RESUMO
Background: Early studies have demonstrated the potential of deep learning in bringing revolutionary changes in medical analysis. However, it is unknown which deep learning based diagnostic pattern is more effective for differentiating malignant and benign breast lesions (BLs) and can assist radiologists to reduce unnecessary biopsies. Methods: A total of 506 malignant BLs and 557 benign BLs were enrolled in this study after excluding incomplete ultrasound images. 396 malignant BLs and 447 benign BLs were included in the training cohort while 110 malignant and 110 benign BLs were included in the validation cohort. All BLs in the training and validation cohort were biopsy-proven. The most common convolutional neural networks (VGG-16 and VGG-19) were applied to identify malignant and benign BLs using grey-scale ultrasound images. Two radiologists determined the malignant (suggestion for biopsy) and benign (suggestion for follow-up) BLs with a 2-step reading session. The first step was based on conventional ultrasound (US) images alone to make a biopsy or follow-up decision. The second step was to take deep learning results into account for the decision adjustment. If a deep learning result of a first-classified benign BL was above the cut-off value, then it was re-classified as malignant. Results: In terms of area under the curve (AUC), the VGG-19 model yielded the best diagnostic performance in both training [0.939, 95% confidence interval (CI): 0.924-0.954] and testing dataset (0.959, 95% CI: 0.937-0.982). With the aid of deep learning models, the AUC of radiologists improved from 0.805 (95% CI: 0.744-0.865) to 0.827 (95% CI: 0.771-0.875, VGG-16) and 0.914 (95% CI: 0.871-0.957, VGG-19). The unnecessary biopsies decreased from 10.0% (11/110) to 8.2% (9/110) (assisted by VGG-16) and 0.9% (1/110) (assisted by VGG-19). Conclusions: The application of deep learning patterns in breast US may improve the diagnostic performance of radiologists by offering a second opinion. And thus, the assist of deep learning algorithm can considerably reduce the unnecessary biopsy rate in the clinical management of breast lesions.
RESUMO
OBJECTIVE: To set up a prediction rule for the pro-operative differential diagnosis of thyroid nodules and evaluate its clinical value. METHODS: All patients of thyroid nodules underwent thyroid operations in Changzheng hospital from June, 1997 to July, 2012 were included in this study. They were randomly divided into the derivation cohort (2/3) and the validation cohort (1/3). A prediction rule was developed based on the logistic regression model and the scoring system was established in accordance with assigning of the value of each variable ß in the model. The prediction consistency, discriminatory power and diagnostic accuracy were conducted to evaluate the clinical value of the scoring system. RESULTS: A total of 13 980 patients were enrolled in the study with 9195 in the derivation cohort and 4785 in the validation cohort. The prediction rule consisted of 18 variables, which were gender, clinical manifestations including fever, neck sore, neck mass, palpitation and sweating, serum level of thyroid stimulating hormone (TSH) , free triiodothyronine (FT3) , thyroid peroxidase antibody (TPOAb) , thyroglobulin antibody (TgAb) , thyroglobulin (Tg) , calcitonin and carcinoembryonic antigen (CEA) , ultrasonography features including nodules number, location, size, boundaries and ethological patterns and the presence and patterns of lymph nodes. The model showed good calibration consistency (P = 0.437) and discrimination power (area under the receiver operating characteristic curve was 0.928) in the derivation cohort. The sensitivity, specificity, accuracy, positive predictive value, negative predictive value, positive likelihood ratio and negative likelihood ratio of the model were 89.3%, 81.5%, 83.2%, 56.8%, 96.6%, 4.83 and 0.13, respectively. CONCLUSION: The prediction rule and its scoring system established in the study are efficacious for the calibration and discrimination of thyroid nodules in Chinese population, which could be a useful tool for the pro-operative risk stratification.
