CoChat
Sign Up
Cited by 19Maternal and fetal healthcareRadiomics and Machine Learning in Medical ImagingPregnancy and preeclampsia studies

Radiomics-based prediction of FIGO grade for placenta accreta spectrum

Open access
Bartels, Helena C., O’Doherty, Jim, Wolsztynski, Eric, Brophy, David P., MacDermott, Roisin, Atallah, David, Saliba, Souha, Young, Constance +5 more·2023·European Radiology Experimental
DOI 10.1186/s41747-023-00369-2PMC PMC10509122
Open this paper in CoChat — free accountOpen PDF

Abstract

Background: Placenta accreta spectrum (PAS) is a rare, life-threatening complication of pregnancy. Predicting PAS severity is critical to individualise care planning for the birth. We aim to explore whether radiomic analysis of T2-weighted magnetic resonance imaging (MRI) can predict severe cases by distinguishing between histopathological subtypes antenatally. Methods: This was a bi-centre retrospective analysis of a prospective cohort study conducted between 2018 and 2022. Women who underwent MRI during pregnancy and had histological confirmation of PAS were included. Radiomic features were extracted from T2-weighted images. Univariate regression and multivariate analyses were performed to build predictive models to differentiate between non-invasive (International Federation of Gynecology and Obstetrics [FIGO] grade 1 or 2) and invasive (FIGO grade 3) PAS using R software. Prediction performance was assessed based on several metrics including sensitivity, specificity, accuracy and area under the curve (AUC) at receiver operating characteristic analysis. Results: Forty-one women met the inclusion criteria. At univariate analysis, 0.64 sensitivity (95% confidence interval [CI] 0.0−1.00), specificity 0.93 (0.38−1.0), 0.58 accuracy (0.37−0.78) and 0.77 AUC (0.56−.097) was achieved for predicting severe FIGO grade 3 PAS. Using a multivariate approach, a support vector machine model yielded 0.30 sensitivity (95% CI 0.18−1.0]), 0.74 specificity (0.38−1.00), 0.58 accuracy (0.40−0.82), and 0.53 AUC (0.40−0.85). Conclusion: Our results demonstrate a predictive potential of this machine learning pipeline for classifying severe PAS cases. Relevance statement: This study demonstrates the potential use of radiomics from MR images to identify severe cases of placenta accreta spectrum antenatally. Key points: • Identifying severe cases of placenta accreta spectrum from imaging is challenging. • We present a methodological approach for radiomics-based prediction of placenta accreta. • We report certain radiomic features are able to predict severe PAS subtypes. • Identifying severe PAS subtypes ensures safe and individualised care planning for birth. Graphical Abstract:

AI snapshot

Written from the paper’s full text.

TL;DR

This bi-centre retrospective analysis of a prospective cohort tested whether radiomic features from T2-weighted MRI can distinguish severe FIGO grade 3 placenta accreta spectrum (PAS) from FIGO grade 1–2 disease. The authors report predictive potential, especially high specificity, but multivariate models showed limited sensitivity and AUC in a small pilot dataset.

Key findings

  • In 41 women with histologically confirmed PAS, univariate radiomic analysis predicted FIGO grade 3 PAS with sensitivity 0.64 (95% CI 0.0–1.00), specificity 0.93 (0.38–1.0), accuracy 0.58 (0.37–0.78), and AUC 0.77 (0.56–0.97).
  • A multivariate support vector machine model yielded sensitivity 0.30 (95% CI 0.18–1.0), specificity 0.74 (0.38–1.00), accuracy 0.58 (0.40–0.82), and AUC 0.53 (0.40–0.85).
  • Radiomic features from inferior and superior placental regions of interest had similar overall performance; image-filtered features were selected more often than original-image features, and no clinical feature ranked as important in multivariate modelling.
  • The study’s Radiomics Quality Score was 38%, and the authors note that image-filtered features had better predictive potential than original-image features alone.

Methods

  • Bi-centre retrospective analysis of a prospective cohort conducted between January 2018 and October 2022; inclusion required MRI during pregnancy and histological confirmation of PAS, yielding 41 participants (18 FIGO grade 1–2, 23 FIGO grade 3).
  • Radiomic features were extracted from T2-weighted MRI using PyRadiomics, producing 1,820 features; after correlation filtering, Near Zero Variance filtering, and removal of features with cross-scanner discrepancies, a final set of 204 features was selected.
  • Univariate logistic regression with bootstrapping and four multivariate classifiers (LASSO, random forest, k-nearest neighbour, and support vector machine) were trained in R; performance was assessed by sensitivity, specificity, accuracy, AUC, calibration curves, and principal component analysis.

