Search

Objectives: This study explored the familiarity, perceptions and confidence of Australian radiology clinicians involved in reading screening mammograms, regarding artificial intelligence (AI) applications in breast cancer detection. Methods: Sixty-five radiologists, breast physicians and radiology trainees participated in an online survey that consisted of 23 multiple choice questions asking about their experience and familiarity with AI products. Furthermore, the survey asked about their confidence in using AI outputs and their preference for AI modes applied in a breast screening context. Participants' responses to questions were compared using Pearson's χ 2 test. Bonferroni-adjusted significance tests were used for pairwise comparisons. Results: Fifty-five percent of respondents had experience with AI in their workplaces, with automatic density measurement powered by machine learning being the most familiar AI product (69.4%). The top AI outputs with the highest ranks of perceived confidence were 'Displaying suspicious areas on mammograms with the percentage of cancer possibility' (67.8%) and 'Automatic mammogram classification (normal, benign, cancer, uncertain)' (64.6%). Radiology and breast physicians preferred using AI as second-reader mode (75.4% saying 'somewhat happy' to 'extremely happy') over triage (47.7%), pre-screening and first-reader modes (both with 26.2%) (P < 0.001). Conclusion: The majority of screen readers expressed increased confidence in utilising AI for highlighting suspicious areas on mammograms and for automatically classifying mammograms. They considered AI as an optimal second-reader mode being the most ideal use in a screening program. The findings provide valuable insights into the familiarities and expectations of radiologists and breast clinicians for the AI products that can enhance the effectiveness of the breast cancer screening programs, benefitting both healthcare professionals and patients alike. What is known about the topic? Artificial intelligence (AI) holds promise in providing computer-aided detection in health care, however, current research suggests that standalone AI applications in clinical practice fall short of matching the accuracy of a single radiologist. What does this paper add? The study showed a significant preference among clinicians for using AI as a supplementary tool, serving as a second-reader. Such an integrated approach, where AI aids in flagging suspicious areas on mammograms or offers automatic classification, reflects the ideal cooperation between breast screening readers and AI systems. What are the implications for practitioners? These insights shed light on clinicians' familiarity with and expectations of AI tools that can boost the effectiveness of breast screening programs. [ABSTRACT FROM AUTHOR]

Simple Summary: Breast cancer is one of the leading causes of cancer-related death in women. The early detection of breast cancer with screening mammograms plays a pivotal role in reducing mortality rates. Although population-based double-reading screening mammograms have reduced mortality by over 31% in women with breast cancer in Europe, continuing this program is difficult due to the shortage of radiologists. Artificial intelligence (AI) is an emerging technology that has provided promising results in medical imaging for disease detection. This study investigates the performance of AI models on an Australian mammographic database, demonstrating how transfer learning from a USA mammographic database to an Australian one, contrast enhancement on mammographic images and the quality of training data according to radiologists' concordance can improve breast cancer diagnosis. Our proposed methodology offers a more efficacious approach for AI to contribute to radiologists' decision making when interpreting mammography images. This paper investigates the adaptability of four state-of-the-art artificial intelligence (AI) models to the Australian mammographic context through transfer learning, explores the impact of image enhancement on model performance and analyses the relationship between AI outputs and histopathological features for clinical relevance and accuracy assessment. A total of 1712 screening mammograms (n = 856 cancer cases and n = 856 matched normal cases) were used in this study. The 856 cases with cancer lesions were annotated by two expert radiologists and the level of concordance between their annotations was used to establish two sets: a 'high-concordances subset' with 99% agreement of cancer location and an 'entire dataset' with all cases included. The area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of Globally aware Multiple Instance Classifier (GMIC), Global-Local Activation Maps (GLAM), I&H and End2End AI models, both in the pretrained and transfer learning modes, with and without applying the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm. The four AI models with and without transfer learning in the high-concordance subset outperformed those in the entire dataset. Applying the CLAHE algorithm to mammograms improved the performance of the AI models. In the high-concordance subset with the transfer learning and CLAHE algorithm applied, the AUC of the GMIC model was highest (0.912), followed by the GLAM model (0.909), I&H (0.893) and End2End (0.875). There were significant differences (p < 0.05) in the performances of the four AI models between the high-concordance subset and the entire dataset. The AI models demonstrated significant differences in malignancy probability concerning different tumour size categories in mammograms. The performance of AI models was affected by several factors such as concordance classification, image enhancement and transfer learning. Mammograms with a strong concordance with radiologists' annotations, applying image enhancement and transfer learning could enhance the accuracy of AI models. [ABSTRACT FROM AUTHOR]

