1.Occupation classification model based on DistilKoBERT: using the 5th and 6th Korean Working Condition Surveys
Tae-Yeon KIM ; Seong-Uk BAEK ; Myeong-Hun LIM ; Byungyoon YUN ; Domyung PAEK ; Kyung Ehi ZOH ; Kanwoo YOUN ; Yun Keun LEE ; Yangho KIM ; Jungwon KIM ; Eunsuk CHOI ; Mo-Yeol KANG ; YoonHo CHO ; Kyung-Eun LEE ; Juho SIM ; Juyeon OH ; Heejoo PARK ; Jian LEE ; Jong-Uk WON ; Yu-Min LEE ; Jin-Ha YOON
Annals of Occupational and Environmental Medicine 2024;36(1):e19-
Accurate occupation classification is essential in various fields, including policy development and epidemiological studies. This study aims to develop an occupation classification model based on DistilKoBERT. This study used data from the 5th and 6th Korean Working Conditions Surveys conducted in 2017 and 2020, respectively. A total of 99,665 survey participants, who were nationally representative of Korean workers, were included. We used natural language responses regarding their job responsibilities and occupational codes based on the Korean Standard Classification of Occupations (7th version, 3-digit codes). The dataset was randomly split into training and test datasets in a ratio of 7:3. The occupation classification model based on DistilKoBERT was fine-tuned using the training dataset, and the model was evaluated using the test dataset. The accuracy, precision, recall, and F1 score were calculated as evaluation metrics. The final model, which classified 28,996 survey participants in the test dataset into 142 occupational codes, exhibited an accuracy of 84.44%. For the evaluation metrics, the precision, recall, and F1 score of the model, calculated by weighting based on the sample size, were 0.83, 0.84, and 0.83, respectively. The model demonstrated high precision in the classification of service and sales workers yet exhibited low precision in the classification of managers. In addition, it displayed high precision in classifying occupations prominently represented in the training dataset. This study developed an occupation classification system based on DistilKoBERT, which demonstrated reasonable performance. Despite further efforts to enhance the classification accuracy, this automated occupation classification model holds promise for advancing epidemiological studies in the fields of occupational safety and health.
2.Effect of preoperative pan-immune-inflammation value on clinical and oncologic outcomes after colorectal cancer surgery: a retrospective study
Yun Ju SEO ; Kyeong Eui KIM ; Woon Kyung JEONG ; Seong Kyu BAEK ; Sung Uk BAE
Annals of Surgical Treatment and Research 2024;106(3):169-177
Purpose:
Surgical resection, the primary treatment for colorectal cancer (CRC), is often linked with postoperative complications that adversely affect the overall survival rates (OS). The pan-immune-inflammation value (PIV), a novel biomarker, is promising in evaluating cancer prognoses. We aimed to explore the impact of preoperative immune inflammation status on postoperative and long-term oncological outcomes in patients with CRC.
Methods:
A retrospective analysis of 203 patients with CRC who underwent surgery (January 2016–June 2020) was conducted. The preoperative PIV was calculated as [(neutrophil count + platelet count + monocyte count) / lymphocyte counts]. The PIV optimal cutoff value was determined based on the OS using the Contal and O’Quigley methods.
Results:
A PIV value ≥155.90 was defined as high. Patients were categorized into low-PIV (n = 85) and high-PIV (n = 118) groups. Perioperative clinical outcomes (total operation time, time to gas out, sips of water, soft diet, and hospital stay) were not significantly different between the groups. The high-PIV group exhibited more postoperative complications (P = 0.024), and larger tumor size compared with the low-PIV group. Multivariate analysis identified that American Society of Anesthesiologists grade III (P = 0.046) and high-PIV (P = 0.049) were significantly associated with postoperative complications. The low-PIV group demonstrated higher OS (P = 0.001) and disease-free survival rates (DFS) (P = 0.021) compared with the high-PIV group. Advanced N stage (P = 0.005) and high-PIV levels (P = 0.047) were the identified independent prognostic factors for OS, whereas advanced N stage (P = 0.045) was an independent prognostic factor for DFS.
