1.A Machine Learning Model for Prostate Cancer Prediction in Korean Men
Sukjung CHOI ; Beomgi SO ; Shane OH ; Hongzoo PARK ; Sang Wook LEE ; Geehyun SONG ; Jong Min LEE ; Jung Ki JO ; Seon Hyeok KIM ; Si Eun LEE ; Eun-Bi CHO ; Jae Hung JUNG ; Jeong Hyun KIM
Journal of Urologic Oncology 2024;22(3):201-210
Purpose:
Unnecessary prostate biopsies for detecting prostate cancer (PCa) should be minimized. Therefore, this study developed a machine learning (ML) model to predict PCa in Korean men and evaluated its usability.
Materials and Methods:
We retrospectively analyzed clinical data from 928 patients who underwent prostate biopsies at Kangwon National University Hospital between May 2013 and May 2023. Of these, 377 (41.6%) were diagnosed with PCa, and 551 (59.4%) did not have cancer. For external validation, clinical data from 385 patients aged 48–89 years who underwent prostate biopsies from September 2005 to September 2023 at Wonju Severance Christian Hospital were also included. Twenty-two clinical features were used to develop an ML model to predict PCa. Features were selected based on their contributions to model performance, leading to the inclusion of 15 features. A meta-learner was constructed using logistic regression to predict the probability of PCa, and the classifier was trained and validated on randomly extracted training and test sets at an 8:2 ratio.
Results:
The prostate health index, prostate volume, age, nodule on digital rectal examination, and prostate-specific antigen were the top 5 features for predicting PCa. The area under the receiver operating characteristic curve (AUC) of the meta-learner logistic regression model was 0.89, and the accuracy, sensitivity, and specificity were 0.828, 0.711, and 0.909, respectively. Our model also showed excellent prediction performance for high-grade PCa, with a Gleason score of 7 or higher and an AUC of 0.903. Furthermore, we evaluated the performance of the model using external cohort clinical data and achieved an AUC of 0.863.
Conclusions
Our ML model excelled in predicting PCa, specifically clinically significant PCa. Although extensive cross-validation in other clinical cohorts is needed, this ML model is a promising option for future diagnostics.
2.A Machine Learning Model for Prostate Cancer Prediction in Korean Men
Sukjung CHOI ; Beomgi SO ; Shane OH ; Hongzoo PARK ; Sang Wook LEE ; Geehyun SONG ; Jong Min LEE ; Jung Ki JO ; Seon Hyeok KIM ; Si Eun LEE ; Eun-Bi CHO ; Jae Hung JUNG ; Jeong Hyun KIM
Journal of Urologic Oncology 2024;22(3):201-210
Purpose:
Unnecessary prostate biopsies for detecting prostate cancer (PCa) should be minimized. Therefore, this study developed a machine learning (ML) model to predict PCa in Korean men and evaluated its usability.
Materials and Methods:
We retrospectively analyzed clinical data from 928 patients who underwent prostate biopsies at Kangwon National University Hospital between May 2013 and May 2023. Of these, 377 (41.6%) were diagnosed with PCa, and 551 (59.4%) did not have cancer. For external validation, clinical data from 385 patients aged 48–89 years who underwent prostate biopsies from September 2005 to September 2023 at Wonju Severance Christian Hospital were also included. Twenty-two clinical features were used to develop an ML model to predict PCa. Features were selected based on their contributions to model performance, leading to the inclusion of 15 features. A meta-learner was constructed using logistic regression to predict the probability of PCa, and the classifier was trained and validated on randomly extracted training and test sets at an 8:2 ratio.
Results:
The prostate health index, prostate volume, age, nodule on digital rectal examination, and prostate-specific antigen were the top 5 features for predicting PCa. The area under the receiver operating characteristic curve (AUC) of the meta-learner logistic regression model was 0.89, and the accuracy, sensitivity, and specificity were 0.828, 0.711, and 0.909, respectively. Our model also showed excellent prediction performance for high-grade PCa, with a Gleason score of 7 or higher and an AUC of 0.903. Furthermore, we evaluated the performance of the model using external cohort clinical data and achieved an AUC of 0.863.
Conclusions
Our ML model excelled in predicting PCa, specifically clinically significant PCa. Although extensive cross-validation in other clinical cohorts is needed, this ML model is a promising option for future diagnostics.
3.A Machine Learning Model for Prostate Cancer Prediction in Korean Men
Sukjung CHOI ; Beomgi SO ; Shane OH ; Hongzoo PARK ; Sang Wook LEE ; Geehyun SONG ; Jong Min LEE ; Jung Ki JO ; Seon Hyeok KIM ; Si Eun LEE ; Eun-Bi CHO ; Jae Hung JUNG ; Jeong Hyun KIM
Journal of Urologic Oncology 2024;22(3):201-210
Purpose:
Unnecessary prostate biopsies for detecting prostate cancer (PCa) should be minimized. Therefore, this study developed a machine learning (ML) model to predict PCa in Korean men and evaluated its usability.
