PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Anal Quant Cytol Histol. Author manuscript; available in PMC 2010 August 1.
Published in final edited form as:
Anal Quant Cytol Histol. 2009 August; 31(4): 197–207.
PMCID: PMC2857332
NIHMSID: NIHMS183687

Phase Contrast Microscopy Analysis of Breast Tissue

Differences in Benign vs. Malignant Epithelium and Stroma
Wendy A. Wells, M.D., M.Sc, F.R.C.Path., Xin Wang, Ph.D., Charles P. Daghlian, Ph.D., Keith D. Paulsen, Ph.D., and Brian W. Pogue, Ph.D.

Abstract

OBJECTIVE

To assess how optical scatter properties in breast tissue, as measured by phase contrast microscopy and interpreted pathophysiologically, might be exploited as a diagnostic tool to differentiate cancer from benign tissue.

STUDY DESIGN

We evaluated frozen human breast tissue sections of adipose tissue, normal breast parenchyma, benign fibroadenoma tumors and noninvasive and invasive malignant cancers by phase contrast microscopy through quantification of grayscale values, using multiple regions of interest (ROI). Student’s t tests were performed on phase contrast measures across diagnostic categories testing data from individual cases; all ROI data were used as separate measures.

RESULTS

Stroma demonstrated significantly higher scatter intensity than did epithelium, with lower scattering in tumor-associated stroma as compared with normal or benign-associated stroma. Measures were comparable for invasive and noninvasive malignant tumors but were higher than those found in benign tumors and were lowest in adipose tissue.

CONCLUSION

Significant differences were found in scatter coefficient properties of epithelium and stroma across diagnostic categories of breast tissue, particularly between benign and malignant-associated stroma. Improved understanding of how scatter properties correlate with morphologic criteria used in routine pathologic diagnoses could have a significant clinical impact as developing optical technology allows macroscopic in situ phase contrast imaging.

Keywords: breast tissue, epithelium, phase contrast microscopy, scatter coefficient properties, stroma, tumor

The diagnostic practice of surgical pathology is mostly based on the cytologic and architectural features of the epithelium, with considerably less attention paid to the stromal properties. While the proteomic and genomic characteristics of benign and malignant stroma associated with epithelium are well documented,14 the morphologic characteristics of the stroma itself are not currently considered to be particularly differentiating in terms of rendering a clinical diagnosis. Developing biomedical imaging technologies that provide direct measurements of different tissue components may be of significant diagnostic value.57 For example, imaging scatter within tissue is now possible both invasively8,9 and noninvasively,10,11 but the direct quantification and correlation of scatter in tissue samples remains a challenge. The initial event in optical scatter occurs when there is a localized change of index of refraction in subcellular and extracellular components of the tissue and there is refraction or reflection of the light.1221 While bulk scatter measurements are generally dominated by the reflected signal from these local index changes, the same index change also leads to light diffraction through the sample. Quantifying the magnitude of diffraction of different tissues through phase contrast microscopy may provide insight into how in situ scatter data can be interpreted.

Phase contrast microscopy is already used to rapidly assess (within seconds) the epidermal margins in unstained cryosections that are generated during Mohs micrographic surgery performed to eradicate recurrent cutaneous malignancies of the skin.22 In phase contrast microscopy the local change in the refractive index causes a phase shift that is then visualized by imaging the forward-diffracted and undiffracted light through a phase plate. Though diffraction is neither a direct nor quantitatively calibrated measure of scatter, the resulting images would likely be monotonically related to overall Mie and Rayleigh scatter from submicroscopic fluctuations of the tissue index of refraction if the index change is the primary cause of scatter coming from the tissue. The current study investigates the magnitude of tissue diffraction through phase contrast microscopy and focuses on the tumor-associated stromal changes, which may be used as a diagnostic tool in routine clinical care.