Limitations

  • The authors state the sample size is small and that multivariate models showed a drop in performance compared with univariate analysis due to loss of statistical power with a small N:P ratio.
  • The authors state that important MRI features such as myometrial thinning and bladder-wall interruption are not captured because only the placenta was segmented, and that manual segmentation was used but radiomic features were stable across readers.
  • Evident from the text: this is a pilot dataset without external validation, and the authors explicitly say future studies are needed to validate and improve the results in larger datasets.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s41747-023-00369-2

Cites

Statistics

Nothing to recompute.

Claims the paper makes

0 of 52 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    At univariate analysis, sensitivity 0.64 (95% CI 0.0-1.00) was achieved for predicting severe FIGO grade 3 PAS.

    “At univariate analysis, 0.64 sensitivity (95% confidence interval [CI] 0.0−1.00), specificity 0.93 (0.38−1.0), 0.58 accuracy (0.37−0.78) and 0.77 AUC (0.56−.097) was achieved for predicting severe FIGO grade 3 PAS.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Using a multivariate approach, a support vector machine model yielded sensitivity 0.30 (95% CI 0.18-1.0) for predicting severe FIGO grade 3 PAS.

    “Using a multivariate approach, a support vector machine model yielded 0.30 sensitivity (95% CI 0.18−1.0]), 0.74 specificity (0.38−1.00), 0.58 accuracy (0.40−0.82), and 0.53 AUC (0.40−0.85).”
    Quote in the paperSubstantial claim
  • Findingown data

    At univariate analysis, accuracy 0.58 (95% CI 0.37-0.78) was achieved for predicting severe FIGO grade 3 PAS.

    “At univariate analysis, 0.64 sensitivity (95% confidence interval [CI] 0.0−1.00), specificity 0.93 (0.38−1.0), 0.58 accuracy (0.37−0.78) and 0.77 AUC (0.56−.097) was achieved for predicting severe FIGO grade 3 PAS.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Using a multivariate approach, a support vector machine model yielded AUC 0.53 (95% CI 0.40-0.85) for predicting severe FIGO grade 3 PAS.

    “Using a multivariate approach, a support vector machine model yielded 0.30 sensitivity (95% CI 0.18−1.0]), 0.74 specificity (0.38−1.00), 0.58 accuracy (0.40−0.82), and 0.53 AUC (0.40−0.85).”
    Quote in the paperSubstantial claim
  • Findingown data

    At univariate analysis, AUC 0.77 (95% CI 0.56-.097) was achieved for predicting severe FIGO grade 3 PAS.

    “At univariate analysis, 0.64 sensitivity (95% confidence interval [CI] 0.0−1.00), specificity 0.93 (0.38−1.0), 0.58 accuracy (0.37−0.78) and 0.77 AUC (0.56−.097) was achieved for predicting severe FIGO grade 3 PAS.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 35
  • May combine two claims. 28

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Forty-one women met the inclusion criteria.

    “Forty-one women met the inclusion criteria.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Forty-one participants met inclusion criteria (34 from site 1, 7 from site 2), including 18 FIGO grade 1-2 and 23 FIGO grade 3 PAS cases.

    “Forty-one participants met inclusion criteria (34 from site 1, 7 from site 2), including 18 FIGO grade 1–2 and 23 FIGO grade 3 PAS cases.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Calibration curves demonstrated reasonable agreement between predicted and observed rates of FIGO grade 3 PAS, indicating the models estimate the probability of PAS appropriately both for FIGO grade 3 and non-grade 3 cases.

    “For classification based on the four multivariate models (LASSO, RF, kNN, and SVM), calibration curves demonstrated reasonable agreement between the predicted and observed rates of FIGO grade 3 PAS, indicating the models estimate the probability of PAS appropriately both for FIGO grade 3 and non-grade 3 cases (Fig. 4).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    After feature filtering, a final feature set of 204 radiomic features was selected.

    “Following removal of features with high cross scanner discrepancies, a final feature set of 204 features were selected.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    After FDR adjustment, there were no significant associations between clinical variables or radiomic features and FIGO grade 1-2 vs 3, with a smallest p value of 0.228.

    “After p value adjustment for FDR, there were no significant associations, with a smallest p value of 0.228.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Clinical features were not of high importance; only number of previous Caesarean sections was included in the top ten accuracies or AUC from univariate logistic regression.

    “Clinical features included in the models were not of high importance, with only the number of previous Caesarean section included in the top ten accuracies or AUC from univariate logistic regression (Fig. 3).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    All four multivariate models had similar overall performance for predicting FIGO grade 3 PAS, with all models having a ROC above 50% and specificity above 60%.

