Performance of imaging interpretation, intra- and inter-reader agreement for diagnosis of pelvic endometriosis: comparison between an abbreviated and full MRI protocol

Silva, Lorena Luryann Cartaxo da; Torres, Ulysses S.; Torres, Lucas Rios; Fong, Marina Shu; Okuyama, Fabio Hiroshi; Caiado, Angela Hissae Motoyama; Chamié, Luciana Pardini; Moura, Ana Paula Carvalhal; Novis, Maria Inês; Warmbrand, Gisele; D'Ippólito, Giuseppe

doi:10.1007/s00261-021-03052-5

Cited by 5 publications

(3 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Indeed, such discrepancies between more and less experienced readers were found in other reproducibility studies in other studies addressing pelvic imaging, for example, especially in MRI of endometriosis. 29,30 Jaramillo-Cardoso et al 29 have found agreement between observers presenting different experience levels, ranging from weak to moderate in the MRI-based qualitative assessment of the extent of pelvic involvement caused by endometriosis. da Silva et al, 30 likewise, found a considerable heterogeneity in the distribution of values of inter-reader agreement for diagnosis of endometriosis according to the anatomical site of involvement.…”

Section: Discussionmentioning

confidence: 96%

Reproducibility of MRI Features of Uterine Leiomyomas: A Study on Interobserver Agreement and Inter-Method Agreement With Surgery

et al. 2021

Self Cite

View full text Add to dashboard Cite

Purpose: To evaluate interobserver agreement in the interpretation of different MRI features of uterine leiomyomas (UL) according to observers’ experience, and to assess the inter-method reproducibility (MRI versus surgery) regarding the International Federation of Gynecology and Obstetrics (FIGO) classification. Methods: Retrospective study including UL patients who underwent MRI and surgical treatment. Four blinded observers (2 vs >10 years of experience) assessed UL regarding dimensions and volume; inner and outer mantles; FIGO classification; vascularization; degeneration; and diffusion-weighted imaging features. Uterine dimensions and volume were calculated. FIGO classification as ascertained by observers was compared to surgical findings. Intraclass correlation coefficient (ICC) estimates were used for interobserver comparison of numerical variables, and kappa statistic for categorical variables. Results: Thirty-five patients (26y-73y) with 61 UL were included in the interobserver analyses, and 31 patients (54 UL) had available data allowing retrospective surgical FIGO classification for assessment of inter-method reproducibility. Both groups of observers had good to excellent agreement in assessing UL (ICC = 0.980-0.994) and uterine volumes (ICC = 0.857-0.914), mantles measurement (ICC = 0.797-0.920), and apparent diffusion coefficient calculation (ICC = 0.787-0.883). There was substantial agreement for both groups regarding FIGO classification (κ = 0.645-0.767). Vascularization, degeneration and restricted diffusion had lower agreement, varying from reasonable to moderate. Inter-method agreement was reasonable (κ = 0.341-0.395). Conclusions: Interobserver agreement of MRI for UL was higher for quantitative than qualitative features, with a little impact of observers’ experience for most features. MRI agreement with surgery was reasonable. Further efforts should be taken to improve interobserver and inter-method reproducibility for MRI in this scenario.

show abstract

Section: Discussionmentioning

confidence: 96%

Reproducibility of MRI Features of Uterine Leiomyomas: A Study on Interobserver Agreement and Inter-Method Agreement With Surgery

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Post-contrast images are not mandatory for DIE evaluation but can be used in cases of complex ovarian or adnexial cysts, or when the possibility of malignant transformation must be ruled out 25 . Recently, we demonstrated that an abreviated protocol including a volumetric coronal T2-weighted fast-spin-echo sequence and axial T1-weighted gradient-echo sequence with fat suppression had similar performance for diagnosing multiple sites of DIE when compared with the full protocol 26 .…”

