• Korean Journal of Radiology
• /
• v.23 no.3
• /
• pp.343-354
• /
• 2022

Objective: To develop and evaluate a deep learning-based artificial intelligence (AI) model for detecting skull fractures on plain radiographs in children. Materials and Methods: This retrospective multi-center study consisted of a development dataset acquired from two hospitals (n = 149 and 264) and an external test set (n = 95) from a third hospital. Datasets included children with head trauma who underwent both skull radiography and cranial computed tomography (CT). The development dataset was split into training, tuning, and internal test sets in a ratio of 7:1:2. The reference standard for skull fracture was cranial CT. Two radiology residents, a pediatric radiologist, and two emergency physicians participated in a two-session observer study on an external test set with and without AI assistance. We obtained the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity along with their 95% confidence intervals (CIs). Results: The AI model showed an AUROC of 0.922 (95% CI, 0.842-0.969) in the internal test set and 0.870 (95% CI, 0.785-0.930) in the external test set. The model had a sensitivity of 81.1% (95% CI, 64.8%-92.0%) and specificity of 91.3% (95% CI, 79.2%-97.6%) for the internal test set and 78.9% (95% CI, 54.4%-93.9%) and 88.2% (95% CI, 78.7%-94.4%), respectively, for the external test set. With the model's assistance, significant AUROC improvement was observed in radiology residents (pooled results) and emergency physicians (pooled results) with the difference from reading without AI assistance of 0.094 (95% CI, 0.020-0.168; p = 0.012) and 0.069 (95% CI, 0.002-0.136; p = 0.043), respectively, but not in the pediatric radiologist with the difference of 0.008 (95% CI, -0.074-0.090; p = 0.850). Conclusion: A deep learning-based AI model improved the performance of inexperienced radiologists and emergency physicians in diagnosing pediatric skull fractures on plain radiographs.

• Korean Journal of Radiology
• /
• v.23 no.12
• /
• pp.1241-1250
• /
• 2022

Objective: To conduct a simulation study to determine whether artificial intelligence (AI)-aided mammography reading can reduce unnecessary recalls while maintaining cancer detection ability in women recalled after mammography screening. Materials and Methods: A retrospective reader study was performed by screening mammographies of 793 women (mean age ± standard deviation, 50 ± 9 years) recalled to obtain supplemental mammographic views regarding screening mammography-detected abnormalities between January 2016 and December 2019 at two screening centers. Initial screening mammography examinations were interpreted by three dedicated breast radiologists sequentially, case by case, with and without AI aid, in a single session. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and recall rate for breast cancer diagnosis were obtained and compared between the two reading modes. Results: Fifty-four mammograms with cancer (35 invasive cancers and 19 ductal carcinomas in situ) and 739 mammograms with benign or negative findings were included. The reader-averaged AUC improved after AI aid, from 0.79 (95% confidence interval [CI], 0.74-0.85) to 0.89 (95% CI, 0.85-0.94) (p < 0.001). The reader-averaged specificities before and after AI aid were 41.9% (95% CI, 39.3%-44.5%) and 53.9% (95% CI, 50.9%-56.9%), respectively (p < 0.001). The reader-averaged sensitivity was not statistically different between AI-unaided and AI-aided readings: 89.5% (95% CI, 83.1%-95.9%) vs. 92.6% (95% CI, 86.2%-99.0%) (p = 0.053), although the sensitivities of the least experienced radiologists before and after AI aid were 79.6% (43 of 54 [95% CI, 66.5%-89.4%]) and 90.7% (49 of 54 [95% CI, 79.7%-96.9%]), respectively (p = 0.031). With AI aid, the reader-averaged recall rate decreased by from 60.4% (95% CI, 57.8%-62.9%) to 49.5% (95% CI, 46.5%-52.4%) (p < 0.001). Conclusion: AI-aided reading reduced the number of recalls and improved the diagnostic performance in our simulation using women initially recalled for supplemental mammographic views after mammography screening.

• Journal of Korean Neurosurgical Society
• /
• v.67 no.1
• /
• pp.94-102
• /
• 2024

Objective : The spontaneous intracerebral hemorrhage (ICH) remains a significant cause of mortality and morbidity throughout the world. The purpose of this retrospective study is to develop multiple models for predicting ICH outcomes using machine learning (ML). Methods : Between January 2014 and October 2021, we included ICH patients identified by computed tomography or magnetic resonance imaging and treated with surgery. At the 6-month check-up, outcomes were assessed using the modified Rankin Scale. In this study, four ML models, including Support Vector Machine (SVM), Decision Tree C5.0, Artificial Neural Network, Logistic Regression were used to build ICH prediction models. In order to evaluate the reliability and the ML models, we calculated the area under the receiver operating characteristic curve (AUC), specificity, sensitivity, accuracy, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR). Results : We identified 71 patients who had favorable outcomes and 156 who had unfavorable outcomes. The results showed that the SVM model achieved the best comprehensive prediction efficiency. For the SVM model, the AUC, accuracy, specificity, sensitivity, PLR, NLR, and DOR were 0.91, 0.92, 0.92, 0.93, 11.63, 0.076, and 153.03, respectively. For the SVM model, we found the importance value of time to operating room (TOR) was higher significantly than other variables. Conclusion : The analysis of clinical reliability showed that the SVM model achieved the best comprehensive prediction efficiency and the importance value of TOR was higher significantly than other variables.

