Purpose To explore the value of tumor intratumoral and peritumoral radiomics models based on dynamic contrast-enhanced MRI in predicting benign and malignant breast tumors.Materials and Methods This retrospective study analyzed clinical data and MRI imaging features from 309 patients with pathologically confirmed solitary breast tumors at the Second Affiliated Hospital of Xiamen Medical College from August 2018 to December 2022.The cohort was randomly divided into training(n=248)and testing(n=61)cohorts in an 8∶2 ratio.Using second-phase dynamic contrast-enhanced MRI images,five distinct peritumoral regions were segmented through 3D-Slicer software:intratumoral region,peritumoral 2 mm region,peritumoral 4 mm region,intratumoral combined with peritumoral 2 mm region and intratumoral combined with peritumoral 4 mm region.Radiomic features were extracted from these regions of interest.Machine learning algorithms integrated with Logistic regression analysis were employed for feature selection,with cross-validation techniques determining the optimal radiomic signature combination.Results Of 309 patients,150 were benign tumors and 159 were malignant tumors.The age,maximum diameter and enhancement type of benign and malignant breast tumors were statistically significant(Z/χ2=-7.695,-5.775,30.248;all P<0.001).For the region of interest of intratumoral combined with peritumoral 4 mm region,the area under curve,sensitivity,specificity,accuracy and F1 scores of the training and validation were 0.893,89.1%,75.0%,82.3%,0.838 and 0.881,74.2%,86.7%,80.3%,0.793,respectively.In the training set,the area under curve of intratumoral combined with peritumoral 4 mm region was significantly higher than that of intratumoral,peritumoral 2 mm,peritumoral 4 mm and intratumoral combined with peritumoral 2 mm regional imaging models(Z=2.506,2.982,3.392,2.157;all P<0.05).The Hosmer-Lemeshow test showed that the calibration curve of the model combined with region of interest of intratumoral combined with peritumoral 4 mm region was highly consistent with the ideal curve(training P=0.381,validation P=0.159).The decision curve indicated that the net benefit of the radiomics model in the region of interest of intratumoral combined with peritumoral 4 mm region was the highest when the risk threshold was between 0 and 1.0.Conclusion The radiomics model has good predictive performance in predicting benign or malignant breast tumors,among them the combined region of interest of intratumoral combined with peritumoral 4 mm region provides the best performance and highest benefit.The technology clinical application will provide a non-invasive predictive method for the differential diagnosis of benign and malignant breast tumors.

Purpose To explore the value of tumor intratumoral and peritumoral radiomics models based on dynamic contrast-enhanced MRI in predicting benign and malignant breast tumors.Materials and Methods This retrospective study analyzed clinical data and MRI imaging features from 309 patients with pathologically confirmed solitary breast tumors at the Second Affiliated Hospital of Xiamen Medical College from August 2018 to December 2022.The cohort was randomly divided into training(n=248)and testing(n=61)cohorts in an 8∶2 ratio.Using second-phase dynamic contrast-enhanced MRI images,five distinct peritumoral regions were segmented through 3D-Slicer software:intratumoral region,peritumoral 2 mm region,peritumoral 4 mm region,intratumoral combined with peritumoral 2 mm region and intratumoral combined with peritumoral 4 mm region.Radiomic features were extracted from these regions of interest.Machine learning algorithms integrated with Logistic regression analysis were employed for feature selection,with cross-validation techniques determining the optimal radiomic signature combination.Results Of 309 patients,150 were benign tumors and 159 were malignant tumors.The age,maximum diameter and enhancement type of benign and malignant breast tumors were statistically significant(Z/χ2=-7.695,-5.775,30.248;all P<0.001).For the region of interest of intratumoral combined with peritumoral 4 mm region,the area under curve,sensitivity,specificity,accuracy and F1 scores of the training and validation were 0.893,89.1%,75.0%,82.3%,0.838 and 0.881,74.2%,86.7%,80.3%,0.793,respectively.In the training set,the area under curve of intratumoral combined with peritumoral 4 mm region was significantly higher than that of intratumoral,peritumoral 2 mm,peritumoral 4 mm and intratumoral combined with peritumoral 2 mm regional imaging models(Z=2.506,2.982,3.392,2.157;all P<0.05).The Hosmer-Lemeshow test showed that the calibration curve of the model combined with region of interest of intratumoral combined with peritumoral 4 mm region was highly consistent with the ideal curve(training P=0.381,validation P=0.159).The decision curve indicated that the net benefit of the radiomics model in the region of interest of intratumoral combined with peritumoral 4 mm region was the highest when the risk threshold was between 0 and 1.0.Conclusion The radiomics model has good predictive performance in predicting benign or malignant breast tumors,among them the combined region of interest of intratumoral combined with peritumoral 4 mm region provides the best performance and highest benefit.The technology clinical application will provide a non-invasive predictive method for the differential diagnosis of benign and malignant breast tumors.

Objective To explore the prevalence of age-related eye diseases (AREDs) among people aged 70 years or above in Hangzhou and evaluate the impact of AREDs on visual function (VF) and visual-related quality of life (QOL) in elders.Methods This study involved a total 2 111 elderly people (≥70 years).All participants received visual acuity and comprehensive eye examinations and complete VF and QoL questionnaires.Results The main cause of visual impairment was AREDs.And the causes were age-related cataract (79.82％),AMD (45.64％),glaucoma (10.95％) and diabetic retinopathy (DR,7.30％).VF and QoL scores declined gradually with age.And the scores declined more rapidly with declined visual acuity among the elders.VF and QoL scores in patients with age-related cataract,AMD,glaucoma and DR were successively lower.After adjusting for age,gender and visual acuity,the elders with AREDs had lower scores across all domains of VF and QoL.Scores for subscales of VF and QoL domains were significantly lower among those with DR and glaucoma compared with those with age-related cataract and AMD.Conclusion Age,presenting vision and AREDs are associated with VF and QoL in this elderly population.And senile people with glaucoma and DR have severe declines in VF and QoL,independent of presenting visual acuity.