Dynamic infrared imaging in identification of breast cancer tissue with combined image processing and frequency analysis

Five combinations of image-processing algorithms were applied to dynamic infrared (IR) images of six breast cancer patients preoperatively to establish optimal enhancement of cancer tissue before frequency analysis. mid-wave photovoltaic (PV) IR cameras with 320x254 and 640x512 pixels were used. The signal-to-noise ratio and the specificity for breast cancer were evaluated with the image-processing combinations from the image series of each patient. Before image processing and frequency analysis the effect of patient movement was minimized with a stabilization program developed and tested in the study by stabilizing image slices using surface markers set as measurement points on the skin of the imaged breast. A mathematical equation for superiority value was developed for comparison of the key ratios of the image-processing combinations. The ability of each combination to locate the mammography finding of breast cancer in each patient was compared. Our results show that data collected with a 640x512-pixel mid-wave PV camera applying image-processing methods optimizing signal-to-noise ratio, morphological image processing and linear image restoration before frequency analysis possess the greatest superiority value, showing the cancer area most clearly also in the match centre of the mammography estimation.     

Imaging of breast cancer with mid- and long-wave infrared camera

In this novel study the breasts of 15 women with palpable breast cancer were preoperatively imaged with three technically different infrared (IR) cameras—micro bolometer (MB), quantum well (QWIP) and photo voltaic (PV)—to compare their ability to differentiate breast cancer from normal tissue. The IR images were processed, the data for frequency analysis were collected from dynamic IR images by pixel-based analysis and from each image selectively windowed regional analysis was carried out, based on angiogenesis and nitric oxide production of cancer tissue causing vasomotor and cardiogenic frequency differences compared to normal tissue. Our results show that the GaAs QWIP camera and the InSb PV camera demonstrate the frequency difference between normal and cancerous breast tissue; the PV camera more clearly. With selected image processing operations more detailed frequency analyses could be applied to the suspicious area. The MB camera was not suitable for tissue differentiation, as the difference between noise and effective signal was unsatisfactory.     

Imaging of breast cancer with mid- and long-wave infrared camera

In this novel study the breasts of 15 women with palpable breast cancer were preoperatively imaged with three technically different infrared (IR) cameras - micro bolometer (MB), quantum well (QWIP) and photo voltaic (PV) - to compare their ability to differentiate breast cancer from normal tissue. The IR images were processed, the data for frequency analysis were collected from dynamic IR images by pixel-based analysis and from each image selectively windowed regional analysis was carried out, based on angiogenesis and nitric oxide production of cancer tissue causing vasomotor and cardiogenic frequency differences compared to normal tissue. Our results show that the GaAs QWIP camera and the InSb PV camera demonstrate the frequency difference between normal and cancerous breast tissue; the PV camera more clearly. With selected image processing operations more detailed frequency analyses could be applied to the suspicious area. The MB camera was not suitable for tissue differentiation, as the difference between noise and effective signal was unsatisfactory.     

A method for determination of optimal image enhancement for the detection of mammographic abnormalities

We present a paradigm for empirical evaluation of digital image enhancement algorithms for mammography that uses psychophysical methods for implementation and analysis of a clinically relevant detection task. In the experiment, the observer is asked to detect and assign to a quadrant, or indicate the absence of, a simulated mammographic structure characteristic of cancer embedded in a background image of normal breast tissue. Responses are indicated interactively on a computer workstation. The parameter values for the enhancement applied to the composite image may be varied on each trial, and structure detection performance is estimated for each enhancement condition. Preliminary investigations have provided insight into an appropriate viewing duration, and furthermore, suggest that nonradiologists may be used under this methodology for the tasks investigated thus far, for predicting parameter values for clinical investigation. We are presently using this method in evaluating several contrast enhancement algorithms of possible benefit in mammography. These methods enable an objective, clinically relevant evaluation, for the purpose of optimal parameter determination or performance assessment, of digital image-processing methods potentially used in mammography.     

