Techniques capable of evaluating human disease in a safe ...



Diagnosis of Breast Cancer Using Diffusive Reflectance and Intrinsic Fluorescence Spectroscopy

Zoya Volynskaya*, Abigail Haka¢, Maryann Fitzmaurice§, Joseph Gardecki*, Robert Shenk§, Nancy Wang§, Jon Nazemi*, Ramachandra Dasari* and Michael Feld*

*Massachusetts Institute of Technology, Cambridge, MA

¢Weill  Medical College of Cornell University

§University Hospitals of Cleveland and Case Western Reserve University, Cleveland, OH

Abstract: We have developed a clinical instrument that combines intrinsic fluorescence spectroscopy (IFS) and diffuse reflectance spectroscopy (DRS) as a clinical tool for the ex vivo diagnosis of breast cancer. Methods: We collected 225 spectra from 105 sites in freshly excised breast biopsies from 25 patients, within 30 minutes of surgical excision. We collected DRS and fluorescence spectra at 10 wavelengths using a clinical instrument and optical fiber probes designed for clinical use. IFS spectra are extracted from the combined fluorescence and DRS spectra and analyzed using multivariate curve resolution with non-negativity constraints. DRS spectra are fit using diffusion theory. Spectroscopy results are compared to pathology diagnosis, and diagnostic algorithms are developed using fit parameters by logistic regression with leave-one-out cross validation. Results: The sensitivity, specificity and overall diagnostic accuracy of the IFS + DRS algorithm are 100%, 96% and 88%, respectively. All invasive breast cancers are correctly diagnosed by our technique. Conclusion: A combination of DRS and IFS yields superior promising results as compared to results achieved from individual methods for the spectroscopic diagnosis of breast cancer.

Background:

Techniques capable of evaluating human disease in a safe, minimally-invasive and reproducible manner are of critical importance for clinical disease diagnosis, risk assessment, therapeutic decision-making, and for evaluating the effects of therapy, in addition to basic investigations of disease pathogenesis and pathophysiology. Among the clinical methods available to diagnose tissue lesions, pathologic examination of cytology preparations, biopsies and surgical specimens is the present day gold standard. Pathologists have traditionally based their diagnoses primarily on tissue morphology. However, as the field of diagnostic pathology has evolved, assessment of tissue morphology has become more sophisticated, including such techniques as morphometry (or quantitative image analysis) [Caruntu 2003, Millot 2000, Wang 2004] and ploidy analysis [Baak 2004, Hall 2004, Kronqvist 2002]. Pathologic diagnosis has also begun to move from complete dependence on morphology to inclusion of a host of adjunct techniques that provide biochemical and molecular information as well. This is particularly true for the diagnosis of cancer, where routine diagnosis begins with morphology but usually also includes such molecular diagnostic techniques[Perez 2004] as immunohistochemistry [Lerwill 2004] and in situ hybridization [Hicks 2005] that identify specific molecular signatures. X-ray mammography is the current gold standard screening technique for early detection of small, non-palpable breast cancers. Mammography quantitatively targets the density of the changes in the breast tissue. Unfortunately, these changes do not uniquely correspond to the breast cancer, thus 70-90% of mammographically suspect lesions are found to be benign upon breast biopsy.

The desire to reduce the number of unnecessary breast biopsies, patient trauma, and time delay, has encouraged scientists to develop accurate, minimally invasive optical methods for early diagnosis of breast cancer [Alfano 1987, Frank 1995]. Our technique, which will be subsequently referred to as TMSDRS/IFS, combines diffuse reflectance and intrinsic fluorescence spectroscopies. The combination of DRS and IFS has several advantages over the individual modalities alone. First, fluorescence spectroscopy provides information about tissue metabolites and fluorophores in the tissue, such as NADH, collagen, tryptophan, elastin, and others [Georgakoudi 2002]. Second, TMSDRS/IFS uses DRS to overcome distortion of fluorescence signatures by the effects of tissue absorption and scattering, and extracts the IFS signature [Georgakoudi 2001]. Third, in addition to its value in extracting IFS, DRS provides critical information and a direct measurement of the tissue absorbers and scatterers themselves, such as hemoglobin and β-carotene [Gupta 1997, Majumder 1998]. The combination of techniques therefore, provides a wealth of information about tissue fluorophores, absorbers and scatterers, which creates a more complete biochemical, morphologic and metabolic tissue profile and lays the groundwork for more robust spectral models and diagnostic algorithms. Specific to cancer diagnosis, IFS and DRS provide information about key cellular metabolites such as NADH and oxy- and deoxy-hemoglobin. As cancer is characterized by rapid cellular proliferation reflected in increased cellular metabolism, TMSDRS/IFS is thus, a natural choice for the diagnosis of cancer.

