Computer-Aided Detection (CAD) in Screening Mammography:Sensitivity of Commercial CAD Systems for Detecting Architectural Distortion

AI-enhanced Mammography With Digital Breast Tomosynthesis for Breast Cancer Detection: Clinical Value and Comparison With Human Performance

Purpose To compare two deep learning-based commercially available artificial intelligence (AI) systems for mammography with digital breast tomosynthesis (DBT) and benchmark them against the performance of radiologists. Materials and Methods This retrospective study included consecutive asymptomatic patients who underwent mammography with DBT (2019-2020). Two AI systems (Transpara 1.7.0 and ProFound AI 3.0) were used to evaluate the DBT examinations. The systems were compared using receiver operating characteristic (ROC) analysis to calculate the area under the ROC curve (AUC) for detecting malignancy overall and within subgroups based on mammographic breast density. Breast Imaging Reporting and Data System results obtained from standard-of-care human double-reading were compared against AI results with use of the DeLong test. Results Of 419 female patients (median age, 60 years [IQR, 52-70 years]) included, 58 had histologically proven breast cancer. The AUC was 0.86 (95% CI: 0.85, 0.91), 0.93 (95% CI: 0.90, 0.95), and 0.98 (95% CI: 0.96, 0.99) for Transpara, ProFound AI, and human double-reading, respectively. For Transpara, a rule-out criterion of score 7 or lower yielded 100% (95% CI: 94.2, 100.0) sensitivity and 60.9% (95% CI: 55.7, 66.0) specificity. The rule-in criterion of higher than score 9 yielded 96.6% sensitivity (95% CI: 88.1, 99.6) and 78.1% specificity (95% CI: 73.8, 82.5). For ProFound AI, a rule-out criterion of lower than score 51 yielded 100% sensitivity (95% CI: 93.8, 100) and 67.0% specificity (95% CI: 62.2, 72.1). The rule-in criterion of higher than score 69 yielded 93.1% (95% CI: 83.3, 98.1) sensitivity and 82.0% (95% CI: 77.9, 86.1) specificity. Conclusion Both AI systems showed high performance in breast cancer detection but lower performance compared with human double-reading. Keywords: Mammography, Breast, Oncology, Artificial Intelligence, Deep Learning, Digital Breast Tomosynthesis © RSNA, 2024.

Can Artificial Intelligence Beat Humans in Detecting Breast Malignancy on Mammograms?

BACKGROUND

The study was aimed at identifying how useful Computer-Aided Detection (CAD) could be in reducing false-negative reporting in mammography and early detection of breast cancer at an early stage as the best protection is early detection.

MATERIALS AND METHODS

This retrospective study was conducted in a tertiary care setup of Atomic Energy Cancer Hospital, Nuclear Medicine, Oncology and Radiotherapy Institute (AECH-NORI), where 33 patients with suspicious findings on mammography and subsequent biopsy-proven malignancy were included. The findings of mammography including the lesion type, breast parenchymal density, and sensitivity of CAD detection, as well as the final biopsy results, were recorded. A second group of 40 normal screening mammograms was also included who had no symptoms, had Breast Imaging-Reporting and Data System category I(BI-RADS I) mammograms, and had no pathology identified on correlative sonomammography as well.

RESULTS

A total of 35 masses, 11 pleomorphic clusters of microcalcification, five clustered foci of macrocalcification, and nine lesions with pleomorphic clusters of microcalcification and two with pleomorphic clusters of microcalcification only were included. The CAD system was able to identify 26 masses (74%), eight lesions with pleomorphic clusters of microcalcification (72%), five foci of macrocalcification (100%), six lesions with pleomorphic clusters of microcalcification (66%), and two pleomorphic clusters of microcalcification without formed mass (100%). The overall sensitivity of the CAD system was 75.8%. CAD was able to identify 13 out of 16 masses with invasive ductal carcinoma (81.3%), eight out of nine lesions proven as invasive ductal carcinoma with ductal carcinoma in situ (DCIS) (88.9%), two out of five masses with invasive lobular carcinoma (40%), four out of four masses with invasive mammary carcinoma (100%), and zero out of one lesion identified as medullary carcinoma (0%). There was 100% detection for pleomorphic clusters of microcalcification without formed mass with CAD marking two out of two mammograms.

CONCLUSION

CAD performed better with combined lesions, accurately marked pleomorphic clusters of microcalcification, and identified small lesions in predominant fibrofatty parenchymal density but was not reliable in dense breast, areas of asymmetric increased density, summation artifacts, edematous breast parenchyma, and retroareolar lesions. It also performed poorly with ill-defined lesions of invasive lobular carcinoma. Human intelligence hence beats CAD for the diagnosis of breast malignancy in mammograms as per our experience.

A Statistical Approach to Assess the Robustness of Radiomics Features in the Discrimination of Mammographic Lesions

Despite mammography (MG) being among the most widespread techniques in breast cancer screening, tumour detection and classification remain challenging tasks due to the high morphological variability of the lesions. The extraction of radiomics features has proved to be a promising approach in MG. However, radiomics features can suffer from dependency on factors such as acquisition protocol, segmentation accuracy, feature extraction and engineering methods, which prevent the implementation of robust and clinically reliable radiomics workflow in MG. In this study, the variability and robustness of radiomics features is investigated as a function of lesion segmentation in MG images from a public database. A statistical analysis is carried out to assess feature variability and a radiomics robustness score is introduced based on the significance of the statistical tests performed. The obtained results indicate that variability is observable not only as a function of the abnormality type (calcification and masses), but also among feature categories (first-order and second-order), image view (craniocaudal and medial lateral oblique), and the type of lesions (benign and malignant). Furthermore, through the proposed approach, it is possible to identify those radiomics characteristics with a higher discriminative power between benign and malignant lesions and a lower dependency on segmentation, thus suggesting the most appropriate choice of robust features to be used as inputs to automated classification algorithms.