Conclusion
Elevated preoperative PIV was associated with an increased incidence of postoperative complications and served as an independent prognostic factor for OS.
3.Occupation classification model based on DistilKoBERT: using the 5th and 6th Korean Working Condition Surveys
Tae-Yeon KIM ; Seong-Uk BAEK ; Myeong-Hun LIM ; Byungyoon YUN ; Domyung PAEK ; Kyung Ehi ZOH ; Kanwoo YOUN ; Yun Keun LEE ; Yangho KIM ; Jungwon KIM ; Eunsuk CHOI ; Mo-Yeol KANG ; YoonHo CHO ; Kyung-Eun LEE ; Juho SIM ; Juyeon OH ; Heejoo PARK ; Jian LEE ; Jong-Uk WON ; Yu-Min LEE ; Jin-Ha YOON
Annals of Occupational and Environmental Medicine 2024;36(1):e19-
Accurate occupation classification is essential in various fields, including policy development and epidemiological studies. This study aims to develop an occupation classification model based on DistilKoBERT. This study used data from the 5th and 6th Korean Working Conditions Surveys conducted in 2017 and 2020, respectively. A total of 99,665 survey participants, who were nationally representative of Korean workers, were included. We used natural language responses regarding their job responsibilities and occupational codes based on the Korean Standard Classification of Occupations (7th version, 3-digit codes). The dataset was randomly split into training and test datasets in a ratio of 7:3. The occupation classification model based on DistilKoBERT was fine-tuned using the training dataset, and the model was evaluated using the test dataset. The accuracy, precision, recall, and F1 score were calculated as evaluation metrics. The final model, which classified 28,996 survey participants in the test dataset into 142 occupational codes, exhibited an accuracy of 84.44%. For the evaluation metrics, the precision, recall, and F1 score of the model, calculated by weighting based on the sample size, were 0.83, 0.84, and 0.83, respectively. The model demonstrated high precision in the classification of service and sales workers yet exhibited low precision in the classification of managers. In addition, it displayed high precision in classifying occupations prominently represented in the training dataset. This study developed an occupation classification system based on DistilKoBERT, which demonstrated reasonable performance. Despite further efforts to enhance the classification accuracy, this automated occupation classification model holds promise for advancing epidemiological studies in the fields of occupational safety and health.
4.Occupation classification model based on DistilKoBERT: using the 5th and 6th Korean Working Condition Surveys
Tae-Yeon KIM ; Seong-Uk BAEK ; Myeong-Hun LIM ; Byungyoon YUN ; Domyung PAEK ; Kyung Ehi ZOH ; Kanwoo YOUN ; Yun Keun LEE ; Yangho KIM ; Jungwon KIM ; Eunsuk CHOI ; Mo-Yeol KANG ; YoonHo CHO ; Kyung-Eun LEE ; Juho SIM ; Juyeon OH ; Heejoo PARK ; Jian LEE ; Jong-Uk WON ; Yu-Min LEE ; Jin-Ha YOON
Annals of Occupational and Environmental Medicine 2024;36(1):e19-
Accurate occupation classification is essential in various fields, including policy development and epidemiological studies. This study aims to develop an occupation classification model based on DistilKoBERT. This study used data from the 5th and 6th Korean Working Conditions Surveys conducted in 2017 and 2020, respectively. A total of 99,665 survey participants, who were nationally representative of Korean workers, were included. We used natural language responses regarding their job responsibilities and occupational codes based on the Korean Standard Classification of Occupations (7th version, 3-digit codes). The dataset was randomly split into training and test datasets in a ratio of 7:3. The occupation classification model based on DistilKoBERT was fine-tuned using the training dataset, and the model was evaluated using the test dataset. The accuracy, precision, recall, and F1 score were calculated as evaluation metrics. The final model, which classified 28,996 survey participants in the test dataset into 142 occupational codes, exhibited an accuracy of 84.44%. For the evaluation metrics, the precision, recall, and F1 score of the model, calculated by weighting based on the sample size, were 0.83, 0.84, and 0.83, respectively. The model demonstrated high precision in the classification of service and sales workers yet exhibited low precision in the classification of managers. In addition, it displayed high precision in classifying occupations prominently represented in the training dataset. This study developed an occupation classification system based on DistilKoBERT, which demonstrated reasonable performance. Despite further efforts to enhance the classification accuracy, this automated occupation classification model holds promise for advancing epidemiological studies in the fields of occupational safety and health.
