Artificial Intelligence in Ultrasound Diagnoses of Ovarian Cancer: A Systematic Review and Meta-Analysis

Mitchell, Sian; Nikolopoulos, Manolis; El-Zarka, Alaa; Al-Karawi, Dhurgham; Al-Zaidi, Shakir; Ghai, Avi; Gaughran, Jonathan E.; Sayasneh, Ahmad

doi:10.3390/cancers16020422

Cited by 3 publications

(1 citation statement)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These could include artificial intelligence approaches and a focus on indeterminate adnexal masses, particularly those with solid components, that are the most difficult to assess even for expert examiners or even where logistic models do not help. 6,33 On the other hand, prospective studies are needed to determine the ideal approach and which 3D vascular indices cut-offs should be used; clearly, a consensus about methodology and diagnostic criteria is needed.…”

Section: Future Research Agendamentioning

confidence: 99%

The Role of Three-dimensional Power Doppler for Detecting Ovarian Cancer in Adnexal Masses: A Systematic Review and Meta-analysis

Sánchez,

Verdú,

Carbonell

et al. 2024

Oncol Adv

View full text Add to dashboard Cite

Background and objectives: Three-dimensional power Doppler (3DPD) ultrasound has been used for assessing adnexal masses, and in this study, we aimed to perform a meta-analysis to evaluate its role in the differential diagnosis of adnexal masses. Methods:A search for primary studies assessing the diagnostic performance of 3DPD in discriminating benign from malignant masses carried out between January 1990 and May 2023 was performed in Medline (PubMed), Scopus, and Web of Science databases with study quality evaluated using QUADAS-2. Results:We identified 404 citations. Ultimately, 18 studies comprising 2,975 women were included, and the mean prevalence of malignant lesions was 37%. In most cases, the quality of studies was moderate. Overall, pooled sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of 3DPD vascular tree assessment for studies including any type of mass were 77% (95% confidence interval [CI] = 52%-91%), 80% (95% CI = 37%-97%), 3.9 (95% CI = 0.7-20.9), and 0.29 (95% CI = 0.10-0.81), respectively. Heterogeneity was high for both sensitivity and specificity. Pooled sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of 3DPD vascular tree assessment for studies including only "complex" or "suspicious" adnexal masses were 90% (95% CI = 82%-94%), 88% (95% CI = 74%-95%), 7.3 (95% CI = 3.2-16.4), and 0.12 (95% CI = 0.06-0.22), respectively. Heterogeneity was moderate for both sensitivity and specificity. We could not perform quantitative synthesis for studies estimating 3D vascular indexes. Conclusions:The diagnostic performance of 3DPD for discriminating benign from malignant adnexal masses is good, and there is great heterogeneity in diagnostic criteria when using this technique.

show abstract

Section: Future Research Agendamentioning

confidence: 99%

The Role of Three-dimensional Power Doppler for Detecting Ovarian Cancer in Adnexal Masses: A Systematic Review and Meta-analysis

Sánchez,

Verdú,

Carbonell

et al. 2024

Oncol Adv

View full text Add to dashboard Cite

show abstract

Intelligent system based on multiple networks for accurate ovarian tumor semantic segmentation

El-khatib,

Popescu,

Teodor

et al. 2024

Heliyon

View full text Add to dashboard Cite

An XAI Approach to Melanoma Diagnosis: Explaining the Output of Convolutional Neural Networks with Feature Injection

Grignaffini,

De Santis,

Frezza

et al. 2024

Information

View full text Add to dashboard Cite

Computer-aided diagnosis (CAD) systems, which combine medical image processing with artificial intelligence (AI) to support experts in diagnosing various diseases, emerged from the need to solve some of the problems associated with medical diagnosis, such as long timelines and operator-related variability. The most explored medical application is cancer detection, for which several CAD systems have been proposed. Among them, deep neural network (DNN)-based systems for skin cancer diagnosis have demonstrated comparable or superior performance to that of experienced dermatologists. However, the lack of transparency in the decision-making process of such approaches makes them “black boxes” and, therefore, not directly incorporable into clinical practice. Trying to explain and interpret the reasons for DNNs’ decisions can be performed by the emerging explainable AI (XAI) techniques. XAI has been successfully applied to DNNs for skin lesion image classification but never when additional information is incorporated during network training. This field is still unexplored; thus, in this paper, we aim to provide a method to explain, qualitatively and quantitatively, a convolutional neural network model with feature injection for melanoma diagnosis. The gradient-weighted class activation mapping and layer-wise relevance propagation methods were used to generate heat maps, highlighting the image regions and pixels that contributed most to the final prediction. In contrast, the Shapley additive explanations method was used to perform a feature importance analysis on the additional handcrafted information. To successfully integrate DNNs into the clinical and diagnostic workflow, ensuring their maximum reliability and transparency in whatever variant they are used is necessary.

show abstract

Artificial Intelligence in Ultrasound Diagnoses of Ovarian Cancer: A Systematic Review and Meta-Analysis

Cited by 3 publications

References 46 publications

The Role of Three-dimensional Power Doppler for Detecting Ovarian Cancer in Adnexal Masses: A Systematic Review and Meta-analysis

The Role of Three-dimensional Power Doppler for Detecting Ovarian Cancer in Adnexal Masses: A Systematic Review and Meta-analysis

Intelligent system based on multiple networks for accurate ovarian tumor semantic segmentation

An XAI Approach to Melanoma Diagnosis: Explaining the Output of Convolutional Neural Networks with Feature Injection

Contact Info

Product

Resources

About