Significance: Current imaging paradigms for differential diagnosis of suspicious breast lesions suffer from high false positive rates that force patients to undergo unnecessary biopsies. Diffuse optical spectroscopic imaging (DOSI) noninvasively probes functional hemodynamic and compositional parameters in deep tissue and has been shown to be sensitive to contrast between normal and malignant tissues.
Aim: DOSI methods are under investigation as an adjunct to mammography and ultrasound that could reduce false positive rates and unnecessary biopsies, particularly in radiographically dense breasts.
Methods: We performed a retrospective analysis of 212 subjects with suspicious breast lesions who underwent DOSI imaging. Physiological tissue parameters were z-score normalized to the patient’s contralateral breast tissue and input to univariate logistic regression models to discriminate between malignant tumors and the surrounding normal tissue. The models were then used to differentiate malignant lesions from benign lesions.
Results: Models incorporating several individual hemodynamic parameters were able to accurately distinguish malignant tumors from both the surrounding background tissue and benign lesions with area under the curve (AUC) ≥0.85. Z-score normalization improved the discriminatory ability and calibration of these predictive models relative to unnormalized or ratio-normalized data.
Conclusions: Findings from a large subject population study show how DOSI data normalization that accounts for normal tissue heterogeneity and quantitative statistical regression approaches can be combined to improve the ability of DOSI to diagnose malignant lesions. This improved diagnostic accuracy, combined with the modality’s inherent logistical advantages of portability, low cost, and nonionizing radiation, could position DOSI as an effective adjunct modality that could be used to reduce the number of unnecessary invasive biopsies.
Ideally, neoadjuvant chemotherapy (NAC) assessment should predict pathologic complete response (pCR), a surrogate clinical endpoint for 5-year survival, as early as possible during typical 3- to 6-month breast cancer treatments. We introduce and demonstrate an approach for predicting pCR within 10 days of initiating NAC. The method uses a bedside diffuse optical spectroscopic imaging (DOSI) technology and logistic regression modeling. Tumor and normal tissue physiological properties were measured longitudinally throughout the course of NAC in 33 patients enrolled in the American College of Radiology Imaging Network multicenter breast cancer DOSI trial (ACRIN-6691). An image analysis scheme, employing z-score normalization to healthy tissue, produced models with robust predictions. Notably, logistic regression based on z-score normalization using only tissue oxygen saturation (StO2) measured within 10 days of the initial therapy dose was found to be a significant predictor of pCR (AUC = 0.92; 95% CI: 0.82 to 1). This observation suggests that patients who show rapid convergence of tumor tissue StO2 to surrounding tissue StO2 are more likely to achieve pCR. This early predictor of pCR occurs prior to reductions in tumor size and could enable dynamic feedback for optimization of chemotherapy strategies in breast cancer.
