The NEST study is a prospective, observational research study designed to collect clinical measurements and image data to develop and evaluate artificial intelligence (AI)-derived algorithms for estimating anthropometric parameters in neonates and young infants. The study focuses on infants from birth up to 6 months of age and aims to assess the accuracy of AI-based estimations of length, weight, and head circumference using photographs and/or video recordings captured during routine clinical care. These AI-derived measurements will be compared against standard clinical measurements obtained by trained healthcare professionals in neonatal and infant care settings.
Infants up from birth up to 6 months of postnatal age (including neonates) who have been admitted to the NICU or SCN at the time of screening
Parent(s) should be able to comprehend the content of the study and be willing for their child to undergo video and photo recording, and to allow access to their blood sampling results (haemoglobin) taken as part of standard clinical practice
Written consent from parents and/or legally acceptable representative
Infants who were born with gestational age of less than 28 weeks of gestational age
Infants who are intubated (i.e., endotracheal, nasotracheal intubation) at the time of screening
The investigator considers for any reason that the participant would not be suitable for the study
The participant has an existing medical condition that would prevent standardised measurement of length and/or head circumference e.g. structural abnormality of the lower limbs, orthopaedic conditions, hydrocephalus
The study population comprises neonates and infants from birth up to 6 months of age who are admitted to a neonatal intensive care unit (NICU) or special care nursery (SCN). Eligible participants include very preterm, moderate to late preterm, and term infants who meet the study's inclusion criteria and whose parents or legally acceptable representatives provide written informed consent. Participants are enrolled during their NICU/SCN stay and followed prospectively for the duration of their hospitalization, up to a maximum of 10 weeks. The study collects anthropometric measurements and image and video data in conjunction with routine clinical care.
Cohort
Prospective
Not listed
To evaluate the accuracy of the algorithm to estimate length (in cm)
The primary outcome is the proof-of-concept accuracy of an artificial intelligence (AI)-based algorithm for estimating infant length in a neonatal intensive care unit (NICU) or special care nursery (SCN) setting. AI-derived length estimates (in centimeters) obtained from supine images and/or videos are compared with standard clinical length measurements performed by trained investigators using World Health Organization (WHO)-recommended techniques. Accuracy is evaluated using a composite metric that includes bias, mean absolute error, mean absolute percentage error, and the distribution of absolute percentage errors at predefined thresholds.
To evaluate the mean absolute error of the algorithm to estimate weight (in kg)
The secondary outcome is the proof-of-concept accuracy of an artificial intelligence (AI)-based algorithm for estimating infant weight in a neonatal intensive care unit (NICU) or special care nursery (SCN) setting. AI-derived weight estimates (in kilograms) generated from supine images and/or videos are compared with standard clinical weight measurements obtained by trained investigators using World Health Organization (WHO)-recommended techniques. Accuracy is assessed using a composite metric that includes bias, mean absolute error, mean absolute percentage error, and the distribution of absolute percentage errors at predefined thresholds.
To evaluate the mean absolute error of the algorithm to estimate head circumference (in cm)
The exploratory outcome evaluates the proof-of-concept accuracy of an artificial intelligence (AI)-based algorithm for estimating infant head circumference in a neonatal intensive care unit (NICU) or special care nursery (SCN) setting. AI-derived head circumference estimates obtained from images and/or videos are compared with standard clinical head circumference measurements performed by trained investigators using World Health Organization (WHO)-recommended techniques. Accuracy is assessed descriptively using appropriate error metrics to explore the feasibility and potential clinical utility of image-based AI estimation methods.
Ease of image collection by investigator for each participant (Very Easy, Easy, Normal, Difficult, Very difficult]
Investigator's assessment on the ease of collecting the images
Expected level of accuracy by the investigator for length [Less than or equal to 1 cm, less than or equal to 2 cm, less than or equal to 3 cm, less than or equal to 4 cm, less than or equal to 5 cm, more than 5 cm]
The expected level of accuracy for the anthropometric measurements for length
To assess the mean absolute error of the Iron algorithm to estimate hemoglobin level (mg/dl) using image recognition technology
This exploratory outcome evaluates the proof-of-concept (POC) accuracy of an artificial intelligence (AI)-based algorithm for estimating haemoglobin levels (Iron AI) in a neonatal intensive care unit (NICU) or special care nursery (SCN) setting. AI-derived haemoglobin estimates (g/dL), generated from images and/or videos collected as part of the study, are compared with haemoglobin values obtained from standard venous blood sampling performed as part of routine clinical care. Accuracy is assessed using a composite metric including bias, mean absolute error, mean absolute percentage error, and predefined percentiles of absolute percentage error. This outcome is exploratory and descriptive in nature.
Click on the lead sponsor to view all of their trials.
Danone Asia Pacific Holdings Pte, Ltd.
Lead sponsor
KK Women's and Children's Hospital
Collaborator
