Identification of suspicious invasive placentation based on clinical MRI data using textural features and automated machine learning

Sun, Huaiqiang; Qu, Haibo; Chen, Lu; Wang, Wei; Liao, Yi; Zou, Ling; Zhou, Ziyi; Wang, Xiaodong; Zhou, Shu

doi:10.1007/s00330-019-06372-9

Cited by 51 publications

(55 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…]2 or 3) for scoring CDs. [15][16][17][18]21 The Weiniger model was developed using a comparatively lower-risk population compared to several other models (Table S2), 15,[17][18][19][20][21]23 which may make it more generalizable, as confirmed by the current study. Of note, a model developed and internally validated by Yang et al 22 in a large cohort of over 8000 patients had high discrimination of 0.93 and was based solely on clinical characteristics (prior CD, vaginal bleeding, medications during pregnancy, praevia).…”

Section: Strengths and Limitationssupporting

confidence: 63%

“…The Weiniger model utilizes US as opposed to MRI. [19][20][21] US is more readily available, and the superiority of MRI over US for diagnosis of PAS has not been established. [33][34][35][36] Risk of PAS increases with each CD, with a 10-fold increase of concurrent PAS with 5 vs. 2 prior CDs.…”

Section: Strengths and Limitationsmentioning

confidence: 99%

“…9 The model had good discrimination by area under the curve (AUC) and may be more generalisable, comprehensive, and easier to use than alternative PAS models. [15][16][17][18][19][20][21][22][23] We aimed to validate this model in two cohorts derived from two tertiary-care hospitals: a variable-risk cohort that included patients with risk factors (prior CD and praevia) and/or US features of PAS (Cohort A) and a higher-risk referral population all of who had US features of PAS and were referred to the PAS service of a tertiary care hospital. Our hypothesis was that this model would demonstrate external validity among the obstetric patients in both tertiary care settings.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A prediction model for placenta accreta spectrum: A multicentre external validation study.

Singh

Carusi

Wang

et al. 2021

Preprint

View full text Add to dashboard Cite

Objective: To validate the Weiniger model, a multivariable prediction model for placenta accreta spectrum (PAS). Design: Multicentre external validation study. Setting: Two tertiary care hospitals in the United States. Population: Cohort A included patients with risk factors (prior caesarean delivery, placenta praevia) and/or ultrasound features of PAS (variable risk) presenting to a tertiary care hospital. Cohort B patients were referred to a tertiary care hospital specifically for ultrasound features of PAS (higher risk). Methods: Weiniger model variables (prior caesarean deliveries, placenta praevia and ultrasound features of PAS) were retrospectively collected from both cohorts and predictive performance of the model was evaluated. Main Outcome Measures: Surgical and/or pathological diagnosis of PAS. Results: The model c-statistic in cohorts A and B was 0.728 (95% CI: 0.662, 0.794) and 0.866 (95% CI: 0.754, 0.977) signifying acceptable and excellent discrimination, respectively. Based on calibration curves, the model underestimated average PAS risk in both cohorts. In both cohorts, high risk was overestimated and low risk underestimated. Use of this model compared to a “treat all” strategy had greater net benefit at a threshold probability of > 0.25 in cohort A, but no net benefit in cohort B. Conclusions: This study provides multicentre external validation of the Weiniger model for PAS prediction, making it a useful triaging tool for management of this high-risk obstetric condition. Clinical usefulness of this model is influenced by the incidence of risk factors and PAS ultrasound features, with better performance in a variable-risk population at threshold probability >25%.

show abstract

Section: Strengths and Limitationssupporting

confidence: 63%

Section: Strengths and Limitationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A prediction model for placenta accreta spectrum: A multicentre external validation study.