Laser Optoacoustic Ultrasonic Imaging System Assembly (LOUISA-3D) was developed in response to demand of diagnostic radiologists for an advanced screening system for the breast to improve on low sensitivity of x-ray based modalities of mammography and tomosynthesis in the dense and heterogeneous breast and low specificity magnetic resonance imaging. It is our working hypothesis that co-registration of quantitatively accurate functional images of the breast vasculature and microvasculature, and anatomical images of breast morphological structures will provide a clinically viable solution for the breast cancer care. Functional imaging is LOUISA-3D is enabled by the full view 3D optoacoustic images acquired at two rapidly toggling laser wavelengths in the near-infrared spectral range. 3D images of the breast anatomical background is enabled in LOUISA-3D by a sequence of B-mode ultrasound slices acquired with a transducer array rotating around the breast. This creates the possibility to visualize distributions of the total hemoglobin and blood oxygen saturation within specific morphological structures such as tumor angiogenesis microvasculature and larger vasculature in proximity of the tumor. The system has four major components: (i) a pulsed dual wavelength laser with fiberoptic light delivery system, (ii) an imaging module with two arc shaped probes (optoacoustic and ultrasonic) placed in a transparent bowl that rotates around the breast, (iii) a multichannel electronic system with analog preamplifiers and digital data acquisition boards, and (iv) computer for the system control, data processing and image reconstruction. The most important advancement of this latest system design compared with previously reported systems is the full breast illumination accomplished for each rotational step of the optoacoustic transducer array using fiberoptic illuminator rotating around the breast independently from rotation of the detector probe. We report here a pilot case studies on one healthy volunteer and on patient with a suspicious small lesion in the breast. LOUISA3D visualized deoxygenated veins and oxygenated arteries of a healthy volunteer, indicative of its capability to visualize hypoxic microvasculature in cancerous tumors. A small lesion detected on optoacoustic image of a patient was not visible on ultrasound, potentially indicating high system sensitivity of the optoacoustic subsystem to small but aggressively growing cancerous lesions with high density angiogenesis microvasculature. The main breast vasculature (0.5-1 mm) was visible at depth of up to 40-mm with 0.3-mm resolution. The results of LOUISA-3D pilot clinical validation demonstrated the system readiness for statistically significant clinical feasibility study.
Exploitation of the optical properties of tissue to characterize biologic composition has created an era of continuous growth over the past decades for optical imaging. These changes enable the identification of functional abnormalities in conjunction with structural changes of biologic tissue. There is currently a wide array of technologies and applications in development and clinical use. The range of different optical hardware choices has led to systems that utilize optical tissue contrast to address specific clinical needs.
We present a framework for characterizing the performance of an experimental imaging technology, diffuse optical spectroscopic imaging (DOSI), in a 2-year multicenter American College of Radiology Imaging Network (ACRIN) breast cancer study (ACRIN-6691). DOSI instruments combine broadband frequency-domain photon migration with time-independent near-infrared (650 to 1000 nm) spectroscopy to measure tissue absorption and reduced scattering spectra and tissue hemoglobin, water, and lipid composition. The goal of ACRIN-6691 was to test the effectiveness of optically derived imaging endpoints in predicting the final pathologic response of neoadjuvant chemotherapy (NAC). Sixty patients were enrolled over a 2-year period at participating sites and received multiple DOSI scans prior to and during 3- to 6-month NAC. The impact of three sources of error on accuracy and precision, including different operators, instruments, and calibration standards, was evaluated using a broadband reflectance standard and two different solid tissue-simulating optical phantoms. Instruments showed <0.0010 mm−1 (10.3%) and 0.06 mm−1 (4.7%) deviation in broadband absorption and reduced scattering, respectively, over the 2-year duration of ACRIN-6691. These variations establish a useful performance criterion for assessing instrument stability. The proposed procedures and tests are not limited to DOSI; rather, they are intended to provide methods to characterize performance of any instrument used in translational optical imaging.
Intraoperative margin assessment to evaluate resected tissue margins for neoplastic tissue is performed to prevent reoperations following breast-conserving surgery. High resolution microendoscopy (HRME) can rapidly acquire images of fresh tissue specimens, but is limited by low image contrast in tissues with high optical scattering. In this study we evaluated two techniques to reduce out-of-focus light: HRME image acquisition with structured illumination (SI-HRME) and topical application of Lugol’s Iodine. Fresh breast tissue specimens from 19 patients were stained with proflavine alone or Lugol’s Iodine and proflavine. Images of tissue specimens were acquired using a confocal microscope and an HRME system with and without structured illumination. Images were evaluated based on visual and quantitative assessment of image contrast. The highest mean contrast was measured in confocal images stained with proflavine. Contrast was significantly lower in HRME images stained with proflavine; however, incorporation of structured illumination significantly increased contrast in HRME images to levels comparable to that in confocal images. The addition of Lugol’s Iodine did not increase mean contrast significantly for HRME or SI-HRME images. These findings suggest that structured illumination could potentially be used to increase contrast in HRME images of breast tissue for rapid image acquisition.
