RESUMO
BACKGROUND: Assessing a patient's suicide risk is challenging for health professionals because it depends on voluntary disclosure by the patient and often has limited resources. The application of novel machine learning approaches to determine suicide risk has clinical utility. OBJECTIVE: This study aimed to investigate cross-sectional and longitudinal approaches to assess suicidality based on acoustic voice features of psychiatric patients using artificial intelligence. METHODS: We collected 348 voice recordings during clinical interviews of 104 patients diagnosed with mood disorders at baseline and 2, 4, 8, and 12 months after recruitment. Suicidality was assessed using the Beck Scale for Suicidal Ideation and suicidal behavior using the Columbia Suicide Severity Rating Scale. The acoustic features of the voice, including temporal, formal, and spectral features, were extracted from the recordings. A between-person classification model that examines the vocal characteristics of individuals cross sectionally to detect individuals at high risk for suicide and a within-person classification model that detects considerable worsening of suicidality based on changes in acoustic features within an individual were developed and compared. Internal validation was performed using 10-fold cross validation of audio data from baseline to 2-month and external validation was performed using data from 2 to 4 months. RESULTS: A combined set of 12 acoustic features and 3 demographic variables (age, sex, and past suicide attempts) were included in the single-layer artificial neural network for the between-person classification model. Furthermore, 13 acoustic features were included in the extreme gradient boosting machine learning algorithm for the within-person model. The between-person classifier was able to detect high suicidality with 69% accuracy (sensitivity 74%, specificity 62%, area under the receiver operating characteristic curve 0.62), whereas the within-person model was able to predict worsening suicidality over 2 months with 79% accuracy (sensitivity 68%, specificity 84%, area under receiver operating characteristic curve 0.67). The second model showed 62% accuracy in predicting increased suicidality in external sets. CONCLUSIONS: Within-person analysis using changes in acoustic features within an individual is a promising approach to detect increased suicidality. Automated analysis of voice can be used to support the real-time assessment of suicide risk in primary care or telemedicine.
Assuntos
Ideação Suicida , Suicídio , Humanos , Tentativa de Suicídio/psicologia , Fatores de Risco , Fala , Inteligência Artificial , Estudos Transversais , Aprendizado de MáquinaRESUMO
This study investigated whether the large language model (LLM) utilizes sufficient domain knowledge to reason about critical medical events such as extubation. In detail, we tested whether the LLM accurately comprehends given tabular data and variable importance and whether it can be used in complement to existing ML models such as XGBoost.
Assuntos
Extubação , Humanos , Processamento de Linguagem Natural , Sistemas de Apoio a Decisões ClínicasRESUMO
BACKGROUND: The aim of this study is to evaluate local control rates after stereotactic body radiotherapy (SBRT) in recurrent spinal metastasis after external beam radiotherapy (EBRT) and new spinal metastatic lesions. MATERIAL AND METHODS: Retrospective review of medical records and radiological data was performed on 54 retreatment and 131 initial SBRT patients. To compare various fractionation schedules, the biologically effective dose (BED) was applied. SBRT dose was calculated with linear-quadratic model and normalized to a 2-Gy equivalent dose (nBED, α/ß =2 Gy for spinal cord, α/ß =10 Gy for tumor). Doses to a point within the spinal cord that received the maximum dose (Pmax) were checked. Local control failure was defined as progression by imaging study. Overall survival, progression free survival, delivered radiation dose to tumor and spinal cord, and spinal cord Pmax nBED were compared in two groups. RESULTS: The mean delivered radiation doses to tumor margin during SBRT were 51.1 Gy2/10 (retreatment) and 50.7 Gy2/10 (initial treatment). Mean survival was 29.6 months (overall)/20.7 months (retreatment)/ 32.4 months (initial treatment). Mean progression free period was 23.9 months (overall)/18.0 months (retreatment)/ 26.0 months (initial treatment). Radiological control rates of retreatment and initial treatment group were 96%/95% at six months, 81%/89% at 12 months and 79%/90% at 24 months. Among 54 retreatment lesions, 13 lesions showed local control failure during follow-up. With regard to spinal cord radiation dose during SBRT, Spinal cord Pmax nBED was 46.2 Gy2/2 (retreatment) and 48.7 Gy2/2 (initial treatment). In retreatment group, total nBED to spinal cord was a mean of 83.4 Gy2/2. There was no case of radiation myelopathy detected. CONCLUSIONS: Retreatment of spinal metastases using SBRT provided effective local control without neurological complications.
