Advanced Search
Now showing items 21-30 of 7283
"Plataforma Láser Chrome"
(Universidad de Chile, 2023)
La medicina estética ha sido una de las áreas más innovadoras en el campo de la salud, y
una de las más demandadas por los pacientes. Según un estudio realizado por “Mordor
Intelligence” el mercado mundial de láseres estéticos está creciendo a una...
Sistema de seguimiento de convenios clínicos de la Facultad de Medicina de la Universidad de Chile
(Universidad de Chile, 2021)
La Facultad de Medicina de la Universidad de Chile, a través de la Dirección Clínica (DC),
mantiene acuerdos de colaboración (convenios) con distintos centros de salud (campos clínicos) a
lo largo del país. Hay 21 hospitales que prestan servicios a...
La otra Chimba: historias del Barrio de San Isidro: Santiago de Chile, 1675-1700
(Universidad de Chile, 2009)
tiempo y gastado mis codos en los bares del sector. Sin embargo, a este informe lo mandan los archivos. Y aterrizándolo un poco más, los papeles de la Biblioteca Patrimonial de la Facultad de Medicina de nuestra casa de estudios. Esta escuela de Medicina...
Historia de los terrenos del Hospital Clínico y la Facultad de Medicina de la Universidad de Chile
(Sociedad Médica de Santiago, 2015)
“Disminuir las inasistencias a tratamientos de rehabilitación ambulatoria en la Unidad de Medicina Física y Rehabilitación del Complejo Hospitalario San José”
(Universidad de Chile, 2022)
consecuencia, mejorar su calidad
de vida". Esto se logrará a través del propósito “bajo número de inasistencias
de los usuarios a los tratamientos de rehabilitación ambulatoria”.
El lugar de intervención es la Unidad de Medicina Física y Rehabilitación del
CDT...
Detección automática de metástasis a distancia descrita en reportes de imagenología mediante el uso de procesamiento de lenguaje natural
(Universidad de Chile, 2022)
Antecedentes: La mortalidad por cáncer se produce principalmente por la progresión
del tumor a la etapa de metástasis a distancia. Uno de los criterios para determinar
progresión a esta etapa es un examen de imagenología o medicina nuclear. Realizar...
Background: Cancer mortality is mainly caused by progression of the tumor to the distant metastatic stage. One of the criteria to determine progression to this stage is an imaging or nuclear medicine examination. Curative treatment often depends on the absence of distant metastases. Having this information allows the management and prioritization of patients on the waiting list for interventions or treatments. Problem: Distant metastasis is not standardized in the electronic health record of the Oncology Institute of the Arturo López Pérez Foundation, but is in free text format. This makes it difficult to access for clinical management, and manually detecting these findings consumes staff hours. On the other hand, the screening for distant metastases is performed with the analysis of each radiology or nuclear medicine report in free text, from where the distant metastasis condition cannot be extracted automatically either. Solution: In this thesis we propose the development of a natural language processing model capable of detecting distant metastasis in radiology and nuclear medicine reports and classifying them according to the presence or absence of it. Methods: A named entity recognition model based on a recurrent neural network was developed based on an annotated corpus with affirmative, denied or uncertain metastasis mentions. This model was capable of automatically extracting distant metastasis findings and classifying each report at document level. The performance, measured in precision, completeness and F1-score, of this model was compared with a rule-based algorithm, which was used as a baseline. Results: It is possible to detect distant metastases of prostate cancer, breast cancer and colorectal cancer in imaging and nuclear medicine reports using natural language processing methods. We were able to detect distant metastasis entities within the clinical text with a balanced mean performance of 0,856 measured in F1-score. In addition, documents were classified using deep learning with maximum F1-score performances of 0.90 for documents without distant metastases (M0) and 0.87 for documents with distant metastases (M1)....
Background: Cancer mortality is mainly caused by progression of the tumor to the distant metastatic stage. One of the criteria to determine progression to this stage is an imaging or nuclear medicine examination. Curative treatment often depends on the absence of distant metastases. Having this information allows the management and prioritization of patients on the waiting list for interventions or treatments. Problem: Distant metastasis is not standardized in the electronic health record of the Oncology Institute of the Arturo López Pérez Foundation, but is in free text format. This makes it difficult to access for clinical management, and manually detecting these findings consumes staff hours. On the other hand, the screening for distant metastases is performed with the analysis of each radiology or nuclear medicine report in free text, from where the distant metastasis condition cannot be extracted automatically either. Solution: In this thesis we propose the development of a natural language processing model capable of detecting distant metastasis in radiology and nuclear medicine reports and classifying them according to the presence or absence of it. Methods: A named entity recognition model based on a recurrent neural network was developed based on an annotated corpus with affirmative, denied or uncertain metastasis mentions. This model was capable of automatically extracting distant metastasis findings and classifying each report at document level. The performance, measured in precision, completeness and F1-score, of this model was compared with a rule-based algorithm, which was used as a baseline. Results: It is possible to detect distant metastases of prostate cancer, breast cancer and colorectal cancer in imaging and nuclear medicine reports using natural language processing methods. We were able to detect distant metastasis entities within the clinical text with a balanced mean performance of 0,856 measured in F1-score. In addition, documents were classified using deep learning with maximum F1-score performances of 0.90 for documents without distant metastases (M0) and 0.87 for documents with distant metastases (M1)....
Historia de la enseñanza de la microbiología en Chile: centros formadores
(Sociedad Chilena de Infectología, 2015)
Maestros de la Medicina Interna Chilena: Dr. Manuel García de los Ríos Álvarez
(SOC MEDICA SANTIAGO, 2007-11)
Dispositivo de apoyo corporal para atención en medicinas complementarias
(Universidad de Chile, 2015)