    “Models had a similar performance overall for predicting FIGO grade 3 PAS, with all models having a ROC of above 50% and specificity above 60% (Fig. 5a, b).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Subsets of individual radiomic features with high specificity were found in both regions, while sensitivity was poorer for both.

    “Subsets of individual radiomic features with high specificity were found in both regions, while sensitivity was poorer for both.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The radiomic profiles of inferior and superior placental regions were different with little correlation between them (Pearson ρ −0.26 and +0.31).

    “The radiomic profiles of inferior and placental regions were, however, different with little correlation between them (Pearson ρ−0.26 and + 0.31).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    The feature subsets for kNN and SVM included only one clinical variable, the number of prior Caesarean sections.

    “which included only one clinical variable, i.e., the number of prior Caesarean sections”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In multivariate models predicting severe FIGO grade 3 PAS, radiomic features with image filters applied were selected more frequently than original image features; no original image feature ranked in the top 20 predictive features for either the inferior or superior placental ROI.

    “For multivariate models, radiomic features with image filters applied were selected at higher frequency than those from the original image, with no radiomic feature from the original image feature ranked in the top 20 predictive features for either the inferior and superior placental ROIs for predicting severe FIGO grade 3 PAS.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Both kNN and SVM tended to use more information from some clusters in particular, i.e., groups of predictors with high information overlap.

    “Both these models tended to use more information from some of the clusters in particular, i.e., groups of predictors with a high level of information overlap.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In multivariate modelling, radiomic features were ranked as highest importance, and no clinical features were ranked as important.

    “radiomic features were ranked as highest importance, with no clinical features ranked as important for modelling”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    kNN yielded variable importance and PCA outputs identical to those obtained from SVM.

    “kNN yielded variable importance and PCA outputs identical to those obtained from support vector machine (SVM)”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    kNN and SVM selected the same feature subsets.

    “k-nearest neighbour and SVM selected the same feature subsets”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    On univariate analysis, radiomic features had reasonable sensitivity and high specificity for identifying severe FIGO grade 3 PAS cases.

    “In this study, radiomic features had a reasonable sensitivity for identifying severe FIGO grade 3 cases on univariate analysis, and a high specificity.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    There is no consensus internationally on optimal management of PAS, with variations in many aspects of clinical care.

    “There is no consensus internationally on optimal management, with variations in many aspects of clinical care [4].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [4] Einerson BD, Silver RM (2019) Multidisciplinary teams in the management of placenta accreta spectrum disorders. Curr Obstet Gynecol Rep Rep 8:80–85 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Ultrasound and MRI are the imaging modalities of choice for diagnosing PAS antenatally.

    “Ultrasound and magnetic resonance imaging (MRI) are the imaging modalities of choice for diagnosing PAS antenatally [5, 6].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [5] Jha P, Pōder L, Bourgioti C et al (2020) Society of Abdominal Radiology (SAR) and European Society of Urogenital Radiology (ESUR) joint consensus statement for MR imaging of placenta accreta spectrum disorders. Eur Radiol 30:2604–2615. https://doi.org/10.1007/s00330-019-06617-7 source

      Not checked in the free run.

    • cites [6] Collins SL, Ashcroft A, Braun T et al (2016) Proposal for standardized ultrasound descriptors of abnormally invasive placenta (AIP). Ultrasound Obstet Gynecol 47:271–275. https://doi.org/10.1002/uog.14952 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    PAS is classified into three grades by FIGO, with FIGO grade 3 being the most severe.

    “PAS is classified into three grades by the International Federation of Gynecology and Obstetrics (FIGO) [2], with FIGO grade 3 being the most severe.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] Jauniaux E, Ayres-de-Campos D, Langhoff-Roos J, Fox KA, Collins S (2019) FIGO classification for the clinical diagnosis of placenta accreta spectrum disorders. Int J Gynaecol Obstet 146:20–24. https://doi.org/10.1002/ijgo.12761 source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Previous work showed that ROIs of the placenta proximal and remote to the prior Caesarean section scar had significant differences in radiomic feature distribution, associated with undergoing caesarean hysterectomy for PAS.

    “The segmentation protocol was informed by a previous work where ROIs of the placenta proximal and remote to the prior Caesarean section scar showed significant differences in the distribution of radiomic features, which were associated with undergoing caesarean hysterectomy for PAS [13].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [13] Do QN, Lewis MA, Xi Y et al (2020) MRI of the placenta accreta spectrum (PAS) disorder: radiomics analysis correlates with surgical and pathological outcome. J Magn Reson Imaging 51:936–946. https://doi.org/10.1002/jmri.26883 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Many radiomics PAS studies were limited by significant heterogeneity in how PAS was defined, with only two studies reporting histological data, the reference standard.

