Background/Aims: Inaccurate prediction of lymph node metastasis (LNM) may lead to unnecessary surgery following endoscopic resection of T1 colorectal cancer (CRC). We aimed to validate the usefulness of artificial intelligence (AI) models for predicting LNM in patients with T1 CRC. Methods: We analyzed the clinical data, laboratory results, pathological reports, and endoscopic findings of patients who underwent radical surgery for T1 CRC. We developed AI models to predict LNM using four algorithms: regularized logistic regression classifier (RLRC), random forest classifier (RFC), CatBoost classifier (CBC), and the voting classifier (VC). Four histological factors and four endoscopic findings were included to develop AI models. Areas under the receiver operating characteristics curves (AUROCs) were measured to distinguish AI model performance in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines. Results: Among 1,386 patients with T1 CRC, 173 patients (12.5%) had LNM. The AUROC values of the RLRC, RFC, CBC, and VC models for LNM prediction were significantly higher (0.673, 0.640, 0.679, and 0.677, respectively) than the 0.525 suggested in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines (vs RLRC, p<0.001; vs RFC, p=0.001; vs CBC, p<0.001; vs VC, p<0.001). The AUROC value was similar between T1 colon versus T1 rectal cancers (0.718 vs 0.615, p=0.700). The AUROC value was also similar between the initial endoscopic resection and initial surgery groups (0.581 vs 0.746, p=0.845). Conclusions AI models trained on the basis of endoscopic findings and pathological features performed well in predicting LNM in patients with T1 CRC regardless of tumor location and initial treatment method.

Background/Aims: Inaccurate prediction of lymph node metastasis (LNM) may lead to unnecessary surgery following endoscopic resection of T1 colorectal cancer (CRC). We aimed to validate the usefulness of artificial intelligence (AI) models for predicting LNM in patients with T1 CRC. Methods: We analyzed the clinical data, laboratory results, pathological reports, and endoscopic findings of patients who underwent radical surgery for T1 CRC. We developed AI models to predict LNM using four algorithms: regularized logistic regression classifier (RLRC), random forest classifier (RFC), CatBoost classifier (CBC), and the voting classifier (VC). Four histological factors and four endoscopic findings were included to develop AI models. Areas under the receiver operating characteristics curves (AUROCs) were measured to distinguish AI model performance in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines. Results: Among 1,386 patients with T1 CRC, 173 patients (12.5%) had LNM. The AUROC values of the RLRC, RFC, CBC, and VC models for LNM prediction were significantly higher (0.673, 0.640, 0.679, and 0.677, respectively) than the 0.525 suggested in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines (vs RLRC, p<0.001; vs RFC, p=0.001; vs CBC, p<0.001; vs VC, p<0.001). The AUROC value was similar between T1 colon versus T1 rectal cancers (0.718 vs 0.615, p=0.700). The AUROC value was also similar between the initial endoscopic resection and initial surgery groups (0.581 vs 0.746, p=0.845). Conclusions AI models trained on the basis of endoscopic findings and pathological features performed well in predicting LNM in patients with T1 CRC regardless of tumor location and initial treatment method.

Background/Aims: Inaccurate prediction of lymph node metastasis (LNM) may lead to unnecessary surgery following endoscopic resection of T1 colorectal cancer (CRC). We aimed to validate the usefulness of artificial intelligence (AI) models for predicting LNM in patients with T1 CRC. Methods: We analyzed the clinical data, laboratory results, pathological reports, and endoscopic findings of patients who underwent radical surgery for T1 CRC. We developed AI models to predict LNM using four algorithms: regularized logistic regression classifier (RLRC), random forest classifier (RFC), CatBoost classifier (CBC), and the voting classifier (VC). Four histological factors and four endoscopic findings were included to develop AI models. Areas under the receiver operating characteristics curves (AUROCs) were measured to distinguish AI model performance in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines. Results: Among 1,386 patients with T1 CRC, 173 patients (12.5%) had LNM. The AUROC values of the RLRC, RFC, CBC, and VC models for LNM prediction were significantly higher (0.673, 0.640, 0.679, and 0.677, respectively) than the 0.525 suggested in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines (vs RLRC, p<0.001; vs RFC, p=0.001; vs CBC, p<0.001; vs VC, p<0.001). The AUROC value was similar between T1 colon versus T1 rectal cancers (0.718 vs 0.615, p=0.700). The AUROC value was also similar between the initial endoscopic resection and initial surgery groups (0.581 vs 0.746, p=0.845). Conclusions AI models trained on the basis of endoscopic findings and pathological features performed well in predicting LNM in patients with T1 CRC regardless of tumor location and initial treatment method.