5.Evaluating the Validity and Reliability of the Korean Version of the Scales for Outcomes in Parkinson’s Disease–Cognition
Jinse PARK ; Eungseok OH ; Seong-Beom KOH ; In-Uk SONG ; Tae-Beom AHN ; Sang Jin KIM ; Sang-Myung CHEON ; Yoon-Joong KIM ; Jin Whan CHO ; Hyeo-Il MA ; Mee Young PARK ; Jong Sam BAIK ; Phil Hyu LEE ; Sun Ju CHUNG ; Jong-Min KIM ; Han-Joon KIM ; Young-Hee SUNG ; Do Young KWON ; Jae-Hyeok LEE ; Jee-Young LEE ; Ji Seon KIM ; Ji Young YUN ; Hee Jin KIM ; Jin Yong HONG ; Mi-Jung KIM ; Jinyoung YOUN ; Hui-Jun YANG ; Won Tae YOON ; Sooyeoun YOU ; Kyum-Yil KWON ; Su-Yun LEE ; Younsoo KIM ; Hee-Tae KIM ; Joong-Seok KIM ; Ji-Young KIM
Journal of Movement Disorders 2024;17(3):328-332
Objective:
The Scales for Outcomes in Parkinson’s Disease–Cognition (SCOPA-Cog) was developed to assess cognition in patients with Parkinson’s disease (PD). In this study, we aimed to evaluate the validity and reliability of the Korean version of the SCOPACog (K-SCOPA-Cog).
Methods:
We enrolled 129 PD patients with movement disorders from 31 clinics in South Korea. The original version of the SCOPA-Cog was translated into Korean using the translation-retranslation method. The test–retest method with an intraclass correlation coefficient (ICC) and Cronbach’s alpha coefficient were used to assess reliability. Spearman’s rank correlation analysis with the Montreal Cognitive Assessment-Korean version (MOCA-K) and the Korean Mini-Mental State Examination (K-MMSE) were used to assess concurrent validity.
Results:
The Cronbach’s alpha coefficient was 0.797, and the ICC was 0.887. Spearman’s rank correlation analysis revealed a significant correlation with the K-MMSE and MOCA-K scores (r = 0.546 and r = 0.683, respectively).
Conclusion
Our results demonstrate that the K-SCOPA-Cog has good reliability and validity.
6.Characteristics and treatment patterns in older patients with locally advanced head and neck cancer (KCSG HN13-01)
Eun Joo KANG ; Yun-Gyoo LEE ; Bhumsuk KEAM ; Jin-Hyuk CHOI ; Jin-Soo KIM ; Keon Uk PARK ; Kyoung Eun LEE ; Hyo Jung KIM ; Keun-Wook LEE ; Min Kyoung KIM ; Hee Kyung AHN ; Seong Hoon SHIN ; Jii Bum LEE ; Jung Hye KWON ; Hye Ryun KIM ; Sung-Bae KIM ; Hwan Jung YUN
The Korean Journal of Internal Medicine 2022;37(1):190-200
Background/Aims:
Treatment decisions for locally advanced head and neck squamous cell carcinoma (LA-HNSCC) are complicated, and multi-modal treatments are usually indicated. However, it is challenging for older patients to complete treatments. Thus, we investigated disease characteristics, real-world treatment, and outcomes in older LA-HNSCC patients.
Methods:
Older patients (aged ≥ 70 years) were selected from a large nationwide cohort that included 445 patients with stage III–IVB LA-HNSCC from January 2005 to December 2015. Their data were retrospectively analyzed and compared with those of younger patients.
