Background and purpose: Breast cancer is the most common cancer among Australian women, with over 20,000 Australians diagnosed with breast cancer each year (representing 28% of all reported cancers in females)1. Artificial Intelligence (AI) has been widely used in medical imaging analysis due to providing promising results in the last decade2. This study aims to evaluate the diagnosis accuracy of AI systems on mammograms with different groups of cancer sizes. Methods: The cancer sizes were reported in histopathological report with 856 biopsy-proven cancer on screening mammograms. The Globally-aware Multiple Instance Classifier (GMIC) and Global-Local Activation Maps (GLAM) models2 were used to detect breast cancer on screening mammograms. Transfer learning of GMIC and GLAM models was conducted using 1712 mammograms (856 cancer and 856 normal) from an Australian database. The cancer sizes were classified as three groups: T1 (<=2cm), T2(>2cm & <5cm), and T3(>=5cm) according to the American Cancer Society. The diagnosis performance of the AI models was evaluated in terms of sensitivity on mammograms with different groups. The ANOVA test was conducted for statistical analysis of the AI models across different size groups. Results: GMIC outperformed GLAM in all three groups of cancer sizes in terms of sensitivity. The sensitivity of GMIC models on T1, T2 and T3 was 0.882, 0.928, and 0.947, respectively. GMIC (P-value=0.022) demonstrated significant difference in sensitivity across three groups of cancer sizes. The sensitivity of GLAM models on T1, T2 and T3 was 0.851, 0.894, and 0.923, respectively. Similarly, GLAM (P-value=0.016) demonstrated significant difference in sensitivity across three groups. Conclusions: The diagnosis accuracy of AI systems is affected by the cancer sizes in mammograms, with larger size of cancer yielding more accurate cancer detection. The findings of this study indicate that the AI models can provide accurate detection of cancers on mammograms, particularly when the cancer sizes are sufficiently large. References 1. Australian Institute of Health and Welfare, Breast cancer in Australia: an overview, AIHW, Australian Government, (2012). 2. Stadnick B, Witowski J, Rajiv V, Chłędowski J, Shamout FE, Cho K, Geras KJ. Meta-repository of screening mammography classifiers. arXiv preprint arXiv:2108.04800, (2021).

Breast microcalcifications are an important primary radiological indicator of breast cancer. However, microcalcification classification and diagnosis may be still challenging for radiologists due to limitations of the standard 2D mammography technique, including spatial and contrast resolution. In this study, we propose an approach to improve the detection of microcalcifications in propagation-based phase-contrast X-ray computed tomography of breast tissues. Five fresh mastectomies containing microcalcifications were scanned at different X-ray energies and radiation doses using synchrotron radiation. Both bright-field (i.e. conventional phase-retrieved images) and dark-field images were extracted from the same data sets using different image processing methods. A quantitative analysis was performed in terms of visibility and contrast-to-noise ratio of microcalcifications. The results show that while the signal-to-noise and the contrast-to-noise ratios are lower, the visibility of the microcalcifications is more than two times higher in the dark-field images compared to the bright-field images. Dark-field images have also provided more accurate information about the size and shape of the microcalcifications.

A method for extracting the dark-field signal in propagation-based phase-contrast imaging is proposed. In the case of objects consisting predominantly of a single material, or several different materials with similar ratios of the real decrement to the imaginary part of the complex refractive index, the proposed method requires a single image for extraction of the dark-field signal in two-dimensional projection imaging. In the case of three-dimensional tomographic imaging, the method needs only one image to be collected at each projection angle. Initial examples using simulated and experimental data indicate that this method can improve visualization of small sharp features inside a larger object, e.g. the visualization of microcalcifications in propagation-based x-ray breast cancer imaging. It is suggested that the proposed approach may be useful in other forms of biomedical imaging, where it can help one to obtain additional small-angle scattering information without increasing the radiation dose to the sample.

PURPOSE:Propagation-based phase-contrast computed tomography (PB-CT) is a method for three-dimensional x-ray imaging that utilizes refraction, as well as absorption, of x rays in the tissues to increase the signal-to-noise ratio (SNR) in the resultant images, in comparison with equivalent conventional absorption-only x-ray tomography (CT). Importantly, the higher SNR is achieved without sacrificing spatial resolution or increasing the radiation dose delivered to the imaged tissues. The present work has been carried out in the context of the current development of a breast CT imaging facility at the Australian Synchrotron.METHODS:Seven unfixed complete mastectomy samples with and without breast cancer lesions have been imaged using absorption-only CT and PB-CT techniques under controlled experimental conditions. The radiation doses delivered to the mastectomy samples during the scans were comparable to those approved for mammographic screening. Physical characteristics of the reconstructed images, such as spatial resolution and SNR, have been measured and compared with the results of the radiological quality assessment of the complete absorption CT and PB-CT image stacks.RESULTS:Despite the presence of some image artefacts, the PB-CT images have outperformed comparable absorption CT images collected at the same radiation dose, in terms of both the measured objective image characteristics and the radiological image scores. The outcomes of these experiments are shown to be consistent with predictions of the theory of PB-CT imaging and previous reported experimental studies of this imaging modality.CONCLUSIONS:The results presented in this paper demonstrate that PB-CT holds a high potential for improving on the quality and diagnostic value of images obtained using existing medical x-ray technologies, such as mammography and digital breast tomosynthesis (DBT). If implemented at suitable synchrotron imaging facilities, PB-CT can be used to complement existing imaging modalities, leading to more accurate breast cancer diagnosis.

An abstract is not available for this content so a preview has been provided. As you have access to this content, a full PDF is available via the ‘Save PDF’ action button.

The aim of this study was to optimise the experimental protocol and data analysis for in-vivo breast cancer x-ray imaging. Results are presented of the experiment at the SYRMEP beamline of Elettra Synchrotron using the propagation-based phase-contrast mammographic tomography method, which incorporates not only absorption, but also x-ray phase information. In this study the images of breast tissue samples, of a size corresponding to a full human breast, with radiologically acceptable x-ray doses were obtained, and the degree of improvement of the image quality (from the diagnostic point of view) achievable using propagation-based phase-contrast image acquisition protocols with proper incorporation of x-ray phase retrieval into the reconstruction pipeline was investigated. Parameters such as the x-ray energy, sample-to-detector distance and data processing methods were tested, evaluated and optimized with respect to the estimated diagnostic value using a mastectomy sample with a malignant lesion. The results of quantitative evaluation of images were obtained by means of radiological assessment carried out by 13 experienced specialists. A comparative analysis was performed between the x-ray and the histological images of the specimen. The results of the analysis indicate that, within the investigated range of parameters, both the objective image quality characteristics and the subjective radiological scores of propagation-based phase-contrast images of breast tissues monotonically increase with the strength of phase contrast which in turn is directly proportional to the product of the radiation wavelength and the sample-to-detector distance. The outcomes of this study serve to define the practical imaging conditions and the CT reconstruction procedures appropriate for low-dose phase-contrast mammographic imaging of live patients at specially designed synchrotron beamlines.