    “Many were limited by the significant heterogeneity in how PAS was defined, with only two studies reporting on histological data, which is considered the reference standard in diagnosing PAS [14, 15].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [14] Stanzione A, Verde F, Cuocolo R et al (2022) Placenta accreta spectrum disorders and radiomics: systematic review and quality appraisal. Eur J Radiol 155:110497. https://doi.org/10.1016/j.ejrad.2022.110497 source

      Not checked in the free run.

    • cites [15] Hecht JL, Baergen R, Ernst LM et al (2020) Classification and reporting guidelines for the pathology diagnosis of placenta accreta spectrum (PAS) disorders: recommendations from an expert panel. Mod Pathol 33:2382–2396. https://doi.org/10.1038/s41379-020-0569-1 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Image segmentation was performed by three independent investigators who were blinded to the results of other imaging and final intraoperative or histopathological outcomes.

    “Images were manually segmented on multiple representative slices with regions of interest (ROIs) by three independent investigators (H.B., D.B., R.M., with 4 over 20 and 6 years of experience). Investigators were aware of the purpose of the study but were blinded to the results of other imaging, or final intraoperative or histopathological outcomes.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The systematic review highlighted varying methodological quality of radiomics PAS studies to date.

    “The review highlighted the varying methodological quality of the radiomics PAS studies to date [14].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [14] Stanzione A, Verde F, Cuocolo R et al (2022) Placenta accreta spectrum disorders and radiomics: systematic review and quality appraisal. Eur J Radiol 155:110497. https://doi.org/10.1016/j.ejrad.2022.110497 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A lack of standardised methodology is an important limitation of all radiomics studies and remains a major challenge in the field.

    “Furthermore, a lack of standardised methodology is an important limitation of all radiomics studies and remains a major challenge in the field [18].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [18] van Timmeren JE, Cester D, Tanadini-Lang S, Alkadhi H, Baessler B (2020) Radiomics in medical imaging-"how-to" guide and critical reflection. Insights Imaging 11:91. https://doi.org/10.1186/s13244-020-00887-2 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Several studies have applied radiomics to MR images in PAS and found radiomic features useful in aiding diagnosis and predicting clinical outcomes such as massive obstetric haemorrhage.

    “Several studies have applied radiomics to MR images in PAS and found radiomic features were useful in aiding diagnosis and predicting clinical outcomes such as massive obstetric haemorrhage [12, 13].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [12] Peng L, Zhang X, Liu J et al (2022) MRI-radiomics-clinical-based nomogram for prenatal prediction of the placenta accreta spectrum disorders. Eur Radiol. https://doi.org/10.1007/s00330-022-08821-410.1007/s00330-022-08821-4. source

      Not checked in the free run.

    • cites [13] Do QN, Lewis MA, Xi Y et al (2020) MRI of the placenta accreta spectrum (PAS) disorder: radiomics analysis correlates with surgical and pathological outcome. J Magn Reson Imaging 51:936–946. https://doi.org/10.1002/jmri.26883 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    MRI was read and reported by specialist radiologists with over 15 and 20 years of experience.

    “MRI, read and reported by specialist radiologists [22] (T.G., with over 15 years of experience; D.B., with over 20 years of experience).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [22] Bartels HC, Mulligan KM, Craven S et al (2021) Maternal morbidity in placenta accreta spectrum following introduction of a multi-disciplinary service compared to standard care: an Irish perspective. Ir J Med Sci 190:1451–1457. https://doi.org/10.1007/s11845-020-02473-3 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    MRI plays an important role in PAS, such as for surgical planning and in the assessment of posterior or lateral defects.

    “MRI plays an important role, such as for surgical planning and in the assessment of posterior or lateral defects [5, 7, 8].”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [5] Jha P, Pōder L, Bourgioti C et al (2020) Society of Abdominal Radiology (SAR) and European Society of Urogenital Radiology (ESUR) joint consensus statement for MR imaging of placenta accreta spectrum disorders. Eur Radiol 30:2604–2615. https://doi.org/10.1007/s00330-019-06617-7 source

      Not checked in the free run.

    • cites [7] Tinari S, Buca D, Cali G et al (2021) Risk factors, histopathology and diagnostic accuracy in posterior placenta accreta spectrum disorders: systematic review and meta-analysis. Ultrasound Obstet Gynecol 57:903–909. https://doi.org/10.1002/uog.22183 source

      Not checked in the free run.