Previously, different number of groups have investigated the use of DRS [Bigio 2000, Yang 1997, Zonios 1999] or fluorescence spectroscopy [Alfano 1987, Nair 2002] for the diagnosis of breast cancer. Moderate success has been achieved [state more about what exactly was done – i.e. do sensitivity and specificities range from 70-80% or something?], i.e. various groups using different optical technologies had reported the sensitivities and specificities of distinguishing malignant tissue from non-malignant tissue in the range of 70 to 90 percent. Palmer et al. [Palmer 2003] examined both fluorescence and diffuse reflectance spectroscopies. The results from this study were promising: multiexcitation fluorescence spectroscopy was successful in discriminating malignant and nonmalignant tissues, with a sensitivity and specificity of 70 and 92 percent, respectively. However, the sensitivity (30 percent) and specificity (78 percent) of diffuse reflectance spectroscopy alone was significantly lower. Combining fluorescence and diffuse reflectance spectra did not improve the classification accuracy of an algorithm based on fluorescence spectra. The difficulty in this study was the inability to determine the biological basis of the differences observed in the fluorescence spectra of malignant and nonmalignant tissues. Further, differences in the clinical protocol, probe geometry, and spectral range (300-600 nm) make it hard to compare this study with our current study.

Gupta and Majumder et al. [Gupta 1997, Majumder 1998] analyzed different data sets collected from the same set of breast tissues ex vivo and showed that the emission spectra at excitation wavelengths of 340 and 488 nm and excitation spectra at emission wavelengths 390 and 460 nm exhibit significant differences between normal, benign and malignant tissues. The fluorescence was attributed to reduced nicotinamide adenine dinucleotide (NADH) and collagen. Spectral differences observed in the fluorescence spectra of normal, benign and malignant breast tissues can also be attributed in part to non-fluorescent absorbers and scatters. According to Majumder et al., the fluorophores responsible for the 340, 390, 440, and 520 nm emission bands are amino acids (tryptophan), structural proteins (collagen and elastin), the co-enzyme (NADH), and flavins, respectively. It is known that the excitation spectra recorded from the breast tissue samples for 340, 390, and 460 nm emission consist of spectral bands with peaks around 290, 335 and 340 nm, which are characteristic excitation peaks for tryptophan, collagen and NADH, respectively. The larger intensities of the 340 nm band in the excitation spectra, corresponding to 460 nm emission, for cancerous tissues would suggest a larger concentration of NADH in cancerous tissues as compared to benign tumors and normal tissues.

Based on this information and the excellent past performance of our TMSDRS/IFS clinical instrument in studies of other cancerous tissues, we embarked on the current study of breast tissue [XXX].

Instrumentation:

A clinical instrument for DRS/IFS studies, the FastEEM, has been developed at the MIT Spectroscopy Laboratory. A schematic of the FastEEM is presented in Figure 1a. This instrument collects white light reflectance and fluorescence excitation-emission matrices (EEMs) within a fraction of a second. The FastEEM delivers a sequence of ten laser pulses (308 – 480 nm) and two white light pulses (300 – 800 nm) to the tissue via an optical fiber probe (Figure 1b). The probe is in the form of a flexible catheter, with an overall length of over 3 m and a diameter of approximately 1.2 mm. The same probe delivers and collects the white light reflectance and fluorescence. The light exiting the fiber probe enters the slit of a diffraction grating spectrometer where it is dispersed onto an intensified CCD detector. All ten laser-induced emission spectra and the two white light reflectance spectra are collected in approximately 0.3 s. Several of these acquisitions can be averaged together to increase the signal-to-noise ratio (SNR). Previously, we found that the acquisition of five measurements provides sufficient SNR in most tissues, making a typical acquisition time on the order of 1.5 s.