Emerging uses of artificial intelligence in breast and axillary ultrasound

Breast ultrasound is a valuable adjunctive tool to mammography in detecting breast cancer, especially in women with dense breasts. Ultrasound also plays an important role in staging breast cancer by assessing axillary lymph nodes. However, its utility is limited by operator dependence, high recall rate, low positive predictive value and low specificity. These limitations present an opportunity for artificial intelligence (AI) to improve diagnostic performance and pioneer novel uses of ultrasound. Research in developing AI for radiology has flourished over the past few years. A subset of AI, deep learning, uses interconnected computational nodes to form a neural network, which extracts complex visual features from image data to train itself into a predictive model. This review summarizes several key studies evaluating AI programs' performance in predicting breast cancer and demonstrates that AI can assist radiologists and address limitations of ultrasound by acting as a decision support tool. This review also touches on how AI programs allow for novel predictive uses of ultrasound, particularly predicting molecular subtypes of breast cancer and response to neoadjuvant chemotherapy, which have the potential to change how breast cancer is managed by providing non-invasive prognostic and treatment data from ultrasound images. Lastly, this review explores how AI programs demonstrate improved diagnostic accuracy in predicting axillary lymph node metastasis. The limitations and future challenges in developing and implementing AI for breast and axillary ultrasound will also be discussed.

Exploiting Patch Sizes and Resolutions for Multi-Scale Deep Learning in Mammogram Image Classification

Recent progress in deep learning (DL) has revived the interest on DL-based computer aided detection or diagnosis (CAD) systems for breast cancer screening. Patch-based approaches are one of the main state-of-the-art techniques for 2D mammogram image classification, but they are intrinsically limited by the choice of patch size, as there is no unique patch size that is adapted to all lesion sizes. In addition, the impact of input image resolution on performance is not yet fully understood. In this work, we study the impact of patch size and image resolution on the classifier performance for 2D mammograms. To leverage the advantages of different patch sizes and resolutions, a multi patch-size classifier and a multi-resolution classifier are proposed. These new architectures perform multi-scale classification by combining different patch sizes and input image resolutions. The AUC is increased by 3% on the public CBIS-DDSM dataset and by 5% on an internal dataset. Compared with a baseline single patch size and single resolution classifier, our multi-scale classifier reaches an AUC of 0.809 and 0.722 in each dataset.

Automated screening of COVID-19 using two-dimensional variational mode decomposition and locally linear embedding

In order to aid imaging physicians to effectively screen chest radiography medical images for presence of Coronavirus Disease 2019 (COVID-19), a novel computer aided diagnosis technology for automatic processing of COVID-19 images is proposed based on two-dimensional variational mode decomposition (2D-VMD) and locally linear embedding (LLE). 2D-VMD algorithm is used to decompose normal and COVID-19 images, and then feature extraction of intrinsic mode functions (IMFs) using Gabor filter. To better extract low-dimensional parameters which are useful for COVID-19 diagnosis, the performance of two dimensionality reduction techniques of principal component analysis (PCA) and LLE are compared, and the LLE is shown to offer satisfactory effect of dimension reduction. Thereafter, the particle swarm optimization-support vector machine (PSO-SVM) algorithm is used to classify. The simulation results show that the proposed technology has achieved accuracy of 99.33%, precision of 100%, recall of 98.63% and F-Measure of 99.31%. Hence, the developed diagnosis technology can be used as an important auxiliary tool to assist diagnosis of imaging physicians.

Architectural distortion detection based on superior-inferior directional context and anatomic prior knowledge in digital breast tomosynthesis

BACKGROUND

In 2020, breast cancer becomes the most leading diagnosed cancer all over the world. The burden is increasing in the prevention and treatment of breast cancer. Accurately detecting breast lesions in screening images is important for early detection of cancer. Architectural distortion (AD) is one of the breast lesions that need to be detected.

PURPOSE

To develop a deep-learning-based computer-aided detection (CADe) model for AD in digital breast tomosynthesis (DBT). This model uses the superior-inferior directional context of DBT and anatomic prior knowledge to reduce false positive (FP). It can identify some negative samples that cannot be distinguished by deep learning features.

METHODS

The proposed CADe model consists of three steps. In the first step, a deep learning detection network detects two-dimensional (2D) candidates of ADs in DBT slices with the inputs preprocessed by Gabor filters and convergence measure. In the second step, three-dimensional (3D) candidates are obtained by stacking 2D candidates along superior-inferior direction. In the last step, FP reduction for 3D candidates is implemented based on superior-inferior directional context and anatomic prior knowledge of breast. DBT data from 99 cases with AD were used as the training set to train the CADe model, and data from 208 cases were used as an independent test set (including 108 cases with AD and 100 cases without AD as the control group). The free-response receiver operating characteristic and mean true positive fraction (MTPF) in the range of 0.05-2.0 FPs per volume are used to evaluate the model.

RESULTS

Compared with the baseline model based on convergence measure, our proposed method demonstrates significant improvement (MTPF: 0.2826 ± 0.0321vs. 0.6640 ± 0.0399). Results of an ablation study show that our proposed context-based and anatomy-based FP reduction methods improve the detection performance. The number of FPs per DBT volume reduces from 2.47 to 1.66 at 80% sensitivity after employing these two schemes.

CONCLUSIONS

The deep learning model demonstrates practical value for AD detection. The results indicate that introducing superior-inferior directional context and anatomic prior knowledge into model can indeed reduce FPs and improve the performance of CADe model. This article is protected by copyright. All rights reserved.

Diagnosis of hypercritical chronic pulmonary disorders using dense convolutional network through chest radiography

