Discriminative diagnosis of ovarian endometriosis cysts and benign mucinous cystadenomas based on the ConvNeXt algorithm.

PURPOSE: The objective of this study was to develop a deep learning model, using the ConvNeXt algorithm, that can effectively differentiate between ovarian endometriosis cysts (OEC) and benign mucinous cystadenomas (MC) by analyzing ultrasound images. The performance of the model in the diagnostic differentiation of these two conditions was also evaluated. METHODS: A retrospective analysis was conducted on OEC and MC patients who had sought medical attention at the Fourth Affiliated Hospital of Harbin Medical University between August 2018 and May 2023. The diagnosis was established based on postoperative pathology or the characteristics of aspirated fluid guided by ultrasound, serving as the gold standard. Ultrasound images were collected and subjected to screening and preprocessing procedures. The data set was randomly divided into training, validation, and testing sets in a ratio of 5:3:2. Transfer learning was utilized to determine the initial weights of the ConvNeXt deep learning algorithm, which were further adjusted by retraining the algorithm using the training and validation ultrasound images to establish a new deep learning model. The weights that yielded the highest accuracy were selected to evaluate the diagnostic performance of the model using the validation set. Receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC) was calculated. Additionally, sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio, negative likelihood ratio, and odds ratio were calculated. Decision curve analysis (DCA) curves were plotted. RESULTS: The study included 786 ultrasound images from 184 patients diagnosed with either OEC or MC. The deep learning model achieved an AUC of 0.90 (95 % CI: 0.85-0.95) in accurately distinguishing between the two conditions, with a sensitivity of 90 % (95 % CI: 84 %-95 %), specificity of 90 % (95 % CI: 77 %-97 %), a positive predictive value of 96 % (95 % CI: 91 %-99 %), a negative predictive value of 77 % (95 % CI: 63 %-88 %), a positive likelihood ratio of 9.27 (95 % CI: 3.65-23.56), and a negative likelihood ratio of 0.11 (95 % CI: 0.06-0.19). The DCA curve demonstrated the practical clinical utility of the model. CONCLUSIONS: The deep learning model developed using the ConvNeXt algorithm exhibits high accuracy (90 %) in distinguishing between OEC and MC. This model demonstrates excellent diagnostic performance and clinical utility, providing a novel approach for the clinical differentiation of these two conditions.

Prediction model of adnexal masses with complex ultrasound morphology.

Background: Based on the ovarian-adnexal reporting and data system (O-RADS), we constructed a nomogram model to predict the malignancy potential of adnexal masses with sophisticated ultrasound morphology. Methods: In a multicenter retrospective study, a total of 430 subjects with masses were collected in the adnexal region through an electronic medical record system at the Fourth Hospital of Harbin Medical University during the period of January 2019-April 2023. A total of 157 subjects were included in the exception validation cohort from Harbin Medical University Tumor Hospital. The pathological tumor findings were invoked as the gold standard to classify the subjects into benign and malignant groups. All patients were randomly allocated to the validation set and training set in a ratio of 7:3. A stepwise regression analysis was utilized for filtering variables. Logistic regression was conducted to construct a nomogram prediction model, which was further validated in the training set. The forest plot, C-index, calibration curve, and clinical decision curve were utilized to verify the model and assess its accuracy and validity, which were further compared with existing adnexal lesion models (O-RADS US) and assessments of different types of neoplasia in the adnexa (ADNEX). Results: Four predictors as independent risk factors for malignancy were followed in the preparation of the diagnostic model: O-RADS classification, HE4 level, acoustic shadow, and protrusion blood flow score (all p < 0.05). The model showed moderate predictive power in the training set with a C-index of 0.959 (95%CI: 0.940-0.977), 0.929 (95%CI: 0.884-0.974) in the validation set, and 0.892 (95%CI: 0.843-0.940) in the external validation set. It showed that the predicted consequences of the nomogram agreed well with the actual results of the calibration curve, and the novel nomogram was clinically beneficial in decision curve analysis. Conclusion: The risk of the nomogram of adnexal masses with complex ultrasound morphology contained four characteristics that showed a suitable predictive ability and provided better risk stratification. Its diagnostic performance significantly exceeded that of the ADNEX model and O-RADS US, and its screening performance was essentially equivalent to that of the ADNEX model and O-RADS US classification.

