Limits...
Reproducibility of aortic intima-media thickness in infants using edge-detection software and manual caliper measurements.

McCloskey K, Ponsonby AL, Carlin JB, Jachno K, Cheung M, Skilton MR, Koleff J, Vuillermin P, Burgner D, Barwon Infant Study investigator gro - Cardiovasc Ultrasound (2014)

Bottom Line: Aortic IMT was measured successfully using edge-detection in 814 infants, and 290 of these infants also had aIMT measured using manual sonographic calipers.Edge-detection measurements were greater than those from manual sonographic calipers (mean aIMT 618 μm (50) versus mean aIMT 563 μm (49) respectively; p < 0.001, mean difference 44 μm, 95% LOA -54, 142).Image quality is not substantially affected by individual and environmental factors.

View Article: PubMed Central - HTML - PubMed

Affiliation: Murdoch Childrens Research Institute, Royal Children's Hospital, Parkville, Australia. peter.vuillermin@deakin.edu.au.

ABSTRACT

Background: Aortic intima-media thickness measured by transabdominal ultrasound (aIMT) is an intermediate phenotype of cardiovascular risk. We aimed to (1) investigate the reproducibility of aIMT in a population-derived cohort of infants; (2) establish the distribution of aIMT in early infancy; (3) compare measurement by edge-detection software to that by manual sonographic calipers; and (4) assess the effect of individual and environmental variables on image quality.

Methods: Participants were term infants recruited to a population-derived birth cohort study. Transabdominal ultrasound was performed at six weeks of age by one of two trained operators. Thirty participants had ultrasounds performed by both operators on the same day. Data were collected on environmental (infant sleeping, presence of a sibling, use of sucrose, timing during study visit) and individual (post-conception age, weight, gender) variables. Two readers assessed image quality and measured aIMT by edge-detection software and a subset by manual sonographic calipers. Measurements were repeated by the same reader and between readers to obtain intra-observer and inter-observer reliability.

Results: Aortic IMT was measured successfully using edge-detection in 814 infants, and 290 of these infants also had aIMT measured using manual sonographic calipers. The intra-reader intra-class correlation (ICC) (n = 20) was 0.90 (95% CI 0.76, 0.96), mean difference 1.5 μm (95% LOA -39, 59). The between reader ICC using edge-detection (n = 20) was 0.92 (95% CI 0.82, 0.97) mean difference 2 μm (95% LOA -45.0, 49.0) and with manual caliper measurement (n = 290) the ICC was 0.84 (95% CI 0.80, 0.87) mean difference 5 μm (95% LOA -51.8, 61.8). Edge-detection measurements were greater than those from manual sonographic calipers (mean aIMT 618 μm (50) versus mean aIMT 563 μm (49) respectively; p < 0.001, mean difference 44 μm, 95% LOA -54, 142). With the exception of infant crying (p = 0.001), no associations were observed between individual and environmental variables and image quality.

Conclusion: In a population-derived cohort of term infants, aIMT measurement has a high level of intra and inter-reader reproducibility. Measurement of aIMT using edge-detection software gives higher inter-reader ICC than manual sonographic calipers. Image quality is not substantially affected by individual and environmental factors.

Show MeSH
aIMT measured using edge-detection software, adapted from [19].
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4061507&req=5

Figure 1: aIMT measured using edge-detection software, adapted from [19].

Mentions: Thirty infants had aortic ultrasounds performed by both ultrasound operators on the same day to allow for measurement of reproducibility between operators.Using edge-detection software (Carotid Analyzer for Research, version 6, Medical Imaging Applications LLC, Iowa), aIMT was measured offline on all infant ultrasounds by one of two readers (who were not ultrasound operators). The best three to five end-diastolic images per infant (timed on the R wave) were analysed. The measurements relied on the edge-detection software identifying the aIMT ‘edge’ along a straight segment of posterior wall, of minimum 4 mm length (Figure 1). The edge-detection software generated a minimum, mean and maximum aIMT for the segment selected. The average of each of these values for an infant’s end-diastolic images was taken as the measure for each infant. In order to calculate the intra-reader (test retest) and inter-reader repeatability of measurements conducted using the edge-detection software, one reader repeated measures in a randomly selected 20 subjects and both readers conducted measures among a separate randomly selected sample of 20 subjects.