Lung-related ailments are prevalent all over the world which majorly includes asthma, chronic obstructive pulmonary disease (COPD), tuberculosis, pneumonia, fibrosis, etc. and now COVID-19 is added to this list. Infection of COVID-19 poses respirational complications with other indications like cough, high fever, and pneumonia. WHO had identified cancer in the lungs as a fatal cancer type amongst others and thus, the timely detection of such cancer is pivotal for an individual's health. Since the elementary convolutional neural networks have not performed fairly well in identifying atypical image types hence, we recommend a novel and completely automated framework with a deep learning approach for the recognition and classification of chronic pulmonary disorders (CPD) and COVID-pneumonia using Thoracic or Chest X-Ray (CXR) images. A novel three-step, completely automated, approach is presented that first extracts the region of interest from CXR images for preprocessing, and they are then used to detects infected lungs X-rays from the Normal ones. Thereafter, the infected lung images are further classified into COVID-pneumonia, pneumonia, and other chronic pulmonary disorders (OCPD), which might be utilized in the current scenario to help the radiologist in substantiating their diagnosis and in starting well in time treatment of these deadly lung diseases. And finally, highlight the regions in the CXR which are indicative of severe chronic pulmonary disorders like COVID-19 and pneumonia. A detailed investigation of various pivotal parameters based on several experimental outcomes are made here. This paper presents an approach that detects the Normal lung X-rays from infected ones and the infected lung images are further classified into COVID-pneumonia, pneumonia, and other chronic pulmonary disorders with an utmost accuracy of 96.8%. Several other collective performance measurements validate the superiority of the presented model. The proposed framework shows effective results in classifying lung images into Normal, COVID-pneumonia, pneumonia, and other chronic pulmonary disorders (OCPD). This framework can be effectively utilized in this current pandemic scenario to help the radiologist in substantiating their diagnosis and in starting well in time treatment of these deadly lung diseases.

Improving the Performance of Radiologists Using Artificial Intelligence-Based Detection Support Software for Mammography: A Multi-Reader Study

Objective

To evaluate whether artificial intelligence (AI) for detecting breast cancer on mammography can improve the performance and time efficiency of radiologists reading mammograms.

Materials and Methods

A commercial deep learning-based software for mammography was validated using external data collected from 200 patients, 100 each with and without breast cancer (40 with benign lesions and 60 without lesions) from one hospital. Ten readers, including five breast specialist radiologists (BSRs) and five general radiologists (GRs), assessed all mammography images using a seven-point scale to rate the likelihood of malignancy in two sessions, with and without the aid of the AI-based software, and the reading time was automatically recorded using a web-based reporting system. Two reading sessions were conducted with a two-month washout period in between. Differences in the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and reading time between reading with and without AI were analyzed, accounting for data clustering by readers when indicated.

Results

The AUROC of the AI alone, BSR (average across five readers), and GR (average across five readers) groups was 0.915 (95% confidence interval, 0.876–0.954), 0.813 (0.756–0.870), and 0.684 (0.616–0.752), respectively. With AI assistance, the AUROC significantly increased to 0.884 (0.840–0.928) and 0.833 (0.779–0.887) in the BSR and GR groups, respectively (p = 0.007 and p < 0.001, respectively). Sensitivity was improved by AI assistance in both groups (74.6% vs. 88.6% in BSR, p < 0.001; 52.1% vs. 79.4% in GR, p < 0.001), but the specificity did not differ significantly (66.6% vs. 66.4% in BSR, p = 0.238; 70.8% vs. 70.0% in GR, p = 0.689). The average reading time pooled across readers was significantly decreased by AI assistance for BSRs (82.73 vs. 73.04 seconds, p < 0.001) but increased in GRs (35.44 vs. 42.52 seconds, p < 0.001).

Conclusion

AI-based software improved the performance of radiologists regardless of their experience and affected the reading time.

Automated Classification of Breast Cancer Lesions for Digitised Mammograms via Computer-Aided Diagnosis System

Women with breast cancer have a high risk of death. Digitised mammograms can be used to detect the early stage of breast cancer. However, digitised mammograms suffer low contrast appearances that may lead to misdiagnosis. This paper proposes a Computer-Aided Diagnosis (CAD) system of automated classification of breast cancer lesions using a modified image processing technique of Fuzzy Anisotropic Diffusion Histogram Equalization Contrast Adaptive Limited (FADHECAL) incorporated with Multilevel Otsu Thresholding on digitised mammograms. Four main blocks were used in this CAD system, namely; (i) Pre-processing and Enhancement block; (ii) Segmentation block; (iii) Region of Interests (ROIs) Extraction block; and (iv) Classification block. The CAD system was tested on 30 digitised mammograms retrieved from the Mini-Mammographic Image Analysis Society (MIAS) database with various degrees of severity and background tissues. The proposed CAD system showed a high accuracy of 96.67% for the detection of breast cancer lesions.

Computer-Aided Detection of Seven Chest Pathologies on Standard Posteroanterior Chest X-Rays Compared to Radiologists Reading Dual-Energy Subtracted Radiographs

RATIONALE AND OBJECTIVES

Retrospective performance evaluation of a computer-aided detection (CAD) system on standard posteroanterior (PA) chest radiographs (PA-CXR) in detection of pulmonary nodules, infectious consolidation, pneumothorax, pleural effusion, aortic calcification, cardiomegaly and rib fractures compared to radiologists analyzing PA-CXR including dual-energy subtraction radiography (further termed as DESR).

MATERIALS AND METHODS

PA-CXR/DESR images of 197 patients were included. All patients underwent chest CT (gold standard) within a short interval (mean 28 hours). All images were evaluated by three blinded readers for the presence of pulmonary nodules, infectious consolidation, pneumothorax, pleural effusion, aortic calcification, cardiomegaly, and rib fractures. Meanwhile PA-CXR were analyzed by a CAD software. CAD results were compared to the majority result of the three readers. Sensitivity and specificity were calculated. McNemar's test was applied to test for significant differences. Interobserver agreement was defined using Cohen's kappa (κ).

RESULTS

Sensitivity of the CAD software was significantly higher (p < 0.05) for detection of infectious consolidation and pulmonary nodules (67.9% vs 26.8% and 54% vs 35.6%, respectively; p < 0.001) compared to radiologists analyzing DESR images. For the residual evaluated pathologies no statistical significant differences could be found. Overall, mean inter observer agreement between the three radiologists was moderate (k = 0.534). The best interobserver agreement could be reached for pneumothorax (k = 0.708) and pleural effusion (k = 0.699), while the worst was obtained for rib fractures (k = 0.412).

CONCLUSION

The CAD system has the potential to improve the detection of infectious consolidation and pulmonary nodules on CXR images.