Background/Aims: Inaccurate prediction of lymph node metastasis (LNM) may lead to unnecessary surgery following endoscopic resection of T1 colorectal cancer (CRC). We aimed to validate the usefulness of artificial intelligence (AI) models for predicting LNM in patients with T1 CRC. Methods: We analyzed the clinical data, laboratory results, pathological reports, and endoscopic findings of patients who underwent radical surgery for T1 CRC. We developed AI models to predict LNM using four algorithms: regularized logistic regression classifier (RLRC), random forest classifier (RFC), CatBoost classifier (CBC), and the voting classifier (VC). Four histological factors and four endoscopic findings were included to develop AI models. Areas under the receiver operating characteristics curves (AUROCs) were measured to distinguish AI model performance in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines. Results: Among 1,386 patients with T1 CRC, 173 patients (12.5%) had LNM. The AUROC values of the RLRC, RFC, CBC, and VC models for LNM prediction were significantly higher (0.673, 0.640, 0.679, and 0.677, respectively) than the 0.525 suggested in accordance with the Japanese Society for Cancer of the Colon and Rectum guidelines (vs RLRC, p<0.001; vs RFC, p=0.001; vs CBC, p<0.001; vs VC, p<0.001). The AUROC value was similar between T1 colon versus T1 rectal cancers (0.718 vs 0.615, p=0.700). The AUROC value was also similar between the initial endoscopic resection and initial surgery groups (0.581 vs 0.746, p=0.845). Conclusions AI models trained on the basis of endoscopic findings and pathological features performed well in predicting LNM in patients with T1 CRC regardless of tumor location and initial treatment method.

Purpose: In Korea, the Broselow tape (BT) is widely used to estimate weight in resuscitation. Validation of BT in Korean children is essential because the tool was developed based on children’s weight and height in the United States. The validation was previously performed in a small-scale dataset. The authors aimed to validate BT using the 2005 Korean nationwide anthropometric survey data. Methods: From the population used for the survey, we sampled children aged 0-12 years. The weights estimated by BT were compared with measured weights of the children using Bland-Altman analysis with results recorded as percentage differences. We measured the accuracy of BT, defined as within a 10% error of the measured weight, and the concordance of the color-coded zones derived from the estimated and measured weights. The accuracy and concordance were further assessed according to the age groups and body mass index-for-age Z-score (< -2, underweight; > 2, overweight or obese). Results: A total of 108,128 children were enrolled. The mean age was 55.2 ± 37.5 months. The bias was –5.4% (P < 0.001), and the limits of agreement were –28.3% and 17.6%, respectively. The accuracy and concordance of BT were 64.4% and 67.2%, respectively. Differences of no more than 1 color-coded zone between estimated and measured weights accounted for 89.8% and 84.1% of the under- and overweight (or obese) children, respectively. Conclusion BT accurately estimates weight in approximately two-thirds of Korean children. In addition, adjustment of 1 color-coded zone may be considered in children with extreme weight.

Central venous catheterization is a routinely performed procedure in clinical practice. While cerebral air embolism after the removal of the central venous catheter is very rare, it is one of the most serious complications that can lead to fatal outcomes. In this report, we present a rare case of a cerebral air embolism after the removal of the central venous catheter in a patient with a patent foramen ovale.

Accessory muscles located in the region of the popliteal fossa are very rare. MRI scan performed in a 52-year-old man with right knee pain revealed an anomalous muscle in the region of the popliteal fossa. Considering the muscle originated from the medial side of the lateral head of the gastrocnemius muscle and attached to the posteromedial articular capsule of the knee joint, it is consistent with the accessory popliteus muscle, previously reported. To our best knowledge, MRI finding about the accessory popliteus muscle has been reported in only one case. We present a case of the accessory popliteus muscle incidentally identified on MRI.

