Dr Oliver Diaz


Research Fellow

Academic and research departments

Department of Electrical and Electronic Engineering.

Biography

My qualifications

2015
Monte Carlo simulations of X-ray imaging and dosimetry (80 hrs) - EUTEMPE-RX training framework
Universitat Politecnica de Catalunya (UPC), Spain.
2015
The use of physical and virtual anthropomorphic phantoms for image quality and patient dose optimization (80 hrs) - EUTEMPE-RX training framework
Technical University of Varna, Bulgaria.
2016
Achieving quality in diagnostic and screening mammography (80 hrs) - EUTEMPE-RX training framework
Dutch reference centre for screening (LRCB), the Netherlands

Research

Research interests

Research projects

Indicators of esteem

Supervision

Completed postgraduate research projects I have supervised

Postgraduate research supervision

My teaching

My publications

Publications

E. García, Y. Diez, O.Diaz, X. Lladó, A. Gubern-Mérida, R. Martí, J.Marti and A.Oliver (2018). 'Multimodal breast parenchymal patterns correlation using a patient-specific biomechanical model'. IEEE Transactions on Medical Imaging, 37 (3): 712-723. DOI: 10.1109/TMI.2017.2749685 [IF 3.942, Q1(16/126) RNMMI]
View abstract View full publication
In this paper, we aim to produce a realistic 2-D projection of the breast parenchymal distribution from a 3-D breast magnetic resonance image (MRI). To evaluate the accuracy of our simulation, we compare our results with the local breast density (i.e., density map) obtained from the complementary full-field digital mammogram. To achieve this goal, we have developed a fully automatic framework, which registers MRI volumes to X-ray mammograms using a subject-specific biomechanical model of the breast. The optimization step modifies the position, orientation, and elastic parameters of the breast model to perform the alignment between the images. When the model reaches an optimal solution, the MRI glandular tissue is projected and compared with the one obtained from the corresponding mammograms. To reduce the loss of information during the ray-casting, we introduce a new approach that avoids resampling the MRI volume. In the results, we focus our efforts on evaluating the agreement of the distributions of glandular tissue, the degree of structural similarity, and the correlation between the real and synthetic density maps. Our approach obtained a high-structural agreement regardless the glandularity of the breast, whilst the similarity of the glandular tissue distributions and correlation between both images increase in denser breasts. Furthermore, the synthetic images show continuity with respect to large structures in the density maps.
R. Agarwal, O. Diaz, X. Llado, A. Gubern-Mérida, J. Vilanova and R. Martí. (2018). 'Lesion segmentation in automated 3D breast ultrasound: volumetric analysis'. Ultrasonic Imaging, 40 (2): 97-112. DOI: 10.1177/0161734617737733 [IF 1.78, Q2 Acoustics (11/31)]
View abstract View full publication
Mammography is the gold standard screening technique in breast cancer, but it has some limitations for women with dense breasts. In such cases, sonography is usually recommended as an additional imaging technique. A traditional sonogram produces a two-dimensional (2D) visualization of the breast and is highly operator dependent. Automated breast ultrasound (ABUS) has also been proposed to produce a full 3D scan of the breast automatically with reduced operator dependency, facilitating double reading and comparison with past exams. When using ABUS, lesion segmentation and tracking changes over time are challenging tasks, as the three-dimensional (3D) nature of the images makes the analysis difficult and tedious for radiologists. The goal of this work is to develop a semi-automatic framework for breast lesion segmentation in ABUS volumes which is based on the Watershed algorithm. The effect of different de-noising methods on segmentation is studied showing a significant impact (p<0.05p<0.05) on the performance using a dataset of 28 temporal pairs resulting in a total of 56 ABUS volumes. The volumetric analysis is also used to evaluate the performance of the developed framework. A mean Dice Similarity Coefficient of 0.69±0.110.69±0.11 with a mean False Positive ratio 0.35±0.140.35±0.14 has been obtained. The Pearson correlation coefficient between the segmented volumes and the corresponding ground truth volumes is r2=0.960r2=0.960 (p=0.05p=0.05). Similar analysis, performed on 28 temporal (prior and current) pairs, resulted in a good correlation coefficient r2=0.967r2=0.967 (p<0.05p<0.05) for prior and r2=0.956r2=0.956 (p<0.05p<0.05) for current cases. The developed framework showed prospects to help radiologists to perform an assessment of ABUS lesion volumes, as well as to quantify volumetric changes during lesions diagnosis and follow-up
E.García, Y.Diez, O.Diaz, X.Lladó, R.Martí, J.Martí and A.Oliver (2018). 'A step-by-step review on patient-specific biomechanical finite elements models for breast MRI to X-ray mammography registration'. Medical Physics, 45(1): e6-e31. DOI: 10.1002/mp.12673 [IF 2.617, Q2(37/126) RNMMI]
View abstract View full publication
Breast magnetic resonance imaging (MRI) and x‐ray mammography are two image modalities widely used for the early detection and diagnosis of breast diseases in women. The combination of these modalities leads to a more accurate diagnosis and treatment of breast diseases. The aim of this paper is to review the registration between breast MRI and x‐ray mammographic images using patient‐specific finite element‐based biomechanical models. Specifically, a biomechanical model is obtained from the patient's MRI volume and is subsequently used to mimic the mammographic acquisition. Due to the different patient positioning and movement restrictions applied in each image modality, the finite element analysis provides a realistic physics‐based approach to perform the breast deformation. In contrast with other reviews, we do not only expose the overall process of compression and registration but we also include main ideas, describe challenges, and provide an overview of the used software in each step of the process. Extracting an accurate description from the MR images and preserving the stability during the finite element analysis require an accurate knowledge about the algorithms used, as well as the software and underlying physics. The wide perspective offered makes the paper suitable not only for expert researchers but also for graduate students and clinicians. We also include several medical applications in the paper, with the aim to fill the gap between the engineering and clinical performance.
E. García, O. Diaz, R. Martí, Y. Diez, A. Gubern-Mérida, M. Sentís, J. Martí, and A. Oliver (2017). 'Local breast density assessment using reacquired mammographic images'. European Journal of Radiology 93: 121-127. DOI: 10.1016/j.ejrad.2017.05.033 [IF 2.593, Q2(45/126) RNMMI]
View abstract View full publication
The aim of this paper is to evaluate the spatial glandular volumetric tissue distribution as well as the density measures provided by Volpara™ using a dataset composed of repeated pairs of mammograms, where each pair was acquired in a short time frame and in a slightly changed position of the breast. We conducted a retrospective analysis of 99 pairs of repeatedly acquired full-field digital mammograms from 99 different patients. The commercial software Volpara™ Density Maps (Volpara Solutions, Wellington, New Zealand) is used to estimate both the global and the local glandular tissue distribution in each image. The global measures provided by Volpara™, such as breast volume, volume of glandular tissue, and volumetric breast density are compared between the two acquisitions. The evaluation of the local glandular information is performed using histogram similarity metrics, such as intersection and correlation, and local measures, such as statistics from the difference image and local gradient correlation measures. Global measures showed a high correlation (breast volume  = 0.99, volume of glandular tissue  = 0.94, and volumetric breast density  = 0.96) regardless the anode/filter material. Similarly, histogram intersection and correlation metric showed that, for each pair, the images share a high degree of information. Regarding the local distribution of glandular tissue, small changes in the angle of view do not yield significant differences in the glandular pattern, whilst changes in the breast thickness between both acquisition affect the spatial parenchymal distribution. This study indicates that Volpara™ Density Maps is reliable in estimating the local glandular tissue distribution and can be used for its assessment and follow-up. Volpara™ Density Maps is robust to small variations of the acquisition angle and to the beam energy, although divergences arise due to different breast compression conditions.