[Machine learning and multiparametric MRI for early diagnosis of prostate cancer]

In the last few years, the early diagnosis of prostate cancer has continued to shift from systematic biopsies to multiparametric MRI (mpMRI)-guided/MRI-transrectal ultrasound (TRUS) fusion biopsies and guidelines are already reflecting these changes. While MRI-TRUS fusion biopsies have already resulted in significant improvements in diagnostic sensitivity and, thus, correct diagnosis of clinically significant prostate cancer (sPC), its use to avoid biopsies in certain men is still controversial. Optimal use of mpMRI requires a high degree of reader expertise due to the difficulty of image interpretation and poses the problem of training sufficient numbers of radiologists while demand is increasing. Recently, artificial intelligence (AI) has been utilized to create fully automatic analysis tools for interpretation of mpMRI of the prostate, rivaling the performance of clinical radiologist interpretation in retrospective research studies, demonstrating the promising potential of AI for diagnostic prostate MRI in the future. This article will provide an overview of machine and deep learning and its application in mpMRI of the prostate for early diagnosis of prostate cancer.

Artificial intelligence and its impact on quality improvement in upper and lower gastrointestinal endoscopy

Artificial intelligence (AI) and its application in medicine has grown large interest. Within gastrointestinal (GI) endoscopy, the field of colonoscopy and polyp detection is the most investigated, however, upper GI follows the lead. Since endoscopy is performed by humans, it is inherently an imperfect procedure. Computer-aided diagnosis may improve its quality by helping prevent missing lesions and supporting optical diagnosis for those detected. An entire evolution in AI systems has been established in the last decades, resulting in optimization of the diagnostic performance with lower variability and matching or even outperformance of expert endoscopists. This shows a great potential for future quality improvement of endoscopy, given the outstanding diagnostic features of AI. With this narrative review, we highlight the potential benefit of AI to improve overall quality in daily endoscopy and describe the most recent developments for characterization and diagnosis as well as the recent conditions for regulatory approval.

WHAT CAN WE ACTUALLY SEE USING COMPUTER AIDED DETECTION IN MAMMOGRAPHY?

The main goal of this study was to compare the results of computer aided detection (CAD) analysis in screening mammography with the results independently obtained by two radiologists for the same samples and to determine the sensitivity and specificity of CAD for breast lesions. A total of 436 mammograms were analyzed with CAD. For each screening mammogram, the changes in breast tissue recognized by CAD were compared to the interpretations of two radiologists. The sensitivity and specificity of CAD for breast lesions were calculated using contingency table. The sensitivity of CAD for all lesions was 54% and specificity 16%. CAD sensitivity for suspicious lesions only was 86%. CAD sensitivity for microcalcifications was 100% and specificity 45%. CAD mainly ‘mistook’ glandular parenchyma, connective tissue and blood vessels for breast lesions, and blood vessel calcifications and axillary folds for microcalcifications. In this study, we confirmed CAD as an excellent tool for recognizing microcalcifications with 100% sensitivity. However, it should not be used as a stand-alone tool in breast screening mammography due to the high rate of false-positive results.

Is mammographic density a marker of breast cancer phenotypes?

PURPOSE

To investigate the association between mammographic density (MD) phenotypes and both clinicopathologic features of breast cancer (BC) and tumor location.

METHODS

MD was measured for 297 BC-affected females using qualitative (visual method) and quantitative (fully automated area-based method) approaches. Radiologists' description, visible external markers, and surgical scar were used to establish the location of tumors. Binary logistic regression models were used to assess the association between MD phenotypes and BC clinicopathologic features.

RESULTS

Categorical and numerical MD measures showed no association with clinicopathologic features of BC (p > 0.05). Participants with higher BI-RADS scores [(51-75% glandular) and (> 75% glandular)] (p < 0.001), and percent density (PD) categories [PD (21-49%) and PD ≥ 50%] (p = 0.01) were more likely to have tumors emanating from dense areas. Additionally, tumors were commonly found in dense regions of the breast among patients with higher medians of PD (p = 0.001), dense area (DA) (p = 0.02), and lower medians of non-dense area (NDA) (p < 0.001). Adjusted logistic regression models showed that high BI-RADS density (> 75% glandular) has an almost fivefold increased odds of tumors developing within dense areas (OR 4.99, 95% CI 0.93-25.9; p = 0.05. PD (OR 1.02, 95% CI 1-1.03, p = 0.002) and NDA (OR 0.99, 95% CI 0.991-0.997, p < 0.001) had very small effect on tumor location. Compared to tumors within non-dense areas, tumors in dense areas tended to exhibit human epidermal growth factor receptor 2 positive (p = 0.05) and carcinoma in situ (p = 0.01) characteristics.

CONCLUSION

MD shows no significant association with clinicopathologic features of BC. However, BC was more likely to originate from dense tissue, with tumors in dense regions having human epidermal growth receptor 2 positive and carcinoma in situ characteristics.

Changes in cancer detection and false-positive recall in mammography using artificial intelligence: a retrospective, multireader study

BACKGROUND

Mammography is the current standard for breast cancer screening. This study aimed to develop an artificial intelligence (AI) algorithm for diagnosis of breast cancer in mammography, and explore whether it could benefit radiologists by improving accuracy of diagnosis.

METHODS

In this retrospective study, an AI algorithm was developed and validated with 170 230 mammography examinations collected from five institutions in South Korea, the USA, and the UK, including 36 468 cancer positive confirmed by biopsy, 59 544 benign confirmed by biopsy (8827 mammograms) or follow-up imaging (50 717 mammograms), and 74 218 normal. For the multicentre, observer-blinded, reader study, 320 mammograms (160 cancer positive, 64 benign, 96 normal) were independently obtained from two institutions. 14 radiologists participated as readers and assessed each mammogram in terms of likelihood of malignancy (LOM), location of malignancy, and necessity to recall the patient, first without and then with assistance of the AI algorithm. The performance of AI and radiologists was evaluated in terms of LOM-based area under the receiver operating characteristic curve (AUROC) and recall-based sensitivity and specificity.

FINDINGS

The AI standalone performance was AUROC 0·959 (95% CI 0·952-0·966) overall, and 0·970 (0·963-0·978) in the South Korea dataset, 0·953 (0·938-0·968) in the USA dataset, and 0·938 (0·918-0·958) in the UK dataset. In the reader study, the performance level of AI was 0·940 (0·915-0·965), significantly higher than that of the radiologists without AI assistance (0·810, 95% CI 0·770-0·850; p<0·0001). With the assistance of AI, radiologists' performance was improved to 0·881 (0·850-0·911; p<0·0001). AI was more sensitive to detect cancers with mass (53 [90%] vs 46 [78%] of 59 cancers detected; p=0·044) or distortion or asymmetry (18 [90%] vs ten [50%] of 20 cancers detected; p=0·023) than radiologists. AI was better in detection of T1 cancers (73 [91%] vs 59 [74%] of 80; p=0·0039) or node-negative cancers (104 [87%] vs 88 [74%] of 119; p=0·0025) than radiologists.