Imaging systems that characterize optical scatter from bulk tissue vary considerably in their technological implementation. Early work in the 1990s showed that elastic scattering from cells could be used to estimate particle size, either through the spectrum or the polarization angle of the remitted light.1316 These observations led directly to applications of spectral elastic scattering to measure features associated with the cell nucleus in suspensions or monolayers.1719 Thin sections of epithelial tissues can be measured across the visible spectrum and in this spectrum can be linked to certain-sized particles and related effects within the tissue.20 For example, Wax et al19 showed that angular dependent scattering could be directly measured and correlated with disease progression. Wilson et al21 found that changes in angular dependent scattering might be used to track response to therapy via an apoptotic loss of mitochondrial integrity that induced short-term mitochondrial swelling. Recently it was reported that elastic scattering spectra can be measured through bulk tissues and that there appears to be scatter contrast in breast tissue pathology relative to the surrounding normal tissue.10 The coupling of this new measurement technique with a magnetic resonance scanner indicates quite dramatically how the data might be used to provide ultra-structural information about tissue in a diagnostic clinical setting. In addition to its potential in diagnostic imaging, use of tissue scatter as a biomarker in surgical pathology is also of significant interest.

In order to determine if the morphologic features are quantifiably different in tumor and tumor-associated areas, this study focused on directly correlating the morphology of different diagnostic groups in breast tissue with their ultrastructural and phase contrast values. The goal was to see if unique tissue compartments, such as benign or malignant stroma, could be better used diagnostically in the clinical setting.

Materials and Methods

Electron Microscopy

As previously reported,23 2-mm3 portions of fresh breast tissue were fixed in 5 mL of 4% buffered glutaraldehyde for at least 2 hours at 4°C and then were placed in 0.1 M sodium cacodylate buffer at 4°C for at least 15 minutes (and up to a week). The institutional review board (IRB) at the Dartmouth-Hitchcock Medical Center (DHMC) approved this de-identified tissue sampling without patient links. Separate areas of fat, connective tissue and glandular tissue were sampled and sent for electron microscopy. The tissue was then placed in 2% osmium tetroxide with 0.1 M sodium cacodylate buffer for 1 hour at 4°C, followed by graded ethanol (50%, 70%, 90% and 100%) and LR White (50%, 75% and 100%) washes (Electron Microscopy Sciences, Hatfield, Pennsylvania, U.S.A.). The tissue was embedded in LR White in covered gelatin capsules and allowed to stand overnight at room temperature in a hood before being placed in a 60°C oven for 22–26 hours.

The 0.5-µm-thick sections were cut and stained with 1% Toluidine Blue (Science Lab, Houston, Texas, U.S.A.). Selected areas were sectioned at 120 ran, placed on 3.0-mm copper grids and then stained with uranyl acetate and lead citrate. These samples were imaged using the FEI Tecnai F20 FEG electron microscope (FEI Company, Hillsboro, Oregon, U.S.A.). The images of fat, collagen and malignant glandular epithelium shown here were specifically included to illustrate the submicroscopic nature of the scatterers present in these tissue compartments (Figure 1).

Figure 1
Transmission electron microscopy images of (a) adipose tissue, (b) stromal tissue and (c) epithelial tissue. The major source of scattering in adipose is thought to be the larger fatty particles apparent here. In stromal tissue the collagen fiber bundles ...

Human Breast Tissues

The Tissue Bank in the Department of Pathology at DHMC snap freezes fresh tissue from any surgical specimen in which there is sufficient tissue over and above that required for a standard-of-care clinical pathologic diagnosis. This tissue is stored at −80°C, and all of the specimens are linked to the pathology diagnostic data in the Laboratory Information System. The Tissue Bank was searched for samples of human breast tissues in the categories of normal, benign neoplastic (fibroadenoma) and malignant neoplastic (noninvasive and invasive carcinomas of both lobular and ductal origin). Sequential frozen sections of each specimen were cut at 4 µm thickness after the frozen tissue had been embedded in OCT (Tissue-Tek OCT Compound for Cryostat Sectioning; Sakura Finetek USA, Inc., Torrance, California, U.S.A.). The IRB at DHMC approved this de-identified tissue sampling without patient links.

