Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods
Background: The prognosis of patients with early diagnosis of malignant endometrial lesions is good. We aimed to identify benign and malignant lesions in endometrial tissue, explore effective methods for assisting diagnosis, and improve the accuracy and precision of identifying en...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IMR Press
2025-01-01
|
| Series: | Clinical and Experimental Obstetrics & Gynecology |
| Subjects: | |
| Online Access: | https://www.imrpress.com/journal/CEOG/52/1/10.31083/CEOG25957 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832586566696960000 |
|---|---|
| author | Shanshan Li Jiali Wang Li Zhou Hui Wang Xiangyu Wang Jian Hu Qingxiu Ai |
| author_facet | Shanshan Li Jiali Wang Li Zhou Hui Wang Xiangyu Wang Jian Hu Qingxiu Ai |
| author_sort | Shanshan Li |
| collection | DOAJ |
| description | Background: The prognosis of patients with early diagnosis of malignant endometrial lesions is good. We aimed to identify benign and malignant lesions in endometrial tissue, explore effective methods for assisting diagnosis, and improve the accuracy and precision of identifying endometrial lesions. Methods: 1142 ultrasound radiomics and 18 clinical features from 1254 patients were analyzed, from which 36 features were selected for machine learning. We sketched the region of interest (ROI) of the abnormalities on the ultrasound images. Then, the radiomics features were extracted. Six common machine learning algorithms, including Support Vector Machine (SVM), Logistic Regression, Decision Tree, Random Forest, Gradient Boosting Tree, and k-Nearest Neighbors, were employed to identify benign and malignant changes in endometrial tissue. Cross-validation and grid search techniques for hyperparameter tuning were utilized to obtain the best model performance. Accuracy, precision, sensitivity, F1-scores, area under the curve (AUC), cross-validation average score and bootstrap average accuracy were also used to evaluate algorithm performance, classification accuracy, and generalization capability. Results: We combined 21 ultrasound characteristics and 15 clinical characteristics to develop and validate six common machine learning algorithms. After internal validation, the best models were the Random Forest models, with accuracy of 89%, precision of 93%, sensitivity of 97%, F1-score of 95%, and AUC of 95%, as well as a 10-fold cross-validation average score of 95% and bootstrap average accuracy of 94%, implying flawless classification in the test set. Conclusions: We identified the clinical and ultrasound features in the early diagnosis of benign or malignant lesions in endometrial tissue. And Random Forest model algorithms have demonstrated excellent performance in identifying benign and malignant changes in endometrial tissue. This is significant for enhancing early diagnostic accuracy and improving treatment outcomes and long-term management. |
| format | Article |
| id | doaj-art-440e13fb1e744b758527dcaae9908b51 |
| institution | Kabale University |
| issn | 0390-6663 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IMR Press |
| record_format | Article |
| series | Clinical and Experimental Obstetrics & Gynecology |
| spelling | doaj-art-440e13fb1e744b758527dcaae9908b512025-01-25T07:14:26ZengIMR PressClinical and Experimental Obstetrics & Gynecology0390-66632025-01-015212595710.31083/CEOG25957S0390-6663(24)02530-2Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning MethodsShanshan Li0Jiali Wang1Li Zhou2Hui Wang3Xiangyu Wang4Jian Hu5Qingxiu Ai6Department of Medical Ultrasound, The Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaDepartment of Medical Ultrasound, The Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaDepartment of Medical Ultrasound, The Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaDepartment of Medical Ultrasound, The Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaDepartment of Medical Ultrasound, The Maternal and Child Health and Family Planning Service Center of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaDepartment of Medical Ultrasound, The Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaDepartment of Medical Ultrasound, The Central Hospital of Enshi Tujia and Miao Autonomous Prefecture, 445000 Enshi, Hubei, ChinaBackground: The prognosis of patients with early diagnosis of malignant endometrial lesions is good. We aimed to identify benign and malignant lesions in endometrial tissue, explore effective methods for assisting diagnosis, and improve the accuracy and precision of identifying endometrial lesions. Methods: 1142 ultrasound radiomics and 18 clinical features from 1254 patients were analyzed, from which 36 features were selected for machine learning. We sketched the region of interest (ROI) of the abnormalities on the ultrasound images. Then, the radiomics features were extracted. Six common machine learning algorithms, including Support Vector Machine (SVM), Logistic Regression, Decision Tree, Random Forest, Gradient Boosting Tree, and k-Nearest Neighbors, were employed to identify benign and malignant changes in endometrial tissue. Cross-validation and grid search techniques for hyperparameter tuning were utilized to obtain the best model performance. Accuracy, precision, sensitivity, F1-scores, area under the curve (AUC), cross-validation average score and bootstrap average accuracy were also used to evaluate algorithm performance, classification accuracy, and generalization capability. Results: We combined 21 ultrasound characteristics and 15 clinical characteristics to develop and validate six common machine learning algorithms. After internal validation, the best models were the Random Forest models, with accuracy of 89%, precision of 93%, sensitivity of 97%, F1-score of 95%, and AUC of 95%, as well as a 10-fold cross-validation average score of 95% and bootstrap average accuracy of 94%, implying flawless classification in the test set. Conclusions: We identified the clinical and ultrasound features in the early diagnosis of benign or malignant lesions in endometrial tissue. And Random Forest model algorithms have demonstrated excellent performance in identifying benign and malignant changes in endometrial tissue. This is significant for enhancing early diagnostic accuracy and improving treatment outcomes and long-term management.https://www.imrpress.com/journal/CEOG/52/1/10.31083/CEOG25957endometrial lesionsearly diagnosisultrasound radiomicsclinical featuresmachine learning |
| spellingShingle | Shanshan Li Jiali Wang Li Zhou Hui Wang Xiangyu Wang Jian Hu Qingxiu Ai Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods Clinical and Experimental Obstetrics & Gynecology endometrial lesions early diagnosis ultrasound radiomics clinical features machine learning |
| title | Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods |
| title_full | Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods |
| title_fullStr | Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods |
| title_full_unstemmed | Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods |
| title_short | Diagnosis of Malignant Endometrial Lesions from Ultrasound Radiomics Features and Clinical Variables Using Machine Learning Methods |
| title_sort | diagnosis of malignant endometrial lesions from ultrasound radiomics features and clinical variables using machine learning methods |
| topic | endometrial lesions early diagnosis ultrasound radiomics clinical features machine learning |
| url | https://www.imrpress.com/journal/CEOG/52/1/10.31083/CEOG25957 |
| work_keys_str_mv | AT shanshanli diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods AT jialiwang diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods AT lizhou diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods AT huiwang diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods AT xiangyuwang diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods AT jianhu diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods AT qingxiuai diagnosisofmalignantendometriallesionsfromultrasoundradiomicsfeaturesandclinicalvariablesusingmachinelearningmethods |