INTERPRETATION

The AI algorithm developed with large-scale mammography data showed better diagnostic performance in breast cancer detection compared with radiologists. The significant improvement in radiologists' performance when aided by AI supports application of AI to mammograms as a diagnostic support tool.

FUNDING

Lunit.

A technical review of artificial intelligence as applied to gastrointestinal endoscopy: clarifying the terminology

Background and aim  The growing number of publications on the application of artificial intelligence (AI) in medicine underlines the enormous importance and potential of this emerging field of research. In gastrointestinal endoscopy, AI has been applied to all segments of the gastrointestinal tract most importantly in the detection and characterization of colorectal polyps. However, AI research has been published also in the stomach and esophagus for both neoplastic and non-neoplastic disorders. The various technical as well as medical aspects of AI, however, remain confusing especially for non-expert physicians. This physician-engineer co-authored review explains the basic technical aspects of AI and provides a comprehensive overview of recent publications on AI in gastrointestinal endoscopy. Finally, a basic insight is offered into understanding publications on AI in gastrointestinal endoscopy.

Outcomes of classic lobular neoplasia diagnosed on breast core needle biopsy: a retrospective multi-center study

Management of classic lobular neoplasia (cLN) diagnosed on core needle biopsy (CNB) is controversial. Our aim in this study was to review cases of cLN diagnosed on CNB to determine the rate and risk factors of an upgrade to ductal carcinoma in situ (DCIS) or invasive carcinoma on excision. All breast CNBs with a diagnosis of atypical lobular hyperplasia (ALH) or classic lobular carcinoma in situ (cLCIS) from three different institutions within a single health care system between 2013 and 2018 were retrieved. Cases with any additional high-risk lesions in the same CNB or discordant radiological-pathological correlation were excluded. Information about age, personal history of prior or concurrent breast cancer (P/CBC), and radiological and histological findings were recorded. A total of 287 cLN cases underwent surgical excision. Analysis of these 287 cLN cases showed 11 (3.8%) upgrade lesions on excision. Among the 172 ALH cases, there were 3 (1.7%) upgrades, which were all invasive lobular carcinomas (ILCs). On the other hand, 8 of 115 (7%) cLCIS cases revealed upgrade on excision (2 ILC, 5 DCIS. and 1 ILC + DCIS). Statistical analysis revealed that cLN cases with P/CBC, radiological asymmetry, or architectural distortion had a statistically significant higher upgrade rate on excision. Our findings revealed a low upgrade rate (3.8%) on the excision of classic lobular neoplasia diagnosed on breast core needle biopsy. Clinicoradiological surveillance can be appropriate when lobular neoplasia is identified on core biopsy with pathological radiological concordance in patients without a history of breast cancer, with the caveat that radiological asymmetry and architectural distortion are associated with a significant increase in an upgrade on excision.

New Frontiers: An Update on Computer-Aided Diagnosis for Breast Imaging in the Age of Artificial Intelligence

OBJECTIVE

The purpose of this article is to compare traditional versus machine learning-based computer-aided detection (CAD) platforms in breast imaging with a focus on mammography, to underscore limitations of traditional CAD, and to highlight potential solutions in new CAD systems under development for the future.

CONCLUSION

CAD development for breast imaging is undergoing a paradigm shift based on vast improvement of computing power and rapid emergence of advanced deep learning algorithms, heralding new systems that may hold real potential to improve clinical care.

Artificial Intelligence in Pathology

As in other domains, artificial intelligence is becoming increasingly important in medicine. In particular,deep learning-based pattern recognition methods can advance the field of pathology byincorporating clinical, radiologic, and genomic data to accurately diagnose diseases and predictpatient prognoses. In this review, we present an overview of artificial intelligence, the brief historyof artificial intelligence in the medical domain, recent advances in artificial intelligence applied topathology, and future prospects of pathology driven by artificial intelligence.

Emergence of "Big Data" and Its Potential and Current Limitations in Medical Imaging

Although electronic imaging was performed in the early 1950s in nuclear medicine, it was the introduction of computed tomography in 1972 that caused a revolution in medical imaging in that it marked the beginning of the inevitable transformation to digital imaging. This transformation is now more or less complete. While initially these CT images were relatively small, comprised of only about 6400 pixels per slice, the steady move toward higher spatial resolution, multislice imaging, digital radiography, and fluoroscopy rapidly increased the size of images and the amount of data required to be stored, processed, displayed, and moved about in a medical imaging department. The more recent introduction of digital pathology with submicron-sized pixels and the need for color further increases these demands. Rising work volumes in hospital, a push for cost containment, and a move toward greater precision in diagnosis and treatment of disease all work together to motivate the development of automated image analysis algorithms and techniques to improve efficiencies in in vivo imaging and pathology. This may require bringing together information from different imaging and nonimaging sources within the institution. While technological development has provided practical means for storage of the burgeoning data load and the use of multiple processors and high-speed networks has enabled more sophisticated analysis locally or in the cloud, challenges remain in terms of the ability to integrate data from different systems, the development of appropriately annotated image bases for training and testing of algorithms, and issues around privacy and ownership in obtaining access to patient-related data.

Machine Learning in Neurooncology Imaging: From Study Request to Diagnosis and Treatment

OBJECTIVE

Machine learning has potential to play a key role across a variety of medical imaging applications. This review seeks to elucidate the ways in which machine learning can aid and enhance diagnosis, treatment, and follow-up in neurooncology.

CONCLUSION

Given the rapid pace of development in machine learning over the past several years, a basic proficiency of the key tenets and use cases in the field is critical to assessing potential opportunities and challenges of this exciting new technology.