MHRA and DES Publications Update

The study focuses on 12 breasts of six breast cancer patients sequential µm-wavelength imaging, taken by two different 3–5 μm wavelength area indium antimony (InSb) photovoltaic cameras. The aim of the study was to compare the functionality of area and pixel-based frequency analyses. Comparisons between these frequency analysis methods were made according to their relevancy to mammographic findings. Another objective of the study was to find reliable imaging conditions by specifying the border conditions for the patient stabilizing imaging bed and managing the imaging situation. According to the results, the match of pixel based frequency analysis to the mammography findings is better than using area frequency analysis. The results also indicate that when the optical axis of the camera in relation to the surface of the breast to be imaged grows to more than 40°, the emissivity changes dramatically and at that point reliable results will not be obtained. Consequently the analysis of the imagined breast requires more images to be fused into one analysis image to cover the whole breast.     

真彩色医学图像处理分析系统与临床应用

通过临床使用自行研制的真彩色通用医学图像处理分析系统发现：该系统充分考虑实验科学研究样本分析的特点，综合医学各科对图像处理分析的不同需求及多种图像分析仪之优点，在微机上实现了处理分析各种医学图像的功能，尤其适合对同一类图像样本的重复处理、分析与测量。系统提供了图像输入／输出，图像增强、编辑、分割，参数测量统计软件的数据接口等功能，可测量图像的几何参数、灰度参数、体视学参数等４０余项。可广泛应用于医学生物学、冶金、机械、材料、地质、石油、化工等领域。     

Scanned-projection digital mammography

The effectiveness of film-screen mammography is limited by tradeoffs between latitude and contrast, film granularity, and the need to increase dose when antiscatter methods are used. We are currently developing a scanned-projection digital mammography (SPDM) system to overcome these limitations. The system consists of a pair of scanning slits, a high-resolution x-ray image intensifier tube, a linear photodiode array, and a digital display. The detective quantum efficiency of the SPDM system at spatial frequencies up to 3 cycles/mm is similar to that of mammographic film-screen combinations, but is lower at high frequencies. For low-contrast objects as small as 0.1 mm in diameter, the signal-to-noise ratio is currently equal to that of optimally exposed mammographic film-screen images for equal dose to the breast and superior for regions which would be underexposed or overexposed on film. This is achieved by the use of a low-noise detector system, geometric magnification, and scatter elimination. Images of a contrast-detail phantom and excised breast tissue illustrate the superior contrast sensitivity of SPDM.     

Contrast Limited Adaptive Histogram Equalization image processing to improve the detection of simulated spiculations in dense mammograms

The purpose of this project was to determine whether Contrast Limited Adaptive Histogram Equalization (CLAHE) improves detection of simulated spiculations in dense mammograms. Lines simulating the appearance of spiculations, a common marker of malignancy when visualized with masses, were embedded in dense mammograms digitized at 50 micron pixels, 12 bits deep. Film images with no CLAHE applied were compared to film images with nine different combinations of clip levels and region sizes applied. A simulated spiculation was embedded in a background of dense breast tissue, with the orientation of the spiculation varied. The key variables involved in each trial included the orientation of the spiculation, contrast level of the spiculation and the CLAHE settings applied to the image. Combining the 10 CLAHE conditions, 4 contrast levels and 4 orientations gave 160 combinations. The trials were constructed by pairing 160 combinations of key variables with 40 backgrounds. Twenty student observers were asked to detect the orientation of the spiculation in the image. There was a statistically significant improvement in detection performance for spiculations with CLAHE over unenhanced images when the region size was set at 32 with a clip level of 2, and when the region size was set at 32 with a clip level of 4. The selected CLAHE settings should be tested in the clinic with digital mammograms to determine whether detection of spiculations associated with masses detected at mammography can be improved.Key Words: mammography, image processing, contrast limited adaptive histogram equalization, observer studies, breast cancer, spiculations     