• ETRI Journal
• /
• v.26 no.3
• /
• pp.229-240
• /
• 2004

This paper reports on our development of a dual-mode transceiver for a CMOS high-rate Bluetooth system-onchip solution. The transceiver includes most of the radio building blocks such as an active complex filter, a Gaussian frequency shift keying (GFSK) demodulator, a variable gain amplifier (VGA), a dc offset cancellation circuit, a quadrature local oscillator (LO) generator, and an RF front-end. It is designed for both the normal-rate Bluetooth with an instantaneous bit rate of 1 Mb/s and the high-rate Bluetooth of up to 12 Mb/s. The receiver employs a dualconversion combined with a baseband dual-path architecture for resolving many problems such as flicker noise, dc offset, and power consumption of the dual-mode system. The transceiver requires none of the external image-rejection and intermediate frequency (IF) channel filters by using an LO of 1.6 GHz and the fifth order onchip filters. The chip is fabricated on a $6.5-mm^{2}$ die using a standard $0.25-{\mu}m$ CMOS technology. Experimental results show an in-band image-rejection ratio of 40 dB, an IIP3 of -5 dBm, and a sensitivity of -77 dBm for the Bluetooth mode when the losses from the external components are compensated. It consumes 42 mA in receive ${\pi}/4-diffrential$ quadrature phase-shift keying $({\pi}/4-DQPSK)$ mode of 8 Mb/s, 35 mA in receive GFSK mode of 1 Mb/s, and 32 mA in transmit mode from a 2.5-V supply. These results indicate that the architecture and circuits are adaptable to the implementation of a low-cost, multi-mode, high-speed wireless personal area network.

• Imaging Science in Dentistry
• /
• v.44 no.3
• /
• pp.213-220
• /
• 2014

Purpose: This study was performed to determine the efficacy of observers' prediction for the need of bone grafting and presence of perioperative complications on the basis of cone-beam computed tomography (CBCT) and panoramic radiographic (PAN) planning as compared to the surgical outcome. Materials and Methods: One hundred and eight partially edentulous patients with a need for implant rehabilitation were referred for preoperative imaging. Imaging consisted of PAN and CBCT images. Four observers carried out implant planning using PAN image datasets, and at least one month later, using CBCT image datasets. Based on their own planning, the observers assessed the need for bone graft augmentation as well as complication prediction. The implant length and diameter, the need for bone graft augmentation, and the occurrence of anatomical complications during planning and implant placement were statistically compared. Results: In the 108 patients, 365 implants were installed. Receiver operating characteristic analyses of both PAN and CBCT preoperative planning showed that CBCT performed better than PAN-based planning with respect to the need for bone graft augmentation and perioperative complications. The sensitivity and the specificity of CBCT for implant complications were 96.5% and 90.5%, respectively, and for bone graft augmentation, they were 95.2% and 96.3%, respectively. Significant differences were found between PAN-based planning and the surgery of posterior implant lengths. Conclusion: Our findings indicated that CBCT-based preoperative implant planning enabled treatment planning with a higher degree of prediction and agreement as compared to the surgical standard. In PAN-based surgery, the prediction of implant length was poor.

• Korean Journal of Optics and Photonics
• /
• v.23 no.1
• /
• pp.6-11
• /
• 2012

A bidirectional TDM-PON link to support 2.5 Gb/s upstream signals of 256 ONUs was considered for an extended transmission distance of 50 km. The power budget of the link was 58 dB for the upstream signal and a SOA was applied as a link extender which had a 25 dB gain. Receiver sensitivity of the upstream signal was -25 dBm for -30 dBm input power to the SOA. When the input power was -10 dBm, pulse overshooting caused by gain transient of the SOA was maximum at 45% and the signal performance degradation gave a power penalty of 1.55 dB for $10^{-12}$ BER. However the penalties diminished rapidly and became negligible as the input power went below -15 dBm. So this input power dynamic range of up to -15 dBm means that it is not positively necessary to use gain control methods for the next generation TDM-PON systems.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.43 no.5 s.347
• /
• pp.182-188
• /
• 2006

In this study, we have designed the millimeter-wave passive imaging system which records energy that is reflected or emitted from the source and produces image. The lens and front-end of receiver appeared to be important in the system to detect input thermal noise signal. The lens for signal focusing has been designed by optical transfer function. Amplifier of the imaging systemhas been set up with 40dB in maximum gain, 5 dB in maximum noise figure, and 10GHz in bandwidth to enhance sensitivity for thermal noise and to receive it in wide-band width as well. The SBD MSS-20 141B10D diode has been used for the detector circuit to convert amplified millimeter-wave signals to DC output.