Purpose

Materials and methods

Results

RRR

Conclusions

A. Mackenzie, D.R. Dance, O. Diaz and K.C. Young (2014). ’Image simulation and a model of noise power spectra across a range of mammographic beam qualities’. Medical Physics, 41, 121901. DOI:10.1118/1.4900819 [IF 3.012, Q1(25/122) RNMMI]
View abstract View full publication
The aim of this work is to create a model to predict the noise power spectra (NPS) for a range of mammographic radiographic factors. The noise model was necessary to degrade images acquired on one system to match the image quality of different systems for a range of beam qualities. Five detectors and x‐ray systems [Hologic Selenia (ASEh), Carestream computed radiography CR900 (CRc), GE Essential (CSI), Carestream NIP (NIPc), and Siemens Inspiration (ASEs)] were characterized for this study. The signal transfer property was measured as the pixel value against absorbed energy per unit area () at a reference beam quality of 28 kV, Mo/Mo or 29 kV, W/Rh with 45 mm polymethyl methacrylate (PMMA) at the tube head. The contributions of the three noise sources (electronic, quantum, and structure) to the NPS were calculated by fitting a quadratic at each spatial frequency of the NPS against . A quantum noise correction factor which was dependent on beam quality was quantified using a set of images acquired over a range of radiographic factors with different thicknesses of PMMA. The noise model was tested for images acquired at 26 kV, Mo/Mo with 20 mm PMMA and 34 kV, Mo/Rh with 70 mm PMMA for three detectors (ASEh, CRc, and CSI) over a range of exposures. The NPS were modeled with and without the noise correction factor and compared with the measured NPS. A previous method for adapting an image to appear as if acquired on a different system was modified to allow the reference beam quality to be different from the beam quality of the image. The method was validated by adapting the ASEh flat field images with two thicknesses of PMMA (20 and 70 mm) to appear with the imaging characteristics of the CSI and CRc systems. The quantum noise correction factor rises with higher beam qualities, except for CR systems at high spatial frequencies, where a flat response was found against mean photon energy. This is due to the dominance of secondary quantum noise in CR. The use of the quantum noise correction factor reduced the difference from the model to the real NPS to generally within 4%. The use of the quantum noise correction improved the conversion of ASEh image to CRc image but had no difference for the conversion to CSI images. A practical method for estimating the NPS at any dose and over a range of beam qualities for mammography has been demonstrated. The noise model was incorporated into a methodology for converting an image to appear as if acquired on a different detector. The method can now be extended to work for a wide range of beam qualities and can be applied to the conversion of mammograms.

Purpose:

Methods:

EE

Results:

Conclusions:

