RESUMEN
BACKGROUND: Although accurate preoperative diagnosis of lymph node metastasis is essential for optimizing treatment strategies for low rectal cancer, the accuracy of present diagnostic modalities has room for improvement. OBJECTIVE: The study aimed to establish a high-precision diagnostic method for lymph node metastasis of low rectal cancer using artificial intelligence. DESIGN: A retrospective observational study. SETTINGS: A single cancer center and a college of engineering in Japan. PATIENTS: Patients with low rectal adenocarcinoma who underwent proctectomy, bilateral lateral pelvic lymph node dissection, and contrast-enhanced multidetector row CT (slice ≤1 mm) between July 2015 and August 2021 were included in the present study. All pelvic lymph nodes from the aortic bifurcation to the upper edge of the anal canal were extracted, regardless of whether within or beyond the total mesenteric excision area, and pathological diagnoses were annotated for training and validation. MAIN OUTCOME MEASURES: Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. RESULTS: A total of 596 pathologically negative nodes and 43 positive nodes from 52 patients were extracted and annotated. Four diagnostic methods, with and without using super-resolution images and with and without using 3-dimensional shape data, were performed and compared. The super-resolution + 3-dimensional shape data method had the best diagnostic ability for the combination of sensitivity, negative predictive value, and accuracy (0.964, 0.966, and 0.968, respectively), whereas the super-resolution only method had the best diagnostic ability for the combination of specificity and positive predictive value (0.994 and 0.993, respectively). LIMITATIONS: Small number of patients at a single center and the lack of external validation. CONCLUSIONS: Our results enlightened the potential of artificial intelligence for the method to become another game changer in the diagnosis and treatment of low rectal cancer. See Video Abstract . DIAGNSTICO POR IMGENES CON INTELIGENCIA ARTIFICIAL MEDIANTE SUPERRESOLUCIN Y FORMA D PARA LA METSTASIS EN LOS GANGLIOS LINFTICOS DEL CNCER DE RECTO BAJO UN ESTUDIO PILOTO DE UN SOLO CENTRO: ANTECEDENTES:Aunque el diagnóstico preoperatorio preciso de metástasis en los ganglios linfáticos es esencial para optimizar las estrategias de tratamiento para el cáncer de recto bajo, la precisión de las modalidades de diagnóstico actuales tiene margen de mejora.OBJETIVO:Establecer un método de diagnóstico de alta precisión para las metástasis en los ganglios linfáticos del cáncer de recto bajo utilizando inteligencia artificial.DISEÑO:Un estudio observacional retrospectivo.AJUSTE:Un único centro oncológico y una facultad de ingeniería en Japón.PACIENTES:En el presente estudio se incluyeron pacientes con adenocarcinoma rectal bajo sometidos a proctectomía, disección bilateral de ganglios linfáticos pélvicos laterales y tomografía computarizada con múltiples detectores con contraste (corte ≤1 mm) entre julio de 2015 y agosto de 2021. Se resecaron todos los ganglios linfáticos pélvicos desde la bifurcación aórtica hasta el borde superior del canal anal, independientemente de si estaban dentro o más allá del área de escisión mesentérica total, y se registraron los diagnósticos patológicos para entrenamiento y validación.PRINCIPALES MEDIDAS DE RESULTADO:Sensibilidad, especificidad, valor predictivo positivo, valor predictivo negativo y precisión.RESULTADOS:Se extrajeron y registraron un total de 596 ganglios patológicamente negativos y 43 positivos de 52 pacientes. Se realizaron y compararon cuatro métodos de diagnóstico, con y sin imágenes de súper resolución y sin datos de imagen en 3D. El método de superresolución + datos de imagen en 3D tuvo la mejor capacidad de diagnóstico para la combinación de sensibilidad, valor predictivo negativo y precisión (0,964, 0,966 y 0,968, respectivamente), mientras que el método de súper resolución solo tuvo la mejor capacidad de diagnóstico para la combinación de especificidad y valor predictivo positivo (0,994 y 0,993, respectivamente).LIMITACIONES:Pequeño número de pacientes en un solo centro y falta de validación externa.CONCLUSIONES:Nuestros resultados iluminan el potencial de la inteligencia artificial para que el método se convierta en otro elemento de cambio en el diagnóstico y tratamiento del cáncer de recto bajo. (Traducción ---Dr. Fidel Ruiz Healy ).