• Asian Pacific Journal of Cancer Prevention
• /
• v.13 no.11
• /
• pp.5483-5488
• /
• 2012

Background: Although nicotine dependence plays a role as a main barrier for smoking cessation, there is still a lack of solid evidence on the validity of biomarkers to determine nicotine dependence in clinical settings. This study aimed to investigate whether urinary cotinine levels could reflect the severity of nicotine dependence in active smokers. Materials and Methods: Data regarding general characteristics and smoking status was collected using a self-administered smoking questionnaire. The Fagerstr$\ddot{o}$m test for nicotine dependence (FTND) was used to determine nicotine dependence of the participants, and a total of 381 participants were classified into 3 groups of nicotine dependence: low (n=205, 53.8%), moderate (n=127, 33.3%), and high dependence groups (n=49, 12.9%). Stepwise multiple linear regression model and receiver operating characteristic (ROC) curves analyses were used to determine the validity of urinary cotinine for high nicotine dependence. Results: In correlation analysis, urinary cotinine levels increased with FTND score (r=0.567, P<0.001). ROC curves analysis showed that urinary cotinine levels predicted the high-dependence group with reasonable accuracy (optimal cut-off value=1,000 ng/mL; AUC=0.82; P<0.001; sensitivity=71.4%; specificity=74.4%). In stepwise multiple regression analysis, the total smoking period (${\beta}$=0.042, P=0.001) and urinary cotinine levels (${\beta}$=0.234, P<0.001) were positively associated with nicotine dependence, whereas an inverse association was observed between highest education levels (>16 years) and nicotine dependence (${\beta}$=-0.573, P=0.034). Conclusions: The results of this study support the validity of using urinary cotinine levels for assessment of nicotine dependence in active smokers.

• The Journal of Internal Korean Medicine
• /
• v.35 no.4
• /
• pp.483-497
• /
• 2014

Objectives: The aims of study were developing cut-off value of Yin-deficiency questionnaire (YDQ) for diagnosis of Yin-deficiency (YD) and compare diagnostic ability between YDQ and Yin-deficiency scale score (YDS) in xerostomia patients. Methods: We recruited 58 xerostomia patients. They were diagnosed YD or non-YD by 3 Korean medicine doctors (KMD). We assessed YD using YDQ and YDS. We evaluated xerostomia using VAS, Dry Mouth Symptom Questionnaire (DMSQ), Salivary Flow Rate (SFR), oral moisture on buccal mucosa and tongue surface (OMB and OMT). We surveyed tongue coatings using Winkel Tongue Coating Index (WTCI). Results: We diagnosed 23 patients YD and 35 patients non-YD. There were no significant differences of age, sex and body mass index between the YD and non-YD groups. Using receiver operating characteristic curve analysis, the optimal cut-off value of YDQ was defined as 304. Sensitivity, specificity and Youden index of YDQ were 86.96%, 71.43% and 1.5839 respectively. Using Cohen's coefficient of agreement, we found that degree of agreement between KMD and YDQ diagnosis was moderate (${\kappa}$=0.524, p<0.001). Using Pearson's correlation analysis, we found concurrent validity of YDQ and YDS were significant correlated. Using area under curve value, we found diagnostic ability between YDQ and YDS were not significantly different (p=0.505), but there were more strong correlations between DMSQ-symptoms and YDQ (r=0.731, p<0.001) than correlations between DMSQ-symptoms and YDS (r=0.418, p<0.01). Conclusions: The cut-off value of YDQ can diagnose YD in xerostomia and diagnostic ability of YDQ in xerostomia is better than YDS.

• Physical Therapy Rehabilitation Science
• /
• v.7 no.3
• /
• pp.102-108
• /
• 2018

Objective: To investigate whether the Minimal Clinically Important Difference (MCID) clinically defines improvement of Berg Balance Scale (BBS) scores in people with acute stroke in response to rehabilitation. Design: Retrospective study. Methods: Seventy-three participants with acute stroke participated in the study. Balance evaluation was performed using the BBS. All patients received rehabilitation with physical therapy for 4 weeks, 5 times a week, for 2 hours and 20 minutes a day. An anchor-based approach using the clinical global impression was used to determine the MCID of the BBS. The MCID was used to define the minimum change in the BBS total score (postintervention-preintervention) that was needed to perceive at least a 3-point improvement on the global rating of change. Receiver operating characteristic (ROC) curves was used to define the cut-off values of the optimal MCID of the BBS in order to discriminate between improvement and no improvement groups. Results: The optimal MCID cut-off point for the BBS change scores was 12.5 points for males with a sensitivity (Sn) of 0.62 and a specificity (Sp) of 0.89, and 12.5 points for females with a Sn of 0.69 and Sp of 0.85. The area under the curve of the ROC curve for all participants were 0.84 (95% confidence interval [CI], 0.72; 0.95, p<0.001), and 0.89 (95% CI, 0.77; 1.00, p<0.001), respectively. Conclusions: The MCID for improvement in balance as measured by the BBS was 13.5 points, indicating that the MCID does clinically detect changes in balance abilities in persons with stroke.