Materials and Methods:
We retrospectively analyzed clinical data from 928 patients who underwent prostate biopsies at Kangwon National University Hospital between May 2013 and May 2023. Of these, 377 (41.6%) were diagnosed with PCa, and 551 (59.4%) did not have cancer. For external validation, clinical data from 385 patients aged 48–89 years who underwent prostate biopsies from September 2005 to September 2023 at Wonju Severance Christian Hospital were also included. Twenty-two clinical features were used to develop an ML model to predict PCa. Features were selected based on their contributions to model performance, leading to the inclusion of 15 features. A meta-learner was constructed using logistic regression to predict the probability of PCa, and the classifier was trained and validated on randomly extracted training and test sets at an 8:2 ratio.
Results:
The prostate health index, prostate volume, age, nodule on digital rectal examination, and prostate-specific antigen were the top 5 features for predicting PCa. The area under the receiver operating characteristic curve (AUC) of the meta-learner logistic regression model was 0.89, and the accuracy, sensitivity, and specificity were 0.828, 0.711, and 0.909, respectively. Our model also showed excellent prediction performance for high-grade PCa, with a Gleason score of 7 or higher and an AUC of 0.903. Furthermore, we evaluated the performance of the model using external cohort clinical data and achieved an AUC of 0.863.
Conclusions
Our ML model excelled in predicting PCa, specifically clinically significant PCa. Although extensive cross-validation in other clinical cohorts is needed, this ML model is a promising option for future diagnostics.
4.Nomogram Using Prostate Health Index for Predicting Prostate Cancer in the Gray Zone:Prospective, Multicenter Study
Jae Hoon CHUNG ; Jeong Hyun KIM ; Sang Wook LEE ; Hongzoo PARK ; Geehyun SONG ; Wan SONG ; Minyong KANG ; Hyun Hwan SUNG ; Hwang Gyun JEON ; Byong Chang JEONG ; Seong IL SEO ; Hyun Moo LEE ; Seong Soo JEON
The World Journal of Men's Health 2024;42(1):168-177
Purpose:
To create a nomogram that can predict the probability of prostate cancer using prostate health index (PHI) and clinical parameters of patients. And the optimal cut-off value of PHI for prostate cancer was also assessed.
Materials and Methods:
A prospective, multi-center study was conducted. PHI was evaluated prior to biopsy in patients requiring prostate biopsy due to high prostate-specific antigen (PSA). Among screened 1,010 patients, 626 patients with clinically suspected prostate cancer with aged 40 to 85 years, and with PSA levels ranging from 2.5 to 10 ng/mL were analyzed.
Results:
Among 626 patients, 38.82% (243/626) and 22.52% (141/626) were diagnosed with prostate cancer and clinically significant prostate cancer, respectively. In the PSA 2.5 to 4 ng/mL group, the areas under the curve (AUCs) of the nomograms for overall prostate cancer and clinically significant prostate cancer were 0.796 (0.727–0.866; p<0.001), and 0.697 (0.598–0.795; p=0.001), respectively. In the PSA 4 to 10 ng/mL group, the AUCs of nomograms for overall prostate cancer and clinically significant prostate cancer were 0.812 (0.783–0.842; p<0.001), and 0.839 (0.810–0.869; p<0.001), respectively.
Conclusions
Even though external validations are necessary, a nomogram using PHI might improve the prediction of prostate cancer, reducing the need for prostate biopsies.
5.A Survey on the Quality of Life of Prostate Cancer Patients in Korean Prostate Cancer Patients Association
Yun-Sok HA ; Kwang Taek KIM ; Wook NAM ; Hongzoo PARK ; Sangjun YOO ; Chan Ho LEE ; Ho Seok CHUNG ; Woo Suk CHOI ; Jiyoun KIM ; Jaeeun SHIN ; Jeong Hyun KIM ; Cheol KWAK
Korean Journal of Urological Oncology 2022;20(4):265-272
Purpose:
We aimed to collect opinions on the diagnostic experiences and quality of life profiles for men with prostate cancer in Korea as part of the “Blue Ribbon Campaign” of the Korean Urological Oncology Society.
Materials and Methods:
Korean Urological Oncology Society conducted an online survey of 212 prostate cancer patients belonging to the Prostate Cancer Patient Association. A survey on diagnostic experience and quality of life based on Expanded Prostate Cancer Index Composite 26 Short Form were conducted.