Reproducibility of aortic intima-media thickness in infants using edge-detection software and manual caliper measurements.

McCloskey K, Ponsonby AL, Carlin JB, Jachno K, Cheung M, Skilton MR, Koleff J, Vuillermin P, Burgner D, Barwon Infant Study investigator gro - Cardiovasc Ultrasound (2014)

aIMT measured using edge-detection software, adapted from [19].
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4061507&req=5

Figure 1: aIMT measured using edge-detection software, adapted from [19].
Mentions: Thirty infants had aortic ultrasounds performed by both ultrasound operators on the same day to allow for measurement of reproducibility between operators.Using edge-detection software (Carotid Analyzer for Research, version 6, Medical Imaging Applications LLC, Iowa), aIMT was measured offline on all infant ultrasounds by one of two readers (who were not ultrasound operators). The best three to five end-diastolic images per infant (timed on the R wave) were analysed. The measurements relied on the edge-detection software identifying the aIMT ‘edge’ along a straight segment of posterior wall, of minimum 4 mm length (Figure 1). The edge-detection software generated a minimum, mean and maximum aIMT for the segment selected. The average of each of these values for an infant’s end-diastolic images was taken as the measure for each infant. In order to calculate the intra-reader (test retest) and inter-reader repeatability of measurements conducted using the edge-detection software, one reader repeated measures in a randomly selected 20 subjects and both readers conducted measures among a separate randomly selected sample of 20 subjects.

Bottom Line: Aortic IMT was measured successfully using edge-detection in 814 infants, and 290 of these infants also had aIMT measured using manual sonographic calipers.Edge-detection measurements were greater than those from manual sonographic calipers (mean aIMT 618 μm (50) versus mean aIMT 563 μm (49) respectively; p < 0.001, mean difference 44 μm, 95% LOA -54, 142).Image quality is not substantially affected by individual and environmental factors.

View Article: PubMed Central - HTML - PubMed

Affiliation: Murdoch Childrens Research Institute, Royal Children's Hospital, Parkville, Australia. peter.vuillermin@deakin.edu.au.

ABSTRACT

Background: Aortic intima-media thickness measured by transabdominal ultrasound (aIMT) is an intermediate phenotype of cardiovascular risk. We aimed to (1) investigate the reproducibility of aIMT in a population-derived cohort of infants; (2) establish the distribution of aIMT in early infancy; (3) compare measurement by edge-detection software to that by manual sonographic calipers; and (4) assess the effect of individual and environmental variables on image quality.

Methods: Participants were term infants recruited to a population-derived birth cohort study. Transabdominal ultrasound was performed at six weeks of age by one of two trained operators. Thirty participants had ultrasounds performed by both operators on the same day. Data were collected on environmental (infant sleeping, presence of a sibling, use of sucrose, timing during study visit) and individual (post-conception age, weight, gender) variables. Two readers assessed image quality and measured aIMT by edge-detection software and a subset by manual sonographic calipers. Measurements were repeated by the same reader and between readers to obtain intra-observer and inter-observer reliability.

Results: Aortic IMT was measured successfully using edge-detection in 814 infants, and 290 of these infants also had aIMT measured using manual sonographic calipers. The intra-reader intra-class correlation (ICC) (n = 20) was 0.90 (95% CI 0.76, 0.96), mean difference 1.5 μm (95% LOA -39, 59). The between reader ICC using edge-detection (n = 20) was 0.92 (95% CI 0.82, 0.97) mean difference 2 μm (95% LOA -45.0, 49.0) and with manual caliper measurement (n = 290) the ICC was 0.84 (95% CI 0.80, 0.87) mean difference 5 μm (95% LOA -51.8, 61.8). Edge-detection measurements were greater than those from manual sonographic calipers (mean aIMT 618 μm (50) versus mean aIMT 563 μm (49) respectively; p < 0.001, mean difference 44 μm, 95% LOA -54, 142). With the exception of infant crying (p = 0.001), no associations were observed between individual and environmental variables and image quality.

Conclusion: In a population-derived cohort of term infants, aIMT measurement has a high level of intra and inter-reader reproducibility. Measurement of aIMT using edge-detection software gives higher inter-reader ICC than manual sonographic calipers. Image quality is not substantially affected by individual and environmental factors.

Show MeSH