Effect of Optical Clearing Agents on the In Vivo Optical Properties of Squamous Epithelial Tissue

Transcription

1 Lasers in Surgery and Medicine 38: (2006) Effect of Optical Clearing Agents on the In Vivo Optical Properties of Squamous Epithelial Tissue Stacy R. Millon, MS, 1 Katherine M. Roldan-Perez, MS, 1 Kristin M. Riching, 1 Gregory M. Palmer, PhD, 2 and Nirmala Ramanujam, PhD 2 * 1 Department of Biomedical Engineering, University of Wisconsin, Madison, Wisconsin Department of Biomedical Engineering, Duke University, Durham, North Carolina Background and Objectives: Optical clearing agents (OCAs) have previously been shown to increase depth penetration within turbid tissue ex vivo. This paper quantifies tissue optical properties of the hamster cheek pouch model in order to provide a means to assess the effect of OCAs quantitatively in vivo. Study Design/Materials and Methods: Diffuse reflectance spectra were obtained from both cheeks of 12 hamsters before and after immersion in dimethyl sulfoxide (DMSO), glycerol or a phosphate buffer saline (PBS) control for 20 minutes. A Monte Carlo model was then utilized to derive the wavelength dependent reduced scattering and absorption coefficients. Results: DMSO caused a statistically significant decrease in the absorption and reduced scattering coefficients derived by the model. Glycerol caused a statistically significant increase in the wavelength dependent absorption coefficient, but no statistically significant changes in the reduced scattering coefficient. Conclusions: DMSO and glycerol act upon tissues differently as reflected by the tissue optical properties, implying that not all OCAs are equally effective in optically clearing tissues. Lasers Surg. Med. 38: , ß 2006 Wiley-Liss, Inc. Key words: absorption; cheek pouch; diffuse reflectance spectroscopy; dimethyl sulfoxide; epithelium; glycerol; hamster; scattering INTRODUCTION The ability to penetrate deep within tissue is advantageous for many light-based diagnostic and therapeutic techniques as it can provide a non-invasive method to interrogate tissue in vivo. However, the penetration depth into tissue is limited by the fact that tissue is a highly scattering medium in the visible and near-infrared wavelengths [1 15]. Optical clearing agents (OCAs) have been demonstrated to be a promising method by which the turbid tissue can be cleared [1 10]. The ability to see through turbid media will be beneficial to many light-based applications. OCAs could clear tissues for high-resolution optical diagnostic techniques such as two-photon microscopy [9,10] and optical coherence tomography or increase the depth of light dosimetry for laser tattoo removal [2,3]. For example, it has been shown that the depth of optical coherence tomography is enhanced with the application of glycerol to allow a human hair placed beneath in vitro hamster skin to be visible in a normal scan [1]. An understanding of how the absorption and scattering properties of tissues are altered by these agents could shed light on the mechanism by which these agents clear the tissue, as well as a comparison of the effect of different OCAs and doses on a specific tissue type. The clearing effect resulting from the application of OCAs is thought to be caused by local dehydration and refractive index matching between the scatterers and the surrounding medium [1]. The local dehydration occurs as a result of water exiting the cell more rapidly than the OCA enters after application. This causes an overall reduction in cellular cytoplasmic volume [13]. Refractive index matching is achieved by selecting hyperosmotic compounds that have similar refractive indices to primary tissue constituents, such as collagen (the refractive index for normally hydrated collagen is 1.47) [14]. Commonly used OCAs include glycerol, dimethyl sulfoxide (DMSO), glucose, trazograph (X-ray contrasting agent), and combinations of polymers such as polypropylene glycol and polyethylene glycol [1 6,8,11,12]. The refractive index of these common OCAs ranges from 1.43 to 1.47 [4]. The change in optical properties due to OCAs has been calculated from diffuse reflectance and transmission measurements of in vitro human and rat dorsal tissues [1,4]. It was found that the reduced scattering coefficient decreases dramatically as a result of optical clearing by glycerol over the 300 1,300 nm wavelength range [1]. The absorption coefficient was also shown to decrease [1]. A second study found that glycerol decreased the reduced scattering coefficient by approximately 300% after application, while DMSO resulted in approximately a 150% reduction in scattering [4]. There have been extensive qualitative in vivo investigations of the ability of the OCAs Contract grant sponsor: The National Science Foundation Graduate Research Fellowship; Contract grant sponsor: NIH NIBIB; Contract grant number: R01-EB *Correspondence to: Nirmala Ramanujam, PhD, Department of Biomedical Engineering, Duke University, 136 Hudson Hall, Box 90281, Campus, Durham, NC nimmi@duke.edu Accepted 27 September 2006 Published online 18 December 2006 in Wiley InterScience ( DOI /lsm ß 2006 Wiley-Liss, Inc.