A review of computer aided detection in mammography

Breast screening with mammography is widely recognized as the most effective method of detecting early breast cancer and has consistently demonstrated a 20-40% decrease in mortality among screened women. Despite this, the sensitivity of mammography ranges between 70 and 90%. Computer aided detection (CAD) is an artificial intelligence (AI) technique that utilizes pattern recognition to highlight suspicious features on imaging and marks them for the radiologist to review and interpret. It aims to decrease oversights made by interpreting radiologists. Here we review the efficacy of CAD and potential future directions.

Pathologic Outcomes of Architectural Distortion on Digital 2D Versus Tomosynthesis Mammography

OBJECTIVE

The purpose of this study is to compare the risk of malignancy associated with architectural distortion detected on 2D digital mammography (DM) versus digital breast tomosynthesis (DBT).

MATERIALS AND METHODS

We performed a retrospective review of architectural distortion cases recommended for biopsy from September 2007 to February 2011, the period before DBT integration (hereafter known as the DM group), and from January 2013 to June 2016, the period after DBT integration (hereafter known as the DBT group). Medical records were reviewed for imaging findings and pathology results.

RESULTS

Architectural distortion was more commonly detected in the DBT group than the DM group (0.14% [274/202,438 examinations] vs 0.07% [121/166,661 examinations]; p < 0.001). The positive predictive value of architectural distortion for malignancy was significantly lower in the DBT group than the DM group (50.7% [139/274 cases] vs 73.6% [89/121 cases]; p < 0.001). Radial scar was the most common nonmalignant finding in both groups, but it was more common in the DBT group (33.2% [91/274] vs 11.6% [14/121]; p < 0.001). In the DBT group, architectural distortion without correlative findings on ultrasound was less likely to represent malignancy than was architectural distortion with correlative findings on ultrasound (29.2% [31/106] vs 66.5% [105/158]; p < 0.001).

CONCLUSION

Architectural distortion is more commonly detected on DBT than DM and is less likely to represent malignancy on DBT. Architectural distortion on DBT is less likely to represent malignancy if there is no sonographic correlate; however, biopsy is warranted even in the absence of a sonographic correlate, given the nearly 30% risk of malignancy in this setting.

MRI in the differential diagnosis of primary architectural distortion detected by mammography

PURPOSE

We aimed to evaluate the diagnostic accuracy of a combination of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and apparent diffusion coefficient (ADC) values in lesions that manifest with architectural distortion (AD) on mammography.

METHODS

All full-field digital mammography (FFDM) images obtained between August 2010 and January 2013 were reviewed retrospectively, and 57 lesions showing AD were included in the study. Two independent radiologists reviewed all mammograms and MRI data and recorded lesion characteristics according to the BI-RADS lexicon. The gold standard was histopathologic results from biopsies or surgical excisions and results of the two-year follow-up. Receiver operating characteristic curve analysis was carried out to define the most effective threshold ADC value to differentiate malignant from benign breast lesions. We investigated the sensitivity and specificity of FFDM, DCE-MRI, FFDM+DCE-MRI, and DCE-MRI+ADC.

RESULTS

Of the 57 lesions analyzed, 28 were malignant and 29 were benign. The most effective threshold for the normalized ADC (nADC) was 0.61 with 93.1% sensitivity and 75.0% specificity. The sensitivity and specificity of DCE-MRI combined with nADC was 92.9% and 79.3%, respectively. DCE-MRI combined with nADC showed the highest specificity and equal sensitivity compared with other modalities, independent of the presentation of calcification.

CONCLUSION

DCE-MRI combined with nADC values was more reliable than mammography in differentiating the nature of disease manifesting as primary AD on mammography.

Characterizing Architectural Distortion in Mammograms by Linear Saliency

Architectural distortion (AD) is a common cause of false-negatives in mammograms. This lesion usually consists of a central retraction of the connective tissue and a spiculated pattern radiating from it. This pattern is difficult to detect due the complex superposition of breast tissue. This paper presents a novel AD characterization by representing the linear saliency in mammography Regions of Interest (ROI) as a graph composed of nodes corresponding to locations along the ROI boundary and edges with a weight proportional to the line intensity integrals along the path connecting any pair of nodes. A set of eigenvectors from the adjacency matrix is then used to extract discriminant coefficients that represent those nodes with higher salient lines. A dimensionality reduction is further accomplished by selecting the pair of nodes with major contribution for each of the computed eigenvectors. The set of main salient lines is then assembled as a feature vector that inputs a conventional Support Vector Machine (SVM). Experimental results with two benchmark databases, the mini-MIAS and DDSM databases, demonstrate that the proposed linear saliency domain method (LSD) performs well in terms of accuracy. The approach was evaluated with a set of 246 RoI extracted from the DDSM (123 normal tissues and 123 AD) and a set of 38 ROI from the mini-MIAS collections (19 normal tissues and 19 AD) respectively. The classification results showed respectively for both databases an accuracy rate of 89 % and 87 %, a sensitivity rate of 85 % and 95 %, and a specificity rate of 93 % and 84 %. Likewise, the area under curve (A _z ) of the Receiver Operating Characteristic (ROC) curve was 0.93 for both databases.

Architectural Distortion on Mammography: Correlation With Pathologic Outcomes and Predictors of Malignancy

OBJECTIVE

The objective of our study was to determine the risk of malignancy associated with architectural distortion and to evaluate the imaging and clinical features that may contribute to the prediction of malignancy in the setting of architectural distortion.

MATERIALS AND METHODS

We performed a retrospective review of architectural distortion cases from January 1, 2004, to December 31, 2013. Imaging findings and pathology outcomes were reviewed.

RESULTS

Over the 10-year study period, architectural distortion that was considered to be suspicious for or highly suggestive of malignancy was present in 435 of 231,051 (0.2%) mammographic examinations. Cases were excluded if the main finding described was a mass with an associated feature of architectural distortion (n = 62) or if no pathology results were available (n = 4). Two hundred seventy-five cases of invasive adenocarcinoma or ductal carcinoma in situ (DCIS) were identified; the positive predictive value (PPV) was therefore 74.5% (275/369). DCIS alone was identified in only 4.1% (15/369). The most common benign finding on pathology was a radial scar or complex sclerosing lesion (27/369, 7.3%). Architectural distortion was less likely to represent malignancy on screening mammography than on diagnostic mammography (67.0% vs 83.1%, respectively; p < 0.001). Architectural distortion without a sonographic correlate was less likely to represent malignancy than architectural distortion with a correlate (27.9% vs 82.9%, respectively; p < 0.001). There was no statistically significant difference in the malignancy rate between pure architectural distortion and architectural distortion with calcifications or asymmetries (73.0% vs 78.8%; p = 0.26).