Phase Contrast Microscopy

A Nikon inverted microscope (Model DIAPHOT-TMD, 811829; Melville, New York, U.S.A.) was used for this study. It had a positive phase plate (Nikon Phase Contrast-2 ELWD 0.3, 206320) for phase contrast imaging below a long working distance condenser lens and a 16-bit CCD camera (CoolSnap Model). The light source was a Nikon super high-pressure mercury lamp. This microscope and phase ring set-up provided phase contrast imaging on a dark background, such that increases in diffraction when passing through the biologic tissue appear as increases in detected light by the CCD. The background of each phase contrast image was nearly black with a 50-count. Tissue that had higher scatter intensity appeared brighter. The acquisition was adjusted to ensure that saturation was not reached with any of the tissue samples and that the intensity of the highest scattering images was below the 16-bit maximum intensity. A low power, ×4 magnification, was used so that images of the same type of tissue were quantified together in terms of a ”bulk” value, rather than looking at the individual subcellular components.

Phase contrast imaging was performed on 1 of the unstained frozen tissue sections from each sample after it had been removed from the freezer and allowed to reach room temperature. A series of digital images were acquired to cover the entire area of the section. These were joined together to form a single montage image of the whole section. Frozen tissue sections were then fixed and stained with hematoxylin-eosin (H-E) to distinguish better the microscopic tissue morphology. A series of digital images of the H-E–stained sections were recorded and concatenated to construct another montage image of the entire field in the same way as the phase contrast sections. A detailed analysis of both images (1 phase contrast image and 1 H-E image) was conducted by pairing the locations of the different tissue types in each under the guidance of a surgical pathologist (W.A.W.). Using the H-E–stained montage, the histologic subtypes in each section were defined, as described in Table I, and the grayscale phase contrast measures were compared with the corresponding morphology.

Table I
Standardized Pathology Terminology Used to Subclassify Breast Tissue into Fat, Epithelium (Benign and Malignant Associations) and Stroma (Benign and Malignant Associations)

Pathology Terminology

A standardized pathology terminology, as shown in Table I, was used to subclassify 3 types of breast tissue: adipose tissue or fat, epithelium and stroma. Benign epithelium (in normal breast glandular tissue or as part of a benign fibroadenoma neoplasm) was distinguished from malignant epithelium (noninvasive and invasive). Stroma (normal collagen) not associated with any epithelium was distinguished from the stroma associated with either benign (normal or fibroadenoma) or malignant (noninvasive and invasive) epithelium. Figure 2 shows images of formalin-fixed, paraffin-embedded, H-E–stained, 4 µm-thick sections of normal breast epithelium and associated stroma (ben-nor E and ben-nor S); abundant intervening stroma not associated with epithelium (nor S); a benign fibroadenoma (ben-FA E and ben-FA S); noninvasive ductal carcinoma in situ (non-inv E and non-inv S); noninvasive and invasive lobular carcinoma (non-inv E, non-inv S, inv E and inv S); and invasive ductal carcinoma (inv E and inv S).

Figure 2
Formalin-fixed, paraffin-embedded, H-E–stained, 4-µm-thick sections of (upper left) normal breast epithelium and associated stroma (ben-nor E and ben-nor S) with abundant intervening stroma not associated with epithelium (nor S) (×40); ...

Phase Contrast Image Analysis

The process of quantifying different kinds of tissue in terms of their phase contrast values was carried out in grayscale using the NIH ImageJ software package (Image Processing and Analysis in Java, National Institutes of Health, Research Services Branch, http://rsb.info.nih.gov/ij/). The grayscale level indicated the relative scattering intensity of the tissue, a higher value corresponding to higher scattered intensity. The background intensity, subject to small variations from image to image, was subtracted off. Using Student’s t tests, the phase contrast measures for each diagnostic subcategory were compared in 2 ways. First, grayscale values were analyzed from approximately 10 regions of interest (ROI) values and their weighted average was taken to be the grayscale value for that tissue type in that section (case values). Second, the grayscale values were analyzed using all ROI data as separate, independent measurements.

Case Composition

The study included a total of 48 human breast cases distributed among the following diagnostic categories (Table II): 3 fat, 16 normal, 6 benign fibroade-nomas, 14 invasive carcinomas (ductal and lobular) and 9 noninvasive carcinomas (ductal and lobular). Measures of normal stroma (not associated with epithelium) and fat were taken from the appropriate morphologic areas in these cases.

