Computer-aided Diagnosis Software Research Articles

Clinical Validation of Computer-Aided Diagnosis Software for Preventing Retained Surgical Sponges.

We previously reported the successful development of a computer-aided diagnosis (CAD) system for preventing retained surgical sponges with deep learning using training data, including composite and simulated radiographs. In this study, we evaluated the efficacy of the CAD system in a clinical setting. A total of 1,053 postoperative radiographs obtained from patients 20 years of age or older who underwent surgery were evaluated. We implemented a foreign object detection application software on the portable radiographic device used in the operating room to detect retained surgical sponges. The results of the CAD system diagnosis were prospectively collected. Among the 1,053 images, the CAD system detected possible retained surgical items in 150 images. Specificity was 85.8%, which is similar to the data obtained during the development of the software. The validation of a CAD system using deep learning in a clinical setting showed similar efficacy as during the development of the system. These results suggest that the CAD system can contribute to the establishment of a more effective protocol than the current standard practice for preventing the retention of surgical items.

A new computer-aided diagnosis tool based on deep learning methods for automatic detection of retinal disorders from OCT images.

Early detection of retinal disorders using optical coherence tomography (OCT) images can prevent vision loss. Since manual screening can be time-consuming, tedious, and fallible, we present a reliable computer-aided diagnosis (CAD) software based on deep learning. Also, we made efforts to increase the interpretability of the deep learning methods, overcome their vague and black box nature, and also understand their behavior in the diagnosis. We propose a novel method to improve the interpretability of the used deep neural network by embedding the rich semantic information of abnormal areas based on the ophthalmologists' interpretations and medical descriptions in the OCT images. Finally, we trained the classification network on a small subset of the online publicly available University of California San Diego (UCSD) dataset with an overall of 29,800 OCT images. The experimental results on the 1000 test OCT images show that the proposed method achieves the overall precision, accuracy, sensitivity, and f1-score of 97.6%, 97.6%, 97.6%, and 97.59%, respectively. Also, the heat map images provide a clear region of interest which indicates that the interpretability of the proposed method is increased dramatically. The proposed software can help ophthalmologists in providing a second opinion to make a decision, and primitive automated diagnoses of retinal diseases and even it can be used as a screening tool, in eye clinics. Also, the improvement of the interpretability of the proposed method causes to increase in the model generalization, and therefore, it will work properly on a wide range of other OCT datasets.

AI for Detection of Tuberculosis: Implications for Global Health.

Tuberculosis, which primarily affects developing countries, remains a significant global health concern. Since the 2010s, the role of chest radiography has expanded in tuberculosis triage and screening beyond its traditional complementary role in the diagnosis of tuberculosis. Computer-aided diagnosis (CAD) systems for tuberculosis detection on chest radiographs have recently made substantial progress in diagnostic performance, thanks to deep learning technologies. The current performance of CAD systems for tuberculosis has approximated that of human experts, presenting a potential solution to the shortage of human readers to interpret chest radiographs in low- or middle-income, high-tuberculosis-burden countries. This article provides a critical appraisal of developmental process reporting in extant CAD software for tuberculosis, based on the Checklist for Artificial Intelligence in Medical Imaging. It also explores several considerations to scale up CAD solutions, encompassing manufacturer-independent CAD validation, economic and political aspects, and ethical concerns, as well as the potential for broadening radiography-based diagnosis to other nontuberculosis diseases. Collectively, CAD for tuberculosis will emerge as a representative deep learning application, catalyzing advances in global health and health equity. Keywords: Computer-aided Diagnosis (CAD), Conventional Radiography, Thorax, Lung, Machine Learning Supplemental material is available for this article. © RSNA, 2024.

