Clinical workflow of sonographers performing fetal anomaly ultrasound scans: deep‐learning‐based analysis

Abstract

What are the novel findings of this work?Using sonography big data and deep learning to describe video content automatically, we studied obstetric ultrasound as a data-science problem. We showed that an anomaly scan can be modeled as a non-ordered sequence of anatomical structure acquisitions. What are the clinical implications of this work?The lack of a universal scanning sequence supports the hypothesis that anomaly scanning is opportunistic by nature, continuously taking advantage of fetal position. Trainees may benefit from appreciating that even experts do not follow guidelines as an ordered list, but rather acquire planes and images according to visibility.

Computational methods for the analysis of early-pregnancy brain ultrasonography: a systematic review

“…made their dataset publicly available, which consists of data of 1792 patients, and contains besides the TC, TT and TV plane, the abdominal, femur and thorax standard planes. 23 The other studies detected the brain, 35 , 65 , 66 , 67 or other standard planes. 28 , 37 Only the study by Bastiaansen et al.…”

“…There were only four (7%) studies using longitudinal data 29 , 48 , 49 , 74 and only two of the studies gave an unreproducible description of their method. 67 , 76 Furthermore, 27 (49%) studies had, besides qualitative and/or quantitative reporting of outcome, additionally multiple raters or compared their result to known clinical outcomes. Regarding analysing the influence of confounders, 19 (35%) studies did not adjust or analyse the influence of at least one of the key confounders (GA, acquisition quality or body mass index) and only four (7%) studies performed an analysis to identify or adjust for additional confounders such as, challenging fetal position, abdominal scarring and uterine fibroid, 67 fetal position, maternal body habitus and prior uterine surgery, 27 maternal age, pregnancy duration, birthweight, number of ultrasound examinations, 49 maternal age and fetal position.…”

Work-Related Musculoskeletal Disorders and Associated Work Systems Factors: Are There Differences Between Sonography Practice Areas?

“…Efforts in sonography workflow have primarily focused on improving exam efficiency through equipment features 19 and computing technologies to improve exam-specific workflow. 20,21 Limited evidence suggests that workflow management systems that consider additional facilitators and barriers beyond exam-specific factors can reduce stress among sonographers. 22 Given the findings of the current analysis that align with other previous studies describing work process challenges, 23 identifying solutions to improve both exam and non-exam workflow is vital to support sonographers’ health and well-being.…”

Sonographer interaction with artificial intelligence: collaboration or conflict?

“…Unlike many other medical‐imaging tasks, the real‐time nature of ultrasound, with simultaneous image acquisition and analysis, means that complete automation of these processes is unlikely to be achieved in the short or medium term. To further complicate matters, there is a wide variation in the number and sequence of fetal structures imaged during a scan, with sonographers taking advantage of fetal position and other factors to optimize scan images 23 . This lack of consistency and huge variation in image appearance, even between images of the same subject taken by different operators, makes the application of AI more challenging.…”