Asunto(s)
Adenocarcinoma , Inteligencia Artificial , Ganglios Linfáticos , Metástasis Linfática , Neoplasias del Recto , Humanos , Neoplasias del Recto/patología , Neoplasias del Recto/diagnóstico por imagen , Metástasis Linfática/diagnóstico por imagen , Metástasis Linfática/patología , Masculino , Femenino , Proyectos Piloto , Estudios Retrospectivos , Persona de Mediana Edad , Anciano , Ganglios Linfáticos/patología , Ganglios Linfáticos/diagnóstico por imagen , Adenocarcinoma/patología , Adenocarcinoma/diagnóstico por imagen , Adenocarcinoma/diagnóstico , Adenocarcinoma/secundario , Proctectomía/métodos , Imagenología Tridimensional/métodos , Escisión del Ganglio Linfático/métodos , Tomografía Computarizada Multidetector/métodos , Valor Predictivo de las Pruebas , Sensibilidad y Especificidad , Pelvis/diagnóstico por imagen , AdultoRESUMEN
BACKGROUND: Although accurate preoperative diagnosis of lymph node metastasis is essential for optimizing treatment strategies for low rectal cancer, the accuracy of present diagnostic modalities has room for improvement. OBJECTIVE: To establish a high-precision diagnostic method for lymph node metastasis of low rectal cancer using artificial intelligence. DESIGN: A retrospective observational study. SETTINGS: A single cancer center and a college of engineering in Japan. PATIENTS: Patients with low rectal adenocarcinoma who underwent proctectomy, bilateral lateral pelvic lymph node dissection, and contrast-enhanced multi-detector row computed tomography (slice ≤1 mm) between July 2015 and August 2021 were included in the present study. All pelvic lymph nodes from the aortic bifurcation to the upper edge of the anal canal were extracted, regardless of whether within or beyond the total mesenteric excision area, and pathological diagnoses were annotated for training and validation. MAIN OUTCOME MEASURES: Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. RESULTS: A total of 596 pathologically negative and 43 positive nodes from 52 patients were extracted and annotated. Four diagnostic methods, with and without using super-resolution images and without using 3D shape data, were performed and compared. The super-resolution + 3D shape data method had the best diagnostic ability for the combination of sensitivity, negative predictive value, and accuracy (0.964, 0.966, and 0.968, respectively), while the super-resolution only method had the best diagnostic ability for the combination of specificity and positive predictive value (0.994 and 0.993, respectively). LIMITATIONS: Small number of patients at a single center and the lack of external validation. CONCLUSIONS: Our results enlightened the potential of artificial intelligence for the method to become another game changer in the diagnosis and treatment of low rectal cancer. See Video Abstract.
RESUMEN
PURPOSE: The purpose of this study was to evaluate the influence of tonometry in detecting the occurrence of glaucoma. METHODS: The subjects, 845 out of 3,488 residents aged 40 years or older, were examined according to standard protocols, including tonometry, slit lamp examination, fundus photography, and automatic perimetry as a recall examination. The intraocular pressure in each subject was measured by both Goldmann applanation tonometer(GAT) and noncontact tonometer CT-70 (NCT). RESULTS: The mean +/- standard deviation intraocular pressure measured by GAT was 15.52 +/- 2.57 mmHg, and 15.03 +/- 2.90 mmHg by NCT. There was a statistically significant correlation(p < 0.0001). The difference between pairs of measurements by GAT and NCT was 0.50 +/- 1.93 mmHg. There was no influence of tonometry in detecting the incidence of glaucoma, which was 4.14%; primary open-angle glaucoma 0.59%, normal tension glaucoma 2.6%, primary angle-closure glaucoma 0.47%, and other types of glaucoma 0.48%. The detection of ocular hypertension was different, and was 2.13% with GAT and 2.72% with NCT. CONCLUSION: In our study, the influence of tonometry in detecting the incidence of glaucoma was very small. A noncontact tonometer is considered to be useful for glaucoma population study.