Detection of early breast cancer: an overview and future prospects

Detection and treatment of breast cancer at an early stage is the only method with proven potential for lowering the death rate from this disease. Detection of early breast cancer is promoted by the American Cancer Society, American College of Radiology, and Canadian Association of Radiologists by encouraging the regular use of three types of screening: breast self-examination, clinical breast examination, and mammography. When all factors are considered, it has been convincingly demonstrated that the potential benefits of mammography far outweigh the minimal, clinically undetected radiation risk incurred by the examination. New technologies, such as computed tomography, magnetic resonance imaging, transillumination diaphanography, ultrasound, thermography, and digital subtraction angiography might offer a wide selection for patient examination. However, none of these procedures, in its present form, is expected to replace mammography as the first-line imaging technique for the detection and diagnosis of benign and malignant breast lesions. Breast cancer is detected now, in most cases, via casual or informed breast self-examination. This first-line of detection is not sufficient, since most tumors may metastasize before they reach a palpable size. Mammography generally shows up tumors no smaller than 1-cm diameter, which in many cases have already metastasized. The more advanced imaging modalities in their current forms suffer from a number of drawbacks that give them a lower overall detection rate than mammography. Understandably, improving breast imaging modalities is a great challenge to diagnostic radiology. The purpose of this article is to provide a comprehensive overview of the detection of early breast cancer. It briefly discusses the understanding of breast cancer, its incidence, and the mortality and survival of patients with breast cancer, as well as screening programs for breast cancer. We review the developments in mammography and other breast imaging modalities over the last several years. Prospects for digital mammography, digital image enhancement, and three-dimensional digital subtraction mammography, which may someday supplant film mammography, are also discussed.     

NADINE: new approaches to detecting breast cancer by sequential μm-wavelength imaging with the aid of novel frequency analysis techniques

The study focuses on 12 breasts of six breast cancer patients sequential μm-wavelength imaging, taken by two different 3-5 μm wavelength area indium antimony (InSb) photovoltaic cameras. The aim of the study was to compare the functionality of area and pixel-based frequency analyses. Comparisons between these frequency analysis methods were made according to their relevancy to mammographic findings. Another objective of the study was to find reliable imaging conditions by specifying the border conditions for the patient stabilizing imaging bed and managing the imaging situation. According to the results, the match of pixel based frequency analysis to the mammography findings is better than using area frequency analysis. The results also indicate that when the optical axis of the camera in relation to the surface of the breast to be imaged grows to more than 40°, the emissivity changes dramatically and at that point reliable results will not be obtained. Consequently the analysis of the imagined breast requires more images to be fused into one analysis image to cover the whole breast.     

Automated detection of microcalcification clusters for digital breast tomosynthesis using projection data only: a preliminary study

Digital breast tomosynthesis (DBT) is a promising modality for breast imaging in which an anisotropic volume image of the breast is obtained. We present an algorithm for computerized detection of microcalcification clusters (MCCs) for DBT. This algorithm operates on the projection views only. Therefore it does not depend on reconstruction, and is computationally efficient. The algorithm was developed using a database of 30 image sets with microcalcifications, and a control group of 30 image sets without visible findings. The patient data were acquired on the first DBT prototype at Massachusetts General Hospital. Algorithm sensitivity was estimated to be 0.86 at 1.3 false positive clusters, which is below that of current MCC detection algorithms for full-field digital mammography. Because of the small number of patient cases, algorithm parameters were not optimized and one linear classifier was used. An actual limitation of our approach may be that the signal-to-noise ratio in the projection images is too low for microcalcification detection. Furthermore, the database consisted of predominantly small MCC. This may be related to the image quality obtained with this first prototype.     

