Towards automated extraction of 2D standard fetal head planes from 3D ultrasound acquisitions: A clinical evaluation and quality assessment comparison

E Skelton,J Matthew,Y Li,B Khanal,J.J Cerrolaza Martinez,N Toussaint,C Gupta,C Knight,B Kainz,J.V Hajnal,M Rutherford

doi:10.1016/j.radi.2020.11.006

E Skelton, J Matthew + Show 9 more

Open Access

https://doi.org/10.1016/j.radi.2020.11.006

Copy DOI

Abstract

IntroductionClinical evaluation of deep learning (DL) tools is essential to compliment technical accuracy metrics. This study assessed the image quality of standard fetal head planes automatically-extracted from three-dimensional (3D) ultrasound fetal head volumes using a customised DL-algorithm. MethodsTwo observers retrospectively reviewed standard fetal head planes against pre-defined image quality criteria. Forty-eight images (29 transventricular, 19 transcerebellar) were selected from 91 transabdominal fetal scans (mean gestational age = 26 completed weeks, range = 20+5–32+3 weeks). Each had two-dimensional (2D) manually-acquired (2D-MA), 3D operator-selected (3D-OS) and 3D-DL automatically-acquired (3D-DL) images. The proportion of adequate images from each plane and modality, and the number of inadequate images per plane was compared for each method. Inter and intra-observer agreement of overall image quality was calculated. ResultsSixty-seven percent of 3D-OS and 3D-DL transventricular planes were adequate quality. Forty-five percent of 3D-OS and 55% of 3D-DL transcerebellar planes were adequate.Seventy-one percent of 3D-OS and 86% of 3D-DL transventricular planes failed with poor visualisation of intra-cranial structures. Eighty-six percent of 3D-OS and 80% of 3D-DL transcerebellar planes failed due to inadequate visualisation of cerebellar hemispheres. Image quality was significantly different between 2D and 3D, however, no significant difference between 3D-modalities was demonstrated (p < 0.005). Inter-observer agreement of transventricular plane adequacy was moderate for both 3D-modalities, and weak for transcerebellar planes. ConclusionThe 3D-DL algorithm can automatically extract standard fetal head planes from 3D-head volumes of comparable quality to operator-selected planes. Image quality in 3D is inferior to corresponding 2D planes, likely due to limitations with 3D-technology and acquisition technique. Implications for practiceAutomated image extraction of standard planes from US-volumes could facilitate use of 3DUS in clinical practice, however image quality is dependent on the volume acquisition technique.

Highlights

Clinical evaluation of deep learning (DL) tools is essential to compliment technical accuracy metrics
These were selected as the goldstandard for comparative analysis against their corresponding image from 3D operator-selected (3D-OS) and 3D-DL
McNemar's tests found significant differences in the overall adequacy of the 2D manually-acquired (2D-MA) planes compared to 3D as rated by both observers, no significant difference between 3D-images was demonstrated (p < 0.005) (Table 5). This evaluation suggests that a 3D-DL algorithm can automatically extract standard planes from fetal head volumes of comparable quality to 3D-image planes selected by an operator from the same volume

Summary

Introduction

Clinical evaluation of deep learning (DL) tools is essential to compliment technical accuracy metrics. The fetal head and brain is examined during 18þ0-20þ6 fetal anomaly ultrasound (US) examinations to assess growth and development of the skull and intracranial structures.[1] Transventricular (TV) and transcerebellar (TC) views are routinely assessed in the basic screening examination These two-dimensional (2D) planes allow identification of intracranial landmarks and acquisition of specific biometric measurements which, if absent or outside expected reference ranges, may be indicative of an anomaly.[2]. Advances in three-dimensional US (3D-US) and multi-planar reconstruction can complement conventional 2D-US and overcome some of its limitations.[4] The 3D-transabdominal “single-shot” technique can be used to acquire a reproducible fetal head USvolume,[5,6] with reconstructed images enabling detailed retrospective review by clinicians when multiple datasets are acquired from different insonation angles (e.g. transverse, sagittal, coronal).[7] Biometric measurements from 3D-images have demonstrated good correlation with 2D-methods.[3]

Objectives

Methods

Results

Discussion

Conclusion