Prediction of benign and malignant ovarian tumors using Resnet34 on ultrasound images.

OBJECTIVE: To develop deep learning (DL) prediction models using transvaginal ultrasound (TVS), transabdominal ultrasound (TAS), and color Doppler flow imaging (CDFI) of TVS (CDFI_TVS) to automatically predict benign or malignant ovarian tumors. METHODS: This retrospective study included women with ovarian tumors who underwent ultrasound between August 2018 and October 2022. Histopathological analysis was used as a reference standard. The dataset was preprocessed by clipping, flipping, and rotating images to generate a larger, more complicated, and diverse dataset to improve accuracy and generalizability. The dataset was then divided into training (80%) and test (20%) sets. The weights of the models, modified from the residual network (ResNet) with the TVS, TAS, and CDFI_TVS images (hereafter, referred to as DLTVS , DLTAS , and DLCDFI_TVS , respectively) were developed. The area under the receiver operating characteristic curve (AUC) analysis in the test set was used to compare the predictive value of DL for malignancy. RESULTS: A total of 2340 images from 1350 women with adnexal masses were included. DLTVS had an AUC of 0.95 (95% CI: 0.93-0.97) for classifying malignant and benign ovarian tumors, comparable with that of DLTAS (AUC, 0.95; 95% CI: 0.91-0.98; p = 0.96) and DLCDFI_TVS (AUC, 0.88; 95% CI: 0.84-0.93; p = 0.02). Decision curve analysis indicated that DLTVS performed better than DLTAS and DLCDFI_TVS . CONCLUSION: We developed DL models based on TVS, TAS, and CDFI_TVS on ultrasound images to predict benign and malignant ovarian tumors with high diagnostic performance. The DLTVS model had the best prediction compared with the DLTAS and DLCDFI_TVS models.

Iodine nutrition and prevalence of sonographic thyroid findings in adults in the Heilongjiang Province, China.

BACKGROUND: Thyroid disease and thyroid nodules are common clinical problems. Iodine nutrition plays an important role in thyroid disease evolution. Here, we aimed to estimate the iodine nutritional status and prevalence of thyroid disease in the adults of the Heilongjiang Province in northeast China. METHODS: We performed a cross-sectional ultrasound (US)-based survey on volunteers aged 20-70 years from 30 regions of the Heilongjiang Province. The participants were recruited using the probability proportional to size (PPS) method, and consent for US screening was obtained from them. The survey was performed by trained technicians using the same US equipment with a 6-15 MHz linear transducer (MyLab 30 cv, Italy) and was hosted in public community locations such as local hospitals and outpatient departments. Information on basic demographic characteristics, such as urinary iodine and iodine intake were collected. The age- and sex-adjusted prevalence of thyroid disease was determined through direct standardization and reported using the province's population in 2016 as reference. RESULTS: From December 12, 2017, to March 10, 2019, 3,754 participants with a mean age of 48.65 (±12.39) years participated in the study. Of them, 3,643 had reliable urinary iodine data. The median urinary iodine and salt iodine concentrations within the normal range were 163.30 µg/L and 24.30 mg/kg, respectively. The age- and sex-adjusted prevalence of thyroid disease was 52.91%. Diffuse thyroid disease (DTD), focal thyroid lesions (FTL), and coexistence of both diseases were prevalent in 8.68%, 36.58%, and 7.65% of the participants, respectively. The prevalence of the five categories according to US-based survey features in the ACR TI-RADS (i.e., TR1, TR2, TR3, TR4, and TR5) was 7.71%, 14.53%, 3.44%, 14.82%, and 3.51%, and the prevalence of nodules that needed fine-needle aspiration was 2.55%. CONCLUSIONS: In Heilongjiang Province, adults aged 20-70 years belong to the optimal iodine status. Further, the salt iodine levels are in the normal range. Thyroid diseases are highly prevalent in this age group; however, the intervention rate is low. We provided population-based estimates of the prevalence of thyroid disease and the iodine status in adults of Heilongjiang Province. These findings are useful to support effective intervention planning for thyroid disease.