Computer-aided diagnosis for screening of lower extremity lymphedema in pelvic computed tomography images using deep learning

Lower extremity lymphedema (LEL) is a common complication after gynecological cancer treatment, which significantly reduces the quality of life. While early diagnosis and intervention can prevent severe complications, there is currently no consensus on the optimal screening strategy for postoperative LEL. In this study, we developed a computer-aided diagnosis (CAD) software for LEL screening in pelvic computed tomography (CT) images using deep learning. A total of 431 pelvic CT scans from 154 gynecological cancer patients were used for this study. We employed ResNet-18, ResNet-34, and ResNet-50 models as the convolutional neural network (CNN) architecture. The input image for the CNN model used a single CT image at the greater trochanter level. Fat-enhanced images were created and used as input to improve classification performance. Receiver operating characteristic analysis was used to evaluate our method. The ResNet-34 model with fat-enhanced images achieved the highest area under the curve of 0.967 and an accuracy of 92.9%. Our CAD software enables LEL diagnosis from a single CT image, demonstrating the feasibility of LEL screening only on CT images after gynecologic cancer treatment. To increase the usefulness of our CAD software, we plan to validate it using external datasets.

Applying the Lifelong Machine Learning Paradigm in Tuberculosis Triage

Tuberculosis (TB) and pneumonia, including pneumonia from SARS-CoV-2 infection, are among the main causes of lower respiratory infections, which are the fourth cause of death worldwide. Recently, the World Health Organization recommended the use of computer-aided diagnosis (CAD) software as a tool to analyze chest radiographs (CXR) for TB screening and triage. Most CAD developed to date aim to screen exclusively for TB. This work applies the lifelong machine learning paradigm to detect both pneumonia and TB through CXRs and evaluate the models’ ability to retain and acquire knowledge. Two well-known lifelong learning models, the Efficient Lifelong Learning Algorithm (ELLA) and Learning without Forgetting (LwF), were applied to two public CXR datasets containing TB and pneumonia samples together with healthy CXR samples. Pneumonia detection was learned first and TB detection was learned as second task. The SP index, a function of sensitivity and specificity, was used to evaluate the models. We concluded that both algorithms were able to retain knowledge about pneumonia detection and were also able to learn TB detection.

A Computer-Aided Diagnosis (CAD) System for Automatic Counting of Ki67 Cells in Meningioma

Meningioma is a type of primary brain tumour where this tumour arises in the three thin layers of tissues, called meninges. Tumour grading is usually used to describe tumour cells' characteristics and behaviours and how they look under a microscope. There were many techniques used for determining the grade of the tumour. Ki67 was the most common proliferation marker used to measure cell proliferation activity. Currently, pathologists used the manual counting technique to count the Ki67 cells before determining tumour grading. However, this technique was time-consuming, tiring and the counting results are often not accurate. Besides that, manual counting has poor reproducibility and discordant between counting values’ among the pathologist. Therefore, this study aimed to develop a Computer-Aided Design (CAD) software that automatically counts the Ki67 cells for determining tumour grading. The purpose of developing this software is to alleviate pathologists’ workload associated with counting Ki67 cells and scoring the Ki67 index. The CAD software was developed through seven stages. Based on Pearson Correlation Coefficient results, there was a good positive correlation between the proposed technique with the manual counting technique in counting positive and negative Ki67 cells with a correlation of 0.99 and 0.72 respectively. The proposed CAD system also showed promising results in computing the Ki67 labeling index with a low percentage absolute error of 1.85%.

S-Thyroid Computer-Aided Diagnosis Ultrasound System of Thyroid Nodules: Correlation Between Transverse and Longitudinal Planes.

