Texture analysis of placental MRI: can it aid in the prenatal diagnosis of placenta accreta spectrum?
- 91 Downloads
To determine if texture analysis can differentiate placenta accreta spectrum (PAS) from normal placenta on MRI.
We performed retrospective image analysis of 80 patients, comprised of 46 patients with PAS and 34 patients without PAS. Histopathology was used as the reference standard. Sagittal single shot fast spin echo T2-weighted MRI sequences acquired from a single institution were analyzed. Placental heterogeneity was quantified using in-house software on a Matlab platform, including the standard deviation of pixel intensity, coefficient of variation, gray-level co-occurrence matrices (GLCM), histogram-oriented gradients (HOG), and fractal analysis with box sizes from 2 to 512. Two-tailed unpaired Student’s t test was used with statistical significance of p < 0.05.
PAS was associated with higher values for standard deviation of pixel intensity and fractal analysis at every box size. Fractal analysis at box sizes 256 (p = 0.011) and 32 (p = 0.021), and standard deviation of pixel intensity (p = 0.023) were the most statistically significant. Fractal values at box size 256 for PAS was 0.13 versus 0.090 for patients without PAS, while standard deviation of pixel intensity was 3.7 for PAS versus 2.5 for patients without PAS. No statistically significant association between PAS and GLCM, coefficient of variation, and HOG was found.
Statistically significant differences were found between normal and abnormal groups using standard deviation of pixel intensity and fractal analysis.
KeywordsPlacenta accreta spectrum Placenta accreta Texture analysis Fractal analysis MRI
The authors acknowledge Dr. Kruti P. Maniar, MD for assistance with clarification of pathological criteria in diagnosis of placenta accreta spectrum, and CCTS support for statistical analysis assistance (Grant Number UL1TR002003).
Compliance with ethical statement
Conflict of interest
The authors declare that they have no conflict of interest.
This study was approved by the IRB of the two main test sites, University of Illinois at Chicago and Northwestern.
- 2.Warshak CR, Eskander R, Hull AD, Scioscia AL, Mattrey RF, Benirschke K, Resnik R (2006) Accuracy of ultrasonography and magnetic resonance imaging in the diagnosis of placenta accreta. Obstet Gynecol 108 (3 Pt 1):573-581. https://doi.org/10.1097/01.aog.0000233155.62906.6d CrossRefGoogle Scholar
- 5.Einerson BD, Rodriguez CE, Kennedy AM, Woodward PJ, Donnelly MA, Silver RM (2018) Magnetic resonance imaging is often misleading when used as an adjunct to ultrasound in the management of placenta accreta spectrum disorders. Am J Obstet Gynecol 218 (6):618.e611-618.e617. https://doi.org/10.1016/j.ajog.2018.03.013 CrossRefGoogle Scholar
- 6.Budorick NE, Figueroa R, Vizcarra M, Shin J (2017) Another look at ultrasound and magnetic resonance imaging for diagnosis of placenta accreta. J Matern Fetal Neonatal Med:1-6. https://doi.org/10.1080/14767058.2016.1252744
- 7.D’Antonio F, Iacovella C, Palacios-Jaraquemada J, Bruno CH, Manzoli L, Bhide A (2014) Prenatal identification of invasive placentation using magnetic resonance imaging: systematic review and meta-analysis. Ultrasound Obstet Gynecol 44 (1):8-16. https://doi.org/10.1002/uog.13327 CrossRefGoogle Scholar