O. Diaz, D.R. Dance, K.C. Young, P. Elangovan, P.R. Bakic and K. Wells (2014). ’Estimation of scattered radiation in digital breast tomosynthesis’. Physics in Medicine and Biology 59 (15): 4375-4390. DOI: 10.1088/0031-9155/59/15/4375 [IF 2.922, Q1 (26/122) RNMMI]
View abstract View full publication
Digital breast tomosynthesis (DBT) is a promising technique to overcome the tissue superposition limitations found in planar 2D x-ray mammography. However, as most DBT systems do not employ an anti-scatter grid, the levels of scattered radiation recorded within the image receptor are significantly higher than that observed in planar 2D x-ray mammography. Knowledge of this field is necessary as part of any correction scheme and for computer modelling and optimisation of this examination. Monte Carlo (MC) simulations are often used for this purpose, however they are computationally expensive and a more rapid method of calculation is desirable. This issue is addressed in this work by the development of a fast kernel-based methodology for scatter field estimation using a detailed realistic DBT geometry. Thickness-dependent scatter kernels, which were validated against the literature with a maximum discrepancy of 4% for an idealised geometry, have been calculated and a new physical parameter (air gap distance) was used to estimate more accurately the distribution of scattered radiation for a series of anthropomorphic breast phantom models. The proposed methodology considers, for the first time, the effects of scattered radiation from the compression paddle and breast support plate, which can represent more than 30% of the total scattered radiation recorded within the image receptor. The results show that the scatter field estimator can calculate scattered radiation images in an average of 80 min for projection angles up to 25° with equal to or less than a 10% error across most of the breast area when compared with direct MC simulations.
P. Elangovan, L. Warren, A. Mackenzie, A. Rashidnasab, O. Diaz, D.R. Dance, K.C. Young, H. Bosmans, C. Strudley, K. Wells (2014). ’Development and validation of a modelling framework for simulating 2D-mammography and breast tomosynthesis images’. Physics in Medicine and Biology 59 (15): 4275. DOI: 10.1088/0031-9155/59/15/4275 [IF 2.922, Q1 (26/122) RNMMI]
View abstract View full publication
Planar 2D x-ray mammography is generally accepted as the preferred screening technique used for breast cancer detection. Recently, digital breast tomosynthesis (DBT) has been introduced to overcome some of the inherent limitations of conventional planar imaging, and future technological enhancements are expected to result in the introduction of further innovative modalities. However, it is crucial to understand the impact of any new imaging technology or methodology on cancer detection rates and patient recall. Any such assessment conventionally requires large scale clinical trials demanding significant investment in time and resources. The concept of virtual clinical trials and virtual performance assessment may offer a viable alternative to this approach. However, virtual approaches require a collection of specialized modelling tools which can be used to emulate the image acquisition process and simulate images of a quality indistinguishable from their real clinical counterparts. In this paper, we present two image simulation chains constructed using modelling tools that can be used for the evaluation of 2D-mammography and DBT systems. We validate both approaches by comparing simulated images with real images acquired using the system being simulated. A comparison of the contrast-to-noise ratios and image blurring for real and simulated images of test objects shows good agreement ( < 9% error). This suggests that our simulation approach is a promising alternative to conventional physical performance assessment followed by large scale clinical trials
R. W. Bouwman, O. Diaz, R.E. van Engen, K.C. Young, G.J. den Heeten, M.J.M Broeders, W.J.H. Veldkamp and D.R. Dance (2013). 'Phantoms for quality control procedures in digital breast tomosynthesis: dose assessment'. Physics in Medicine and Biology 58 (13): 4423-4438. DOI: 10.1088/0031-9155/58/13/4423 [IF 2.922, Q1 (26/122) RNMMI]
View abstract View full publication
The recent introduction of digital breast tomosynthesis into clinical practice requires quality control procedures. In this study we have investigated whether the assessment of the average glandular dose for modelled standard breasts can be performed using a combination of polymethyl methacrylate (PMMA) and polyethylene (PE) slabs that matches standard breast thicknesses. For this purpose the energies absorbed per unit area of the image receptor when imaging standard breasts and PMMA-PE slabs have been matched taking account of both primary and scattered photons. To achieve this a two-step approach was used. Firstly, the behaviour of the scatter-to-primary ratio (SPR) of PMMA-PE phantoms and standard breasts was investigated using Monte Carlo simulations for various conditions. For imaging without an anti-scatter grid, it was found that the values of standard breast and phantom SPR were significantly different and it follows that these differences are relevant when matching the absorbed energy. In the second part, a set of PMMA-PE combinations is proposed which, for dosimetric purposes, can be used to simulate standard breasts in the thickness range 20 to 100 mm. The dosimetric error when using these PMMA-PE slabs was found to be below 6% for thicknesses up to 7 cm and increases to 10% for 10 cm thickness
A. Rashidnasab, P. Elangovan, M. Yip, O. Diaz, D. R. Dance, K.C. Young and K. Wells (2013). 'Simulation and assessment of realistic breast lesions using fractal growth models'. Physics in Medicine and Biology 58 (16): 5613-5627. DOI: 10.1088/0031-9155/58/16/5613 [IF 2.922, Q1 (26/122) RNMMI]
View abstract View full publication
A new method of generating realistic three dimensional simulated breast lesions known as diffusion limited aggregation (DLA) is presented, and compared with the random walk (RW) method. Both methods of lesion simulation utilize a physics-based method for inserting these simulated lesions into 2D clinical mammogram images that takes into account the polychromatic x-ray spectrum, local glandularity and scatter. DLA and RW masses were assessed for realism via a receiver operating characteristic (ROC) study with nine observers. The study comprised 150 images of which 50 were real pathology proven mammograms, 50 were normal mammograms with RW inserted masses and 50 were normal mammograms with DLA inserted masses. The average area under the ROC curve for the DLA method was 0.55 (95% confidence interval 0.51–0.59) compared to 0.60 (95% confidence interval 0.56–0.63) for the RW method. The observer study results suggest that the DLA method produced more realistic masses with more variability in shape compared to the RW method. DLA generated lesions can overcome the lack of complexity in structure and shape in many current methods of mass simulation.
D. Rodriguez Gutierrez, K. Wells, O. Diaz Montesdeoca, A. Moran Santana, I. A. Mendichovszky and I. Gordon (2010). 'Partial volume effects in dynamic contrast magnetic resonance renal studies'. European Journal of Radiology 75(2): 221-229. DOI: 10.1016/j.ejrad.2009.04.073 [IF 2.941, Q2 (30/113) RNMMI]
View abstract View full publication
This is the first study of partial volume effect in quantifying renal function on dynamic contrast enhanced magnetic resonance imaging. Dynamic image data were acquired for a cohort of 10 healthy volunteers. Following respiratory motion correction, each voxel location was assigned a mixing vector representing the ‘overspilling’ contributions of each tissue due to the convolution action of the imaging system's point spread function. This was used to recover the true intensities associated with each constituent tissue. Thus, non-renal contributions from liver, spleen and other surrounding tissues could be eliminated from the observed time–intensity curves derived from a typical renal cortical region of interest. This analysis produced a change in the early slope of the renal curve, which subsequently resulted in an enhanced glomerular filtration rate estimate. This effect was consistently observed in a Rutland–Patlak analysis of the time–intensity data: the volunteer cohort produced a partial volume effect corrected mean enhancement of 36% in relative glomerular filtration rate with a mean improvement of 7% in 2 fitting of the Rutland–Patlak model compared to the same analysis undertaken without partial volume effect correction. This analysis strongly supports the notion that dynamic contrast enhanced magnetic resonance imaging of kidneys is substantially affected by the partial volume effect, and that this is a significant obfuscating factor in subsequent glomerular filtration rate estimation.
r
R. Agarwal, O. Diaz, X. Lladó, R. Marti (2018). 'Mass detection in mammograms using pre-trained deep learning models'. Proc. SPIE 10718, 14th International Workshop on Breast Imaging (IWBI 2018): 107181F; doi: 10.1117/12.2317681
View abstract View full publication