Introduction: We compare the differences in the diagnostic results of S-thyroid, a computer-aided diagnosis (CAD) software, based on two mutually perpendicular planes. Methods: Initially, 149 thyroid nodules confirmed by surgical pathology were enrolled in our study. CAD in our study was based on the ACR TI-RADS lexicon. t test, rank-sum test, and Chi-square test were used. The interclass correlation coefficient and Cohen’s kappa were used to explore the correlation between CAD features. Receiver operating characteristic was plotted for different combinations of CAD features. Results: The patient’s age, transverse diameter, longitudinal diameter, shape, margin, echogenicity, echogenic foci, composition, TI-RADS classification, and risk probability of nodules in the transverse and longitudinal planes were related to thyroid cancer (p < 0.05). The AUC (95%CI) of TI-RADS classification in the transverse plane of CAD is better than that of the longitudinal plane [0.90 (0.84–0.95) vs. 0.83 (0.77–0.90), p = 0.04]. The AUC (95%CI) of risk probability of nodules in the transverse planes shows no difference from that in the longitudinal plane statistically [0.90 (0.85–0.95) vs. 0.88 (0.82–0.94), p = 0.52]. The AUC (95% CI), specificity, sensitivity, and accuracy [TI-RADS classification (transverse plane) + TI-RADS classification (longitudinal plane) + risk (transverse plane) + risk (longitudinal plane)] are 0.93 (0.89–0.97), 86.15%, 90.48%, and 88.59%, respectively. Conclusion: The diagnosis of thyroid cancer in the CAD transverse plane was superior to that in the CAD longitudinal plane when using the TI-RADS classification, but there was no difference in the diagnosis between the two planes when using risk. However, the combination of CAD transverse and longitudinal planes had the best diagnostic ability.

Deep Learning-Based Classification of Cancer Cell in Leptomeningeal Metastasis on Cytomorphologic Features of Cerebrospinal Fluid.

BackgroundIt is a critical challenge to diagnose leptomeningeal metastasis (LM), given its technical difficulty and the lack of typical symptoms. The existing gold standard of diagnosing LM is to use positive cerebrospinal fluid (CSF) cytology, which consumes significantly more time to classify cells under a microscope.ObjectiveThis study aims to establish a deep learning model to classify cancer cells in CSF, thus facilitating doctors to achieve an accurate and fast diagnosis of LM in an early stage.MethodThe cerebrospinal fluid laboratory of Xijing Hospital provides 53,255 cells from 90 LM patients in the research. We used two deep convolutional neural networks (CNN) models to classify cells in the CSF. A five-way cell classification model (CNN1) consists of lymphocytes, monocytes, neutrophils, erythrocytes, and cancer cells. A four-way cancer cell classification model (CNN2) consists of lung cancer cells, gastric cancer cells, breast cancer cells, and pancreatic cancer cells. Here, the CNN models were constructed by Resnet-inception-V2. We evaluated the performance of the proposed models on two external datasets and compared them with the results from 42 doctors of various levels of experience in the human-machine tests. Furthermore, we develop a computer-aided diagnosis (CAD) software to generate cytology diagnosis reports in the research rapidly.ResultsWith respect to the validation set, the mean average precision (mAP) of CNN1 is over 95% and that of CNN2 is close to 80%. Hence, the proposed deep learning model effectively classifies cells in CSF to facilitate the screening of cancer cells. In the human-machine tests, the accuracy of CNN1 is similar to the results from experts, with higher accuracy than doctors in other levels. Moreover, the overall accuracy of CNN2 is 10% higher than that of experts, with a time consumption of only one-third of that consumed by an expert. Using the CAD software saves 90% working time of cytologists.ConclusionA deep learning method has been developed to assist the LM diagnosis with high accuracy and low time consumption effectively. Thanks to labeled data and step-by-step training, our proposed method can successfully classify cancer cells in the CSF to assist LM diagnosis early. In addition, this unique research can predict cancer’s primary source of LM, which relies on cytomorphologic features without immunohistochemistry. Our results show that deep learning can be widely used in medical images to classify cerebrospinal fluid cells. For complex cancer classification tasks, the accuracy of the proposed method is significantly higher than that of specialist doctors, and its performance is better than that of junior doctors and interns. The application of CNNs and CAD software may ultimately aid in expediting the diagnosis and overcoming the shortage of experienced cytologists, thereby facilitating earlier treatment and improving the prognosis of LM.