Diffuse reflectance spectroscopy (DRS) provides information about the morphology and biochemistry of the stromal tissue and epithelium. Incident white light undergoes many scattering and absorption events as it propagates through the tissue, and the emerging (“diffusely reflected”) light exhibits prominent spectral features caused by the interplay of scattering and absorption. DRS employs a mathematical model based on the diffusion approximation of light propagation in tissue to determine values of the absorption and reduced scattering coefficients, µa(λ) and µs’(λ), respectively.

The collected fluorescence and reflectance spectra can be used to extract the intrinsic fluorescence spectra (i.e., the fluorescence unaffected by tissue absorption and scattering) [XXX]. Intrinsic fluorescence spectroscopy (IFS) yields the relative contributions of endogenous tissue fluorophores (e.g. NADH and collagen).

Calibration was performed every day prior to data collection by collecting fluorescence spectra of a Rhodamine B dye to correct for time-dependant changes in the FastEEM instrument; and by obtaining reflectance spectra of Spectralon in order to correct for wavelength-dependant system response; water spectra to correct for background.

Clinical study:

The study was conducted at University Hospitals Cleveland in collaboration with Dr. Robert Shenk, a surgical oncologist and Medical Director of the Breast Center at University Hospitals of Cleveland. The study was performed on fresh surgical biopsies within 30 minutes of surgical resection, in the Frozen Section Room of the Mather Surgical Pavilion at University Hospitals of Cleveland. Most of the 30 minute time delay was due to inking and sectioning of the specimen performed as part of the routine pathology consultation performed on these specimens for intra-operative margin assessment. 225 IFS and diffuse reflectance spectra were obtained from a total of 105 fresh breast tissues from 25 patients. Specimens from patients with pre-operative chemotherapy or who underwent repeat excisional biopsy are excluded from the study. Once the spectra were acquired, the exact spot of probe placement was marked with colloidal ink for registration with histopathology. The breast specimens were then fixed and submitted for routine pathology examination, which is performed by an experienced breast pathologist blinded to the spectroscopy results. The histopathology diagnoses are: 32 normal, 55 fibrocystic change, 9 fibroadenoma and 9 invasive carcinoma (all infiltrating ductal carcinoma) [Volynskaya 2005].

Data Analysis:

Tissue is characterized by μa, μs’, and index of reflection, n, which for the soft biological tissue has typical value of 1.35 – 1.45. Diffuse reflectance spectroscopy (DRS) provides information about the morphology and biochemistry of the stromal tissue and epithelium. Incident white light undergoes many scattering and absorption events as it propagates through the tissue, and the emerging (“diffusely reflected”) light exhibits prominent spectral features caused by the interplay of scattering and absorption. DRS employs a mathematical model based on the diffusion approximation of light propagation in tissue to determine values of the absorption and reduced scattering coefficients, µa(λ) and µs’(λ), respectively.

Reflectance spectra, collected from excited tissue, are analyzed using the diffusion approximation to extract tissue morphological properties such as scattering, oxyhemoglobin concentration, and β-carotene concentration. For reduced scattering coefficient,diffusive scattering ((s',), wavelength dependence of the form A(-B (inverse power law) is used. It should be stated that oxygen saturation values for ex vivo breast tissue is typically higher than 90%, which is why for this ex vivo study we used only oxyhemoglobin and not β-hemoglobin or dioxyhemoglobin as absorbers. Two absorbers, oxyhemoglobin and (-carotene, are required to model the extracted absorption coefficient (a. The maximum absorption coefficient, μa, as determined from DRS, is highly correlated with concentration of oxyhemoglobin. Indeed, only one of these parameters can be used as a diagnostic. Therefore, DRS provided, among other parameters, the amplitude of the scattering coefficient, A, and the concentration of oxyhemoglobin.

The intrinsic fluorescence photon-migration model is used to correct the fluorescence spectrum for distortions introduced by tissue absorption and scattering. IFS spectra are extracted from the combined fluorescence and DRS and are analyzed using multivariate curve resolution (MCR) with non-negativity constraints, a standard chemometric method [Navea 2002], to extract the contributions of the biochemical tissue constituents NADH and collagen at each excitation wavelength.