Results:
Older patients accounted for 18.7% (83/445) of all patients with median age was 73 years (range, 70 to 89). Proportions of primary tumors in the hypopharynx and larynx were higher in older patients and older patients had a more advanced T stage and worse performance status. Regarding treatment strategies of older patients, 44.5% of patients received concurrent chemoradiotherapy (CCRT), 41.0% underwent surgery, and 14.5% did not complete the planned treatment. Induction chemotherapy (IC) was administered to 27.7% (23/83) of older patients; the preferred regimen for IC was fluorouracil and cisplatin (47.9%). For CCRT, weekly cisplatin was prescribed 3.3 times more often than 3-weekly cisplatin (62.2% vs. 18.9%). Older patients had a 60% higher risk of death than younger patients (hazard ratio, 1.6; p = 0.035). Oral cavity cancer patients had the worst survival probability.
Conclusions
Older LA-HNSCC patients had aggressive tumor characteristics and received less intensive treatment, resulting in poor survival. Further research focusing on the older population is necessary.
7.Pagetoid Bowen’s Disease on Face
Jung Eun SEOL ; Sang Woo AHN ; Seung Hee JANG ; Seong Min HONG ; Jong Uk KIM ; Ji Young YUN ; Hyojin KIM
Korean Journal of Dermatology 2022;60(10):701-712
8.Short-Term Outcomes of Laparoscopic Proximal Gastrectomy With DoubleTract Reconstruction Versus Laparoscopic Total Gastrectomy for Upper Early Gastric Cancer: A KLASS 05Randomized Clinical Trial
Sun-Hwi HWANG ; Do Joong PARK ; Hyung-Ho KIM ; Woo Jin HYUNG ; Hoon HUR ; Han-Kwang YANG ; Hyuk-Joon LEE ; Hyoung-Il KIM ; Seong-Ho KONG ; Young Woo KIM ; Han Hong LEE ; Beom Su KIM ; Young-Kyu PARK ; Young-Joon LEE ; Sang-Hoon AHN ; In-Seob LEE ; Yun-Suhk SUH ; Ji-Ho PARK ; Soyeon AHN ; Sang-Uk HAN
Journal of Gastric Cancer 2022;22(2):94-106
Purpose:
Laparoscopic proximal gastrectomy with double-tract reconstruction (LPG-DTR) is a function-preserving procedure performed for treating upper early gastric cancer (EGC).However, few studies have compared the outcomes of LPG-DTR with those of laparoscopic total gastrectomy (LTG). This study aimed at comparing the short-term outcomes of LPGDTR between LTG and upper EGC.
Materials and Methods:
For upper-third EGC, a multicenter, prospective, randomized trial was performed to compare those who underwent LPG-DTR with those who underwent LTG. Short-term outcomes, including clinicopathologic results, morbidity, mortality, and postoperative courses, were evaluated using a full analysis set based on the intention-to-treat principle and the per-protocol set.
Results:
Of the patients, 138 who fulfilled the criteria were randomized to each group. One patient in the LPG-DTR group withdrew consent. Sixty-eight patients underwent LPGDTR and 69 underwent LTG. The operative time (LPG-DTR=219.4 minutes; LTG=201.8 minutes; P=0.085), estimated blood loss (LPG-DTR=76.0 mL; LTG=66.1 mL; P=0.413), and the morbidity rate (LPG-DTR=23.5%; LTG=17.4%; P=0.373) between the groups were not significantly different. No mortality occurred in either of the study groups. Two weeks post operation, the Visick scores for postprandial symptoms, including reflux symptoms, were not significantly different between the groups (P=0.749). Laboratory findings on postoperative day 5 were not significantly different between the groups.
Conclusions
The short-term outcomes of LPG-DTR for upper EGC were comparable to those of LTG.Trial Registration: ClinicalTrials.gov Identifier: NCT02892643
9.Induction Chemotherapy as a Prognostication Index and Guidance for Treatment of Locally Advanced Head and Neck Squamous Cell Carcinoma: The Concept of Chemo-Selection (KCSG HN13-01)
Yun-Gyoo LEE ; Eun Joo KANG ; Bhumsuk KEAM ; Jin-Hyuk CHOI ; Jin-Soo KIM ; Keon Uk PARK ; Kyoung Eun LEE ; Hyo Jung KIM ; Keun-Wook LEE ; Min Kyoung KIM ; Hee Kyung AHN ; Seong Hoon SHIN ; Hye Ryun KIM ; Sung-Bae KIM ; Hwan Jung YUN
Cancer Research and Treatment 2022;54(1):109-117
Purpose:
Certain patient subgroups who do not respond to induction chemotherapy (IC) show inherent chemoresistance in locally advanced head and neck squamous cell carcinoma (LA-HNSCC). This study aimed to assess the prognostic value of IC, and role of IC in guiding the selection of a definitive locoregional therapy.