Novel diagnostic model for bone metastases in renal cell carcinoma patients based on bone scintigraphy analyzed by computer-aided diagnosis software and bone turnover markers

BackgroundComputer-assisted diagnosis (CAD) systems for bone scans have been introduced as clinical quality assurance tools, but few studies have reported on its utility for renal cell carcinoma (RCC) patients. The aim of this study was to assess the diagnostic validity of the CAD system for bone scans and to construct a novel diagnostic system for bone metastases in RCC patients.MethodsWe evaluated bone scan images of 300 RCC patients. Artificial neural network (ANN) values, which represent the probability of abnormality, were calculated by BONENAVI, the CAD software for bone scans. By analyzing ANN values, we assessed the diagnostic validity of BONENAVI. Next, we selected 108 patients who underwent measurements of bone turnover markers and assessed the combined diagnostic validity of BONENAVI and bone turnover markers.ResultsForty-three out of 300 RCC patients had bone metastases. The AUC of ANN values was 0.764 and the optimum sensitivity and specificity were 83.7 and 62.7%. By logistic analysis of 108 cases, we found that ICTP, a bone resorption marker, could be a diagnostic marker. The AUC of ICTP was 0.776 and the optimum sensitivity and specificity were 57.1 and 86.8%. Subsequently, we developed a novel diagnostic model based on ANN values and ICTP. Using this model, the AUC was 0.849 and the optimum sensitivity and specificity were 76.2 and 80.7%.ConclusionBy combining the high sensitivity provided by BONENAVI and the high specificity provided by ICTP, we constructed a novel, high-accuracy diagnostic model for bone metastases in RCC patients.

The diagnostic performance of ultrasound computer-aided diagnosis system for distinguishing breast masses: a prospective multicenter study.

To evaluate the diagnostic value of computer-aided diagnosis (CAD) software on ultrasound in distinguishing benign and malignant breast masses and avoiding unnecessary biopsy. This prospective, multicenter study included patients who were scheduled for pathological diagnosis of breast masses between April 2019 and November 2020. Ultrasound images, videos, CAD analysis, and BI-RADS were obtained. The AUC, accuracy, sensitivity, specificity, PPV, and NPV were calculated and compared with radiologists. Overall, 901 breast masses in 901 patients were enrolled in this study. The accuracy, sensitivity, specificity, PPV and NPV of CAD software were 89.6%, 94.2%, 87.0%, 80.4%, and 96.3, respectively, in the long-axis section; 89.0%, 91.4%, 87.7%, 80.8%, and 94.7%, respectively, in the short-axis section. With BI-RADS 4a as the cut-off value, CAD software has a higher AUC (0.906 vs 0.734 vs 0.696, all p < 0.001) than both experienced and less experienced radiologists. With BI-RADS 4b as the cut-off value, CAD software showed better AUC than less experienced radiologists (0.906 vs 0.874, p < 0.001), but not superior to experienced radiologists (0.906 vs 0.883, p = 0.057). After the application of CAD software, the unnecessary biopsy rate of BI-RADS categories 4 and 5 was significantly decreased (33.0% vs 11.9%, 37.8% vs 14.5%), and the malignant rate of biopsy in category 4a was significantly increased (11.6% vs 40.7%, 7.4% vs 34.9%, all p < 0.001). CAD software on ultrasound can be used as an effective auxiliary diagnostic tool for differential diagnosis of benign and malignant breast masses and reducing unnecessary biopsy. ClinicalTrials.gov (NCT03887598) KEY POINTS: • Prospective multicenter study showed that computer-aided diagnosis software provides greater diagnostic confidence for differentiating benign and malignant breast masses. • Computer-aided diagnosis software can help radiologists reduce unnecessary biopsy. • The management of patients with breast masses becomes more appropriate.

Diagnostic validation of a deep learning nodule detection algorithm in low-dose chest CT: determination of optimized dose thresholds in a virtual screening scenario.