Section: Patient Preparation and Imaging Techniquementioning

confidence: 99%

Imaging diagnosis of endometriosis

Chamié¹

2022

JMEXFRI

View full text Add to dashboard Cite

Endometrosis is a common gynecologic disease that affects women of reproductive age and commonly causes pelvic pain and infertility. The most common types are superficial peritoneal implants, ovarian endometriotic cysts (endometriomas), and deep-infiltrating subperitoneal lesions. Diagnosis is often delayed, and up to 65% of women are initially misdiagnosed. Imaging methods play a pivotal role in patient counseling and clinical management of the disease, and have been replacing diagnostic laparoscopy in specialized centers worldwide. Comprehensive imaging mapping is required for adequate surgical planning and to assist fertility doctors in determining the appropriate treatment options. Transvaginal ultrasound (TVU) is the first-line imaging modality and is a widely available tool that enables accurate diagnosis of endometriosis when a dedicated protocol is used. TVU is the best imaging method to investigate multiple bowel lesions and small peritoneal implants. Magnetic resonance imaging (MRI) is an excellent multiplanar method for evaluating the pelvic cavity and extrapelvic sites in endometriosis. It is the method of choice to differentiate ovarian cysts and investigate endometriosis affecting the ureters, pelvic nerves, pelvic floor, and diaphragm. This In-Depth Review describes the imaging protocol and findings of TVU and MRI to investigate endometriosis.

show abstract

“…Medical images are routinely used in clinical practice to detect and to diagnose diseases including cancer. However, reading and interpreting medical images is often a difficult and time-consuming task for radiologists, which does not only reduce diagnostic accuracy but also generates large intra- and inter-reader variability [ 1 ]. For example, full-field digital mammography (FFDM) is the most popular imaging modality used in the general population-based breast cancer screening in order to detect breast cancer at an early stage.…”

Section: Introductionmentioning

confidence: 99%

A Comparison of Computer-Aided Diagnosis Schemes Optimized Using Radiomics and Deep Transfer Learning Methods

et al. 2022

View full text Add to dashboard Cite

Objective: Radiomics and deep transfer learning are two popular technologies used to develop computer-aided detection and diagnosis (CAD) schemes of medical images. This study aims to investigate and to compare the advantages and the potential limitations of applying these two technologies in developing CAD schemes. Methods: A relatively large and diverse retrospective dataset including 3000 digital mammograms was assembled in which 1496 images depicted malignant lesions and 1504 images depicted benign lesions. Two CAD schemes were developed to classify breast lesions. The first scheme was developed using four steps namely, applying an adaptive multi-layer topographic region growing algorithm to segment lesions, computing initial radiomics features, applying a principal component algorithm to generate an optimal feature vector, and building a support vector machine classifier. The second CAD scheme was built based on a pre-trained residual net architecture (ResNet50) as a transfer learning model to classify breast lesions. Both CAD schemes were trained and tested using a 10-fold cross-validation method. Several score fusion methods were also investigated to classify breast lesions. CAD performances were evaluated and compared by the areas under the ROC curve (AUC). Results: The ResNet50 model-based CAD scheme yielded AUC = 0.85 ± 0.02, which was significantly higher than the radiomics feature-based CAD scheme with AUC = 0.77 ± 0.02 (p < 0.01). Additionally, the fusion of classification scores generated by the two CAD schemes did not further improve classification performance. Conclusion: This study demonstrates that using deep transfer learning is more efficient to develop CAD schemes and it enables a higher lesion classification performance than CAD schemes developed using radiomics-based technology.

show abstract

Performance of imaging interpretation, intra- and inter-reader agreement for diagnosis of pelvic endometriosis: comparison between an abbreviated and full MRI protocol

Cited by 5 publications

References 36 publications

Reproducibility of MRI Features of Uterine Leiomyomas: A Study on Interobserver Agreement and Inter-Method Agreement With Surgery

Reproducibility of MRI Features of Uterine Leiomyomas: A Study on Interobserver Agreement and Inter-Method Agreement With Surgery

Imaging diagnosis of endometriosis

A Comparison of Computer-Aided Diagnosis Schemes Optimized Using Radiomics and Deep Transfer Learning Methods

Contact Info

Product

Resources

About