    • cites [8] Palacios-Jaraquemada JM, Fiorillo A, Hamer J, Martínez M, Bruno C (2022) Placenta accreta spectrum: a hysterectomy can be prevented in almost 80% of cases using a resective-reconstructive technique. J Maternal Fetal Neonatal Med 35:275–282. https://doi.org/10.1080/14767058.2020.1716715 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    The current literature is limited by the lack of standardised definitions for PAS and the varying quality of radiomic methodology.

    “Therefore, the current literature is limited by the lack of standardised definitions for PAS and the varying quality of the radiomic methodology.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A recent systematic review included 10 studies on radiomics in PAS.

    “These were summarised in a recent systematic review, which included 10 studies [14].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [14] Stanzione A, Verde F, Cuocolo R et al (2022) Placenta accreta spectrum disorders and radiomics: systematic review and quality appraisal. Eur J Radiol 155:110497. https://doi.org/10.1016/j.ejrad.2022.110497 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The diagnostic accuracy of MRI is highly dependent on reader expertise and can incorrectly classify PAS cases in up to 30% of cases.

    “However, the diagnostic accuracy of MRI is highly dependent on reader expertise and can incorrectly classify cases in up to 30% of cases [10].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10] 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 218:618.e611-618.e617. https://doi.org/10.1016/j.ajog.2018.03.013 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Features with high cross-scanner discrepancies (adjusted p < 0.05) were removed.

    “Features with high cross-scanner discrepancies, defined as an adjusted p value < 0.05, were removed.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [28] Zhovannik I, Bussink J, Traverso A et al (2019) Learning from scanners: Bias reduction and feature correction in radiomics. Clin Transl Radiat Oncol 19:33–38. https://doi.org/10.1016/j.ctro.2019.07.003 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Univariate logistic regression models were trained using bootstrapping and tested on out-of-bag bootstrap points.

    “Univariate logistic regression models trained using bootstrapping, and tested on the out-of-bag bootstrap points, were used to identify the best performing radiomic features.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Previous studies have used radiomic features to predict clinical outcomes such as undergoing caesarean hysterectomy and blood loss.

    “Previous studies have used radiomic features to predict clinical outcomes such as undergoing caesarean hysterectomy and blood loss [13, 37]”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [13] Do QN, Lewis MA, Xi Y et al (2020) MRI of the placenta accreta spectrum (PAS) disorder: radiomics analysis correlates with surgical and pathological outcome. J Magn Reson Imaging 51:936–946. https://doi.org/10.1002/jmri.26883 source

      Not checked in the free run.

    • cites [37] Wu Q, Yao K, Liu Z et al (2019) Radiomics analysis of placenta on T2WI facilitates prediction of postpartum haemorrhage: a multicentre study. EBioMedicine 50:355–365. https://doi.org/10.1016/j.ebiom.2019.11.010 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Using haemorrhage as an outcome for predictive modelling is limited because blood loss measurement is highly heterogeneous and often inaccurate, complicating comparisons between studies.

    “using haemorrhage as an outcome for predicting modelling is limited as blood loss measurement is highly heterogenous and often inaccurate, hence complicating comparisons between studies using this as an outcome [39, 40]”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [39] Doctorvaladan SV, Jelks AT, Hsieh EW, Thurer RL, Zakowski MI, Lagrew DC (2017) Accuracy of blood loss measurement during cesarean delivery. AJP Rep 7:e93–e100. https://doi.org/10.1055/s-0037-1601382 source

      Not checked in the free run.

    • cites [40] Liu Y, Zhu W, Shen Y, Qiu J, Rao L, Li H (2021) Assessing the accuracy of visual blood loss estimation in postpartum hemorrhage in Shanghai hospitals: a web-based survey for nurses and midwives. J Clin Nurs 30:3556–3562. https://doi.org/10.1111/jocn.15860 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Some studies reported that use of pre-processing image filters improved predictive performance of models.

    “Previously, some studies reported the use of pre-processing image filters was shown to improve predictive performance of models [47, 48].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [47] Demircioğlu A (2022) The effect of preprocessing filters on predictive performance in radiomics. Eur Radiol Exp 6:40. https://doi.org/10.1186/s41747-022-00294-w source

      Not checked in the free run.

    • cites [48] Teng X, Zhang J, Zwanenburg A et al (2022) Building reliable radiomic models using image perturbation. Sci Rep 12:10035. https://doi.org/10.1038/s41598-022-14178-x source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Previous radiomic studies in PAS found that radiomic features from image filters were more predictive and had higher diagnostic accuracy than those from the original image.

    “Our results support the finding of previous radiomic studies in PAS, where radiomic features from image filters were more predictive and had higher diagnostic accuracy than those from the original image [45, 46].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [45] Chu C, Liu M, Zhang Y et al (2022) MRI-based radiomics analysis for intraoperative risk assessment in gravid patients at high risk with placenta accreta spectrum. Diagnostics (Basel) 12(2):485. https://doi.org/10.3390/diagnostics12020485 source

      Not checked in the free run.