Mammography is a gold standard imaging modality and is widely used for breast cancer screening. With recent advances in the field of deep learning, the use of deep convolution neural networks (CNNs) in medical image analysis has become very encouraging. The aim of this study is to exploit CNNs for mass detection in mammograms using pre-trained networks. We use the resnet-50 CNN architecture pre-trained with the ImageNet database to perform mass detection on two publicly available image datasets: CBIS-DDSM and INbreast. We demonstrate that the CNN model pretrained using natural image database (ImageNet) can be effectively finetuned to yield better results, compared to randomly initialized models. Further, the benefit of applying transfer learning on a smaller dataset is demonstrated by using the best model obtained from CBIS-DDSM training to finetune on the INbreast database. We analyzed the adaptability of the CNN’s last fully connected (FC) layer and the all convolutional layers to detect masses. The results showed a testing accuracy of 0.92 and an area under the receiver operating characteristic curve (AUC) of 0.98 for the model finetuned on all convolutional layers, while testing accuracy of 0.86 and AUC=0.93 when the model is trained only on the last FC layer.
M. Chevalier, C. Viloria, P. Squair, M.S. Nogueira, O. Diaz (2018). 'Digital breast tomosynthesis: impact of a new beam quality on dose to patients'. Proc. SPIE 10718, 14th International Workshop on Breast Imaging (IWBI 2018): 107181M; doi: 10.1117/12.2317689
View abstract View full publication
In this study we analyze the impact of new x-ray beam spectra on the mean glandular doses (MGD) delivered by a digital breast tomosynthesis system. The new polyenergetic spectra are generated with a rhodium (Rh) target and a 30 μm silver (Ag) filter. To evaluate the influence of the new spectra on patient doses, we compare the MGD values with those delivered with a regular Rh/Rh target/filter combination. Individual glandularity (%) of the patients in the study was estimated using the commercial software Volpara. Median of MGD values for CC and MLO views are around 38% and 46% lower with the Rh/Ag combination than with the Rh/Rh combination. Results suggest that the new spectra, with reduced dose properties, could be very useful in breast cancer screening programs.
E. García, A. Oliver, O. Diaz, Y. Díez, A. Gubern-Mérida, J. Martí, R. Martí (2018). 'Changes in breast density over time using automatic density measures: preliminary analysis'. Proc. SPIE 10718, 14th International Workshop on Breast Imaging (IWBI 2018): 107181I; doi: 10.1117/12.2317912
View abstract View full publication
Breast density is an important risk factor for the development of breast cancer. During the women lifetime, the breast glandularity varies due to hormonal changes. In particular, around menopause, the glandular tissue tends to decrease. The aim of this paper is to evaluate temporal breast density changes using density maps, provided by the commercial software VolparaTM. The dataset is composed of 563 mammograms from 55 patients (aged between 24 and 75 years old). The time frame between two acquisitions varies from less than one year to 4 years. Pairs of mammograms are registered using the morphons registration algorithm, in order to evaluate the structural similarity of the parenchymal distribution between the two acquisitions. To provide a fair comparison, the results are divided considering the patient age during the first mammographic acquisition and the time between the two studies. To evaluate the changes in breast density, local and global measures, such as the rate of change of the volumetric breast density, the histogram intersection between two density maps and the normalized cross-correlation after the registration, are considered. The results show significant differences in the statistics, mainly focused on patients younger than 30 years old and ranged between 56 and 65 years old with respect to those in the adulthood (between 30 and 55 years old). Similarly, the time between the two mammographic acquisitions shows a significant difference for patients older than 56 years old considering one and two year of difference between the two studies.
A. Malet, D. Garcia-Pinto, J. Fernandez, R. Marti, O. Diaz. (2018). 'Breast Tomosynthesis reconstruction using software tool TIGRE' (2018). Proc. SPIE 10718, 14th International Workshop on Breast Imaging (IWBI 2018): 107181L; doi: 10.1117/12.2317933
View abstract View full publication
This article shows the feasibility of using the open source Tomographic Iterative GPU-based Reconstruction (TIGRE) Toolbox, originally developed for cone-beam x-ray computed tomography (CBCT), to reconstruct images from a Digital Breast Tomosynthesis (DBT) system. We present reconstructed images of simple simulated phantoms as well as the commercially available breast phantoms CIRS models 013 and 073; acquired by a Hologic Selenia Dimensions system. Initial results have shown the ability of TIGRE to reconstruct images using several reconstruction algorithms (FDK, OSSART, MLEM), although a wider variety of iterative algorithms could be also considered. This is the first work that uses the TIGRE reconstruction tool for DBT geometries, opening new possibilities for free, fast and reliable reconstruction algorithms to other research groups
G. Trovini, C. Napoli, R. Marti, A. Martin, A. Bria, C. Marrocco, M. Molinara, F. Tortorella, O. Diaz (2018). 'A deep learning framework for micro-calcification detection in 2D mammography and C-view'. Proc. SPIE 10718, 14th International Workshop on Breast Imaging (IWBI 2018): 1071811; doi: 10.1117/12.2318023
View abstract View full publication
The aim of this paper is to propose a deep learning framework for micro-calcification detection in 2D mammography and in 2D synthetic mammography (C-view) from digital breast tomosynthesis (DBT). The dataset analyzed for 2D mammograms is the INbreast dataset that consists of 410 digital images and we used 360 images with annotated micro-calcifications. For the synthetic views in DBT, we used a private dataset of 245 images, where micro-calcifications were validated by an experienced radiologist. The network is trained in a patch-based fashion, where micro-calcifications are considered positive samples, while patches containing other breast tissues are considered negative. For evaluating the entire dataset, a 2-fold cross validation was performed. In addition, a sliding window method was used to classify new patches within an image with those from the trained model. Considering 5,656 positive samples and 18,000,000 of negative samples, results for the 2D mammography, on the entire dataset, showed an area under the curve (AUC) of 0.9998 and a logarithmic partial area under the curve (logPAUC), in the interval (10−6 , 1), of 0.8252. Results for the C-View, considering 3,420 positive samples and 11,395,939 of negative samples, showed an AUC, on the entire dataset, of 0.9997 and a logPAUC, in the interval (10−6 , 1), of 0.8178. In this paper, we illustrate the applied methodologies, the network architecture used for training and test, and the results obtained.
S. Kazemi, O. Diaz, P. Elangovan, K. Wells and A. Lohstroh (2018). 'Validation and Application of a New Image Reconstruction Software tool (TIGRE) for Breast Cone-Beam Computed Tomography'. Proc. SPIE 10573, Medical Imaging 2018: Physics of Medical Imaging: 105735E. DOI: 10.1117/12.2293200
View abstract View full publication
A new image reconstruction software toolbox TIGRE (Tomographic Iterative GPU-based Reconstruction) has been evaluated for use in breast cone-beam computed tomography (CBCT) studies. This new software toolbox TIGRE has been compared to a standard Matlab-based implementation previously validated for X-ray mammography imaging. In particular, the image projection generator algorithm in the TIGRE toolbox, which is based on the Siddon ray-tracing algorithm, has been studied. The quantitative evaluation in terms of histograms and profile analyses, illustrates that TIGRE’s image projection show good agreement with our in-house validated X-ray ray tracing tool. In addition, it has been observed that since TIGRE uses GPU-based calculations, it produces projections approximately 90 times faster than CPU-based algorithms, dependent on choice of GPU. The breast CT images have also been reconstructed and evaluated using the two projection tools. The analyses show that the projections taken by TIGRE and our in-house developed Siddon algorithm, yield systematically similar results. To further investigate the differences between these two algorithms, the reconstructed images have been compared to each other. The correlation coefficients for an entire 3D reconstructed breast volumes using the two methods studied is 0.99±3.64x10-12 (mean ±standard deviation), the peak signal noise ratio is 117.17, the mean square error is 1.92x10-12 and the similarity index is 1.00.
S. Kazemi, O. Diaz, P. Elangovan, K. Wells and A. Lohstroh (2018). 'Comparison of Breast Cone-Beam Computed Tomography, 2D Mammography and Digital Breast Tomosynthesis Imaging Modalities, using 4-Alternative Forced Choice Human Observation Study'. Proc. SPIE 10573, Medical Imaging 2018: Physics of Medical Imaging: 105735I. DOI: 10.1117/12.2293201.
View abstract View full publication