Optimization of exposure parameters in full field digital mammography

Optimization of exposure parameters (target, filter, and kVp) in digital mammography necessitates maximization of the image signal-to-noise ratio (SNR), while simultaneously minimizing patient dose. The goal of this study is to compare, for each of the major commercially available full field digital mammography (FFDM) systems, the impact of the selection of technique factors on image SNR and radiation dose for a range of breast thickness and tissue types. This phantom study is an update of a previous investigation and includes measurements on recent versions of two of the FFDM systems discussed in that article, as well as on three FFDM systems not available at that time. The five commercial FFDM systems tested, the Senographe 2000D from GE Healthcare, the Mammomat Novation DR from Siemens, the Selenia from Hologic, the Fischer Senoscan, and Fuji's 5000MA used with a Lorad M-IV mammography unit, are located at five different university test sites. Performance was assessed using all available x-ray target and filter combinations and nine different phantom types (three compressed thicknesses and three tissue composition types). Each phantom type was also imaged using the automatic exposure control (AEC) of each system to identify the exposure parameters used under automated image acquisition. The figure of merit (FOM) used to compare technique factors is the ratio of the square of the image SNR to the mean glandular dose. The results show that, for a given target/filter combination, in general FOM is a slowly changing function of kVp, with stronger dependence on the choice of target/filter combination. In all cases the FOM was a decreasing function of kVp at the top of the available range of kVp settings, indicating that higher tube voltages would produce no further performance improvement. For a given phantom type, the exposure parameter set resulting in the highest FOM value was system specific, depending on both the set of available target/filter combinations, and on the receptor type. In most cases, the AECs of the FFDM systems successfully identified exposure parameters resulting in FOM values near the maximum ones, however, there were several examples where AEC performance could be improved.     

Cytologic identification of clinically occult proliferative breast disease in women with a family history of breast cancer

A cytologic method for sampling the normal breast by fine-needle aspiration (FNA) was used to determine the frequency of clinically inapparent proliferative breast disease (PBD) in women with family histories of breast cancer. The authors attempted to obtain specimens from each quadrant of both breasts in 51 female first-degree relatives of breast cancer patients. The study group had no detectable masses by physical examination or mammography. Samples were prepared on membrane filters, Papanicolaou stained, and evaluated cytomorphologically. Three hundred seventy-eight of 408 (92.6%) possible quadrants were sampled; cellular material was obtained from 290 (76.7%) quadrants. PBD was identified in 20 of the 51 women (39.2%). When epithelium was obtained, nuclear area, perimeter, and diameter were measured with the use of computerized image analysis. Nuclei in samples containing atypical hyperplasia showed significant differences in these parameters when compared with cells from samples containing normal epithelium or benign hyperplasia. The authors' findings indicate that FNA sampling and computerized image analysis are useful in the detection and characterization of clinically inapparent PBD.     

Theoretical optimization of dual-energy x-ray imaging with application to mammography

Detection of a target object in a radiological image is often impeded by an obscuring background "clutter" resulting from the contrast between various materials in the neighborhood of the target. Dual-energy techniques can reduce or remove this clutter. In order for the target to be detectable in the image after dual-energy processing, the signal-to-noise ratio (SNR), defined as the difference between the target and the background divided by the photon noise in the difference, must exceed some threshold. A given SNR may be obtained for a wide range of the energies of the two x-ray beams and the ratio of their fluences. A theoretical model is developed which permits the choice of beams to be optimized with respect to some critical parameter--in this case, patient dose. The analysis is applied to the detection of calcifications in mammography. For an ideal imaging system, we predict that the optimum beam energies are 19 and 68 keV. A dose of 0.42 cGy is required to obtain an SNR of 5 for detection of a 0.02-cm cubic calcification in the resulting clutter-free image. This can be reduced to 0.16 cGy if the higher energy image is smoothed, prior to dual-energy processing, such that its variance is reduced to one-fourth of its unsmoothed value.     

AEC for scanning digital mammography based on variation of scan velocity

A theoretical evaluation of nonuniform x-ray field distributions in mammography was conducted. An automatic exposure control (AEC) is proposed for a scanning full field digital mammography system. It uses information from the leading part of the detector to vary the scan velocity dynamically, thus creating a nonuniform x-ray field in the scan direction. Nonuniform radiation fields were also created by numerically optimizing the scan velocity profile to each breast's transmission distribution, with constraints on velocity and acceleration. The goal of the proposed AEC is to produce constant pixel signal-to-noise ratio throughout the image. The target pixel SNR for each image could be set based on the breast thickness, breast composition, and the beam quality as to achieve the same contrast-to-noise ratio between images for structures of interest. The results are quantified in terms of reduction in entrance surface air kerma (ESAK) and scan time relative to a uniform x-ray field. The theoretical evaluation was performed on a set of 266 mammograms. The performance of the different methods to create nonuniform fields decreased with increased detector width, from 18% to 11% in terms of ESAK reduction and from 30% to 25% in terms of scan time reduction for the proposed AEC and detector widths from 10 to 60 mm. Some correlation was found between compressed breast thickness and the projected breast area onto the image field. This translated into an increase of the ESAK and decrease of the scan time reduction with breast thickness. Ideally a nonuniform field in two dimensions could reduce the entrance dose by 39% on average, whereas a field nonuniform in only the scanning dimension ideally yields a 20% reduction. A benefit with the proposed AEC is that the risk of underexposing the densest region of the breast can be virtually eliminated.     