    • cites [46] Romeo V, Ricciardi C, Cuocolo R et al (2019) Machine learning analysis of MRI-derived texture features to predict placenta accreta spectrum in patients with placenta previa. Magn Reson Imaging 64:71–76. https://doi.org/10.1016/j.mri.2019.05.017 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    There is currently no consensus on how to perform image pre-processing steps for MR images in radiomics, and it is one of the open challenges in radiomic processing.

    “There is currently no consensus on how to perform these steps for MR images in radiomics, and is one of the many open challenges in radiomic processing [18, 44].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [18] van Timmeren JE, Cester D, Tanadini-Lang S, Alkadhi H, Baessler B (2020) Radiomics in medical imaging-"how-to" guide and critical reflection. Insights Imaging 11:91. https://doi.org/10.1186/s13244-020-00887-2 source

      Not checked in the free run.

    • cites [44] Moradmand H, Aghamiri SMR, Ghaderi R (2020) Impact of image preprocessing methods on reproducibility of radiomic features in multimodal magnetic resonance imaging in glioblastoma. J Appl Clin Med Phys 21:179–190. https://doi.org/10.1002/acm2.12795 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    When datasets are small, limiting analysis to radiomic features from the original image only showed similar results to when all radiomic features with image filters were included.

    “However, when datasets are small, limiting the analysis to radiomic features from the original image only showed similar results to when all radiomic features with image filters were included [47].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [47] Demircioğlu A (2022) The effect of preprocessing filters on predictive performance in radiomics. Eur Radiol Exp 6:40. https://doi.org/10.1186/s41747-022-00294-w source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted).

    Premisecitation

    The inferior placental ROI is the area where abnormal uterine remodelling from a prior Caesarean section has resulted in abnormal placentation.

    “this is the area where abnormal uterine remodelling from a prior Caesarean section has resulted in abnormal placentation [42]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [42] Jauniaux E, Jurkovic D, Hussein AM, Burton GJ (2022) New insights into the etiopathology of placenta accreta spectrum. Am J Obstet Gynecol 227:384–391. https://doi.org/10.1016/j.ajog.2022.02.038 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Well-described MR imaging features such as abnormal vascularisation of the placental bed and focal exophytic mass are signs of severe PAS and are usually seen at the level of the bladder, thus included in the inferior placental ROI.

    “Well-described MR imaging features, such as abnormal vascularisation of the placental bed and focal exophytic mass, are signs of severe PAS; they are usually seen at the level of the bladder, and are therefore included in the inferior placental ROI [5].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [5] Jha P, Pōder L, Bourgioti C et al (2020) Society of Abdominal Radiology (SAR) and European Society of Urogenital Radiology (ESUR) joint consensus statement for MR imaging of placenta accreta spectrum disorders. Eur Radiol 30:2604–2615. https://doi.org/10.1007/s00330-019-06617-7 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Many radiomic studies score poorly when assessed externally.

    “many when assessed externally score poorly”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    This study is penalised by the RQS in a number of areas because it does not assess oncologic data.

    “This study, as with many studies not assessing oncologic data, is penalised by the RQS in a number of areas.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The RQS was developed to assess the quality of oncologic radiomics studies and not studies exploring conditions such as PAS.
    • Rests on an assumption: Open source images, segmentations and code are not currently available for PAS.

  • May combine two claims.

    Premisecitation

    The applicability of trained algorithms to new datasets is currently limited and often results in failure of accurate segmentations.

    “the applicability of trained algorithms to new datasets is currently limited and often results in failure of accurate segmentations [18]”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [18] van Timmeren JE, Cester D, Tanadini-Lang S, Alkadhi H, Baessler B (2020) Radiomics in medical imaging-"how-to" guide and critical reflection. Insights Imaging 11:91. https://doi.org/10.1186/s13244-020-00887-2 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Unlike in cancer care, there is no indication to repeat imaging to assess treatment response in PAS.

    “Hence, unlike in cancer care, there is no indication to repeat imaging to assess treatment response.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: In PAS, the condition is limited to a distinct period of time—pregnancy—and once the pregnancy is completed and the placenta removed, the condition can be considered as “treated”.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Few radiomic studies to date have included a RQS.

    “Few radiomic studies to date have included a RQS in their work [20]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [20] Park JE, Kim HS, Kim D et al (2020) A systematic review reporting quality of radiomics research in neuro-oncology: toward clinical utility and quality improvement using high-dimensional imaging features. BMC Cancer 20:29. https://doi.org/10.1186/s12885-019-6504-5 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Automated segmentations may be considered ideal because they minimise inter-observer variability.