Results:
About half of all respondents (50.5%) answered, “I experienced symptoms of urine leakage more than once a week,” during the last four weeks, 85% of the respondents said their sexual function level was “weak,” and 64.2% said, “very weak.” When asked about the level of erectile dysfunction, 58 percent of the respondents answered, “I never had an erection when I wanted one.” Of the respondents, 47.1% of men said that clinical stage at initial presentation was prostate cancer stage 3–4 and 99.1% of the respondents hoped that the prostate-specific antigen (PSA) test would be included in the national cancer screening.
Conclusions
Through this survey of patients, we were able to confirm the difficulties of the low quality of life currently experienced by prostate cancer patients and what they want to do with prostate cancer treatment. All patients are eager to include a PSA test in the national cancer screening so that prostate cancer can be detected early and patients can receive proper treatment at an appropriate time.
6.Inhibition of voltage-dependent K+ channels by antimuscarinic drug fesoterodine in coronary arterial smooth muscle cells
Seojin PARK ; Minji KANG ; Ryeon HEO ; Seo-Yeong MUN ; Minju PARK ; Eun-Taek HAN ; Jin-Hee HAN ; Wanjoo CHUN ; Hongzoo PARK ; Won Sun PARK
The Korean Journal of Physiology and Pharmacology 2022;26(5):397-404
Fesoterodine, an antimuscarinic drug, is widely used to treat overactive bladder syndrome. However, there is little information about its effects on vascular K+ channels. In this study, voltage-dependent K+ (Kv) channel inhibition by fesoterodine was investigated using the patch-clamp technique in rabbit coronary artery. In whole-cell patches, the addition of fesoterodine to the bath inhibited the Kv currents in a concentration-dependent manner, with an IC50 value of 3.19 ± 0.91 μM and a Hill coefficient of 0.56 ± 0.03. Although the drug did not alter the voltage-dependence of steady-state activation, it shifted the steady-state inactivation curve to a more negative potential, suggesting that fesoterodine affects the voltage-sensor of the Kv channel. Inhibition by fesoterodine was significantly enhanced by repetitive train pulses (1 or 2 Hz). Furthermore, it significantly increased the recovery time constant from inactivation, suggesting that the Kv channel inhibition by fesoterodine is use (state)-dependent. Its inhibitory effect disappeared by pretreatment with a Kv 1.5 inhibitor. However, pretreatment with Kv2.1 or Kv7 inhibitors did not affect the inhibitory effects on Kv channels. Based on these results, we conclude that fesoterodine inhibits vascular Kv channels (mainly the Kv1.5 subtype) in a concentration- and use (state)-dependent manner, independent of muscarinic receptor antagonism.
7.Inhibitory effects of the atypical antipsychotic, clozapine, on voltage-dependent K+ channels in rabbit coronary arterial smooth muscle cells
Minji KANG ; Ryeon HEO ; Seojin PARK ; Seo-Yeong MUN ; Minju PARK ; Eun-Taek HAN ; Jin-Hee HAN ; Wanjoo CHUN ; Kwon-Soo HA ; Hongzoo PARK ; Won-Kyo JUNG ; Il-Whan CHOI ; Won Sun PARK
The Korean Journal of Physiology and Pharmacology 2022;26(4):277-285
To investigate the adverse effects of clozapine on cardiovascular ion channels, we examined the inhibitory effect of clozapine on voltage-dependent K+(Kv) channels in rabbit coronary arterial smooth muscle cells. Clozapine-induced inhibition of Kv channels occurred in a concentration-dependent manner with an halfinhibitory concentration value of 7.84 ± 4.86 µM and a Hill coefficient of 0.47 ± 0.06.Clozapine did not shift the steady-state activation or inactivation curves, suggesting that it inhibited Kv channels regardless of gating properties. Application of train pulses (1 and 2 Hz) progressively augmented the clozapine-induced inhibition of Kv channels in the presence of the drug. Furthermore, the recovery time constant from inactivation was increased in the presence of clozapine, suggesting that clozapineinduced inhibition of Kv channels is use (state)-dependent. Pretreatment of a Kv1.5 subtype inhibitor decreased the Kv current amplitudes, but additional application of clozapine did not further inhibit the Kv current. Pretreatment with Kv2.1 or Kv7 subtype inhibitors partially blocked the inhibitory effect of clozapine. Based on these results, we conclude that clozapine inhibits arterial Kv channels in a concentrationand use (state)-dependent manner. Kv1.5 is the major subtype involved in clozapineinduced inhibition of Kv channels, and Kv2.1 and Kv7 subtypes are partially involved.