This study was conducted to evaluate the effect of dose reduction on the performance of a deep learning (DL)-based computer-aided diagnosis (CAD) system regarding pulmonary nodule detection in a virtual screening scenario. Sixty-eight anthropomorphic chest phantoms were equipped with 329 nodules (150 ground glass, 179 solid) with four sizes (5mm, 8mm, 10mm, 12mm) and scanned with nine tube voltage/current combinations. The examinations were analyzed by a commercially available DL-based CAD system. The results were compared by a comparison of proportions. Logistic regression was performed to evaluate the impact of tube voltage, tube current, nodule size, nodule density, and nodule location. The combination with the lowest effective dose (E) and unimpaired detection rate was 80kV/50 mAs (sensitivity: 97.9%, mean false-positive rate (FPR): 1.9, mean CTDIvol: 1.2 ± 0.4mGy, mean E: 0.66mSv). Logistic regression revealed that tube voltage and current had the greatest impact on the detection rate, while nodule size and density had no significant influence. The optimal tube voltage/current combination proposed in this study (80kV/50 mAs) is comparable to the proposed combinations in similar studies, which mostly dealt with conventional CAD software. Modification of tube voltage and tube current has a significant impact on the performance of DL-based CAD software in pulmonary nodule detection regardless of their size and composition. • Modification of tube voltage and tube current has a significant impact on the performance of deep learning-based CAD software. • Nodule size and composition have no significant impact on the software's performance. • The optimal tube voltage/current combination for the examined software is 80kV/50 mAs.

Comparative Analysis of Computer-Aided Diagnosis and Computer-Assisted Subjective Assessment in Thyroid Ultrasound

The value of computer-aided diagnosis (CAD) and computer-assisted techniques equipped with different TIRADS remains ambiguous. Parallel diagnosis performances of computer-assisted subjective assessments and CAD were compared based on AACE, ATA, EU, and KSThR TIRADS. CAD software computed the diagnosis of 162 thyroid nodule sonograms. Two raters (R1 and R2) independently rated the sonographic features of the nodules using an online risk calculator while blinded to pathology results. Diagnostic efficiency measures were calculated based on the final pathology results. R1 had higher diagnostic performance outcomes than CAD with similarities between KSThR (SEN: 90.3% vs. 83.9%, p = 0.57; SPEC: 46% vs. 51%, p = 0.21; AUROC: 0.76 vs. 0.67, p = 0.02), and EU (SEN: 85.5% vs. 79%, p = 0.82; SPEC: 62% vs. 55%, p = 0.27; AUROC: 0.74 vs. 0.67, p = 0.06). Similarly, R2 had higher AUROC and specificity but lower sensitivity than CAD (KSThR-AUROC: 0.74 vs. 0.67, p = 0.13; SPEC: 61% vs. 46%, p = 0.02 and SEN: 75.8% vs. 83.9%, p = 0.31, and EU-AUROC: 0.69 vs. 0.67, p = 0.57, SPEC: 64% vs. 55%, p = 0.19, and SEN: 71% vs. 79%, p = 0.51, respectively). CAD had higher sensitivity but lower specificity than both R1 and R2 with AACE for 114 specified nodules (SEN: 92.5% vs. 88.7%, p = 0.50; 92.5% vs. 79.3%, p = 0.02, and SPEC: 26.2% vs. 54.1%, p = 0.001; 26.2% vs. 62.3%, p < 0.001, respectively). All diagnostic performance outcomes were comparable for ATA with 96 specified nodules. Computer-assisted subjective interpretation using KSThR is more ideal for ruling out papillary thyroid carcinomas than CAD. Future larger multi-center and multi-rater prospective studies with a diverse representation of thyroid cancers are necessary to validate these findings.