PURPOSE: Leber hereditary optic neuropathy (LHON) is one of the most common hereditary optic neuropathies caused by mutations of mitochondrial DNA. Three common mitochondrial mutations causing LHON are m.3460, m.11778, and m.14484. We report a rare mutation of the mitochondrial tRNA (Leu [UUR]) gene (MT-TL1) (m.3268 A > G) in a patient with bilateral optic atrophy. CASE SUMMARY: A 59-year-old female diagnosed with glaucoma 3 years earlier at a community eye clinic presented to our neuro-ophthalmology clinic. On examination, her best corrected visual acuity was 0.4 in the right eye and 0.7 in the left eye, and optic atrophy was noticed in both eyes. Optical coherence tomography revealed retinal nerve fiber layer (RNFL) thinning in both eyes; average RNFL thickness was 52 µm in the right eye and 44 µm in the left eye, but the papillomacular bundle was relatively preserved in both eyes. Goldmann perimetry demonstrated peripheral visual field defects, mostly involving superotemporal visual field in both eyes. Mitochondrial DNA mutation test showed an unusual mutation in MT-TL1 gene seemingly related to this optic neuropathy. CONCLUSIONS: We found a rare mutation (m.3268 A > G) of the mitochondrial DNA in a patient having bilateral optic atrophy, which led to the diagnosis of LHON. There have been previous reports about mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) and infantile myopathy caused by MT-TL1 mutation, but this is the first case of LHON associated with the same mutation. In this case of LHON associated with MT-TL1 mutation, atypical clinical features were observed with a relatively mild phenotype and peripheral visual field defects.
Diagnosis ; DNA, Mitochondrial ; Female ; Glaucoma* ; Humans ; MELAS Syndrome ; Middle Aged ; Muscular Diseases ; Nerve Fibers ; Optic Atrophy ; Optic Atrophy, Hereditary, Leber* ; Optic Nerve Diseases ; Phenotype ; Retinaldehyde ; RNA, Transfer ; Tomography, Optical Coherence ; Visual Acuity ; Visual Field Tests ; Visual Fields

PURPOSE: To assess the validity of central corneal thickness (CCT) and corrected intraocular pressure (IOP) values obtained by tono-pachymetry in non-surgical and post-photorefractive keratectomy (PRK) eyes. METHODS: For the study, 108 young healthy participants and 108 patients who had PRK were enrolled. Measurements were randomly performed by tono-pachymetry, ultrasonic (US) pachymetry, and Goldmann applanation tonometry (GAT). CCT measurement by tono-pachymetry was compared to that of US pachymetry. The corrected IOP value obtained by tono-pachymetry was compared to that obtained by US pachymetry and GAT. The corrected IOP from US pachymetry and GAT was calculated using the identical compensation formula built into the tono-pachymetry. Bland-Altman plot and paired t-test were conducted to evaluate the between-method agreements. RESULTS: The mean CCT measurement using tono-pachymetry was significantly greater by 7.3 µm in non-surgical eyes (p < 0.001) and 17.8 µm in post-PRK eyes (p < 0.001) compared with US pachymetry. Differences were significant in both Bland-Altman plotand paired t-test. The mean difference of corrected IOP values obtained by tono-pachymetry and calculated from measurements by US pachymetry and GAT was 0.33 ± 0.87 mmHg in non-surgical eyes and 0.57 ± 1.08 mmHg in post-PRK eyes. The differences in the Bland-Altman plot were not significant. CONCLUSIONS: The CCT measurement determined using tono-pachymetrywas significantly thicker than that of US pachymetry. The difference in CCT was greater in post-PRK eyes than in non-surgical eyes. However, the corrected IOP value obtained by tono-pachymetry showed reasonable agreement with that calculated from US pachymetry and GAT measurements.
Compensation and Redress ; Corneal Pachymetry ; Healthy Volunteers ; Humans ; Intraocular Pressure* ; Manometry ; Photorefractive Keratectomy ; Ultrasonics

OBJECTIVE: The aim of this study was to elucidate the epidemiology of pediatric patients transported by the National 119 Rescue Services in Korea. METHODS: We enrolled all pediatric patients (<16 years old) who used the National 119 Rescue Services in Korea between January 2006 and December 2008, and analyzed the 119 ambulance patient care record databases. RESULTS: The total number of the cases was 238,644 for 3 years. The median age was 6 years old and 59.0% were male, and the 2- to 5-year-old group was the largest (31.0%). The peak transport times were in the afternoon (from 12:00 p.m. to 17:59 p.m., 36.3%), on Saturday and Sunday (15.9% and 15.7%), and in summer (June to August, 27.3%). The ratio of disease versus injury as the cause of the transports was 42.3% vs. 57.7%. Among the 16 metropolitan cities and provinces, Gyeonggi (25.7%), Seoul (17.6%), and Incheon (7.0%) account for almost half of the all transported children. Regarding the annual transport rates per 100,000 children standardized by age, and gender to the Korean child population, Jeju was the largest (1,650.2) followed by Gangwon (1,201.3), and Jeonnam (1,178.1). CONCLUSION: This report presents comprehensive epidemiologic data of pediatric patients transported by 119 rescue services in Korea.
Ambulances ; Child* ; Child, Preschool ; Emergencies* ; Emergency Medical Services* ; Epidemiology* ; Gangwon-do ; Gyeonggi-do ; Humans ; Incheon ; Jeollanam-do ; Korea ; Male ; Patient Care ; Seoul