Materials and Methods:
Out of the 445 patients in multi-institutional LA-HNSCC cohort, 158 (36%) receiving IC were enrolled. The study outcome was to assess overall survival (OS) through IC responsiveness and its role to select subsequent treatments.
Results:
Among 135 patients who completed subsequent treatment following IC, 74% responded to IC (complete response in 17% and partial response in 58%). IC-non-responders showed 4.5 times higher risk of mortality than IC-responders (hazard ratio, 4.52; 95% confidence interval, 2.32 to 8.81; p < 0.001). Among IC-responders, 84% subsequently received definitive concurrent chemoradiotherapy (CCRT) and OS was not differed by surgery or CCRT (p=0.960). Regarding IC-non-responders, 54% received CCRT and 46% underwent surgery, and OS was poor in CCRT (24-month survival rate of 38%) or surgery (24-month survival rate of 63%).
Conclusion
Response to IC is a favorable prognostic factor. For IC-responders, either surgery or CCRT achieved similar survival probabilities. For IC-non-responder, multidisciplinary approach was warranted reflecting patients’ preference, morbidity, and prognosis.
10.Outcomes and Biomarkers of Immune Checkpoint Inhibitor Therapy in Patients with Refractory Head and Neck Squamous Cell Carcinoma: KCSG HN18-12
Yun-Gyoo LEE ; Hyun CHANG ; Bhumsuk KEAM ; Sang Hoon CHUN ; Jihyun PARK ; Keon Uk PARK ; Seong Hoon SHIN ; Ho Jung AN ; Kyoung Eun LEE ; Keun-Wook LEE ; Hye Ryun KIM ; Sung-Bae KIM ; Myung-Ju AHN ; In Gyu HWANG
Cancer Research and Treatment 2021;53(3):671-677
Purpose:
This study was conducted to determine the effectiveness of immune checkpoint inhibitors (ICIs) in recurrent or metastatic head and neck squamous cell carcinoma (R/M HNSCC) after platinum-containing chemotherapy. We also identified clinical biomarkers which may be predictive of patient prognosis.
Materials and Methods:
We analyzed 125 patients with R/M HNSCC who received ICIs, retrospectively. Overall response rate (ORR) was the primary study outcome. Overall survival (OS) and progression-free survival (PFS) were the secondary study outcomes.
Results:
The patients received anti–programmed cell death protein-1 (PD-1) (n=73, 58%), anti–programmed death-ligand 1 (PD-L1) (n=24, 19%), or a combination of anti–PD-1/PD-L1 and anti–cytotoxic T-lymphocyte antigen 4 (n=28, 22%). The median age was 57 years (range, 37 to 87). The location of the primary tumor was in the oral cavity in 28% of the cases, followed by oropharynx (27%), hypopharynx (20%), and larynx (12%). The ORR was 15% (19/125). With 12.3 months of median follow-up, median PFS was 2.7 months. Median OS was 10.8 months. A neutrophil-to-lymphocyte ratio (NLR) > 4 was significantly associated with poor response to ICIs (odds ratio, 0.30; p=0.022). A sum of the target lesions > 40 mm (hazard ratio [HR], 1.53; p=0.046] and a NLR > 4 (HR, 1.75; p=0.009) were considered to be predictive markers of short PFS. A poor performance status (HR, 4.79; p < 0.001), a sum of target lesions > 40 mm (HR, 1.93; p=0.025), and an NLR > 4 (HR, 3.36; p < 0.001) were the significant predictors for poor survival.
Conclusion
ICIs exhibited favorable antitumor activity in R/M HNSCC. Clinically, our findings can be used to recognize patients benefit from receiving ICI.

Result Analysis
Print
Save
E-mail