Novel Computer-Aided Diagnosis Software for the Prevention of Retained Surgical Items

Retained surgical items are a serious human error. Surgical sponges account for 70% of retained surgical items. To prevent retained surgical sponges, it is important to establish a system that can identify errors and avoid the occurrence of adverse events. To date, no computer-aided diagnosis software specialized for detecting retained surgical sponges has been reported. We developed a software program that enables easy and effective computer-aided diagnosis of retained surgical sponges with high sensitivity and specificity using the technique of deep learning, a subfield of artificial intelligence. In this study, we developed the software by training it through deep learning using a dataset and then validating the software. The dataset consisted of a training set and validation set. We created composite x-rays consisting of normal postoperative x-rays and surgical sponge x-rays for a training set (n= 4,554) and a validation set (n= 470). Phantom x-rays (n= 12) were prepared for software validation. X-rays obtained with surgical sponges inserted into cadavers were used for validation purposes (formalin: Thiel's method= 252:117). In addition, postoperative x-rays without retained surgical sponges were used for the validation of software performance to determine false-positive rates. Sensitivity, specificity, and false positives per image were calculated. In the phantom x-rays, both the sensitivity and specificity in software image interpretation were 100%. The software achieved 97.7% sensitivity and 83.8% specificity in the composite x-rays. In the normal postoperative x-rays, 86.6% specificity was achieved. In reading the cadaveric x-rays, the software attained both sensitivity and specificity of >90%. Software with high sensitivity for diagnosis of retained surgical sponges was developed successfully.

Applications of Artificial Intelligence in Mammography from a Development and Validation Perspective

유방촬영술은 유방암 검진 및 진단을 위한 기본적인 영상 검사이지만, 판독이 어려우며 높은 숙련도를 필요로 한다고 잘 알려져 있다. 이러한 어려움을 극복하기 위해 최근 몇 년 사이에 인공지능을 이용한 유방암 검출 알고리즘들이 활발히 연구되고 있다. 본 종설에서 저자는 고전적인 computer-aided detection 소프트웨어 대비 최근 많이 사용되는 딥러닝의 특징을 알아보고, 딥러닝 알고리즘의 개발 방법과 임상적 검증 방법에 대해서 기술하였다. 또한 딥러닝 기반의 검진 유방촬영술의 판독 방법 분류, 유방 치밀도 평가, 그리고 유방암 위험도 예측 모델 등을 위한 딥러닝 연구들도 소개하였다. 마지막으로 유방촬영술 관련 인공지능 기술들에 대한 영상의학과 전문의의 관심과 의견의 필요성을 기술하였다.

Diagnostic Performance of Ultrasound Computer-Aided Diagnosis Software Compared with That of Radiologists with Different Levels of Expertise for Thyroid Malignancy: A Multicenter Prospective Study

The aim of the work described here was to evaluate the diagnostic performance of ultrasound thyroid computer-aided diagnosis (CAD) software. This multicenter prospective study included 494 patients (565 thyroid nodules) who underwent surgery or biopsy after ultrasonography at four hospitals from January 2019 to September 2019. The diagnostic performance metrics of different readers were calculated and compared with the pathologic results. The sensitivity of CAD was outstanding and was equivalent to that of a senior radiologist (90.51% vs. 88.47%, p > 0.05). The area under the curve of CAD was equivalent to that of a junior radiologist (0.748 vs. 0.739, p > 0.05). However, the specificity was only 49.63%, which was lower than those of the three radiologists (75.56%, 85.93% and 90.37% for the junior, intermediate and senior radiologists, respectively). The diagnostic performance of the junior radiologist was significantly improved with the aid of CAD (junior + CAD). The sensitivity and area under the curve of junior + CAD were improved from 72.20% to 89.93% and from 0.739 to 0.816, respectively (both p values <0.05), and the positive predictive value, negative predictive value and κ coefficient improved from 76.3% to 78.6%, 82.0% to 86.8% and 0.394 to 0.511, respectively. Though specificity slightly decreased from 75.56% to 73.33%, the difference was not statistically significant (p > 0.05). In general, the clinical application value of CAD is promising, and its instrumental value for junior radiologists is significant.

