摘要
Abstract
Objective To explore the relationship between ultrasound radiomics and clinicopathological characteristics of ovarian cancer and its predictive value for positive expression of Ki-67.Methods A retrospective cohort of 126 ovarian cancer patients treated at a certain hospital between July 2021 and July 2024 was included into a training set.Based on the Ki-67 expression status,the training set was divided into a positive group(n=93)and a negative group(n=33).An additional 80 ovarian cancer patients treated at the same hospital during the same period were enrolled into a test set.The clinical data of the training and test sets were compared,and inter-group comparison between the positive and negative groups was carried out in terms of clinicopathological and conventional ultrasound characteristics,radiomics characteristics.The factors influencing Ki-67 positive expression were investigated with the multivariate regression analysis,which were incorporated into a logistic regression model for fitting to yield a clinical parameter model.The ultrasound radiomics characteristics with intra-group correlation coefficients>0.8 were included in least absolute shrinkage and selection operator(LASSO)regression for dimensionality reduction and final feature selection,then the selected characteristics were integrated into multiple logistic regression analysis to construct an ultrasound radiomics model.A combined model was established based on the ultrasound radiomics scores(Rad-score)and clinicopathological characteristics.Receiver operating characteristic(ROC)curves,calibration curves and clinical decision curves were plotted to evaluate the predictive performance of the three models.Spearman's correlation analysis was used to examine the relationship between Rad-score and clinicopathological characteristics.Statistical analysis was performed using SPSS 22.0 software.Results There was no statistically significant difference in clinical data between the training set and the test set(P>0.05).In the positive group the patients with FIGO stage Ⅲ-Ⅳ cancer,lymph node metastasis,distant or peripheral tissue infiltration,low differentiation and tumor maximum diameter≥10 cm were more than those in the negative group,with the difference being significant(P<0.05).There were more patients with unclear boundaries,irregular morphology and blood flow signals in the positive group than in the negative group,and the patients in the positive group had higher resistance indexes than those in the negative group,with the differences being statistically significant(P<0.05).Multivariate logistic regression analysis revealed the risk factors for Ki-67 positive expression in ovarian cancer patients included distant or peripheral tissue infiltration,lymph node metastasis,FIGO stage Ⅲ-Ⅳ cancer,tumor maximum diameter≥10 cm and low differentiation(P<0.05).The ROC curves showed the AUC values for the clinical parameter model,ultrasound radiomics model and combined model were 0.805,0.856 and 0.927 respectively,all demonstrating high discriminatory power,and the combined model outperformed the individual models;the calibration curves proved the predicted values from all the three models had high agreement with the observed values,and the combined model gained higher accuracy than the others;the clinical decision curve analysis indicated the three models were gifted with high net benefit values,and the combined model behaved better than the others.Rad-scores was negatively correlated with clinicopathological characteristics,and the Rad-score showed high correlations with FIGO staging(r=-0.59,P<0.001),presence of lymph node metastasis(r=-0.87,P<0.001),presence of peripheral tissue infiltration(r=-0.82,P<0.001),differentiation(r=-0.58,P<0.001)and tumor maximum diameter(r=-0.78,P<0.001).Conclusion Ultrasound radiomics characteristics in ovarian cancer patients correlate with clinicopathological characteristics,and the combined model incorporating clinical parameters and ultrasound radiomics demonstrates high predictive value for Ki-67 positive expression.[Chinese Medical Equipment Journal,2026,47(4):73-81]关键词
超声影像组学/卵巢癌/Ki-67阳性表达/临床病理特征Key words
ultrasound radiomics/ovarian cancer/Ki-67 positive expression/clinicopathological characteristic分类
医药卫生