X-ray mammography is the gold standard for detecting malignancies in a breast cancer screening context. However, limited angle tomosynthesis has now started to be used in screening due to its ability to remove overlying image clutter. However, breast CT is a method, which can potentially remove all overlying clutter through the use of tomographic image reconstruction.   The aim of this work is to investigate whether breast cone-beam computed tomography (CBCT) can provide better lesion detectability compared to 2D mammography or digital breast tomosynthesis (DBT).   Lesions with a diameter of 4 mm, 5 mm and 6 mm have been inserted in a simulated breast phantom. In total 180 images are analysed, out of which 90 images contain lesions (equally divided between the 4 mm, 5mm and 6mm diameter lesions) and the rest represent normal breast tissues. The TIGRE (Tomographic Iterative GPU-based Reconstruction) has been used to simulate 360 projections and to reconstruct the images using the FeldKamp, Davis and Kress (FDK) algorithm. Scattered radiation and Poisson noise have also been added to the projections prior the image reconstruction.   In total 10 observers, some with, and some without experience of mammography images, have been used as observers for this preliminary 4AFC study. The analysis of the 4AFC study shows that the mean minimum detectable lesion size for the breast CBCT is 2.96±0.23 mm with a 95% confidence intervals of [2.73, 3.19].
E. García, A. Oliver, Y. Díez, O. Diaz, X. Llado, R. Martí and J. Marti (2017). 'Similarity metrics for intensity-based registration using breast density maps'. In Pattern Recognition and Image Analysis, LNCS vol. 10255, pp. 217-225. DOI: 10.1007/978-3-319-58838-4_24
View abstract View full publication
Intensity-based registration algorithms have been widely used in medical image applications. This type of registration algorithms uses an object function to compute a transformation and optimizes a measure of similarity between the images being registered. The most common similarity metrics used in registration are sum of squared differences, mutual information and normalized cross-correlation. This paper aims to compare these similarity metrics, using common registration algorithms applied to breast density maps registration. To evaluate the results, we use the protocols for evaluation of similarity measures proposed by Škerl et al. They consist in defining a set of random directions in the parameter space of the registration algorithm and compute statistical measures, such as the accuracy, capture range, number of maxima and risk of non-convergence, along these directions. The obtained results show a better performance corresponding to normalized cross-correlation for the rigid registration algorithm, while the sum of squared difference obtains the best result for the B-Spline method.
O. Dı́az, E. Garcı́a, A. Oliver, J. Martı́ and R. Martı́ (2017). 'Scattered radiation in DBT geometries with flexible breast compression paddles: a Monte Carlo simulation study'. Proc. SPIE 10132, Medical Imaging 2017: Physics of Medical Imaging, 101324G. DOI: 10.1117/12.2255722
View abstract View full publication
Scattered radiation is an undesired signal largely present in most digital breast tomosynthesis (DBT) projection images as no physically rejection methods, i.e. anti-scatter grids, are regularly employed, in contrast to full- field digital mammography. This scatter signal might reduce the visibility of small objects in the image, and potentially affect the detection of small breast lesions. Thus accurate scatter models are needed to minimise the scattered radiation signal via post-processing algorithms. All prior work on scattered radiation estimation has assumed a rigid breast compression paddle (RP) and reported large contribution of scatter signal from RP in the detector. However, in this work, flexible paddles (FPs) tilting from 0° to 10° will be studied using Monte Carlo simulations to analyse if the scatter distribution differs from RP geometries. After reproducing the Hologic Selenia Dimensions geometry (narrow angle) with two (homogeneous and heterogeneous) compressed breast phantoms, results illustrate that the scatter distribution recorded at the detector varies up to 22% between RP and FP geometries (depending on the location), mainly due to the decrease in thickness of the breast observed for FP. However, the relative contribution from the paddle itself (3-12% of the total scatter) remains approximately unchanged for both setups and their magnitude depends on the distance to the breast edge.
E. Marimon, H. Nait-Charif, A. Khan, P. A. Marsden and O. Diaz (2017). 'Scatter reduction for grid-less mammography using the convolution-based image post-processing technique'. Proc. SPIE 10132, Medical Imaging 2017: Physics of Medical Imaging, 101324D. DOI: 10.1117/12.2255558
View abstract View full publication
X-ray Mammography examinations are highly affected by scattered radiation, as it degrades the quality of the image and complicates the diagnosis process. Anti-scatter grids are currently used in planar mammography examinations as the standard physical scattering reduction technique. This method has been found to be inefficient, as it increases the dose delivered to the patient, does not remove all the scattered radiation and increases the price of the equipment. Alternative scattering reduction methods, based on post-processing algorithms, are being investigated to substitute anti-scatter grids. Methods such as the convolution-based scatter estimation have lately become attractive as they are quicker and more flexible than pure Monte Carlo (MC) simulations. In this study we make use of this specific method, which is based on the premise that the scatter in the system is spatially diffuse, thus it can be approximated by a two-dimensional low-pass convolution filter of the primary image. This algorithm uses the narrow pencil beam method to obtain the scatter kernel used to convolve an image, acquired without anti-scatter grid. The results obtained show an image quality comparable, in the worst case, to the grid image, in terms of uniformity and contrast to noise ratio. Further improvement is expected when using clinically-representative phantoms.
E. Garcı́a, A. Oliver, O. Diaz, Y. Diez, R. Martı́ and J. Martı́ (2017). 'Mapping 3D breast lesions from full-field digital mammograms using subject-specific finite element models'. Proc. SPIE 10135, Medical Imaging 2017: Image-Guided Procedures, Robotic Interventions, and Modeling, 1013504. DOI: 10.1117/12.2255957
View abstract View full publication
Patient-specific finite element (FE) models of the breast have received increasing attention due to the potential capability of fusing images from different modalities. During the Magnetic Resonance Imaging (MRI) to X-ray mammography registration procedure, the FE model is compressed mimicking the mammographic acquisition. Subsequently, suspicious lesions in the MRI volume can be projected into the 2D mammographic space. However, most registration algorithms do not provide the reverse information, avoiding to obtain the 3D geometrical information from the lesions localized in the mammograms. In this work we introduce a fast method to localize the 3D position of the lesion within the MRI, using both cranio-caudal (CC) and medio-lateral oblique (MLO) mammographic projections, indexing the tetrahedral elements of the biomechanical model by means of an uniform grid. For each marked lesion in the Full-Field Digital Mammogram (FFDM), the X-ray path from source to the marker is calculated. Barycentric coordinates are computed in the tetrahedrons traversed by the ray. The list of elements and coordinates allows to localize two curves within the MRI and the closest point between both curves is taken as the 3D position of the lesion. The registration errors obtained in the mammographic space are 9.89 ± 3.72 mm in CC- and 8.04 ± 4.68 mm in MLO-projection and the error in the 3D MRI space is equal to 10.29 ± 3.99 mm. Regarding the uniform grid, it is computed spending between 0.1 and 0.7 seconds. The average time spent to compute the 3D location of a lesion is about 8 ms.
O. Diaz, A. Oliver, S. Ganau, E. García, J. Martí, M. Sentís and R. Martí (2016). 'Feasibility of Depth Sensors to Study Breast Deformation During Mammography Procedures'. In chapter Breast Imaging, Lectures Notes in Computer Science 9699, pp 446-453. DOI: 10.1007/978-3-319-41546-8_56
View abstract View full publication
Virtual clinical trials (VCT) currently represent key tools for breast imaging optimisation, especially in two-dimensional planar mammography and digital breast tomosynthesis. Voxelised breast models are a crucial part of VCT as they allow the generation of synthetic image projections of breast tissue distribution. Therefore, realistic breast models containing an accurate representation of women breasts are needed. Current voxelised breast models show, in their compressed version, a very round contour which might not be representative of the entire population. This work pretends to develop an imaging framework, based on depth cameras, to investigate breast deformation during mammographic compression. Preliminary results show the feasibility of depth sensors for such task, however post-processing steps are needed to smooth the models. The proposed framework can be used in the future to produce more accurate compressed breast models, which will eventually generate more realistic images in VCT
E. Marimon, H. Nait-Charif, A. Khan, P.A. Marsden and O. Diaz (2016). 'Detailed Analysis of Scatter Contribution from Different Simulated Geometries of X-ray Detectors'. In chapter Breast Imaging, Lectures Notes in Computer Science 9699, pp 203-210. DOI: 10.1007/978-3-319-41546-8_27