Singh

Carusi

Wang

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…• In [19], a CI based prediction system for the best delivery timing on a pregnant woman duffering of SLE was proposed. • Placental disorders identification was the focus in [178], [179]. ML, CI and VIS techniques were employed for this purpose.…”

Section: Other Pregnancy Processesmentioning

confidence: 99%

Artificial Intelligence in Pregnancy: A Scoping Review

et al. 2020

View full text Add to dashboard Cite

Artificial Intelligence has been widely applied to a majority of research areas, including health and medicine. Certain complications or disorders that can appear during pregnancy can endanger the life of both mother and fetus. There is enough scientific literature to support the idea that emotional aspects can be a relevant risk factor in pregnancy (such as anxiety, stress or depression, for instance). This paper presents a scoping review of the scientific literature from the past 12 years (2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020) to identify which methodologies, techniques, algorithms and frameworks are used in Artificial Intelligence and Affective Computing for pregnancy health and well-being. The methodology proposed by Arksey and O'Malley, in conjunction with PRISMA-ScR framework has been used to create this review. Despite the relevance that emotional status can have as a risk factor during pregnancy, one of the main findings of this study is that there is still not a significant amount of literature on automatic analysis of emotion. Health enhancement and well-being for pregnant women can be achieved with artificial intelligence or affective computing based devices, hence future work on this topic is strongly suggested.

show abstract

“…As for MRI, another important diagnostic tool for prenatal diagnosis, the latest systematic review and meta-analysis of MRI in identifying invasive placentation showed a sensitivity of 94.4%, a speci city of 84.0%, a positive likelihood ratio of 5.91, and a negative likelihood ratio of 0.07 [13]. Therefore, MRI can predict depth of invasion and topography of invasion in PAS with high accuracy through speci c signs on the T2-weighted image, such as uterine bulging, heterogeneous signal intensity, focal interruption of the myometrium, bladder tenting, and dark intraplacental bands [13,14]. However, compared MRI to ultrasonography, the review found no difference in sensitivity or speci city in identifying PAS [13].…”

Section: Introductionmentioning

confidence: 99%

Analysis of Related Influencing Factors of Hysterectomy in Patients With Placenta Accreta Spectrum Disorders After Uterine Artery Embolization

Zheng

Shen

Liu

et al. 2020

Preprint

View full text Add to dashboard Cite

Objective: To estimate the related factors of hysterectomy in patients with Placenta accreta spectrum (PAS) after uterine artery embolization (UAE), and try to evaluate the effectiveness and safety of UAE in patients with PAS.Methods: From January 2012 to July 2020, a retrospective analysis was performed in 85 patients undergoing TAE for PAS. Information regarding clinical data, angiography as well as embolization details, and clinical outcomes was obtained. Univariate and multivariate analyses were performed to determine the factors related to hysterectomy.Results: Bleeding greater than or equal to 500ml during the delivery(p = 0.037), the placenta type by MR or US(placenta increta vs placenta percreta, P = 0.01) and the type of ovarian artery(No vs Bilateral, P = 0.005; Unilateral vs Bilateral, P = 0.01) were independent risk factors of hysterectomy in PAS patients treated with UAE. The area under the curve (AUC) of the predictive model that incorporated the independent risk factors was 0.844. Abnormal collateral vessels communicating with uterine artery were observed on angiography in 24 patients (28.2%) with 31 abnormal collateral vessels. The major abnormal collateral vessel was the abnormal branches of the internal iliac artery(n = 13), followed by the inferior vesical artery (n = 11), internal pudendal artery(n = 3), obturator artery (n = 2), vaginal artery(n = 1) and the abnormal branches of the external iliac artery(n = 1).Conclusions: TAE is safe and effective for patients with PAS. Bleeding greater than or equal to 500ml during the delivery, the placenta type by MR or US and the type of ovarian artery were related to the hysterectomy. For patients with hypertrophic ovarian-uterine artery anastomosis and no fertility requirements, Ovarian artery embolization(OAE) could be a feasible choice.

show abstract

Identification of suspicious invasive placentation based on clinical MRI data using textural features and automated machine learning

Cited by 51 publications

References 32 publications

A prediction model for placenta accreta spectrum: A multicentre external validation study.

A prediction model for placenta accreta spectrum: A multicentre external validation study.

Artificial Intelligence in Pregnancy: A Scoping Review

Analysis of Related Influencing Factors of Hysterectomy in Patients With Placenta Accreta Spectrum Disorders After Uterine Artery Embolization

Contact Info

Product

Resources

About