Assuntos
Neoplasias da Mama/cirurgia , Neoplasias da Próstata/cirurgia , Radiocirurgia , Neoplasias da Coluna Vertebral/secundário , Neoplasias da Coluna Vertebral/cirurgia , Neoplasias da Mama/mortalidade , Neoplasias da Mama/patologia , Feminino , Seguimentos , Humanos , Masculino , Prontuários Médicos , Pessoa de Meia-Idade , Prognóstico , Neoplasias da Próstata/mortalidade , Neoplasias da Próstata/patologia , Lesões por Radiação , Tolerância a Radiação , Eficiência Biológica Relativa , Retratamento , Estudos Retrospectivos , Neoplasias da Coluna Vertebral/mortalidade , Taxa de SobrevidaRESUMO
Background: Depression and suicide are critical social problems worldwide, but tools to objectively diagnose them are lacking. Therefore, this study aimed to diagnose depression through machine learning and determine whether it is possible to identify groups at high risk of suicide through words spoken by the participants in a semi-structured interview. Methods: A total of 83 healthy and 83 depressed patients were recruited. All participants were recorded during the Mini-International Neuropsychiatric Interview. Through the suicide risk assessment from the interview items, participants with depression were classified into high-suicide-risk (31 participants) and low-suicide-risk (52 participants) groups. The recording was transcribed into text after only the words uttered by the participant were extracted. In addition, all participants were evaluated for depression, anxiety, suicidal ideation, and impulsivity. The chi-square test and student's T-test were used to compare clinical variables, and the Naive Bayes classifier was used for the machine learning text model. Results: A total of 21,376 words were extracted from all participants and the model for diagnosing patients with depression based on this text confirmed an area under the curve (AUC) of 0.905, a sensitivity of 0.699, and a specificity of 0.964. In the model that distinguished the two groups using statistically significant demographic variables, the AUC was only 0.761. The DeLong test result (p-value 0.001) confirmed that the text-based classification was superior to the demographic model. When predicting the high-suicide-risk group, the demographics-based AUC was 0.499, while the text-based one was 0.632. However, the AUC of the ensemble model incorporating demographic variables was 0.800. Conclusion: The possibility of diagnosing depression using interview text was confirmed; regarding suicide risk, the diagnosis accuracy increased when demographic variables were incorporated. Therefore, participants' words during an interview show significant potential as an objective and diagnostic marker through machine learning.
RESUMO
Both minor and major depression have high prevalence and are important causes of social burden worldwide; however, there is still no objective indicator to detect minor depression. This study aimed to examine if voice could be used as a biomarker to detect minor and major depression. Ninety-three subjects were classified into three groups: the not depressed group (n = 33), the minor depressive episode group (n = 26), and the major depressive episode group (n = 34), based on current depressive status as a dimension. Twenty-one voice features were extracted from semi-structured interview recordings. A three-group comparison was performed through analysis of variance. Seven voice indicators showed differences between the three groups, even after adjusting for age, BMI, and drugs taken for non-psychiatric disorders. Among the machine learning methods, the best performance was obtained using the multi-layer processing method, and an AUC of 65.9%, sensitivity of 65.6%, and specificity of 66.2% were shown. This study further revealed voice differences in depressive episodes and confirmed that not depressed groups and participants with minor and major depression could be accurately distinguished through machine learning. Although this study is limited by a small sample size, it is the first study on voice change in minor depression and suggests the possibility of detecting minor depression through voice.
RESUMO
OBJECT: Differentiation between malignant and benign vertebral compression fractures (VCFs) is important but sometimes difficult, especially in elderly cancer patients. The authors investigated the findings of MR imaging and FDG-PET/CT for the differentiation of VCFs. METHODS: Between 2007 and 2008, the authors evaluated and treated 102 VCFs in 96 patients. The final diagnosis, based on biopsy results or clinical follow-up, was benign fracture in 67 lesions in 65 patients and malignant fracture in 35 lesions in 31 patients. Magnetic resonance images were obtained in all patients, and FDG-PET/CT was performed in 17 patients in the benign fracture group and 20 in the malignant fracture group. The prevalence of 3 significant MR imaging findings (posterior cortical bulging, epidural mass formation, and pedicle enhancement) and the presence of radiotracer uptake on FDG-PET/CT were evaluated in the 2 groups. The maximum standardized uptake value (SUV(max)) on FDG-PET/CT was compared between the 2 groups, and diagnostic threshold value was sought to confirm malignancy. The diagnostic accuracy of MR imaging and FDG-PET/CT was compared in the differentiation of malignant from benign VCFs. RESULTS: Posterior cortical bulging was seen in 26 (74%) of 35 malignant lesions and 30 (45%) of 67 benign ones, epidural mass formation in 27 (77%) of the malignant lesions and 25% of the benign ones, and pedicle enhancement in 30 (91%) of the 33 malignant lesions and 18 (39%) of the 46 benign ones evaluated with Gd-enhanced MR imaging. These differences were statistically significant for each feature. Sensitivity and specificity for predicting malignancy were, respectively, 74% and 55% for posterior cortical bulging, 77% and 74% for epidural mass formation, and 90% and 61% for pedicle enhancement. Simultaneous occurrence of 3 significant features was found in 21 (64%) of the 33 malignant and 8 (17%) of the 46 benign lesions for which complete MR imaging data were available and showed sensitivity of 64% and specificity of 83%. The presence of radiotracer uptake on FDG-PET/CT was seen in all 20 (100%) of the 20 malignant lesions and 12 (71%) 17 of the benign lesions evaluated by FDG-PET/CT and showed a sensitivity of 100% and specificity of 29%. There was a significant difference in mean (± SD) SUV(max) for the malignant (6.29 ± 3.50) and benign (2.38 ± 1.90) lesions (p < 0.001). The most reliable threshold for SUV(max) was found to be 4.25, which yielded a sensitivity of 85% and a specificity of 71%. CONCLUSIONS: When MR imaging findings are equivocal, FDG-PET/CT can be considered as an adjunctive diagnostic method for differentiating malignant from benign VCFs. In comparison with MR imaging, FDG-PET/CT showed slightly higher sensitivity and lower specificity.