Artificial Intelligence Applied to Breast MRI for Improved Diagnosis.

Background Recognition of salient MRI morphologic and kinetic features of various malignant tumor subtypes and benign diseases, either visually or with artificial intelligence (AI), allows radiologists to improve diagnoses that may improve patient treatment. Purpose To evaluate whether the diagnostic performance of radiologists in the differentiation of cancer from noncancer at dynamic contrast material-enhanced (DCE) breast MRI is improved when using an AI system compared with conventionally available software. Materials and Methods In a retrospective clinical reader study, images from breast DCE MRI examinations were interpreted by 19 breast imaging radiologists from eight academic and 11 private practices. Readers interpreted each examination twice. In the "first read," they were provided with conventionally available computer-aided evaluation software, including kinetic maps. In the "second read," they were also provided with AI analytics through computer-aided diagnosis software. Reader diagnostic performance was evaluated with receiver operating characteristic (ROC) analysis, with the area under the ROC curve (AUC) as a figure of merit in the task of distinguishing between malignant and benign lesions. The primary study end point was the difference in AUC between the first-read and the second-read conditions. Results One hundred eleven women (mean age, 52 years ± 13 [standard deviation]) were evaluated with a total of 111 breast DCE MRI examinations (54 malignant and 57 nonmalignant lesions). The average AUC of all readers improved from 0.71 to 0.76 (P = .04) when using the AI system. The average sensitivity improved when Breast Imaging Reporting and Data System (BI-RADS) category 3 was used as the cut point (from 90% to 94%; 95% confidence interval [CI] for the change: 0.8%, 7.4%) but not when using BI-RADS category 4a (from 80% to 85%; 95% CI: -0.9%, 11%). The average specificity showed no difference when using either BI-RADS category 4a or category 3 as the cut point (52% and 52% [95% CI: -7.3%, 6.0%], and from 29% to 28% [95% CI: -6.4%, 4.3%], respectively). Conclusion Use of an artificial intelligence system improves radiologists' performance in the task of differentiating benign and malignant MRI breast lesions. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Krupinski in this issue.

Computer-aided diagnostic system for thyroid nodule sonographic evaluation outperforms the specificity of less experienced examiners.

Computer-aided diagnosis (CAD) may improve interobserver agreement in the risk stratification of thyroid nodules. This study aims to evaluate the performance of the Korean Thyroid Imaging Reporting and Data System (K-TIRADS) classification as estimated by an expert radiologist, a senior resident, a medical student, and a CAD system, as well as the interobserver agreement among them. Between July 2016 and 2018, 107 nodules (size 5-40mm, 27 malignant) were classified according to the K-TIRADS by an expert radiologist and CAD software. A third-year resident and a medical student with basic imaging training, both blinded to previous findings, retrospectively estimated the K-TIRADS classification. The diagnostic performance was calculated, including sensitivity, specificity, positive and negative predictive values, and the area under the receiver operating characteristic curve. The CAD system and the expert achieved a sensitivity of 70.37% (95% CI 49.82-86.25%) and 81.48% (61.92-93.7%) and a specificity of 87.50% (78.21-93.84%) and 88.75% (79.72-94.72%), respectively. The specificity of the student was significantly lower (76.25% [65.42-85.05%], p = 0.02). In our opinion, the CAD evaluation of thyroid nodules stratification risk has a potential role in a didactic field and does not play a real and effective role in the clinical field, where not only images but also specialistic medical practice is fundamental to achieve a diagnosis based on family history, genetics, lab tests, and so on. The CAD system may be useful for less experienced operators as its specificity was significantly higher.