View abstract View full publication
Scattering is one of the main issues left in planar mammography examinations, as it degrades the quality of the image and complicates the diagnostic process. Although widely used, anti-scatter grids have been found to be inefficient, increasing the dose delivered, the equipment price and not eliminating all the scattered radiation. Alternative scattering reduction methods, based on post-processing algorithms using Monte Carlo (MC) simulations, are being developed to substitute anti-scatter grids. Idealized detectors are commonly used in the simulations for the purpose of simplification. In this study, the scatter distribution of three detector geometries is analyzed and compared: Case 1 makes use of idealized detector geometry, Case 2 uses a scintillator plate and Case 3 uses a more realistic detector simulation, based on the structure of an indirect mammography X-ray detector. This paper demonstrates that common configuration simplifications may introduce up to 14 % of underestimation of the scatter in simulation results.
E. García, A. Oliver, Y. Diez, O. Diaz, A.Gubern-Mérida, X. Lladó and J. Martí (2016). 'Comparison of Four Breast Tissue Segmentation Algorithms for Multi-modal MRI to X-ray Mammography Registration'. In chapter Breast Imaging, Lectures Notes in Computer Science 9699, pp 493-500. DOI: 10.1007/978-3-319-41546-8_62
View abstract View full publication
Breast MRI to X-ray mammography registration usingpatient-specific biomechanical models is one challenging task in medical imaging. To solve this problem, the accurate knowledge about internal and external factors of the breast, such as internal tissues distribution, is needed for modelling a suitable physical behavior. In this work, we compare four different tissue segmentation algorithms, two intensity-based segmentation algorithms (Fuzzy C-means and Gaussian mixture model) and two improvements that incorporate spatial information (Kernelized Fuzzy C-means and Markov Random Fields, respectively), and analyze their effect to the multi-modal registration. The overall framework consists on using a density estimation software (VolparaTMTM) to extract the glandular tissue from full-field digital mammograms, meanwhile, a biomechanical model is used to mimic the mammographic acquisition from the MRI, computing the glandular tissue traversed by the X-ray beam. Results with 40 patients show a high agreement between the amount of glandular tissue computed for each method
O. Díaz, R. Agarwal, A. Gubern-Mérida, J. van Zelst, Y. Díez and R. Martí (2016). 'Automated volumetric lesion quantification in automated 3D breast ultrasound: comparison of 5 breast lesion segmentation algorithms'. Insights Imaging (2016) 7 (Suppl 1): S407, European Congress of Radiology (ECR) 2016. DOI: 10.1007/s13244-016-0475-8
View abstract View full publication
: Lesion growth is an important feature used for breast lesion characterisation. In automated 3D breast ultrasound (ABUS), lesion size is typically performed by measuring the longest lesion axis without accounting for more accurate measurements such lesion volume. This work presents a comparison of five breast lesion segmentation algorithms for ABUS that can be used in clinical practice to quantify lesion volume. : 20 ABUS volumes with a total of 16 benign lesions (BI-RADS 2) and 4 malignant lesions (BI-RADS 4/5) were used. The average lesion size was 9.03 mm. Five segmentation algorithms were investigated: an in-house Markov Random Field-Maximum A Posteriori (MRF-MAP) segmentation framework and four widely known image segmentation methods that were optimised for ABUS lesion segmentation: confidence connected (CC), connected threshold (CT), neighbourhood connected (NC) and watershed (WAT) algorithm. Dice Similarity Coefficient (DSC) was used to quantify the overlap between automated and ground truth (GT) segmentations. Pearson correlation between lesion volumes computed with the most accurate segmentation algorithm and GT was calculated.   : Best results were obtained for the WAT algorithm with a mean DSC value of 74±10.4%. MRF-MAP, CC, CT and NC obtained a DSC of 51.7±30.1%, 55.9±22.3%, 45.3±24.6% and 2±5.3%, respectively. Correlation between lesion volumetric measures computed on WAT and GT segmentations was 0.97.   : The watershed algorithm showed the most similarities with the radiologists’ annotations. The integration of such an algorithm into clinical workflow has the potential to aid radiologists in quantitatively measuring tumour changes over time in ABUS.
PurposeMaterials and MethodsResultsConclusions
Y. Díez A. Maroto, O. Díaz, A. Gubern-Mérida and R. Martí (2016). 'A study of rigid methods for ABUS temporal studies'. European Congress of Radiology (ECR) 2016 DOI: 10.1594/ecr2016/C-0532
E. García, A. Oliver, Y. Díez, O. Diaz, J. Georgii, R. Martí and J. Martí (2015). 'Comparing regional breast density using Full-Field Mammograms and Magnetic Resonance Imaging: A preleminary study'. MICCAI-BIA 2015, Proceeding of the 3rd MICCAI Workshop on Breast Image Analysis, 33-40
View abstract View full publication
Breast density is well established as an important risk factor for the development of breast cancer. Therefore, its objective estimation has been the focus of research in the past decades. In addition to global volumetric measures, the local distribution and patterns of this density are currently being investigated to determine whether they can provide complementary information for risk assessment. This paper proposes a framework to evaluate the correlation between local spatial distribution of dense tissue in full-field digital mammograms (FFDM) using a density estimation software (VolparaTM) and magnetic resonance imaging (MRI). Initial results with 51 patients (204 images) showed a significant correlation using several local measures, the largest being 0.81. This indicates that local density patterns estimated in FFDM correlate well with those in MRI. However, pixelwise measures failed to yield the same degree of correlation. This may indicate that the areas where tissue densities are located in both approaches are comparable, but small variations in pixelwise tissue distribution between both approaches exist.
L. Wang, A. Gubern-Mérida, O. Diaz, Y. Diez, R. Mann, S. Diekmann, F. Zöhrer, H. Laue and J. Schwaab (2015). 'Automated assessment of motion in breast MRI to assess study quality and prevent unnecessary call-backs'. Presented at European Congress of Radiology (ECR), Poster C-1845. doi: 10.1594/ecr2015/C-1845
View abstract View full publication
Dynamic contrast enhanced (DCE) MRI is used for breast cancer screening examinations for women of high risk for developing breast cancer. Motion caused e.g. by muscle relaxation or coughing during image acquisition can reduce the interpretability of breast MRI. For scans with strong motion, it might be necessary to repeat the scan, but in typical screening workflow this is only detected by the radiologist when the woman has already left the clinic. As a consequence, the woman might need to be recalled for a repeated scan. In this work, a fully automated tool based on image-processing is proposed to detect and quantify motion for unambiguous scan quality evaluation before the woman leaves the clinic.
O. Diaz, P. Elangovan, S. Enshaeifar, P. Seller, S. Pani and K. Wells (2013). 'Breast CT image simulation framework for optimisation for lesion visualisation'. Accepted for IEEE Nuclear Seience Symposium and Medical Imaging Conference Record (NSS/MIC), Seoul, South Korea, October 2013. doi: 10.1109/NSSMIC.2013.6829158
View abstract View full publication
Although X-ray mammography is the gold standard technique for breast cancer detection, it suffers from limitations due to tissue superposition which could either obscure or mimic a breast lesion. Dedicated breast computed-tomography(BrCT) represents an alternative technology with the potential to overcome these limitations. However, this technology is still under investigation in order to study and improve certain parameters (e.g. dose, scattered radiation, etc.). In this work, an image simulation framework is proposed to generate realistic BrCT images and spectral imaging analysis is explored to enhance the contrast of breast lesions. Results illustrated an improvement in contrast between 5 and 10% when the final image is reconstructed using X-ray photons with energies between 21 and 30 keV, in comparison with the reconstructed image from the polychromatic energy spectrum recorded within the image receptor.
A. Rashidnasab, P. Elangovan, O. Diaz, A. Mackenzie, K.C. Young, D.R. Dance and K. Wells (2013). 'Simulation of 3D DLA masses in digital breast tomosynthesis'. In Proc. of SPIE Medical Imaging 8668, 86680Y-1-9. doi:10.1117/12.2008333
View abstract View full publication