Table II
Number of Cases and ROIs for Each Tissue Classification and the Variance Parameters for Box and Whisker Plots of the Phase Contrast Grayscale Values of All ROIs for Each Tissue Classification

Results

Electron Microscopy

Electron microscopy images of fat, collagenous stroma and glandular epithelium (Figure 1) have been used to examine the size and morphology of the constituents that might contribute to optical scattering.23 The initial event in a scattering process occurs when there is a change of index of refraction among the many tissue compartments, such as interstitial fluid, cellular membranes, cytoplasm, nuclei, nucleoli, mitochondria, golgi apparatus, lysosomes, chromatin, rough and smooth endoplasmic reticulum, collagen and actin filaments. The electron microscopy staining process relies upon the intercalation of high atomic number atoms into the membranes and stromal regions of the tissue. The submicroscopic features observed have a refractive index change in the visible, near-infrared wave-lengths relative to the surrounding fluid that leads to scatter. Fat (Figure 1A) is composed of amorphous fat particles that have diameters in the range of 10–100 nm, with some up to 1 µm in size. The average size of these particles, determined by counting methods, has been reported as 110 nm.23 The collagenous stromal tissue (Figure 1B) is arranged in both longitudinal and transverse bundles, and its scattering is dominated by its cross-sectional diameter, which falls between the Mie and Rayleigh scatter ranges. While the average cross-sectional diameter of collagen fibers—determined by the same counting methodology—is 50 nm, the fiber length is too large to be assessed using this quantitative methodology. Epithelial cells (Figure 1C) are characterized by a haphazard arrangement of membrane-bound organelles in the cytoplasm and the nucleotide material in the nucleus. There is a mixture of small and large particulate material with an average size determined previously to be 210 ± 20 nm.23

Phase Contrast Imaging of Scattering

Figure 2 shows images of formalin-fixed, paraffin-embedded, H-E–stained, 4-µm-thick sections of normal breast epithelium and associated stroma (ben-nor E and ben-nor S) with abundant intervening stroma not associated with epithelium (normal S), a benign fibroadenoma (ben-FA E and ben-FA S), non-invasive ductal carcinoma in situ (non-inv E and non-inv S), invasive ductal carcinoma (inv E and inv S) and noninvasive and invasive lobular carinoma (non-inv E, non-inv S, inv E and inv S). Figure 3 presents a set of matching low-power montaged phase contrast images with corresponding spatially correlated H-E–stained frozen section images of the same cases as those displayed in Figure 2.

Figure 3
A detailed analysis of the montage images (H-E image on left, phase contrast image on right) correlates the different tissue-type morphologies with the grayscale phase contrast measures (×40). Stroma shows the highest scatter in the phase contrast ...

In the phase contrast images corresponding to normal breast epithelium and associated stroma, ben-nor E appears dark gray, ben-nor S appears light gray and nor S looks bright. This indicates that nor S has the highest scattering intensity and that ben-nor S has a scattering intensity between ben-nor E and nor S. The phase contrast images of a fibroadenoma show distinct differences in intensity between the epithelium and stroma (ben-FA E and ben-FA S) as compared with the adjacent normal breast parenchyma (ben-nor E, ben-nor S and nor S). The phase contrast images corresponding to the malignant tumors (invasive and noninvasive) reveal a disproportionate increase in the amount of malignant epithelium (inv E and non-inv E) as compared with the associated stroma (inv S and non-inv S).

The phase contrast images were quantified in grayscale intensity values, after background correction, using software NIH ImageJ according to the pathology terminology in Table I. In the first analysis (Table III), the grayscale values for each diagnostic subcategory were recorded from approximately 10 ROI values and their weighted average was taken to be the grayscale value for that tissue type in that section (case values). Of these 45 cases, there were 16 ben-nor E and S; 6 ben-FA E and S; 9 non-inv E and S and 14 inv-E and S. In the second analysis (Table IV), the phase contrast value of each tissue type was analyzed based on the phase contrast values of all ROIs in that tissue type. As shown in Table IV, there were a total of 130 ROI values of inv-E; 92 ROI values of inv-S; 39 ROI values of non-inv E; 13 ROI values of non-inv S; 49 ROI values of ben-FA E; 43 ROI values of ben-FA S; 129 ROI values of ben-nor E; 114 ROI values of ben-nor S; 54 ROI values of nor S; and 13 ROI values of fat. The parameters for box and whisker plots are listed in Table II and shown graphically in Figure 4. The Student’s t test results for grayscale values of individual cases and using all ROI data as separate measurements are provided in Table III and andIV,IV, respectively.