Examples of DRS and IFS spectra are displayed in Figure 2. MCR calculates basis spectra by minimizing the fitting error of a given spectrum using an initial guess spectrum as the input. The resulting MCR-generated spectral components at 340 nm are shown in Figure 3a and Figure 3b, and are thought to represent NADH and collagen, respectively, because they are similar to their measured IFS spectra. The spectra are similar, but not identical, as both the lineshape and wavelength maximum of a fluorescence peak obtained from a solution of a pure component is known to be different than that obtained from the same component in a different chemical environment, such as tissue [Shafer-Peltier 2001]. It is expected that spectra of the fluorophores in the tissue are broader and red-shifted than spectra of the same fluorophores as pure components. In addition, when MCR is used to extract more than two basis spectra, the concentration of the third spectrum is found to be negligible compared to the first two. This suggests that only contributions of the first two basis spectra, identified as NADH and collagen, are significant in the breast tissue.

Figure 4 displays average IFS data from each of the pathologies encountered in this study. Differences are observed even without a detailed analysis. The full width at half-maximum (FWHM) is largest for IDC and then decreases for fibroadenoma, DCIS, FCC and is smallest for normal tissue. A large FWHM possibly indicates a higher concentration of NADH.

Three different excitation wavelengths (308, 340, and 360 nm) are analyzed in order to reveal different fluorophores that could each provide information useful to a diagnostic algorithm. Table 1 presents an overview of fluorophores that may be present in the breast tissue and fluoresce at particular excitation wavelengths. However, upon analysis it is discovered that some of these fluorophores are not present at high enough levels in our samples to be detected. These include tryptophan excited at 308 nm, elastin excited at 340 nm, and porphyrin excited at 360 nm [Shafer-Peltier 2001].

Representative spectra of DRS and IFS collected with different excitation wavelengths are illustrated in Figure 5.[can’t see axis labels – must make them large enough to be readable – one way to do this is to remove x-labels from all spectra except the bottom row and remove y-labels from all spectra except left column of DRA and left column of IFS.] Not all of the data collected are subsequently used for analysis. Specifically, DRS data with overall reflectance less than 1 percent are excluded because of the inability to use this information to process the fluorescence data to obtain the intrinsic fluorescence. Furthermore, though most of the fits are observed to be adequate, in cases of high oxyhemoglobin concentration, the region between 660 – 750 nm fits weakly. We believe this to be a result of the fitting procedure and is a topic that requires further investigation.

The DRS/IFS algorithm is developed using leave-one-out cross validation and logistic regression. The desired algorithm must be able to distinguish among the 4 major pathologies. From an examination of breast histopathology, it is known that normal breast tissue consists mostly of adipocytes (fat) while the progression of malignancy includes an increase in the amount of collagen. Therefore, we expect that normal tissue can be separated from the remaining pathologies by the relative presence of collagen and β-carotene, which is fat-soluble. Also from histopathology, fibroadenoma displays an increased cellular density. Because the parameter A is representative of the number of scatterers in the tissue, we expect fibroadenomas to have a relatively high A parameter. Furthermore, we expect the NADH contribution, which is representative of cellular metabolism, to be less than that from cancerous tissue. By maximizing the sensitivity and specificity of each stage of the algorithm we are able to identify the diagnostic parameters that can distinguish between pathologies in the breast tissue.

As expected, the maximum absorption coefficient, μa, as determined from DRS, is highly correlated with concentration of oxyhemoglobin. Therefore, only one of these parameters can be used as a diagnostic. The reduced scattering coefficient, μs’, is very different for different pathologies, with highest values for IDC specimens (Figure 6).

FinallyT, the diagnostically-relevant parameters from DRS are found to be β-carotene contribution, oxyhemoglobin hemoglobin, and the scattering A parameter. The diagnostically-relevant parameters from IFS are found to be the fit coefficients for NADH at 340 nm excitation and the fit coefficients for collagen at both 340 and 360 nm excitation wavelengths. However, careful examination of the collagen fit coefficients obtained with 340 and 360 nm excitation revealed that only the results from one wavelength are necessary. Because of the slight difference in wavelength between 340 nm and 360 nm, the penetration depth of the light is not sufficient to result in different sampling volumes. However, there is enough variation between the fit coefficients at both wavelengths that averaging them does not provide benefit. This finding is quite important as the number of diagnostic parameters should be minimized in order to prevent overfitting the data set. Therefore, the diagnostic parameters from IFS are reduced to NADH and collagen at 340 nm excitation, as NADH has maximum emission at this wavelength, as was stated previously.