Digital breast tomosynthesis (DBT) is suggested to have superior performance compared to 2D mammography in terms of cancer visibility, especially in the case of dense breasts. However, the overall performance of tomosynthesis for screening applications, and the manner in which tomosynthesis should be optimally used for screening remains unclear. This motivates the development of software tools that can insert user-defined synthetic pathology of realistic appearance into clinical tomosynthesis images for subsequent use in virtual clinical trials. We present a method for inserting lesions grown using Diffusion Limited Aggregation, previously validated in 2D mammograms, into clinical DBT images. A preliminary pilot study was used to validate the realism of the masses, wherein three readers each viewed 19 cases and rated the realism of the inserted masses. Each case included a simulated mass inserted in the tomosynthesis projections and the counterpart digital 2D mammogram. These results show that masses can be successfully embedded in the tomosynthesis projections and can produce visually authentic DBT images containing synthetic pathology. These results will be used to further optimize the appearance of these masses in DBT for an upcoming validation.
M. Esposito, T. Anaxagoras, O. Diaz, K. Wells and N. M. Allinson (2012). Radiation Hardness of a Large Area CMOS Active Pixel Sensor for Bio-medical applications'. In IEEE Nuclear Science Symposium and Medical Imaging Conference Record (NSS/MIC), N14-183, pp 1300-1304. doi:10.1109/NSSMIC.2012.6551318
View abstract View full publication
A wafer scale CMOS Active Pixel Sensor has been designed employing design techniques of transistor enclosed geometry and P+ doped guard rings to offer ionizing radiation tolerance. The detector was irradiated with 160 kVp X-rays up to a total dose of 94 kGy(Si) and remained functional. The radiation damage produced in the device has been studied, resulting in a dark current density increase per decade of 96±S pA/cm 2 /decade and a damage threshold of 204 Gy(Si). The damage produced in the detector has been compared with a commercially available CMOS APS, showing a radiation tolerance about 100 times higher. Moreover Monte Carlo simulations have been performed to evaluate primary and secondary energy deposition in each of the detector stages.
R. Bouwman, O. Diaz, K.C. Young, R. van Engen , W. Veldkamp, D.R. Dance (2012). 'Phantoms for quality control procedures of digital breast tomosynthesis'. IWDM 2012, Lecture Notes in Computer Science 7361, pp 322-329. doi:10.1007/978-3-642-31271-7_42
View abstract View full publication
For quality control (QC) protocols in full field digital mammography polymethyl methacrylate (PMMA) phantoms are generally used. The possibility of using alternative materials has been investigated for digital breast tomosynthesis (DBT) because of the increased importance of scatter and more complex imaging geometries. We have investigated the use of PMMA in combination with polyethylene (PE) to simulate a range of typical breasts using a computation model of the imaging system. The scatter-to-primary ratios (SPRs) of both breast and phantom were also investigated and a difference up to 18% is found. Neglecting this difference in SPR in designing phantoms for DBT may lead to dosimetry errors. Taking into account estimated SPR values and relevant X-ray spectra, a combination of PMMA-PE slabs has been proposed to simulate typical breasts of thicknesses 30, 60 and 90 mm. The dosimetric error associated with using these phantoms for relevant X-ray spectra is less than 10%.
P. Elangovan, A. Mackenzie, O. Diaz, A. Rashidnasab, D.R. Dance, K.C. Young, L. M. Warren, E. Shaheen, H. Bosmans, P.R. Bakic, K. Wells (2012). 'A Modelling Framework for Evaluation of 2D-Mammography and Breast Tomosynthesis systems'. IWDM 2012, Lecture Notes in Computer Science 7361, pp 338-345. doi:10.1007/978-3-642-31271-7_44
View abstract View full publication
Planar 2D X-ray mammography is the most common screening technique used for breast cancer detection. Digital breast tomosynthesis (DBT) is a new and emerging technology that overcomes some of the limitations of conventional planar imaging. However, it is important to understand the impact of these two modalities on cancer detection rates and patient recall. Since it is difficult to adequately evaluate different modalities clinically, a collection of modeling tools is introduced in this paper that can be used to emulate the image acquisition process for both modalities. In this paper, we discuss image simulation chains that can be used for the evaluation of 2D-mammography and DBT systems in terms of both technical factors and observer studies.
A. Mackenzie, D.R. Dance, O. Diaz, A. Barnard, K.C. Young (2012). 'Converting One Set of Mammograms to Simulate a Range of Detector Imaging Characteristics for Observer Studies'. IWDM 2012, Lecture Notes in Computer Science 7361, pp 394-401. doi:10.1007/978-3-642-31271-7_51
View abstract View full publication
A methodology for adjusting mammographic images taken on a given imaging system to simulate their appearance if taken on a different system for use in observer studies is presented. The process involves adjusting the image sharpness and noise, which takes into account the detector, breast thickness, and beam quality. The method has been tested by converting images acquired using an a-Se detector of a CDMAM test object and ‘Rachel’ anthropomorphic breast phantom. They were degraded to appear as if acquired using a computed radiography (CR) detector. Good agreement was achieved in the resulting threshold gold thickness for the simulated CR images with measured real values for CDMAM images. Power spectra comparisons of real and simulated images of the ‘Rachel’ phantom agree with an average difference of 4%. This tool in conjunction with observer studies can be used to understand the effects of the detector characteristics on cancer detection in mammography.
O. Diaz, D.R. Dance, K.C. Young, P. Elangovan, P.R. Bakic and K. Wells (2012). 'A fast scatter field estimator for digital breast tomosynthesis'. In Proc. SPIE Medical Imaging 8313, 831305. doi:10.1117/12.911494
View abstract View full publication
Digital breast tomosynthesis (DBT) is a promising alternative approach to overcome the limitations of tissue superposition found in full-field 2D digital mammography. However, due to the absence of anti-scatter grids in DBT, accurate scatter estimation for each projection is necessary for modelling the image reconstruction stage. In this work we identify the limitations associated with scatter estimation using spatial invariant scatter kernels, in particular at the edge region where such methods result in scatter overestimation. Such approaches show an overestimation of scatter-to-primary ratio of over 50% at the edges when compared with results from direct Monte Carlo simulation. This problem was found to increase with projection angle. Simulation work presented here shows that this overestimation in scatter is largely due to air gap between the lower curved breast edge and the detector. We propose a new fast, accurate scatter field estimator for use in DBT which not only considers the breast thickness and primary incidence angle, but also accounts for scatter exiting the breast edge region and traversing an air gap prior to absorption in the detector. The new proposed scatter estimator represents an alternative approach to this problem which reduces discrepancies at the edge of a breast phantom. Moreover, the time required for generating scatter has dropped from approximately 12 hours using Monte Carlo simulations for 1010 photons to just a few minutes per projection. The insertion of scatter from the compression paddle to aforementioned methodologies is also discussed.
A. Rashidnasab, P. Elangovan, D.R. Dance, K.C. Young, M. Yip, O. Diaz and K. Wells (2012). 'Realistic simulation of breast mass appearance using random walk'. In Proc. SPIE Medical Imaging 8313, 83130L. doi:10.1117/12.911641
View abstract View full publication
The aim of the present work was to develop a method for simulating breast lesions in digital mammographic images. Based on the visual appearance of real masses, three dimensional masses were created using a 3D random walk method where the choice of parameters (number of walks and number of steps) enables one to control the appearance of the simulated structure. This work is the first occasion that the random walk results have been combined with a model of digital mammographic imaging systems. This model takes into account appropriate physical image acquisition processes representing a particular digital X-ray mammography system. The X-ray spectrum, local glandularity above the insertion site and scatter were all taken account during the insertion procedure. A preliminary observer study was used to validate the realism of the masses. Seven expert readers each viewed 60 full field mammograms and rated the realism of the masses they contained. Half of the images contained real, histologically-confirmed masses, and half contained simulated lesions. The ROC analysis of the study (average AUC of 0.58±0.06) suggests that, on the average, there is evidence that the radiologists could distinguish, somewhat, between real and simulated masses.
A. Rashidnasab, P. Elangovan, D.R. Dance, K.C. Young, O. Diaz and K. Wells (2012). Modelling realistic breast lesions using diffusion limited aggregation'. In Proc. Medical Imaging SPIE 8313, 83134L. doi:10.1117/12.911512
View abstract View full publication