CONCLUSION

The PPV of architectural distortion for malignancy is 74.5%. Architectural distortion is less likely to represent malignancy if detected on screening mammography than on diagnostic mammography or if there is no sonographic correlate.

In the digital era, architectural distortion remains a challenging radiological task

AIM

To compare readers' performance in detecting architectural distortion (AD) compared with other breast cancer types using digital mammography.

MATERIALS AND METHODS

Forty-one experienced breast screen readers (20 US and 21 Australian) were asked to read a single test set of 30 digitally acquired mammographic cases. Twenty cases had abnormal findings (10 with AD, 10 non-AD) and 10 cases were normal. Each reader was asked to locate and rate any abnormalities. Lesion and case-based performance was assessed. For each collection of readers (US; Australian; combined), jackknife free-response receiver operating characteristic (JAFROC), figure of merit (FOM), and inferred receiver operating characteristic (ROC), area under curve (Az) were calculated using JAFROC v.4.1 software. Readers' sensitivity, location sensitivity, JAFROC, FOM, ROC, Az scores were compared between cases groups using Wilcoxon's signed ranked test statistics.

RESULTS

For lesion-based analysis, significantly lower location sensitivity (p=0.001) was shown on AD cases compared with non-AD cases for all reader collections. The case-based analysis demonstrated significantly lower ROC Az values (p=0.02) for the first collection of readers, and lower sensitivity for the second collection of readers (p=0.04) and all-readers collection (p=0.008), for AD compared with non-AD cases.

CONCLUSIONS

The current work demonstrates that AD remains a challenging task for readers, even in the digital era.

Bilateral Image Subtraction and Multivariate Models for the Automated Triaging of Screening Mammograms

Mammography is the most common and effective breast cancer screening test. However, the rate of positive findings is very low, making the radiologic interpretation monotonous and biased toward errors. This work presents a computer-aided diagnosis (CADx) method aimed to automatically triage mammogram sets. The method coregisters the left and right mammograms, extracts image features, and classifies the subjects into risk of having malignant calcifications (CS), malignant masses (MS), and healthy subject (HS). In this study, 449 subjects (197 CS, 207 MS, and 45 HS) from a public database were used to train and evaluate the CADx. Percentile-rank (p-rank) and z-normalizations were used. For the p-rank, the CS versus HS model achieved a cross-validation accuracy of 0.797 with an area under the receiver operating characteristic curve (AUC) of 0.882; the MS versus HS model obtained an accuracy of 0.772 and an AUC of 0.842. For the z-normalization, the CS versus HS model achieved an accuracy of 0.825 with an AUC of 0.882 and the MS versus HS model obtained an accuracy of 0.698 and an AUC of 0.807. The proposed method has the potential to rank cases with high probability of malignant findings aiding in the prioritization of radiologists work list.

Breast Cancer: Computer-aided Detection with Digital Breast Tomosynthesis

PURPOSE

To evaluate a commercial tomosynthesis computer-aided detection (CAD) system in an independent, multicenter dataset.

MATERIALS AND METHODS

Diagnostic and screening tomosynthesis mammographic examinations (n = 175; cranial caudal and mediolateral oblique) were randomly selected from a previous institutional review board-approved trial. All subjects gave informed consent. Examinations were performed in three centers and included 123 patients, with 132 biopsy-proven screening-detected cancers, and 52 examinations with negative results at 1-year follow-up. One hundred eleven lesions were masses and/or microcalcifications (72 masses, 22 microcalcifications, 17 masses with microcalcifications) and 21 were architectural distortions. Lesions were annotated by radiologists who were aware of all available reports. CAD performance was assessed as per-lesion sensitivity and false-positive results per volume in patients with negative results.

RESULTS

Use of the CAD system showed per-lesion sensitivity of 89% (99 of 111; 95% confidence interval: 81%, 94%), with 2.7 ± 1.8 false-positive rate per view, 62 of 72 lesions detected were masses, 20 of 22 were microcalcification clusters, and 17 of 17 were masses with microcalcifications. Overall, 37 of 39 microcalcification clusters (95% sensitivity, 95% confidence interval: 81%, 99%) and 79 of 89 masses (89% sensitivity, 95% confidence interval: 80%, 94%) were detected with the CAD system. On average, 0.5 false-positive rate per view were microcalcification clusters, 2.1 were masses, and 0.1 were masses and microcalcifications.

CONCLUSION

A digital breast tomosynthesis CAD system can allow detection of a large percentage (89%, 99 of 111) of breast cancers manifesting as masses and microcalcification clusters, with an acceptable false-positive rate (2.7 per breast view). Further studies with larger datasets acquired with equipment from multiple vendors are needed to replicate the findings and to study the interaction of radiologists and CAD systems.

A novel image toggle tool for comparison of serial mammograms: automatic density normalization and alignment-development of the tool and initial experience

PURPOSE

The purpose is to develop a new image toggle tool with automatic density normalization (ADN) and automatic alignment (AA) for comparing serial digital mammograms (DMGs).

MATERIALS AND METHODS

We developed an ADN and AA process to compare the images of serial DMGs. In image density normalization, a linear interpolation was applied by taking two points of high- and low-brightness areas. The alignment was calculated by determining the point of the greatest correlation while shifting the alignment between the current and prior images. These processes were performed on a PC with a 3.20-GHz Xeon processor and 8 GB of main memory. We selected 12 suspected breast cancer patients who had undergone screening DMGs in the past. Automatic processing was retrospectively performed on these images. Two radiologists subjectively evaluated them.

RESULTS

The process of the developed algorithm took approximately 1 s per image. In our preliminary experience, two images could not be aligned approximately. When they were aligned, image toggling allowed detection of differences between examinations easily.

CONCLUSIONS

We developed a new tool to facilitate comparative reading of DMGs on a mammography viewing system. Using this tool for toggling comparisons might improve the interpretation efficiency of serial DMGs.