Remote Non-invasive Stereoscopic Imaging of Blood Vessels: First In-vivo Results of a New Multispectral Contrast Enhancement Technology

We describe a contactless optical technique selectively enhancing superficial blood vessels below variously pigmented intact human skin by combining images in different spectral bands. Two CMOS-cameras, with apochromatic lenses and dual-band LED-arrays, simultaneously streamed Left (L) and Right (R) image data to a dual-processor PC. Both cameras captured color images within the visible range (VIS, 400–780 nm) and grey-scale images within the near infrared range (NIR, 910–920 nm) by sequentially switching between LED-array emission bands. Image-size-settings of 1280 × 1024 for VIS & 640 × 512 for NIR produced 12 cycles/s (1 cycle = 1 VIS L&R-pair + 1 NIR L&R-pair). Decreasing image-size-settings (640 × 512 for VIS and 320 × 256 for NIR) increased camera-speed to 25 cycles/s. Contrasts from below the tissue surface were algorithmically distinguished from surface shadows, reflections, etc. Thus blood vessels were selectively enhanced and back-projected into the stereoscopic VIS-color-image using either a 3D-display or conventional shutter glasses.As a first usability reconnaissance we applied this custom-built mobile stereoscopic camera for several clinical settings:• blood withdrawal;• vein inspection in dark skin;• vein detection through iodide;• varicose vein and nevi pigmentosum inspection.Our technique improves blood vessel visualization compared to the naked eye, and supports depth perception.     

High-temporal resolution DCE-MRI improves assessment of intra- and peri-breast lesions categorized as BI-RADS 4

Grading of cancer histopathology slides requires more pathologists and expert clinicians as well as it is time consuming to look manually into whole-slide images. Hence, an automated classification of histopathological breast cancer sub-type is useful for clinical diagnosis and therapeutic responses. Recent deep learning methods for medical image analysis suggest the utility of automated radiologic imaging classification for relating disease characteristics or diagnosis and patient stratification. To develop a hybrid model using the convolutional neural network (CNN) and the long short-term memory recurrent neural network (LSTM RNN) to classify four benign and four malignant breast cancer subtypes. The proposed CNN-LSTM leveraging on ImageNet uses a transfer learning approach in classifying and predicting four subtypes of each. The proposed model was evaluated on the BreakHis dataset comprises 2480 benign and 5429 malignant cancer images acquired at magnifications of 40×, 100×, 200× and 400×. The proposed hybrid CNN-LSTM model was compared with the existing CNN models used for breast histopathological image classification such as VGG-16, ResNet50, and Inception models. All the models were built using three different optimizers such as adaptive moment estimator (Adam), root mean square propagation (RMSProp), and stochastic gradient descent (SGD) optimizers by varying numbers of epochs. From the results, we noticed that the Adam optimizer was the best optimizer with maximum accuracy and minimum model loss for both the training and validation sets. The proposed hybrid CNN-LSTM model showed the highest overall accuracy of 99% for binary classification of benign and malignant cancer, and, whereas, 92.5% for multi-class classifier of benign and malignant cancer subtypes, respectively. To conclude, the proposed transfer learning approach outperformed the state-of-the-art machine and deep learning models in classifying benign and malignant cancer subtypes. The proposed method is feasible in classification of other cancers as well as diseases.     