Figure 4
Phase contrast data is shown for the categories of fat, epithelium and stroma, with the latter two groups subdivided into the list in Table I. Mean values are indicated with the box centerline, the inner quartiles are represented by the boxes and the ...
Table III
The p Values are Shown from t Tests to Determine if the Values from Phase Contrast of Each Tissue Subtype are Significantly Different from Each Other: Individual Case Analysis
Table IV
The p Values are Shown from t Tests to Determine if the Values from Phase Contrast of Each Tissue Subtype are Significantly Different from Each Other: ROI Data Measurements

There are significant differences in phase contrast between fat, epithelium and stroma (some associated with epithelium, some not). When comparing stroma not associated with epithelium (nor S) with epithelial-associated stroma in benign and malignant tumors (ben-FA S, non-inv S and inv S) in individual cases (Table III), the p values all show significant differences (<0.001, 0.005 and 0.002, respectively). There is no significant difference between stroma not associated with epithelium (nor S) and stroma associated with benign epithelium (ben-nor S). When comparing stroma not associated with epithelium (nor S) to any epithelial-associated stroma (ben-nor S, ben-FA S, non-inv S and inv S) using all ROI data as separate measurements (Table IV), the p values all show significant differences (0.005, <0.001, 0.001 and 0.002, respectively).

Fat has the smallest phase contrast value. All epithelial tissues (ben-nor E, ben-FA E, non-inv E and inv E) have smaller phase contrast values as compared with the associated stromal tissue (ben-nor S, ben-FA S, non-inv S and inv S). Normal breast epithelium (ben-nor E) and its associated stroma (ben-nor S) had the highest phase contrast intensity of all the associated epithelial and stromal subtypes, respectively. Measures for both the epithelium and stroma were comparable for invasive and noninvasive malignant tumors and were generally higher than those for benign tumors (fibroadenomas). The significant differences in the phase contrast measures of benign vs. malignant-associated stroma strongly suggest that the scatter of stroma could be diagnostically useful.

Discussion

This study quantifiably correlates the morphology of different diagnostic groups in breast tissue with their ultrastructural properties and phase contrast measures of diffraction. The results are important because they reveal statistically significant patterns of phase contrast that may be used to diagnostically distinguish benign and malignant lesions by differentiating the epithelial and stromal components of those lesions. Across all diagnostic categories, fat had the lowest phase contrast intensity, and stroma showed higher values when compared with its associated epithelial tissue. Of the epithelium measured, normal breast had the highest phase contrast intensity and the benign epithelium of fibroadenoma had the lowest values. Stroma either associated with normal epithelium or not associated with any epithelium had the highest values. The phase contrast intensity of fibroadenoma-associated stroma was significantly lower than malignant-associated (invasive and noninvasive) stroma. If phase contrast microscopy is able to distinguish these clinically relevant diagnostic categories, then the development of new optical technology allowing phase contrast imaging on a macroscopic, in situ scale would have an important clinical impact.

Imaging technology based upon phase contrast microscopy is already used to rapidly assess (within seconds) the epidermal margins in unstained cryosections generated during Mohs micrographic surgery, performed to eradicate recurrent cutaneous malignancies of the skin. This method saves time and effort as compared with the labor-intensive evaluation of numerous sequential, freshly cut, H-E–stained cryosections.22 The imaging distinction between the stroma associated with malignant (noninvasive and invasive) tumors, as compared with benign (fibroadenoma) tumors, is not well documented and may be an interesting observation to potentially exploit when imaging cancers and fibroadenomas in vivo.

There is also an overwhelming need for new macroscopic diagnostic tools to image resected tissue or the resection cavity intraoperatively, while in the surgical suite, with a level of specificity that is superior to simple white light imaging. The diffraction patterns observed here may lead to improved methods for the rapid evaluation of resection margin tissue during oncologic surgery, given the limitations of frozen and touch prep analysis on fatty breast tissue. A national reexcision rate of 32–63% for breast cancer reflects the considerable clinical impact of a positive resection margin after lumpectomy and the inadequacy of current methods for evaluating specimen margin status intraoperatively. If a surgeon could evaluate the presence of residual cancer at a resection margin, that would allow for directed, additional tissue to be taken at the time of the original excision rather than during a separate re-excision procedure.