The reduced scattering coefficient, μs’, is very different for different pathologies, with highest values for IDC specimens (Figure 6).

.

Figure 7 shows the mean and standard deviations of the extracted diagnostic parameters. [I thought you were going to normalize these values to the highest value in each so that you can actually see the bars. 1-3 don’t mean anything since they’re so small. And the y-axis doesn’t have a label anyway, so it should definitely be normalized.] In this plot, as expected, β-carotene is found to be high in normal breast tissue. This finding agrees well with histopathology in that normal tissue consists mostly of adipose tissue - fat cells that contain large amounts of lipid-soluble β-carotene. Iit can be seen that amount of oxyhemoglobin in cancerous specimens is higher than in the rest of the pathologies. Also, β-carotene is found to be high in normal breast tissue. This finding agrees well with histopathology in that normal tissue consists mostly of adipose tissue - fat cells that contain large amounts of lipid-soluble β-carotene.

The DRS/IFS diagnostic algorithm is developed using logistic regression and leave-one-out cross validation, and the analysis performed in sequential fashion. Scatter plots and decision lines for each step of the diagnostic algorithm are depicted in Figure 8. Normal tissue is identified using the collagen and (-carotene contributions extracted from intrinsic fluorescence at 340 nm excitation wavelength and diffuse reflectance, respectively (Figure 8a). [You should use symbols that can be easily distinguished with a black-and-white photocopier. The stars versus circles work well, or you could do open circles and closed squares. Since the algorithm is sequential, you can use the same two symbols for each plot]. The finding of low fit coefficients for collagen and (-carotene correlates with histopathology., as normal breast tissue consists largely of adipose tissue, mostly fat cells which contain large amounts of lipid-soluble (-carotene . After the normal tissue is excluded, fibroadenoma is discriminated from fibrocystic change and invasive breast cancer, using the DRS scattering parameter A and IFS NADH fit coefficient (Figure 8b). Fibrocystic change is distinguished from invasive breast cancer using the DRS oxyhemoglobin and IFS collagen fit coefficients at 340 nm (Figure 8c). In the latter plot, three samples (6 spectra) classified as FCC are misdiagnosed as cancer. Two out of those three data samples are previously misclassified as normals.

This diagnostic algorithm achieves the goal of an algorithm with contributions from both the cells (NADH) and the stroma (collagen). However, it is unclear why the fit coefficient for collagen and scattering parameter A should be lower for fibroadenoma than for invasive carcinoma and fibrocystic change, or the fit coefficients for oxyhemoglobin should be higher for an invasive breast cancer than for fibrocystic change. Currently, we are still working to understand the physics behind those findings. A comparison of the DRS/IFS spectral diagnoses and histopathology diagnoses is shown in Table 2. The total efficiency (correct prediction of each of the pathologies) is 87.6% (92/105). The sensitivity and specificity for the separation of cancerous and non-cancerous pathologies are 100% and 95.8%, respectively. All of the invasive carcinomas are diagnosed correctly by TMSDRS/IFS and only 4 normals or fibrocystic changes are misclassified as invasive carcinoma.

This pilot study is the first from our laboratory to use diffuse reflectance and intrinsic fluorescence spectroscopies to examine breast cancer ex vivo. This study clearly demonstrates the feasibility potential usage of DRS/IFSDRS/IFS as a clinical tool for breast cancer diagnosis.

Discussions and Conclusions:

We have presented the results from a pilot ex vivo study to determine the feasibility of using diffuse reflectance and intrinsic fluorescence to distinguish different pathologies of breast tissue. The study was performed on 25 patients undergoing mastectomy and lumpectomy at University Hospitals of Cleveland. Our diagnostic algorithm is based on physically meaningful parameters, which include the concentration of β-carotene, oxyhemoglobin, scattering A parameter, and the relative amounts of NADH and collagen fluorescence excited at 340 nm. A leave-one-out cross validation was performed in order to find the best possible diagnostic algorithm to distinguish among pathologies. The algorithm resulted in 100% and 96% sensitivity and specificity, respectively.