MuDeRN: Multi-category classification of breast histopathological image using deep residual networks

MotivationIdentifying carcinoma subtype can help to select appropriate treatment options and determining the subtype of benign lesions can be beneficial to estimate the patients’ risk of developing cancer in the future. Pathologists’ assessment of lesion subtypes is considered as the gold standard, however, sometimes strong disagreements among pathologists for distinction among lesion subtypes have been previously reported in the literature.ObjectiveTo propose a framework for classifying hematoxylin-eosin stained breast digital slides either as benign or cancer, and then categorizing cancer and benign cases into four different subtypes each.Materials and methodsWe used data from a publicly available database (BreakHis) of 81 patients where each patient had images at four magnification factors (×40, ×100, ×200, and ×400) available, for a total of 7786 images. The proposed framework, called MuDeRN (MUlti-category classification of breast histopathological image using DEep Residual Networks) consisted of two stages. In the first stage, for each magnification factor, a deep residual network (ResNet) with 152 layers has been trained for classifying patches from the images as benign or malignant. In the next stage, the images classified as malignant were subdivided into four cancer subcategories and those categorized as benign were classified into four subtypes. Finally, the diagnosis for each patient was made by combining outputs of ResNets’ processed images in different magnification factors using a meta-decision tree.ResultsFor the malignant/benign classification of images, MuDeRN’s first stage achieved correct classification rates (CCR) of 98.52%, 97.90%, 98.33%, and 97.66% in ×40, ×100, ×200, and ×400 magnification factors respectively. For eight-class categorization of images based on the output of MuDeRN’s both stages, CCRs in four magnification factors were 95.40%, 94.90%, 95.70%, and 94.60%. Finally, for making patient-level diagnosis, MuDeRN achieved a CCR of 96.25% for eight-class categorization.ConclusionsMuDeRN can be helpful in the categorization of breast lesions.     

Basal lamina visualization using color image processing and pattern recognition.

In histologic assessment, the absence of basal lamina is a useful feature for distinguishing invasive malignancy from benign and in situ lesions. As this feature is not possible to assess in routine H&E sections, pathologists have instead relied on histochemical and immunohistochemical stains to show components of the basal lamina such as laminin or type IV collagen. Standard image-processing software with the necessary image-processing toolbox (Matlab v5, Mathworks, Natick, MA) was used in a unique combination of color image processing and pattern recognition techniques to accentuate the collagenous stroma surrounding glands, which approximates basal lamina, in a series of benign, in situ, and invasive breast proliferations. Distinct differences in pattern were found between benign and invasive lesions, and also between in situ and malignant lesions, corresponding to that observed with type IV collagen immunostaining. Compared with immunostaining, this computer-generated method had a sensitivity of 0.96, specificity of 0.89, positive predictive value of 0.92, negative predictive value of 0.89, positive likelihood ratio of 9.1, and negative likelihood ratio of 0.042. Digital image processing serves as a less expensive and faster way of visualizing basal lamina and represents a useful adjunct to identify invasive malignancy in routinely stained sections. In addition, digital visualization of basal lamina is readily amenable to quantitative assessment, and the method provides a basis for the development of computer-based cancer diagnosis.     

Effect of area x-ray beam equalization on image quality and dose in digital mammography

In mammography, thick or dense breast regions persistently suffer from reduced contrast-to-noise ratio (CNR) because of degraded contrast from large scatter intensities and relatively high noise. Area x-ray beam equalization can improve image quality by increasing the x-ray exposure to under-penetrated regions without increasing the exposure to other breast regions. Optimal equalization parameters with respect to image quality and patient dose were determined through computer simulations and validated with experimental observations on a step phantom and an anthropomorphic breast phantom. Three parameters important in equalization digital mammography were considered: attenuator material (Z = 13-92), beam energy (22-34 kVp) and equalization level. A Mo/Mo digital mammography system was used for image acquisition. A prototype 16 x 16 piston driven equalization system was used for preparing patient-specific equalization masks. Simulation studies showed that a molybdenum attenuator and an equalization level of 20 were optimal for improving contrast, CNR and figure of merit (FOM = CNR2/dose). Experimental measurements using these parameters showed significant improvements in contrast, CNR and FOM. Moreover, equalized images of a breast phantom showed improved image quality. These results indicate that area beam equalization can improve image quality in digital mammography.