The limitations of intraoperative frozen-section analysis of resection margins include the time taken to render a diagnosis, the freezing artifact that limits interpretation, the wide range of reported positive predictive rates (from 68% to 97%)2425 and the high rates of diagnoses being ”deferred” to the better-quality permanent tissue sections.26 Some institutions perform touch imprints of resection margins that are fixed in 95% methanol and rapidly stained with H-E.27 Although this technique is used to evaluate a larger area of the margin per imprint, the reported correlation between touch preparation cytology and histologic margins is variable, ranging from 99.1% to 37.5% sensitivity and 100% to 85.1% specificity.28 Given the current standard of care to assess margin status, new imaging technology that can provide an immediate intraoperative evaluation of the entire surgical margin for the presence of residual tumor would allow additional but targeted tissue to be removed at the time of the original excision. The resultant substantial benefit to the patient (e.g., reduced surgical risk, improved cosmetic outcome and reduced psychological stress) would have a major impact on the health care system (e.g., reduced costs and recovery time).

Phase contrast x-ray imaging appears to offer enhanced image quality compared with traditional x-ray methods at lower doses than required by conventional absorption techniques.29 Potential bio-medical uses include the noninvasive imaging of cartilage and bone in the development of treatments for degenerative joint disease, breast imaging using an analyzer crystal30,31 and lung imaging. But the current lack of suitable x-ray sources, optics and detectors remain obstacles to the widespread application of these techniques.9,32

Further study of the link between phase contrast diffraction, as measured here, and in situ scatter signals that are thought to be from reflected light should be further studied. Pilot studies have already demonstrated the role of scatter spectroscopy in the diagnosis of breast cancer. Elastic scattering spectroscopy, mediated by fiber optic probes, has been used to assess tumor resection margins during surgery8,9 using 2 artificial intelligence methods of spectral classification. Artificial neural networks yielded sensitivity and specificity values of 69% and 85%, respectively, whereas hierarchical cluster analysis produced sensitivity and specificity values of 67% and 79%, respectively. Though these results appear promising, the methodology requires multiple point measures, making it impractical for a rapid intraoperative evaluation of an excision cavity. Additionally, because the tumor-associated stroma and the tumor epithelium can be highly heterogeneous, it is not evident when studies based upon intrasurgical or biopsy-guided point measurements represent 1 or both of these tissue types. Imaging of scatter in diagnostically distinct tissue types may alleviate these problems and lead to more workable solutions for guiding surgery based upon scatter signatures. The basic studies of diffraction intensity examined here provide correlative pathobiologic information that could be used to examine the links between bulk tissue scattering signatures and the fundamental index of refraction changes occurring in that tissue.

Acknowledgments

The authors acknowledge the technical expertise of Maudine Waterman, HT(ASCP), in the Department of Pathology and Katherine Connolly in the Dartmouth Electron Microscope Facility, Dartmouth College, New Hampshire.

This work was funded by the National Cancer Institute funding of the Network for Translational Research in Optical Imaging (NTROI) (grant U54 CA105480) and the Program Project Grant PO1CA80139.

Footnotes

Financial Disclosure: The authors have no connection to any companies or products mentioned in this article.