Stellate masses and histologic grades in breast cancer

Breast masses with a radiologic stellate pattern often transform into malignancies, but their tendency to be of low histologic grade yields a better survival rate compared with tumors with other patterns on mammography screening. This study was designed to investigate the correlation of histologic grade with stellate features extracted from the coronal plane of 3-D ultrasound images. A pre-processing method was proposed to facilitate the extraction of stellate features. Extracted features were statistically measured to derive a set of indices that quantitatively represent the stellate pattern. These indices then went through a selection procedure to build proper decision trees. The splitting rules of decision trees indicated that stellate tumors are associated with low grade. A set of indices from the low grade-associated rules has the potential to represent the stellate feature. Further investigation of the hypoechoic region of peripheral tissue is essential to establishment of a complete discriminating model for tumor grades.

Detection of breast cancer with a computer-aided detection applied to full-field digital mammography

A study was conducted to evaluate the sensitivity of computer-aided detection (CAD) with full-field digital mammography in detection of breast cancer, based on mammographic appearance and histopathology. Retrospectively, CAD sensitivity was assessed in total group of 152 cases for subgroups based on breast density, mammographic presentation, lesion size, and results of histopathological examination. The overall sensitivity of CAD was 91 % (139 of 152 cases). CAD detected 100 % (47/47) of cancers manifested as microcalcifications; 98 % (62/63) of those manifested as non-calcified masses; 100 % (15/15) of those manifested as mixed masses and microcalcifications; 75 % (12/16) of those manifested as architectural distortions, and 69 % (18/26) of those manifested as focal asymmetry. CAD sensitivity was 83 % (10/12) for cancers measuring 1-10 mm, 92 % (37/40) for those measuring 11-20 mm, and 92 % (92/100) for those measuring >20 mm. There was no significant difference in CAD detection efficiency between cancers in dense breasts (88 %; 69/78) and those in non-dense breasts (95 %; 70/74). CAD showed a high sensitivity of 91 % (139/152) for the mammographic appearance of cancer and 100 % sensitivity for identifying cancers manifested as microcalcifications. Sensitivity was not influenced by breast density or lesion size. CAD should be effective for helping radiologists detect breast cancer at an earlier stage.

Computer-aided detection of breast cancer on mammograms: a swarm intelligence optimized wavelet neural network approach

Breast cancer is the second leading cause of cancer death in women. Accurate early detection can effectively reduce the mortality rate caused by breast cancer. Masses and microcalcification clusters are an important early signs of breast cancer. However, it is often difficult to distinguish abnormalities from normal breast tissues because of their subtle appearance and ambiguous margins. Computer aided diagnosis (CAD) helps the radiologist in detecting the abnormalities in an efficient way. This paper investigates a new classification approach for detection of breast abnormalities in digital mammograms using Particle Swarm Optimized Wavelet Neural Network (PSOWNN). The proposed abnormality detection algorithm is based on extracting Laws Texture Energy Measures from the mammograms and classifying the suspicious regions by applying a pattern classifier. The method is applied to real clinical database of 216 mammograms collected from mammogram screening centers. The detection performance of the CAD system is analyzed using Receiver Operating Characteristic (ROC) curve. This curve indicates the trade-offs between sensitivity and specificity that is available from a diagnostic system, and thus describes the inherent discrimination capacity of the proposed system. The result shows that the area under the ROC curve of the proposed algorithm is 0.96853 with a sensitivity 94.167% of and specificity of 92.105%.

Role of computer-aided detection in very small screening detected invasive breast cancers

This study aims to assess computer-aided detection (CAD) performance with full-field digital mammography (FFDM) in very small (equal to or less than 1 cm) invasive breast cancers. Sixty-eight invasive breast cancers less than or equal to 1 cm were retrospectively studied. All cases were detected with FFDM in women aged 49-69 years from our breast cancer screening program. Radiological characteristics of lesions following BI-RADS descriptors were recorded and compared with CAD sensitivity. Age, size, BI-RADS classification, breast density type, histological type of the neoplasm, and role of the CAD were also assessed. Per-study specificity and mass false-positive rate were determined by using 100 normal consecutive studies. Thirty-seven (54.4 %) masses, 17 (25 %) calcifications, 6 (8.8 %) masses with calcifications, 7 (10.3 %) architectural distortions, and 1 asymmetry (1.5 %) were found. CAD showed an overall sensitivity of 86.7 % (masses, 86.5 %; calcifications, 100 %; masses with calcifications, 100 %; and architectural distortion, 57.14 %), CAD failed to detect 9 out of 68 cases: 5 of 37 masses, 3 of 7 architectural distortions, and 1 of 1 asymmetry. Fifteen out of 37 masses were hyperdense, and all of them were detected by CAD. No association was seen among mass morphology or margins and detectability. Per-study specificity and CAD false-positive rate was 26 % and 1.76 false marks per study. In conclusion, CAD shows a high sensitivity and a low specificity. Lesion size, histology, and breast density do not influence sensitivity. Mammographic features, mass density, and thickness of the spicules in architectural distortions do influence.

Detection of architectural distortion in prior mammograms via analysis of oriented patterns

We demonstrate methods for the detection of architectural distortion in prior mammograms of interval-cancer cases based on analysis of the orientation of breast tissue patterns in mammograms. We hypothesize that architectural distortion modifies the normal orientation of breast tissue patterns in mammographic images before the formation of masses or tumors. In the initial steps of our methods, the oriented structures in a given mammogram are analyzed using Gabor filters and phase portraits to detect node-like sites of radiating or intersecting tissue patterns. Each detected site is then characterized using the node value, fractal dimension, and a measure of angular dispersion specifically designed to represent spiculating patterns associated with architectural distortion. Our methods were tested with a database of 106 prior mammograms of 56 interval-cancer cases and 52 mammograms of 13 normal cases using the features developed for the characterization of architectural distortion, pattern classification via quadratic discriminant analysis, and validation with the leave-one-patient out procedure. According to the results of free-response receiver operating characteristic analysis, our methods have demonstrated the capability to detect architectural distortion in prior mammograms, taken 15 months (on the average) before clinical diagnosis of breast cancer, with a sensitivity of 80% at about five false positives per patient.