Background: Computerized image analysis is a viable technique for evaluating image quality as a complement to human observers. Purpose: To systematically review the image analysis software used in the assessment of 2D image quality using mammography phantoms. Material and Methods: A systematic search of multiple databases was performed from inception to July 2020 for articles that incorporated computerized analysis of 2D images of physical mammography phantoms to determine image quality. Results: A total of 26 studies were included, 12 were carried out using direct digital imaging and 14 using screen film mammography. The ACR phantom (model-156) was the most frequently evaluated phantom, possibly due to the lack of accepted standard software. In comparison to the inter-observer variations, the computerized image analysis was more consistent in scoring test objects. The template matching method was found to be one of the most reliable algorithms, especially for high-contrast test objects, while several algorithms found low-contrast test objects to be harder to distinguish due to the smaller contrast variations between test objects and their backgrounds. This was particularly true for small object sizes. Conclusion: Image analysis software was in agreement with human observers but demonstrated higher consistency and reproducibility of quality evaluation. Additionally, using computerized analysis, several quantitative metrics such as contrast-to-noise ratio (CNR) and the signal-to-noise ratio (SNR) could be used to complement the conventional scoring method. Implementing a computerized approach for monitoring image quality over time would be crucial to detect any deteriorating mammography system before clinical images are impacted. [ABSTRACT FROM AUTHOR]

Introduction: Phase‐contrast imaging (PCI) is a novel technology that can visualise variations in X‐ray refraction (phase contrast) in addition to differences in X‐ray attenuation (absorption contrast). Compared to radiography using conventional methods (i.e. absorption‐based imaging), PCI techniques can potentially produce images with higher contrast‐to‐noise ratio and superior spatial resolution at the same or lower radiation doses. This has led PCI to be explored for implementation in medical imaging. While interest in this research field is increasing, the whole body of PCI research in medical imaging has been under‐investigated. This paper provides an overview of PCI literature and then focusses on evaluating its development within the scope of medical imaging. Methods: Bibliographic data between 1995 and 2018 were used to visualise collaboration networks between countries, institutions and authors. Social network analysis techniques were implemented to measure these networks in terms of centrality and cohesion. These techniques also assisted in the exploration of underlying research paradigms of clinical X‐ray PCI investigations. Results: Forty‐one countries, 592 institutions and 2073 authors contributed 796 investigations towards clinical PCI research. The most influential contributors and network collaboration characteristics were identified. Italy was the most influential country, with the European Synchrotron Radiation Facility being the most influential institution. At an author level, F. Pfeiffer was found to be the most influential researcher. Among various PCI techniques, grating interferometry was the most investigated, while computed tomography was the most frequently examined modality. Conclusions: By gaining an understanding of collaborations and trends within clinical X‐ray PCI research, the links between existing collaborators were identified, which can aid future collaborations between emerging and established collaborators. Moreover, exploring the paradigm of past investigations can shape future research – well‐researched PCI techniques may be studied to bring X‐ray PCI closer to clinical implementation, or the potential of seldom‐investigated techniques may be explored. [ABSTRACT FROM AUTHOR]