References

1. Mylona E, Nomikos A, Magkou C, Kamberou M, Papas-sideri I, Keramopoulos A, Nakopoulou L. The clinicopatho-logical and prognostic significance of membrane type 1 matrix metalloproteinase (MT1-MMP) and MMP-9 according to their localization in invasive breast carcinoma. Histopathology. 2007;50:338–347. [PubMed]
2. Kalembeyi I, Inada H, Nishiura R, Imanaka-Yoshida K, Sakakura T, Yoshida T. Tenascin-C upregulates matrix metalloproteinase-9 in breast cancer cells: Direct and synergistic effects with transforming growth factor betal. Int J Cancer. 2003;105:53–60. [PubMed]
3. West RB, Nuyten DS, Subramanian S, Nielsen TO, Corless CL, Rubin BP, Montgomery K, Zhu S, Patel R, Hernandez-Boussard T, Goldblum JR, Brown PO, van de Vijver M, van de Rijn M. Determination of stromal signatures in breast carcinoma. PLoS Biol. 2005;3:el87. [PMC free article] [PubMed]
4. Netti PA, Berk DA, Swartz MA, Grodzinsky AJ, Jain RK. Role of extracellular matrix assembly in interstitial transport in solid tumors. Cancer Res. 2000;60:2497–2503. [PubMed]
5. Poplack SP, Paulsen KD, Hartov A, Meaney P, Pogue B, Tosteson TD, Grove MR, Soho SK, Wells WA. Electromagnetic breast imaging: Average tissue property values in women with negative clinical findings. Radiology. 2004;231:571–580. [PubMed]
6. Poplack SP, Tosteson TD, Wells WA, Pogue BW, Meaney PM, Hartov A, Kogel CA, Soho SK, Gibson JJ, Paulsen KD. Electromagnetic breast imaging: Results of a pilot study in women with abnormal mammograms. Radiology. 2007;243:350–359. [PubMed]
7. Wells WA, Daghlian CP, Tosteson TD, Grove MR, Poplack SP, Knowlton-Soho S, Paulsen KD. Analysis of the microvasculature and tissue type ratios in normal vs. benign and malignant breast tissue: Morphometric correlates in electromagnetic breast imaging. Anal Quant Cytol Histol. 2004;26:166–174. [PubMed]
8. Bigio II, Bown SG, Briggs G, Kelley C, Lakhani S, Pickard D, Ripley PM, Rose IG, Saunders C. Diagnosis of breast cancer using elastic-scattering spectroscopy: Preliminary clinical results. J Biomed Opt. 2000;5:221–228. [PubMed]
9. Johnson K, Chicken DW, Pickard DC, Lee AC, Briggs G, Falzon M, Bigio IJ, Keshtgar MR, Bown SG. Elastic scattering spectroscopy for intraoperative determination of sentinel lymph node status in the breast. J Biomed Opt. 2004;9:1122–1128. [PubMed]
10. Carpenter CM, Pogue BW, Jiang S, Dehghani H, Wang X, Paulsen KD, Wells WA, Forero J, Kogel C, Weaver JB, Poplack SP, Kaufman PA. Image-guided optical spectroscopy provides molecular-specific information in vivo: MRI-guided spectroscopy of breast cancer hemoglobin, water, and scatterer size. Opt Lett. 2007;32:933–935. [PubMed]
11. Cuccia DJ, Bevilacqua F, Durkin AJ, Tromberg BJ. Modulated imaging: Quantitative analysis and tomography of turbid media in the spatial-frequency domain. Opt Lett. 2005;30:1354–1356. [PubMed]
12. Streibl N. Three-dimensional imaging by a microscope. J Opt Soc Am A. 1985;2:121–127.
13. Hielscher AH, Mourant JR, Bigio IJ. Influence of particle size and concentration on the diffuse backscattering of polarized light from tissue phantoms and biological cell suspensions. Appl Opt. 1997;36:125–135. [PubMed]
14. Mourant JR, Freyer JP, Hielscher AH, Eick AA, Shen D, Johnson TM. Mechanisms of light scattering from biological cells relevant to noninvasive optical-tissue diagnostics. Appl Opt. 1998;37:3586–3593. [PubMed]
15. Perelman LT, Backman V, Wallace M, Zonios G, Manoharan R, Nusrat A, Shields S, Seiler M, Lima C, Hamano T, Itzkan I, Van Darn J, Crawford JM, Feld MS. Observation of periodic fine structure in reflectance from biological tissue: A new technique for measuring nuclear size distribution. Phys Rev Lett. 1998;80:627–630.
16. Wax A, Yang C, Backman V, Badizadegan K, Boone CW, Dasari RR, Feld MS. Cellular organization and substructure measured using angle-resolved low-coherence interferometry. Biophys J. 2002;82:2256–2264. [PubMed]