8.Investigation of Information Acquisition Channel for Prostate Cancer High-Risk Group
Yun-Sok HA ; Kwang Taek KIM ; Wook NAM ; Hongzoo PARK ; Sangjun YOO ; Chan Ho LEE ; Ho Seok CHUNG ; Woo Suk CHOI ; Jiyoun KIM ; Jaeeun SHIN ; Jeong Hyun KIM ; Cheol KWAK
Korean Journal of Urological Oncology 2021;19(3):174-182
Purpose:
The survey was conducted on Korean men to examine information acquisition channel for prostate cancer high risk group as part of the “Blue Ribbon Campaign” of the Korean Urological Oncology Society.
Materials and Methods:
An online survey of 500 men aged 50 years old or older was completed to query investigation of the status of prostate cancer awareness and information acquisition from February 4 to February 9, 2021.
Results:
Most men in their 50s and older are well aware that prostate cancer can also occur in young men in their 40s, so the rate of misunderstanding of the timing of prostate cancer screening after their 60s is very low. Two-thirds of all respondents (67.2%) were also confirmed that prostate cancer had no initial symptoms and was not included in the national cancer screening. Seventy-five percent of people look up information on their own in case of suspected prostate cancer, and 51.6% seek out knowledge on their own to prevent prostate cancer. Of the respondents, 27.4% of men contacted prostate cancer-related information within the past year, and the percentage of people contacted through ‘Internet/Phone,’ ‘People Around’ and ‘Television’ was high. The most trusted channel among prostate cancer information channels was ‘medical professionals,’ but the experience rate was not high, and the channel with high experience rate and reliability was shown as ‘television.’
Conclusions
Much effort is still needed to understand the information acquisition behavior of Korean men and to improve awareness of early screening for prostate cancer.
9.Investigation of Information Acquisition Channel for Prostate Cancer High-Risk Group
Yun-Sok HA ; Kwang Taek KIM ; Wook NAM ; Hongzoo PARK ; Sangjun YOO ; Chan Ho LEE ; Ho Seok CHUNG ; Woo Suk CHOI ; Jiyoun KIM ; Jaeeun SHIN ; Jeong Hyun KIM ; Cheol KWAK
Korean Journal of Urological Oncology 2021;19(3):174-182
Purpose:
The survey was conducted on Korean men to examine information acquisition channel for prostate cancer high risk group as part of the “Blue Ribbon Campaign” of the Korean Urological Oncology Society.
Materials and Methods:
An online survey of 500 men aged 50 years old or older was completed to query investigation of the status of prostate cancer awareness and information acquisition from February 4 to February 9, 2021.
Results:
Most men in their 50s and older are well aware that prostate cancer can also occur in young men in their 40s, so the rate of misunderstanding of the timing of prostate cancer screening after their 60s is very low. Two-thirds of all respondents (67.2%) were also confirmed that prostate cancer had no initial symptoms and was not included in the national cancer screening. Seventy-five percent of people look up information on their own in case of suspected prostate cancer, and 51.6% seek out knowledge on their own to prevent prostate cancer. Of the respondents, 27.4% of men contacted prostate cancer-related information within the past year, and the percentage of people contacted through ‘Internet/Phone,’ ‘People Around’ and ‘Television’ was high. The most trusted channel among prostate cancer information channels was ‘medical professionals,’ but the experience rate was not high, and the channel with high experience rate and reliability was shown as ‘television.’
Conclusions
Much effort is still needed to understand the information acquisition behavior of Korean men and to improve awareness of early screening for prostate cancer.
10.De Novo Papillary Urothelial Carcinoma at a Previous Ureteroneocystostomy Site for Benign Ureteral Injury
Korean Journal of Urological Oncology 2020;18(1):68-72
Ureteroneocystostomy is a good treatment option for iatrogenic ureteral injury. Common complications at ureteroneocystostomy sites are strictures, stone formation, urinary infections, fistulas, and ureteral leaks. Here, we report a rare occurrence of urothelial carcinoma occurring at the site of a previous ureteral reimplantation. A 57-year-old female presented in the Emergency Department with left flank pain and chills. She had undergone a left ureteroneocystostomy with Boari flap due to iatrogenic ureteral obstruction during a laparoscopic left ovarian cystic mass excision 2 years ago. Computed tomography revealed left ureteral obstruction by the tumor at the neo-ureterovesical junction site. Both anterograde and retrograde ureteral catheterization approaches failed. We conducted a left percutaneous nephrostomy and administered antibiotics. Urine cytology was negative. We performed a left ureterovesical obstructive mass excision and Yang-Monti ileal ureter reconstruction. Biopsy of the ureteral-obstructing tumor revealed a low-grade papillary urothelial carcinoma. The patient’s symptoms and signs improved after surgery. To the best of our knowledge, this is the first report of a de novo urothelial carcinoma at the site of previous ureterovesical junction surgery. Urothelial carcinoma should be considered as one of the causes of stricture after ureteroneocystostomy. (Korean J Urol Oncol 2020;18:68-72)

Result Analysis
Print
Save
E-mail