Breast cancer management could be improved by developing real-time imaging tools to assess tissue architecture without extensive processing. We sought to determine whether confocal fluorescence microscopy (CFM) provides sufficient information to identify neoplasia in breast tissue. Breast tissue specimens were imaged following proflavine application. Regions of interest (ROIs) were selected in histologic slides and in the corresponding region on confocal images, and then divided into sets for training and validation. Readers reviewed images in the training set and evaluated images in the validation set for the presence of neoplasia. Accuracy was assessed using histologic diagnosis as the gold standard. Seventy tissue specimens from 31 patients were imaged; 235 ROIs were identified and diagnosed as neoplastic or non-neoplastic. A training set was assembled using 23 matched ROIs; 49 matched ROIs were assembled into a validation set. Neoplasia was identified in histologic images: 93% sensitivity, 97% specificity [area under the curve (AUC=0.987 )] and in confocal images: 93% sensitivity 93% specificity (AUC=0.957 ). CFM produced images of architectural features in breast tissue comparable with conventional histology, while requiring little processing. Potential applications include assessment of excised tissue margins and evaluation of tissue adequacy for bio-banking and genomic studies.
To study the effects of overlapping anatomy on microcalcification detection at various incident exposure levels. Images
of an anthropomorphic breast phantom (RMI 169) overlapping with simulated microcalcifications ranging from 150 to
212 μm in size placed in two breast density regions, fatty and heterogeneously dense, were acquired with an a-Si/a-Se
flat panel based digital mammography system (Selenia) operated with Mo-Mo target/filter combination at 28 kVp. The
mammograms were exposed with 20, 30, 40, 60, 80, 120, 160, 240 and 325 mAs for varying the exposure level. A 4-AFC study was performed for evaluation of the detection performance. Four 400×400-pixel images were displayed as 2×2 array on a LCD flat panel based review workstation. One of the four images contained a cluster of five microcalcifications and was randomly placed in one of the four quadrants. A physicist was asked to select the image
containing the microcalcifications and to report the number of visible microcalcifications. The fraction of correct
responses was computed with two different criteria: (1) the selected images contained one or more microcalcifications,
and (2) the selected images contained 4 or 5 visible microcalcifications. The statistical significance of the differences in fractions for different exposure levels and regions was evaluated. The results showed that, if visibility of one or more
microcalcifications is required, the fractions of correct responses were 1 for all size groups and most exposure levels in
both fatty and heterogeneously dense regions. If a visibility of 80% or more of the microcalcifications was required, the
fractions of correct responses significantly decreased in both regions. The results indicated that microcalcification
detection in the fatty region appeared to be mainly limited by the quantum noise, and that in the heterogeneously dense region may be limited by both the anatomic noise and the quantum noise.
Breast density has been recognized as one of the major risk factors for breast cancer. However, breast
density is currently estimated using mammograms which are intrinsically 2D in nature and cannot
accurately represent the real breast anatomy. In this study, a novel technique for measuring breast density
based on the segmentation of 3D cone beam CT (CBCT) images was developed and the results were
compared to those obtained from 2D digital mammograms. 16 mastectomy breast specimens were imaged
with a bench top flat-panel based CBCT system. The reconstructed 3D CT images were corrected for the
cupping artifacts and then filtered to reduce the noise level, followed by using threshold-based
segmentation to separate the dense tissue from the adipose tissue. For each breast specimen, volumes of the
dense tissue structures and the entire breast were computed and used to calculate the volumetric breast
density. BI-RADS categories were derived from the measured breast densities and compared with those
estimated from conventional digital mammograms. The results show that in 10 of 16 cases the BI-RADS
categories derived from the CBCT images were lower than those derived from the mammograms by one
category. Thus, breasts considered as dense in mammographic examinations may not be considered as
dense with the CBCT images. This result indicates that the relation between breast cancer risk and true
(volumetric) breast density needs to be further investigated.