17. Mourant JR, Johnson TM, Freyer JP. Characterizing mammalian cells and cell phantoms by polarized backscattering fiber-optic measurements. Appl Opt. 2001;40:5114–5123. [PubMed]
18. Backman V, Gopal V, Kalashnikov M, Badizadegan K, Gurjar R, Wax A, Georgakoudi I, Mueller M, Boone CW, Dasari RR, Feld MS. Measuring cellular structure at submicrometer scale with light scattering spectroscopy. IEEE J Sel Top Quantum Electron. 2001;7:887–893.
19. Wax A, Yang C, Müller MG, Nines R, Boone CW, Steele VE, Stoner GD, Dasari RR, Feld MS. In situ detection of neoplastic transformation and chemopreventive effects in rat esophagus epithelium using angle-resolved low-coherence interferometry. Cancer Res. 2003;63:3556–3559. [PubMed]
20. Wallace MB, Perelman LT, Backman V, Crawford JM, Fitzmaurice M, Seiler M, Badizadegan K, Shields SJ, Itzkan I, Dasari RR, Van Dam J, Feld MS. Endoscopic detection of dysplasia in patients with Barrett’s esophagus using light-scattering spectroscopy. Gastroenterology. 2000;119:677–682. [PubMed]
21. Wilson JD, Bigelow CE, Calkins DJ, Foster TH. Light scattering from intact cells reports oxidative-stress-induced mitochondrial swelling. Biophys J. 2005;88:2929–2938. [PubMed]
22. Sato K, Kennard CD. Phase contrast microscopy for identifying epidermal margins in unstained cryosections for Mohs micrographic surgery. J Dermatol Surg Oncol. 1993;19:869–874. [PubMed]
23. Bartek M, Wang X, Wells W, Paulsen KD, Pogue BW. Estimation of subcellular particle size histograms with electron microscopy for prediction of optical scattering in breast tissue. J Biomed Opt. 2006;11:064007. [PubMed]
24. Tinnemans JG, Wobbes T, Holland R, Hendriks JH, van der Sluis RF, Lubbers EJ, de Boer HH. Mammographic and histopathologic correlation of nonpalpable lesions of the breast and reliability of frozen section diagnosis. Surg Gynecol Obstet. 1987;165:523–529. [PubMed]
25. Ferreiro JA, Gisvold JJ, Bostwick DG. Accuracy of frozen-section diagnosis of mammographically detected breast biopsies. Results of 1,490 consecutive cases. Am J Surg Path. 1995;19:1267–1271. [PubMed]
26. Fessia L, Ghiringhello B, Arisio R, Botta G, Aimone V. Accuracy of frozen section diagnosis in breast cancer detection: A review of 4436 biopsies and comparison with cytodiagnosis. Pathol Res Pract. 1984;179:61–66. [PubMed]
27. Klimberg VS, Westbrook KC, Korourian S. Use of touch preps for diagnosis and evaluation of surgical margins in breast cancer. Ann Surg Oncol. 1998;5:220–226. [PubMed]
28. Saarela AO, Paloneva TK, Rissanen TJ, Kiviniemi HO. Determinants of positive histologic margins and residual tumor after lumpectomy for early breast cancer: A prospective study with special reference to touch preparation cytology. J Surg Oncol. 1997;66:248–253. [PubMed]
29. Arfelli F, Assante M, Bonvicini V, Bravin A, Cantatore G, Castelli E, Dalla Palma L, Di Michiel M, Longo R, Olivo A, Pani S, Pontoni D, Poropat P, Prest M, Rashevsky A, Tromba G, Vacchi A, Vallazza E, Zanconati F. Low-dose phase contrast x-ray medical imaging. Phys Med Biol. 1998;43:2845–2852. [PubMed]
30. Pisano ED, Johnston RE, Chapman D, Geradts J, Iacocca MV, Livasy CA, Washburn DB, Sayers DE, Zhong Z, Kiss MZ, Thomlinson WC. Human breast cancer specimens: Diffraction-enhanced imaging with histologic correlation—Improved conspicuity of lesion detail compared with digitial radiography. Radiology. 2000;214:895–901. [PubMed]
31. Ingal VN, Beliaevskaya EA, Brianskaya AP, Merkurieva RD. Phase mammography: A new technique for breast investigation. Phys Med Biol. 1998;43:2555–2567. [PubMed]
32. Lewis RA. Medical phase contrast x-ray imaging: Current status and future prospects. Phys Med Biol. 2004;49:3573–3583. [PubMed]