The concentration of collagen, as expected, is significantly lower in normal breast tissue as compared to the rest of the pathologies. Our results are also in agreement with the recent work of Palmer et. al [Palmer 2006] that reports the reduced scattering coefficient is higher for malignant tissue than for normal tissue. The concentration of oxyhemoglobin is highest for normal specimens, but also high for specimens classified as IDC as compared to FCC and fibroadenoma lesions. This fact provides an opportunity to distinguish IDC from FCC. It should be noted that oxygen saturation values for ex vivo breast tissue is typically higher than 90%, which is why we used only oxyhemoglobin and not β-hemoglobin or dioxyhemoglobin as absorbers.

This study is based on 105 specimens that contained only 9 malignant specimens, diagnosed as IDC. The limited number of cancerous specimens is in line with the positive identification rate of biopsied tumors. Our diagnostic algorithm should be validated by another ex vivo or in vivo independent clinical study.

This study was performed at the University Hospitals of Cleveland under the National Institute of Health funding. It was also approved by COUHES (Committee On the Use of Humans as Experimental Subjects) of Massachusetts Institute of Technology.

References:

Alfano RR, Tang G, Pradham A, Lam W, Choy DSJ and Opher E, "Fluorescence spectra from cancerous and normal human breast and lung tissues", IEEE Journal of Quantum Electronics, 23(1806-1811): (1987)

Baak JP and Janssen E, "DNA ploidy analysis in histopathology. Morphometry and DNA cytometry reproducibility conditions and clinical applications", Histopathology, 44(6): 603-14 (2004)

Bigio IJ, Bown SG, Briggs G, Kelley C, Lakhani S, Pickard D, Ripley PM, Rose IG and Saunders C, "Diagnosis of breast cancer using elastic-scattering spectroscopy: preliminary clinical results", J Biomed Opt, 5(2): 221-8 (2000)

Caruntu ID, "Microscopic morphometry--a modern approach", Rev Med Chir Soc Med Nat Iasi, 107(1): 9-18 (2003)

Frank CJ, McCreery RL and Redd DC, "Raman spectroscopy of normal and diseased human breast tissues", Anal Chem, 67(5): 777-83 (1995)

Georgakoudi I, Jacobson BC, Muller MG, Sheets EE, Badizadegan K, Carr-Locke DL, Crum CP, Boone CW, Dasari RR, Van Dam J and Feld MS, "NAD(P)H and collagen as in vivo quantitative fluorescent biomarkers of epithelial precancerous changes", Cancer Res, 62(3): 682-7 (2002)

Georgakoudi I, Jacobson BC, Van Dam J, Backman V, Wallace MB, Muller MG, Zhang Q, Badizadegan K, Sun D, Thomas GA, Perelman LT and Feld MS, "Fluorescence, reflectance, and light-scattering spectroscopy for evaluating dysplasia in patients with Barrett's esophagus", Gastroenterology, 120(7): 1620-9 (2001)

Gupta PK, Majumder SK and Uppal A, "Breast cancer diagnosis using N2 laser excited autofluorescence spectroscopy", Lasers Surg Med, 21(5): 417-22 (1997)

Hall PA, "DNA ploidy analysis in histopathology. DNA ploidy studies in pathology-a critical appraisal", Histopathology, 44(6): 614-20 (2004)

Hicks DG and Tubbs RR, "Assessment of the HER2 status in breast cancer by fluorescence in situ hybridization: a technical review with interpretive guidelines", Hum Pathol, 36(3): 250-61 (2005)

Kronqvist P, Kuopio T, Jalava P and Collan Y, "Morphometrical malignancy grading is a valuable prognostic factor in invasive ductal breast cancer", Br J Cancer, 87(11): 1275-80 (2002)

Lerwill MF, "Current practical applications of diagnostic immunohistochemistry in breast pathology", Am J Surg Pathol, 28(8): 1076-91 (2004)

Majumder S, Gupta P, Jain B and Uppal A, "UV excited autofluorescence spectroscopy of human breast tissues for discriminating cancerous tissue from benign tumor and normal tissue", Lasers Life Sci, 00(1-16 (1998)

Millot C and Dufer J, "Clinical applications of image cytometry to human tumor analysis", Histol Histopathol, 15(4): 1185-200 (2000)

Nair MS, Ghosh N, Raju NS and Pradhan A, "Determination of optical parameters of human breast tissue from spatially resolved fluorescence: a diffusion theory model", Appl Opt, 41(19): 4024-35 (2002)

