabolic syndrome on the prevalence of prostate cancer: historical cohort study using the national health insurance service database

PURPOSE

To estimate the effect of metabolic syndrome (MetS) on the prevalence of prostate cancer using a large retrospective cohort with a 5-year follow-up duration.

METHODS

National Health Insurance Service health checkup cohort was used for the study. In total, 130,342 men included in the health checkup cohort in 2009 were divided into two groups according to the presence of prostate cancer. The prevalence of prostate cancer from 2009 to 2013 was cumulatively calculated from 2003. A generalized estimating equation was used to assess the effect of MetS and its component on the prevalence of prostate cancer after adjusting for other variables.

RESULTS

Prostate cancer was present in 2369 men (1.8%) in 2009. The prevalence of prostate cancer was significantly higher in patients with MetS than in those without MetS throughout the entire follow-up duration. Multivariable analysis showed that in addition to year at evaluation and age, the presence of MetS was associated with an increased prevalence of prostate cancer. Alcohol consumption and smoking levels were negatively associated with the prevalence of prostate cancer. Among MetS components, decreased high density lipoprotein (HDL)-cholesterolemia and central obesity were associated with an increased prevalence of prostate cancer after adjusting for other variables.

CONCLUSION

MetS and its components, especially decreased HDL-cholesterol levels and central obesity, were related to the increased prevalence of prostate cancer. Preventing MetS, maintaining high HDL-cholesterol level, and maintaining low waist circumference might be useful ways for decreasing the prevalence of prostate cancer.

Voided volume < 150 mL on initial uroflowmetry in men with storage symptoms: Is it an unreliable test result or a sign of severe storage symptoms?

Purpose

To assess the utility of voided volume on initial uroflowmetry in predicting symptom severity in men with storage symptoms. In addition, we assessed the probability that men would show a voided volume of ≥ 150 mL on uroflowmetry if the examination is repeated.

Methods

Between 2016 and 2017, 352 men with storage symptoms were enrolled in the study. The patients completed the frequency–volume chart and were divided into four groups based on the voided volume. A multivariate analysis was performed to determine the variables affecting voided volume.

Results

The VV was < 68.8 mL in 88 patients (25.0%), 68.9–150 in 89 (25.3%), 150–250 in 87 (24.7%), and ≥ 250 in 88 (25.0%) patients. Although the International Prostate Symptom Score did not differ among the groups, the storage subscore was significantly higher in patients with decreased voided volume (p = 0.010). The total number of voids was greater in patients with decreased voided volume (p < 0.001), as was the number of nocturnal voids (p = 0.007). The maximum voided volume (p < 0.001) and 24-h urine output (p = 0.003) decreased as voided volume decreased. The proportions of patients with a mean daytime urine output ≥ 150 mL were 30.1%, 43.0%, 64.7%, and 66.7% in each group, respectively (p < 0.001). Older age and decreased maximum voided volume significantly affected the voided volume on initial uroflowmetry.

Conclusions

A decreased voided volume on initial uroflowmetry may be a sign of severe storage symptoms in men. This finding is related to older age and decreased functional capacity. In these patients, it is better to perform a careful clinical assessment to diagnose and treat overactive bladder.