Rationale and Objectives: Propagation-based phase-contrast CT (PB-CT) is an advanced X-ray imaging technology that exploits both refraction and absorption of the transmitted X-ray beam. This study was aimed at optimizing the experimental conditions of PB-CT for breast cancer imaging and examined its performance relative to conventional absorption-based CT (AB-CT) in terms of image quality and radiation dose.Materials and Methods: Surgically excised breast mastectomy specimens (n = 12) were scanned using both PB-CT and AB-CT techniques under varying imaging conditions. To evaluate the radiological image quality, visual grading characteristics (VGC) analysis was used in which 11 breast specialist radiologists compared the overall image quality of PB-CT images with respect to the corresponding AB-CT images. The area under the VGC curve was calculated to measure the differences between PB-CT and AB-CT images.Results: The highest radiological quality was obtained for PB-CT images using a 32 keV energy X-ray beam and by applying the Homogeneous Transport of Intensity Equation phase retrieval with the value of its parameter γ set to one-half of the theoretically optimal value for the given materials. Using these optimized conditions, the image quality of PB-CT images obtained at 4 mGy and 2 mGy mean glandular dose was significantly higher than AB-CT images at 4 mGy (AUCVGC = 0.901, p = 0.001 and AUCVGC = 0.819, p = 0.011, respectively).Conclusion: PB-CT achieves a higher radiological image quality compared to AB-CT even at a considerably lower mean glandular dose. Successful translation of the PB-CT technique for breast cancer imaging can potentially result in improved breast cancer diagnosis. [ABSTRACT FROM AUTHOR]

Rationale and Objectives: This study employs clinical/radiological evaluation in establishing the optimum imaging conditions for breast cancer imaging using the X-ray propagation-based phase-contrast tomography.Materials and Methods: Two series of experiments were conducted and in total 161 synchrotron-based computed tomography (CT) reconstructions of one breast mastectomy specimen were produced at different imaging conditions. Imaging factors include sample-to-detector distance, X-ray energy, CT reconstruction method, phase retrieval algorithm applied to the CT projection images and maximum intensity projection. Observers including breast radiologists and medical imaging experts compared the quality of the reconstructed images with reference images approximating the conventional (absorption) CT. Various radiological image quality attributes in a visual grading analysis design were used for the radiological assessments.Results: The results show that the application of the longest achievable sample-to-detector distance (9.31 m), the lowest employed X-ray energy (32 keV), the full phase retrieval, and the maximum intensity projection can significantly improve the radiological quality of the image. Several combinations of imaging variables resulted in images with very high-quality scores.Conclusion: The results of the present study will support future experimental and clinical attempts to further optimize this innovative approach to breast cancer imaging. [ABSTRACT FROM AUTHOR]

This study presents a quantitative analysis of global breast imaging research over the last eight decades. A dedicated Sino-Australian case study via a social network analysis (SNA) is included as China and Australia have a recent rapidly increasing number of research partnerships and strategic education/economic connections. Bibliographic data was extracted via Scopus and analysed for the social network parameters of degree centrality, closeness centrality, betweenness centrality and multiple cohesion measures in order to explore research collaboration networks at the organisational level. Within the last three decades there has been a tremendous increase in the publication rate within the scientific domain of breast imaging research, however, there is a significant lag in the development of this research area in China compared with Australia. Breast imaging research in China is considerably more insular, with less international collaboration and reduced variation between collaborators than Australia. The impact of national breast screening programs and novel cancer technologies upon collaboration networks is discussed alongside the ability of networks paradigm to reveal both frailties in research connections and to highlight networking strategies. [ABSTRACT FROM AUTHOR]

