RESUMEN
Though consistently shown to detect mammographically occult cancers, breast ultrasound has been noted to have high false-positive rates. In this work, we present an AI system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images. Developed on 288,767 exams, consisting of 5,442,907 B-mode and Color Doppler images, the AI achieves an area under the receiver operating characteristic curve (AUROC) of 0.976 on a test set consisting of 44,755 exams. In a retrospective reader study, the AI achieves a higher AUROC than the average of ten board-certified breast radiologists (AUROC: 0.962 AI, 0.924 ± 0.02 radiologists). With the help of the AI, radiologists decrease their false positive rates by 37.3% and reduce requested biopsies by 27.8%, while maintaining the same level of sensitivity. This highlights the potential of AI in improving the accuracy, consistency, and efficiency of breast ultrasound diagnosis.
Asunto(s)
Algoritmos , Inteligencia Artificial , Neoplasias de la Mama/diagnóstico por imagen , Mama/diagnóstico por imagen , Detección Precoz del Cáncer , Ultrasonografía/métodos , Adulto , Anciano , Neoplasias de la Mama/diagnóstico , Femenino , Humanos , Mamografía/métodos , Persona de Mediana Edad , Curva ROC , Radiólogos/estadística & datos numéricos , Reproducibilidad de los Resultados , Estudios RetrospectivosRESUMEN
OBJECTIVE: Maternal obesity is associated with poor breastfeeding outcomes, yet no intervention has been developed to improve them. To ascertain whether increased breastfeeding support or provision of a breast pump is a feasible, effective intervention to improve breastfeeding, we enrolled obese women who intended to breastfeed in two randomized trials. METHODS: In Bassett Improving Breastfeeding Study (BIBS) 1, 40 women received targeted breastfeeding support in the hospital and via telephone or usual care. Information regarding breastfeeding was collected via telephone for 7 days after delivery and at 30 and 90 days postpartum. In BIBS 2, 34 obese mothers received a manual or electric breast pump to use for 10-14 days or no pump; data collection was similar. RESULTS: In both experiments, randomization failed to distribute women of differing postpartum body mass index adequately among the treatment groups. When analyses were adjusted for this, there was no difference in BIBS 1 between targeted and usual care groups and in BIBS 2 among the treatment groups in the proportion of women still breastfeeding at the times studied. CONCLUSIONS: In future studies of obese women, stratified randomization may be necessary. Further development of interventions to help obese women achieve optimal breastfeeding outcomes is required.