Navea S, de Juan A and Tauler R, "Detection and resolution of intermediate species in protein folding processes using fluorescence and circular dichroism spectroscopies and multivariate curve resolution", Analytical Chemistry, 74(23): 6031-6039 (2002)

Palmer GM, Zhu C, Breslin TM, Xu F, Gilchrist KW and Ramanujam N, "Comparison of multiexcitation fluorescence and diffuse reflectance spectroscopy for the diagnosis of breast cancer (March 2003)", IEEE Trans Biomed Eng, 50(11): 1233-42 (2003)

Palmer GM, Zhu C, Breslin TM, Xu F, Gilchrist KW and Ramanujam N, "Monte Carlo-based inverse model for calculating tissue optical properties. Part II: Application to breast cancer diagnosis", Appl Opt, 45(5): 1072-8 (2006)

Perez EA, Pusztai L and Van de Vijver M, "Improving patient care through molecular diagnostics", Semin Oncol, 31(5 Suppl 10): 14-20 (2004)

Shafer-Peltier K,"Chemical Basis for Breast Cancer Diagnosis Using Raman Spectroscopy", 2001, PhD, Cambridge, MIT

Volynskaya Z,"Multi-modal spectroscopy of breast tissue", 2005, MS, Cambridge, MIT

Wang KK, "Detection and staging of esophageal cancers", Curr Opin Gastroenterol, 20(4): 381-5 (2004)

Yang Y, Katz A, Celmer EJ, Zurawska-Szczepaniak M and Alfano RR, "Fundamental differences of excitation spectrum between malignant and benign breast tissues", Photochem Photobiol, 66(4): 518-22 (1997)

Zonios G, Perelman L, Backman V, Manoharan R, Fitzmaurice M., Crowe J, Dasari RR and Feld M, "Diffuse Reflectance Spectroscopy of Human Adenomatous Colon Polyps In Vivo", Appl. Opt., 38(6628-6637 (1999)

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

or

Figure 8

Table 1

Table 2

-----------------------

[pic][pic]

Figure 7. The mean and the standard deviation of the extracted diagnostic parameters.

1 – β-caroteneA, 2 – Oxyhemoglobin, 3 – β-carotene“, 4 – NADH, 5 – OxyhemoglobinCollagen.

Normal-blue, FCC-green, fibroadenoma – pink, IDC – red.

[pic]Figure 7. The mean and standard deviations of the extracted diagnostic parameters. 1 – A, 2 – Oxyhemoglobin, 3 – β-carotene, 4 – NADH, 5 – Collagen.

[pic]

Table 2. Comparis湯漠⁦剄⽓䙉⁓湡⁤楨瑳灯瑡潨潬楧⁣楤条潮楳⁳on of DRS/IFS and histopathologic diagnosis for ex vivo study of fresh surgical breast biopsies. The TMSDRS/IFS diagnostic algorithm had an overall accuracy of 87.6% (92/105).

| |Excitation wavelength |

| |308 nm 340 nm 360 nm |

|Fluorophores | |

| |NADH |NADH |NADH |

| |Collagen |Collagen |Collagen |

| |Tryptophan |Elastin |FAD |

| | | |Porphyrin |

Table 1. Expected fluorophores for different excitation wavelengths.

[pic][pic]

Figure 5. Representative spectra of DRS and IFS of different excitation wavelengths for all pathologies.

c)

b)

a)

[pic]

Figure 2. a) Representative DRS spectrum and fit; b) Representative IFS spectra obtained with 340 nm excitation. The original spectrum acquired from breast tissue is in blue and the contributions of NADH (green) and collagen (black) were found via multivariate curve resolution (MCR).

[pic]

Figure 1. a) FastEEM clinical spectrophotometer. b) Schematic diagram of the distal tip of the optical fiber probe.

B

A

[pic]

Figure 3. Comparison of basis spectra vs. MCR components excited at 340 nm. Basis spectra (blue); MCR (red).

[pic]a) b)

[pic]

Figure 6. a) Average r

b) Reduced scattering coefficient extracted, μs’, from DRS for all 4 different pathologies.

[pic]

[pic]Figure 8. Discriminating of pathologies using parameters extracted by DRS and IFS.

[pic]

Figure 4. Normalized intrinsic fluorescence spectra of breast tissue at 340 nm excitation. Peak at 680 nm represents second order laser peak.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download