UDRSNet: An unsupervised deformable registration module based on image structure similarity.

BACKGROUND: Image registration is a challenging problem in many clinical tasks, but deep learning has made significant progress in this area over the past few years. Real-time and robust registration has been made possible by supervised transformation estimation. However, the quality of registrations using this framework depends on the quality of ground truth labels such as displacement field. PURPOSE: To propose a simple and reliable method for registering medical images based on image structure similarity in a completely unsupervised manner. METHODS: We proposed a deep cascade unsupervised deformable registration approach to align images without reliable clinical data labels. Our basic network was composed of a displacement estimation module (ResUnet) and a deformation module (spatial transformer layers). We adopted l 2 $l_2$ -norm to regularize the deformation field instead of the traditional l 1 $l_1$ -norm regularization. Additionally, we utilized structural similarity (ssim) estimation during the training stage to enhance the structural consistency between the deformed images and the reference images. RESULTS: Experiments results indicated that by incorporating ssim loss, our cascaded methods not only achieved higher dice score of 0.9873, ssim score of 0.9559, normalized cross-correlation (NCC) score of 0.9950, and lower relative sum of squared difference (SSD) error of 0.0313 on CT images, but also outperformed the comparative methods on ultrasound dataset. The statistical t $t$ -test results also proved that these improvements of our method have statistical significance. CONCLUSIONS: In this study, the promising results based on diverse evaluation metrics have demonstrated that our model is simple and effective in deformable image registration (DIR). The generalization ability of the model was also verified through experiments on liver CT images and cardiac ultrasound images.

Coplanarity Constrained Ultrasound Probe Calibration Based on N-Wire Phantom.

OBJECTIVE: N-wire phantom-based ultrasound probe calibration has been used widely in many freehand tracked ultrasound imaging systems. The calibration matrix is obtained by registering the coplanar point cloud in ultrasound space and non-coplanar point cloud in tracking sensor space based on the least squares method. This method is sensitive to outliers and loses the coplanar information of the fiducial points. In this article, we describe a coplanarity-constrained calibration algorithm focusing on these issues. METHODS: We verified that the out-of-plane error along the oblique wire in the N-wire phantom followed a normal distribution and used it to remove the experimental outliers and fit the plane with the Levenberg-Marquardt algorithm. Then, we projected the points to the plane along the oblique wire. Coplanarity-constrained point cloud registration was used to calculate the transformation matrix. RESULTS: Compared with the other two commonly used methods, our method had the best calibration precision and achieved 25% and 36% improvement of the mean calibration accuracy than the closed-form solution and in-plane error method respectively at depth 16. Experiments at different depths revealed that our algorithm had better performance in our setup. CONCLUSION: Our proposed coplanarity-constrained calibration algorithm achieved significant improvement in both precision and accuracy compared with existing algorithms with the same N-wire phantom. It is expected that calibration accuracy will improve when the algorithm is applied to all other N-wire phantom-based calibration procedures.

Assessment of the statistical optimization strategies and clinical evaluation of an artificial intelligence-based automated diagnostic system for thyroid nodule screening.