To investigate how the radiation dose level affects the detection of microcalcifications (MCs) in cone beam breast CT (CBCT), simulated MCs were embedded in simulated breast tissue and imaged with an experimental CBCT system. The system employs a 30 x 40 cm2 a-Si/CsI based flat panel detector with a pixel size of 194 microns. Three 5 x 5 clusters of simulated calcifications (212-224, 250-280, and 300-355 μm) were embedded in a stack of 11 cm diameter lunch meat and positioned at the center of each slice of lunch meat. 300 projection images over 360 degrees were acquired in the non-binning mode at various dose levels (4.2, 6, 12, 18, and 24 mGy) three times, and were reconstructed with the Feldkamp algorithm. After that, 767 x 767 x 9 volume data were extracted from the fifteen reconstructed images for each size group, resulting in 45 CBCT MC phantom images. An observer experiment was performed by counting the number of visible MCs for each MC phantom image. The phantom images were displayed on a review workstation with a 1600 x 1200 CRT monitor and reviewed by six readers independently. The order of the images was randomized for each reader. The ratios of the visible MCs were averaged over all readers and plotted as a function of the dose level. The CNR was calculated for each MC size and each doe level as well. The results showed that the performance of the reconstructed images acquired with 4.2 mGy was similar to the images acquired with 6 mGy, and the images acquired with 18 mGy performed similarly to those acquired with 24 mGy.
Overlapping fibroglandular tissue structures may obscure small calcifications, essential to the early detection of breast cancer. Dual-energy digital mammography (DEDM), where separate low- and high-energy images are acquired and synthesized to cancel the tissue structures, may improve the ability to detect and visualize calcifications amidst fibroglandular structures. We have developed and implemented a DEDM technique under full-field imaging conditions using a commercially available flat-panel based digital mammography system. We have developed techniques to suppress residual structures due to scatter contamination and non-uniformity in the x-ray field and detector response in our DEDM implementation. The total mean-glandular dose from the low- and high-energy images was constrained to be similar to screening examination levels. The low- and high-energy images were combined using a calibrated nonlinear (cubic) mapping function to generate the calcification images. To evaluate the dual-energy calcification images, we have designed a special phantom with calcium carbonate crystals to simulate calcifications of different sizes superimposed with a 5 cm thick breast-tissue-equivalent material with a continuously varying glandular-tissue ratio from 0.0 to 1.0. The suppression of tissue-structure background by dual-energy imaging comes with the cost of increased noise in the dual-energy images. We report on the effects of different image processing techniques on the dual-energy image signal and noise levels. The effects of image processing on the calcification contrast-to-noise ratios are also presented.
The purpose of this study is to compare the detection performance of three different mammography systems: screen/film (SF) combination, a-Si/CsI flat-panel (FP-), and charge-coupled device (CCD-) based systems. A 5-cm thick 50% adipose/50% glandular breast tissue equivalent slab phantom was used to provide an uniform background. Calcium carbonate grains of three different size groups were used to simulate microcalcifications (MCs): 112-125, 125-140, and 140-150 μm overlapping with the uniform background. Calcification images were acquired with the three mammography systems. Digital images were printed on hardcopy films. All film images were displayed on a mammographic viewer and reviewed by 5 mammographers. The visibility of the MC was rated with a 5-point confidence rating scale for each detection task, including the negative controls. Scores were averaged over all readers for various detectors and size groups. Receiver operating characteristic (ROC) analysis was performed and the areas under the ROC curves (Az’s) were computed for various imaging conditions. The results shows that (1) the FP-based system performed significantly better than the SF and CCD-based systems for individual size groups using ROC analysis (2) the FP-based system also performed significantly better than the SF and CCD-based systems for individual size groups using averaged confidence scale, and (3) the results obtained from the Az’s were largely correlated with these from confidence level scores. However, the correlation varied slightly among different imaging conditions.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.