    “automated segmentations may be considered the ideal approach as it minimises the inter-observability between segmentations”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [18] van Timmeren JE, Cester D, Tanadini-Lang S, Alkadhi H, Baessler B (2020) Radiomics in medical imaging-"how-to" guide and critical reflection. Insights Imaging 11:91. https://doi.org/10.1186/s13244-020-00887-2 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Initiatives such as RQS and Image Biomarker Standardisation Initiative attempt to harmonise radiomic studies and produce results that can be validated and applied to clinical settings.

    “Initiatives such as the RQS [20] and Image Biomarker Standardisation Initiative [49] attempt to harmonise radiomic studies and produce results which can be validated and applied to clinical settings.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [20] Park JE, Kim HS, Kim D et al (2020) A systematic review reporting quality of radiomics research in neuro-oncology: toward clinical utility and quality improvement using high-dimensional imaging features. BMC Cancer 20:29. https://doi.org/10.1186/s12885-019-6504-5 source

      Not checked in the free run.

    • cites [49] Zwanenburg A, Vallières M, Abdalah MA et al (2020) The image biomarker standardization initiative: standardized quantitative radiomics for high-throughput image-based phenotyping. Radiology 295:328–338. https://doi.org/10.1148/radiol.2020191145 source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductreasoning

    The ROIs presented here include only the placenta, so important MRI features of PAS such as myometrial thinning and interruption of the bladder wall are not captured.

    “Some important MRI features of PAS such as myometrial thinning and interruption of the bladder wall will not be captured by the ROIs presented here, which include only the placenta.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Attempts to limit inter-observer variability in this study included calculation of ICC between readers for a subset of segmentations, a clearly defined segmentation protocol, and three investigators independently performing segmentations with cross-reference.

    “Attempts to limit inter-observer variability in this study included calculation of the ICC between readers for a subset of segmentations, a clearly defined segmentation protocol and three investigators independently performing segmentations with cross-reference.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    Efforts to increase sample size were employed by using data available from another centre.

    “Efforts to increase our sample size were employed, by using data available from another centre.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Radiomics has clinical potential to detect and rule out severe PAS.

    “This demonstrates the clinical potential of using radiomics to detect and rule out severe PAS.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Ultrasound and MRI both rely heavily on reader expertise of the clinician.

    “Ultrasound and MRI both rely heavily on reader expertise of the clinician [5, 10].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [5] Jha P, Pōder L, Bourgioti C et al (2020) Society of Abdominal Radiology (SAR) and European Society of Urogenital Radiology (ESUR) joint consensus statement for MR imaging of placenta accreta spectrum disorders. Eur Radiol 30:2604–2615. https://doi.org/10.1007/s00330-019-06617-7 source

      Not checked in the free run.

    • cites [10] 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 218:618.e611-618.e617. https://doi.org/10.1016/j.ajog.2018.03.013 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    PAS remains a rare complication of pregnancy, but its incidence is increasing as a result of the rise in Caesarean section rate.

    “PAS remains a rare complication of pregnancy, however the incidence is increasing as a result of the rise in the Caesarean section rate [50].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [50] Jauniaux E, Chantraine F, Silver RM, Langhoff-Roos J (2018) FIGO consensus guidelines on placenta accreta spectrum disorders: epidemiology. Int J Gynaecol Obstet 140:265–273. https://doi.org/10.1002/ijgo.12407 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Radiomics to predict disease severity may assist clinicians in individualising care for women with PAS.

    “Radiomics to predict disease severity may assist clinicians in individualising care for women with PAS.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Identifying severe PAS subtypes may allow interventions associated with additional morbidity, such as elective preterm delivery and interventional radiology techniques, to be reserved for these cases.

    “By identifying severe PAS subtypes, interventions associated with additional morbidity, such as elective preterm delivery and interventional radiology techniques, may be reserved for these cases.”
    Quote in the paperNot a substantial claim
55 of 55

  • The paper relies on this without support.

    Identifying or predicting severe PAS from imaging antenatally is challenging.

    “Predicting disease severity antenatally remains a major challenge.”
    “Identifying severe cases of placenta accreta spectrum from imaging is challenging.”

  • The paper relies on this without support.

    MRI features such as placental heterogeneity and T2-dark bands are seen throughout the whole placenta.

    “MRI features such as placental heterogeneity and T2-dark bands are seen throughout the whole placenta.”

  • The paper relies on this without support.

    The RQS was developed to assess the quality of oncologic radiomics studies and not studies exploring conditions such as PAS.