Objectives: Propagation‐based phase‐contrast computed tomography (PB‐CT) is a new imaging technique that exploits refractive and absorption properties of X‐rays to enhance soft tissue contrast and improve image quality. This study compares image quality of PB‐CT and absorption‐based CT (AB‐CT) for breast imaging while exploring X‐ray energy and radiation dose. Methods: Thirty‐nine mastectomy samples were scanned at energy levels of 28‐34keV using a flat panel detector at radiation dose levels of 4mGy and 2mGy. Image quality was assessed using signal‐to‐noise ratio (SNR), contrast‐to‐noise ratio (CNR), spatial resolution (res) and visibility (vis). Statistical analysis was performed to compare PB‐CT images against their corresponding AB‐CT images scanned at 32keV and 4mGy. Results: The PB‐CT images at 4mGy, across nearly all energy levels, demonstrated superior image quality than AB‐CT images at the same dose. At some energy levels, the 2mGy PB‐CT images also showed better image quality in terms of CNR/Res and vis compared to the 4mGy AB‐CT images. At both investigated doses, SNR and SNR/res were found to have a statistically significant difference across all energy levels. The difference in vis was statistically significant at some energy levels. Conclusion: This study demonstrates superior image quality of PB‐CT over AB‐CT, with X‐ray energy playing a crucial role in determining image quality parameters. Advances in knowledge: Our findings reveal that standard dose PB‐CT outperforms standard dose AB‐CT across all image quality metrics. Additionally, we demonstrate that low dose PB‐CT can produce superior images compared to standard dose AB‐CT in terms of CNR/Res and vis. [ABSTRACT FROM AUTHOR]

The aim of this study was to highlight the advantages that propagation‐based phase‐contrast computed tomography (PB‐CT) with synchrotron radiation can provide in breast cancer diagnostics. For the first time, a fresh and intact mastectomy sample from a 60 year old patient was scanned on the IMBL beamline at the Australian Synchrotron in PB‐CT mode and reconstructed. The clinical picture was described and characterized by an experienced breast radiologist, who underlined the advantages of providing diagnosis on a PB‐CT volume rather than conventional two‐dimensional modalities. Subsequently, the image quality was assessed by 11 breast radiologists and medical imaging experts using a radiological scoring system. The results indicate that, with the radiation dose delivered to the sample being equal, the accuracy of a diagnosis made on PB‐CT images is significantly higher than one using conventional techniques. [ABSTRACT FROM AUTHOR]

Materials and Methods: In this paper, we propose a theoretical model based upon previous studies about personal and social network dynamics of job performance. We provide empirical support for this model using real-world data within the context of the Australian radiology profession. An examination of radiologists’ professional network topology through structural-positional and relational dimensions and radiologists’ personal characteristics in terms of knowledge, experience and self-esteem is provided. Thirty one breast imaging radiologists completed a purpose designed questionnaire regarding their network characteristics and personal attributes. These radiologists also independently read a test set of 60 mammographic cases: 20 cases with cancer and 40 normal cases. A Jackknife free response operating characteristic (JAFROC) method was used to measure the performance of the radiologists’ in detecting breast cancers. Results: Correlational analyses showed that reader performance was positively correlated with the social network variables of degree centrality and effective size, but negatively correlated with constraint and hierarchy. For personal characteristics, the number of mammograms read per year and self-esteem (self-evaluation) positively correlated with reader performance. Hierarchical multiple regression analysis indicated that the combination of number of mammograms read per year and network’s effective size, hierarchy and tie strength was the best fitting model, explaining 63.4% of the variance in reader performance. The results from this study indicate the positive relationship between reading high volumes of cases by radiologists and expertise development, but also strongly emphasise the association between effective social/professional interactions and informal knowledge sharing with high performance. [ABSTRACT FROM AUTHOR]

The population growth, technological improvements, and the need for repairing old or installing new utilities result in a high demand for trenching and drilling activities. However, penetrating the subsurface incurs the risk of damaging existing underground facilities because they were not properly documented, if at all. Ground Penetrating Radar (GPR) constitutes a well-established technology that uses electromagnetic waves to identify objects underground by detecting their reflections. The work presented in this paper focuses on the timing and other characteristics of radar pulses reflected from the buried utilities. It is hypothesised that integrating the knowledge of construction practice, geophysical principles, and electromagnetic wave propagation behaviour in various soil conditions will improve the reliability and accuracy of GPR. This paper presents the results of field experiments that studied the effects of large void such as sinkholes or drainage pipes in several undergrounds. It provides important insights into the features and patterns that can be used to improve current methods. [ABSTRACT FROM AUTHOR]