Background: Thyroid cancer is the most common endocrine cancer in the world. Accurately distinguishing between benign and malignant thyroid nodules is particularly important for the early diagnosis and treatment of thyroid cancer. This study aimed to investigate the best possible optimization strategies for an already-trained artificial intelligence (AI)-based automated diagnostic system for thyroid nodule screening and, in addition, to scrutinize the clinically relevant limitations using stratified analysis to better standardize the application in clinical workflows. Methods: We retrospectively reviewed a total of 1,092 ultrasound images associated with 397 thyroid nodules collected from 287 patients between April 2019 and January 2021, applying postoperative pathology as the gold standard. We applied different statistical approaches, including averages, maximums, and percentiles, to estimate per-nodule-based malignancy scores from the malignancy scores per image predicted by AI-SONIC Thyroid v. 5.3.0.2 (Demetics Medical Technology Ltd., Hangzhou, China) system, and we assessed its diagnostic efficacy on nodules of different sizes or tumor types with per-nodule analysis using performance metrics. Results: Of the 397 thyroid nodules, 272 thyroid nodules were overrepresented by malignant nodules according to the results of the surgical pathological examinations. Taking the median of the malignancy scores per image to estimate the nodule-based score with a cutoff value of 0.56 optimized for the means of sensitivity and specificity, the AI-based automated detection system demonstrated slightly lower sensitivity, significantly higher specificity (almost independent of nodule size), and similar accuracy to that of the senior radiologist. Both the AI system and the senior radiologist demonstrated higher sensitivity in diagnosing smaller nodules (≤25 mm) and comparable diagnostic performances for larger nodules. The mean diagnostic time per nodule of the AI system was 0.146 s, which was in sharp contrast to the 2.8 to 4.5 min of the radiologists. Conclusions: Using our optimization strategy to achieve nodule-based diagnosis, the AI-SONIC Thyroid automated diagnostic system demonstrated an overall diagnostic accuracy equivalent to that of senior radiologists. Thus, it is expected that it can be used as a reliable auxiliary diagnostic method by radiologists for the screening and preoperative evaluation of malignant thyroid nodules.

An artificial intelligence ultrasound system's ability to distinguish benign from malignant follicular-patterned lesions.

Objectives: To evaluate the application value of a generally trained artificial intelligence (AI) automatic diagnosis system in the malignancy diagnosis of follicular-patterned thyroid lesions (FPTL), including follicular thyroid carcinoma (FTC), adenomatoid hyperplasia nodule (AHN) and follicular thyroid adenoma (FTA) and compare the diagnostic performance with radiologists of different experience levels. Methods: We retrospectively reviewed 607 patients with 699 thyroid nodules that included 168 malignant nodules by using postoperative pathology as the gold standard, and compared the diagnostic performances of three radiologists (one junior, two senior) and that of AI automatic diagnosis system in malignancy diagnosis of FPTL in terms of sensitivity, specificity and accuracy, respectively. Pairwise t-test was used to evaluate the statistically significant difference. Results: The accuracy of the AI system in malignancy diagnosis was 0.71, which was higher than the best radiologist in this study by a margin of 0.09 with a p-value of 2.08×10-5. Two radiologists had higher sensitivity (0.84 and 0.78) than that of the AI system (0.69) at the cost of having much lower specificity (0.35, 0.57 versus 0.71). One senior radiologist showed balanced sensitivity and specificity (0.62 and 0.54) but both were lower than that of the AI system. Conclusions: The generally trained AI automatic diagnosis system can potentially assist radiologists for distinguishing FTC from other FPTL cases that share poorly distinguishable ultrasonographical features.

Techniques and Algorithms for Hepatic Vessel Skeletonization in Medical Images: A Survey.

Hepatic vessel skeletonization serves as an important means of hepatic vascular analysis and vessel segmentation. This paper presents a survey of techniques and algorithms for hepatic vessel skeletonization in medical images. We summarized the latest developments and classical approaches in this field. These methods are classified into five categories according to their methodological characteristics. The overview and brief assessment of each category are provided in the corresponding chapters, respectively. We provide a comprehensive summary among the cited publications, image modalities and datasets from various aspects, which hope to reveal the pros and cons of every method, summarize its achievements and discuss the challenges and future trends.

A facile fabrication of alginate microbubbles using a gas foaming reaction.

Microbubble particles have been extensively utilized as temporal templates for various biomedical applications. This study proposes a facile strategy to obtain microbubble-containing alginate particles (i.e., microbubbles inside alginate gel particles, called alginate microbubbles). The chemical reaction of sodium bicarbonate and hydrogen peroxide to produce gaseous carbon dioxide and oxygen was utilized to form microbubbles within alginate particles. Uniform alginate particles were obtained by a stable needle-based droplet formation process. Kinetic reaction of gas formation was monitored for 2% alginate particles. The gas formation increased with the concentrations of sodium bicarbonate (1-5 wt%) and hydrogen peroxide (0-36.5 wt%).