Morphometric traits to estimate brain and liver weight and their ratio for the diagnosis of intrauterine growth restriction in newborn piglets

Abstract

Intrauterine growth restriction (IUGR) is defined as inadequate foetal growth during gestation. In response to placenta insufficiency, IUGR piglets prioritise brain development as a survival mechanism. This adaptation leads to a higher brain-to-liver weight ratio (BrW/LW) at birth. This study assessed the potential of using morphometric traits to estimate brain (BrW) and liver (LW) weights, enabling non-invasive diagnosis of IUGR in newborn piglets. At birth, body weight (BtW) of individual piglets (n = 144) was recorded. One day (± 1) after birth, BrW and LW were measured with computed tomography (n = 94) or by weighing the organs after natural death or euthanasia (n = 50). Additionally, 20 morphometric traits were captured from images of each piglet and correlated with the BrW and LW. The morphometric traits that showed a r ≥ 0.70 in linear correlation with the BrW or LW were selected. Each selected trait was combined as an independent variable with BtW to develop multiple linear regression models to predict the BrW and LW. Six models were chosen based on the highest adjusted R² value: three for estimating BrW and three for LW. The dataset was then randomly divided into a training (75% of the data) and a testing (remaining 25%) subsets. Within the training subset, three equations to predict the BrW and three to predict the LW were extrapolated from the six selected models. The equations were then applied to the testing subset. The accuracy of the equations in predicting organ weight was assessed by calculating mean absolute and mean absolute percentage error (MAE and MAPE) between predicted and actual BrW and LW. To predict the BrW/LW, an equation including BtW and the two morphometric traits which better predicted BrW and LW was used. In the testing dataset, the equation combining ear distance and BtW better estimated the BrW. The equation performed with a MAE of 1.95 and a MAPE of 0.06 between the true and estimated weight of the brain. For the liver, the equation combining the abdominal area delimited by a square and BtW displayed the best performance, with a MAE of 9.29 and a MAPE of 0.17 between the true and estimated weight. Finally, the MAE and MAPE between the actual and estimated BrW/LW were 0.14 and 0.17, respectively. These findings suggest that specific morphometric traits can be used to estimate brain and liver weights, facilitating accurate and non-invasive identification of IUGR in newborn piglets.