Computer-aided diagnosis for characterization of colorectal lesions: comprehensive software that includes differentiation of serrated lesions

Endoscopy guidelines recommend adhering to policies such as resect and discard only if the optical biopsy is accurate. However, accuracy in predicting histology can vary greatly. Computer-aided diagnosis (CAD) for characterization of colorectal lesions may help with this issue. In this study, CAD software developed at the University of Adelaide (Australia) that includes serrated polyp differentiation was validated with Japanese images on narrow-band imaging (NBI) and blue-laser imaging (BLI). CAD software developed using machine learning and densely connected convolutional neural networks was modeled with NBI colorectal lesion images (Olympus 190 series - Australia) and validated for NBI (Olympus 290 series) and BLI (Fujifilm 700 series) with Japanese datasets. All images were correlated with histology according to the modified Sano classification. The CAD software was trained with Australian NBI images and tested with separate sets of images from Australia (NBI) and Japan (NBI and BLI). An Australian dataset of 1235 polyp images was used as training, testing, and internal validation sets. A Japanese dataset of 20 polyp images on NBI and 49 polyp images on BLI was used as external validation sets. The CAD software had a mean area under the curve (AUC) of 94.3% for the internal set and 84.5% and 90.3% for the external sets (NBI and BLI, respectively). The CAD achieved AUCs comparable with experts and similar results with NBI and BLI. Accurate CAD prediction was achievable, even when the predicted endoscopy imaging technology was not part of the training set.

Computer‐aided diagnosis of breast ultrasound images using ensemble learning from convolutional neural networks

Breast ultrasound and computer aided diagnosis (CAD) has been used to classify tumors into benignancy or malignancy. However, conventional CAD software has some problems (such as handcrafted features are hard to design; conventional CAD systems are difficult to confirm overfitting problems, etc.). In our study, we propose a CAD system for tumor diagnosis using an image fusion method combined with different image content representations and ensemble different CNN architectures on US images. The CNN-based method proposed in this study includes VGGNet, ResNet, and DenseNet. In our private dataset, there was a total of 1687 tumors that including 953 benign and 734 malignant tumors. The accuracy, sensitivity, specificity, precision, F1 score and the AUC of the proposed method were 91.10%, 85.14%, 95.77%, 94.03%, 89.36%, and 0.9697 respectively. In the open dataset (BUSI), there was a total of 697 tumors that including 437 benign lesions, 210 malignant tumors, and 133 normal images. The accuracy, sensitivity, specificity, precision, F1 score, and the AUC of the proposed method were 94.62%, 92.31%, 95.60%, 90%, 91.14%, and 0.9711. In conclusion, the results indicated different image content representations that affect the prediction performance of the CAD system, more image information improves the prediction performance, and the tumor shape feature can improve the diagnostic effect.

Breast elastography: diagnostic performance of computer-aided diagnosis software and interobserver agreement.

ObjectiveTo determine the best cutoff value for classifying breast masses by ultrasound elastography, using dedicated software for strain elastography, and to determine the level of interobserver agreement.Materials and MethodsWe enrolled 83 patients with 83 breast masses identified on ultrasound and referred for biopsy. After B-mode ultrasound examination, the lesions were manually segmented by three radiologists with varying degrees of experience in breast imaging, designated reader 1 (R1, with 15 years), reader 2 (R2, with 2 years), and reader 3 (R3, with 8 years). Elastography was performed automatically on the best image with computer-aided diagnosis (CAD) software. Cutoff values of 70%, 75%, 80%, and 90% of hard areas were applied for determining the performance of the CAD software. The best cutoff value for the most experienced radiologists was then compared with the visual assessment. Interobserver agreement for the best cutoff value was determined, as were the interclass correlation coefficient and concordance among the radiologists for the areas segmented.ResultsThe best cutoff value of the proportion of hard area within a breast mass, for experienced radiologists, was found to be 75%. At a cutoff value of 75%, the interobserver agreement was excellent between R1 and R2, as well as between R1 and R3, and good between R2 and R3. The interclass concordance coefficient among the three radiologists was 0.950. When assessing the segmented areas by size, we found that the level of agreement was higher among the more experienced radiologists.ConclusionThe best cutoff value for a quantitative CAD system to classify breast masses was 75%.