Predicting the efficacy of microwave ablation of benign thyroid nodules from ultrasound images using deep convolutional neural networks

Abstract

Background: Thyroid nodules are frequent in clinical settings, and their diagnosis in adults is growing, with some persons experiencing symptoms. Ultrasound-guided thermal ablation can shrink nodules and alleviate discomfort. Because the degree and rate of lesion absorption vary greatly between individuals, there is no reliable model for predicting the therapeutic efficacy of thermal ablation.

Methods: Five convolutional neural network models including VGG19, Resnet 50, EfficientNetB1, EfficientNetB0, and InceptionV3, pre-trained with ImageNet, were compared for predicting the efficacy of ultrasound-guided microwave ablation (MWA) for benign thyroid nodules using ultrasound data. The patients were randomly assigned to one of two data sets: training (70%) or validation (30%). Accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve (AUC) were all used to assess predictive performance.

Results: In the validation set, fine-tuned EfficientNetB1 performed best, with an AUC of 0.85 and an ACC of 0.79.

Conclusions: The study found that our deep learning model accurately predicts nodules with VRR < 50% after a single MWA session. Indeed, when thermal therapies compete with surgery, anticipating which nodules will be poor responders provides useful information that may assist physicians and patients determine whether thermal ablation or surgery is the preferable option. This was a preliminary study of deep learning, with a gap in actual clinical applications. As a result, more in-depth study should be undertaken to develop deep-learning models that can better help clinics. Prospective studies are expected to generate high-quality evidence and improve clinical performance in subsequent research.

Keywords: Benign thyroid nodules; Convolutional neural networks; Deep learning; Thermal ablation; Volume reduction rate.

Fig. 1

(A) Schematic diagram of the patient selection. (B) An original random image (C) an augmented random image. VRR, volume reduction rate

Fig. 2

Flowchart of deep learning model construction. Workflow of deep learning model construction for predicting efficacy of thermal ablation in patients with thyroid nodules

Fig. 3

ROC curve of the algorithms used in building the Deep learning models. Green and blue lines represent the validation and train area respectively. ROC, receiver operating characteristic. (A) InceptionV3 (B) VGG19 (C) ResNet50 (D) EfficientNetB1 (E) EfficientNetB0 (F) EfficientNetB1 (fine tune)

Fig. 4

Accuracy curve of the algorithms used in building the deep learning models. Blue and yellow curves represent the training and validation cohort respectively. (A) InceptionV3 (B) VGG19 (C) ResNet50 (D) EfficientNetB1 (E) EfficientNetB0

Fig. 5

Loss curves of the algorithms used in building the deep learning models. Blue and yellow curves represent the training and validation cohort respectively. (A) InceptionV3 (B) VGG19 (C) ResNet50 (D) EfficientNetB1 (E) EfficientNetB0

Fig. 6

Decision curve analysis of the deep learning models for the training cohort (A) and the validation cohort (B)

Fig. 7

Confusion matrix. The 2 × 2 contingency table reports the number of true positives, false positives, false negatives, and true negatives; training cohort (A) and the validation cohort (B)

Fig. 8

Accuracy and loss curves of EfficientNetB1 model after fine-tuning. Blue and yellow curves represent the training and validation cohorts respectively and the green line represents where fine tune starts