    “this score was developed to assess the quality of oncologic radiomics studies and not studies exploring conditions such as PAS.”

  • The paper relies on this without support.

    In PAS, the condition is limited to a distinct period of time—pregnancy—and once the pregnancy is completed and the placenta removed, the condition can be considered as “treated”.

    “In PAS, the condition is limited to a distinct period of time—pregnancy—and once the pregnancy is completed and the placenta removed, the condition can be considered as “treated”.”

  • The paper relies on this without support.

    Open source images, segmentations and code are not currently available for PAS.

    “To our knowledge, these are not currently available for PAS.”
5 of 5

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • The integration of deep learning techniques consistently outperformed manual diagnostic methods for PAS when trained on quality datasets.

    “The integration of deep learning techniques consistently outperformed manual diagnostic methods when trained on quality datasets [24-26].”

    From Machine Learning in Maternity: Early Diagnosis of Placenta Accreta Spectrum (2026) · cites this paper as [25]

    The Discussion (page unavailable) reports that a radiomics-based machine-learning approach showed potential for predicting severe PAS, but says the pilot findings need external validation. It does not show that deep learning consistently outperformed manual diagnosis or establish that result for quality datasets; the paper’s comparison of performance concerns different radiomic image features, not manual methods. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Elective-delivery percentage is wrong for FIGO grade 3

    Minor

    Table 1 gives seven elective deliveries among 23 FIGO grade 3 participants, which is 30.4%, not 100.0%.

    Reported 100 · recomputed 30.434783

  • Abstract and Results give different AUC confidence limits

    Minor

    The abstract prints the upper confidence limit for the 0.77 AUC as .097, below both the estimate and the lower limit. The Results give the corresponding upper limit as 0.97.

  • Elective-delivery count exceeds the FIGO grade 1-2 group

    Minor

    Table 1 reports 30 elective deliveries in a group of 18 and labels this 75.0%. Thirty divided by 18 is 166.7%, so the count and percentage cannot describe that group.

    Reported 75 · recomputed 166.666667

  • Univariate models are called both logistic and linear regression

    Minor

    The Methods identify the bootstrapped univariate models as logistic regression, whereas the Fig. 3 caption calls them linear regression models.

4 of 4

The citation lists haven’t been fetched yet

References 48 · Cited by 19

Research works printed but not linked: 1

  1. SimkoTWaV (2021) R package ’corrplot’Visualization of a Correlation Matrix (Version 0.92)

Authors
Bartels, H. C., O’Doherty, J., Wolsztynski, E., Brophy, D. P., MacDermott, R., Atallah, D., Saliba, S., Young, C., Downey, P., Donnelly, J., Geoghegan, T., Brennan, D. J., & Curran, K. M.
Affiliations
  • University College Dublin (IE)
  • University College Cork (IE)
  • St. Vincent's University Hospital (IE)
  • Saint Joseph University (LB)
  • National Maternity Hospital (IE)
  • Rotunda Hospital (IE)
  • Mater Misericordiae University Hospital (IE)
Corresponding author
  • Bartels, Helena C.
  • Curran, Kathleen M.
ORCID
Venue
European Radiology Experimental
Publisher
Springer Science and Business Media LLC
Published
2023-09-20
Volume
7
Issue
1
Pages
54-54
ISSN
2509-9280 (ISSN-L)
DOI
10.1186/s41747-023-00369-2
PMC
PMC10509122
OpenAlex
W4386852543
Semantic Scholar
f0961288fe5ebb09303aa8fe7125c4ad854b3a92
Cited by
19
References
48
Citations by year
2026: 7 · 2025: 4 · 2024: 6 · 2023: 2
FWCI
7.00
Type
journal article
Language
en
Update notices
correction (2023-11-22)
Open access
Licence
Locations
  • European Radiology Experimental (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Lenus, The Irish Health Repository (Dr Steevens Hospital Library) (submittedVersion)
Funders
  • Medical Fund from National Maternity Hospital, Dublin, Ireland (No award number)
  • Science Foundation Ireland, co-funded under European Regional Development Fund (12/RC/2289-P2)
Topics
  • Health Sciences › Medicine › Pediatrics, Perinatology and Child Health › Maternal and fetal healthcare
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
  • Health Sciences › Medicine › Obstetrics and Gynecology › Pregnancy and preeclampsia studies
Keywords
placenta accreta spectrum, FIGO grade, radiomics, T2-weighted MRI, prenatal prediction, support vector machine, pregnancy complications, histopathological grading
MeSH terms
Machine Learning, Female, Humans, Placenta Accreta, Pregnancy, Prospective Studies, Research Design, Retrospective Studies

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review