Synthesizing the appearance of malignant masses and inserting these into digital mammograms can be used as part of a wider framework for investigating the radiological detection task in X-ray mammography. However, the randomness associated with cell division within cancerous masses and the associated complex morphology challenges the realism of the modeling process. In this paper, Diffusion Limited Aggregation (DLA), a type of fractal growth process is proposed and utilized for modeling breast lesions. Masses of different sizes, shapes and densities were grown by controlling DLA growth parameters either prior to growth, or dynamically updating these during growth. A validation study was conducted by presenting 30 real and 30 simulated masses in a random order to a team of radiologists. The results from the validation study suggest that the observers found it difficult to differentiate between the real and simulated lesions.
O. Diaz, M. Yip, J. Cabello, D. R. Dance, K.C. Young and K. Wells (2010). 'Monte Carlo simulation of scatter field for calculation of contrast of discs in synthetic CDMAM images'. IWDM 2010, Lecture Notes in Computer Science 6136, pp 628-635. doi:10.1007/978-3-642-13666-5_85
View abstract View full publication
This paper reports on a further development of an image simulation chain, and in particular, the inclusion of contrast degradation across an image using scatter to primary ratios calculated using Monte Carlo simulation. The Monte Carlo technique, using the Geant4 toolkit, has been implemented to model the scatter conditions when imaging the CDMAM phantom with commercial digital mammography. Observed differences between linear and cellular anti scatter grid are presented and discussed. These results support previous assumptions taken by Yip et al.[1].
D. Rodriguez Gutierrez, O. Diaz Montesdeoca, A. Moran Santana, K.Wells, J. Cabello, I. A. Mendichovszky, I. Gordon (2007). 'MR-Based renography as a replacement for radionuclide diagnostic renography studies'. In Nuclear Science Symposium Conference Record, 2007; 6: 4556-4563. doi:10.1109/NSSMIC.2007.4437125
View abstract View full publication
This study explores the importance of the partial volume effect (PVE) in quantifying renal function on magnetic dynamic contrast enhanced magnetic resonance imaging (DCE- MRI). DCE-MRI image data were acquired for a healthy volunteer and after motion correction a PV correction step was applied to remove non-renal contributions to time-intensity curves derived from a typical renal cortical region of interest (ROI). PV correction consisted on the assignment of a mixing vector to each voxel location, representing the contributions of each tissue into a given voxel due to the convolution action of the point spread function (PSF) of the acquisition sequence. These mixing vectors were then used to recover the true intensities that correspond to the unmixed (i.e. pure) signals associated with each constituent tissue, eliminating contributions from liver, spleen and other surrounding tissues from the renal component. The result was an increased slope in the filtration part of the enhancement curve for the renal component. Quantitatively, PV correction resulted in an increase of glomerular filtration rate (GFR) estimated through a Rutland-Patlak analysis. Using a common DCE-MRI sequence, the contribution to a typical renal cortical ROI from non-renal tissues was found to be ~ 25%. The removal of these component resulted in a 32% (left) and 37% (right kidney) increase in relative GFR and an increased R2 for the Rutland Patlak model compared to the same analysis undertaken with no PVE correction.
P. Elangovan, A. Mackenzie, D.R. Dance, K. Wells, O. Diaz, L. M. Warren, K. C. Young (2019). 'Validation of modelling tools for simulating wide-angle DBT systems'. Accepted for presentation at the SPIE Medical Imaging Conference 2019: Physics of Medical Imaging (Conference 10948). Paper 10948-85
View abstract View full publication
The OPTIMAM image simulation toolbox contains a suite of tools that can used to simulate visually and clinically realistic images for VCTs. Recently, tools for simulating a wide-angle DBT system were added to the toolbox. In this paper, we present the simulation methodology and validation results for a wide-angle DBT system. The validation was performed by simulating images of standard test objects and comparing these with real images acquired using identical settings on the simulated real system. The comparison of the contrast-to-noise ratios, geometrical distortion (z-resolution) and image blurring for real and simulated images of test objects showed good agreement. This suggests that the images of a wide-angle DBT system produced using our simulation approach are comparable to real images.
C. Fedon, M. Caballo, O. Diaz, I. Sechopoulos (2019). Characterization of fibroglandular tissue distribution in compressed breasts. Accepted for presentation at the SPIE Medical Imaging Conference 2019: Physics of Medical Imaging (Conference 10948). Paper 10948-145.
View abstract View full publication
In this study, 42 breast CT images were automatically classified into four material categories (i.e. air, skin, adipose and fibroglandular tissue) and their mechanical compression simulated to match the compressed breast geometry during mammography or tomosynthesis. Each resulting breast volume was divided into 30 regions for each view (i.e. coronal, axial and sagittal), resulting in a 3D matrix with dimension 30x30x30. We observed that the fibroglandular tissue tends to be concentrated close to the patient chest-wall (rather than to the nipple region), laterally centred (i.e. no medial-lateral preferred orientation) and towards the bottom part of the breast in the axial direction. Our work suggests a different axial distribution (i.e. not symmetric about the centerline), pointing to the need for further investigation on a larger patient cohort, and to the extension of this analysis to consider other parameters.
O. Diaz, P. Elangovan, D.R. Dance, K. Wells, K.C. Young (2019). Can breast models be simplified to estimate scattered radiation in breast tomosythesis?. Accepted for presentation at the SPIE Medical Imaging Conference 2019: Physics of Medical Imaging (Conference 10948). Paper 10948-201.
View abstract View full publication
Scattered radiation can represent a large portion of the total signal recorded at the image receptor in digital breast tomosynthesis (DBT). This work investigates if scattered radiation produced from homogeneous breast models can be used to simulate the scatter field from an heterogeneous model. This can help to develop faster and simplest scatter estimation approaches, which are highly demanded in virtual clinical trials (VCT) strategies. Results have suggested that homogeneous phantoms, with appropriate glandularity, can approximate the scattered radiation produced by a heterogeneous phantom in the majority of the breast region, showing a median error below 2% across the central area.
C. Fedon, C. Rabin, M. Caballo, O. Diaz, E. Garcia, A. Rodriguez-Ruiz, G. Gonzalez-Sprinberg, I. Sechopoulos (2019). 'Monte Carlo study on optimal breast voxel resolution for dosimetry estimates in Digital Breast Tomosynthesis'. Accepted for publication in Physics in Medicine and Biology. DOI: 10.1088/1361-6560/aaf453 [IF 2.665, Q2 (42/129) RNMMI]
View abstract View full publication
Digital Breast Tomosynthesis (DBT) is currently used as an adjunct technique to Digital Mammography (DM) for breast cancer imaging. Being a quasi-3D image, DBT is capable of providing depth information on the internal breast glandular tissue distribution, which may be enough to obtain an accurate patient-specific radiation dose estimate. However, for this, information regarding the location of the glandular tissue, especially in the vertical direction (i.e. x-ray source to detector), is needed. Therefore, a dedicated reconstruction algorithm designed to localize the amount of glandular tissue, rather than for optimal diagnostic value, could be desirable. Such a reconstruction algorithm, or, alternatively, a reconstructed DBT image classification algorithm, could benefit from the use of larger voxels, rather than the small sizes typically used for the diagnostic task. In addition, the Monte Carlo (MC) based dose estimates would be accelerated by the representation of the breast tissue with fewer and larger voxels. Therefore, in this study we investigate the optimal DBT reconstructed voxel size that allows accurate dose evaluations (i.e. within 5%) using a validated Geant4-based MC code. For this, sixty patient-based breast models, previously acquired using dedicated breast computed tomography (BCT) images, were deformed to reproduce the breast during compression under a given DBT scenario. Two re-binning approaches were applied to the compressed phantoms, leading to isotropic and anisotropic voxels of different volumes. MC DBT simulations were performed reproducing the acquisition geometry of a SIEMENS Mammomat Inspiration system. Results show that isotropic cubic voxels of 2.73 mm size provide a dose estimate accurate to within 5% for 51/60 patients, while a comparable accuracy is obtained with anisotropic voxels of dimension 5.46 x 5.46 x 2.73 mm3. In addition, the MC simulation time is reduced by more than half in respect to the original voxel dimension of 0.273 x 0.273 x 0.273 mm3 when either of the proposed re-binning approaches is used. No significant differences in the effect of binning on the dose estimates are observed (Wilcoxon-Mann-Whitney test, p-value &gt; 0.4) between the 0° the 23° (i.e. the widest angular range) exposure.