2 EFFECT OF OPTICAL CLEARING AGENTS IN VIVO 921 to clear tissue [3 5,12]. However, there has not been a quantitative assessment of the effect of OCAs on the intrinsic absorption and scattering properties of in vivo tissues. The goal of this study was to quantify the effect of two commonly used OCAs, glycerol and DMSO on the intrinsic absorption and scattering properties of the in vivo epithelial layer of the inner hamster cheek pouch, which is a model of stratified squamous epithelial tissue. The hamster cheek model was chosen because of its similar architecture to many human epithelial tissue systems (e.g., skin, cervix, and the oral cavity). Glycerol and DMSO were chosen as the OCAs because they have been previously demonstrated to be effective OCAs in vitro and in vivo [2]. A concentration of 100% has been used in many previous in vivo and ex vivo studies and has been proven effective [1,2,5 9,12]. This same concentration was utilized for our in vivo study in order to be able to relate our results more directly to those reported previously in the literature. The two OCAs have a refractive index of 1.47 [2]. DMSO is often used to dissolve therapeutic and toxic chemicals and is known to enhance penetration of solutions through biological membranes [16]. Glycerol is commonly used as a cosmetic, pharmaceutical agent, and food ingredient. It occurs naturally in the body when stored body fat is oxidized and has the ability to bind water in the vascular system [11]. The OCAs were applied to the inner epithelial surface of the hamster cheek pouch. OCAs have previously had difficulty in penetrating the keratin layer in epithelial tissues such as skin. The hamster cheek pouch has been shown to have a relatively thin keratin layer which facilitates the penetration of the OCA to the epithelial layer more readily [17]. Diffuse reflectance spectra between 400 and 460 nm were obtained from both cheeks of 12 hamsters before and after treatment with DMSO, glycerol or a phosphate buffer saline (PBS) control for 20 minutes. A Monte Carlo model of diffuse reflectance [18] was utilized to derive the wavelength dependent absorption and reduced scattering coefficients. DMSO was found to statistically decrease the reduced scattering coefficient and absorption coefficient. Glycerol caused a statistically significant increase in the absorption coefficient. Glycerol decreased the reduced scattering coefficient, but this value was not statistically significant. DMSO and glycerol act upon tissue in different ways as indicated by the change in the absorption and scattering properties. This variability suggests that OCAs should be investigated separately for a given target tissue type. MATERIALS AND METHODS Treatment of the Animal Model With the Optical Clearing Agent Both cheek pouches in a total of twelve male Golden Syrian hamsters ( g) were examined in this study. Animal care and procedures were in accordance with the guidelines in the U.S. Department of Health and Human Services and NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee at the University of Wisconsin. Two OCAs were chosen, glycerol (13 M) and DMSO (14 M), because of their proven efficacy for optically clearing tissues [1,2,4,7 12]. Phosphate buffered saline (PBS) was used as the control agent. Of the 24 cheek pouches in 12 animals, 9 were painted with DMSO, 9 were painted with glycerol, and 6 were painted with PBS. All animals were anesthetized prior to any experimentation with an intra-peritoneal injection of a mixture of acepromazine/ketamine/xylazine at 2.5/ /5 10 mg/kg. The hair on the outside of the cheek was removed to circumvent any possible interference on the optical measurements (in the event the entire cheek pouch was transparent), although this was later found to be unnecessary because the tissues were optically thick even after application of the OCA (see Results). Next, the cheek pouch was inverted and stretched over a post, such that the inner part of the cheek pouch was exposed to the air. The cheek was wiped with saline and the fiber-optic probe was gently pressed on the tissue such that the tip was flush with the tissue surface, and the probe was fixed in position using a clamp assembly. A baseline diffuse reflectance measurement was then taken on a single site on the inner epithelium of the cheek pouch. Following the baseline measurement, the same part of the cheek pouch was inverted and pinned onto a cork covered with aluminum to prevent the cork from coloring and/or contaminating the cheek. Next, a test tube was half-filled with the agent to be used and petroleum jelly was applied to the rim of the test tube. The inner part of the cheek pouch was placed on the rim of the test tube which was then inverted by positioning the hamster on its back. The petroleum jelly formed a seal between the test tube and the tissue to prevent the agent from being ingested by the animal. The cheek pouch was exposed to the agent for 20 minutes. After 20 minutes of exposure to the agent, a second diffuse reflectance measurement was then taken using the same setup as described earlier. This procedure was repeated for the opposite cheek pouch. The 20-minute application time was optimized after running a series of experiments over different time points ranging from 5 to 45 minutes in which the effect of the OCAs was visually observed. It was noted that a minimum of 20 minutes was required to make the tissue translucent. However, increasing the time beyond 20 minutes increased the rigidity of the tissue thus making it hard to handle during experimental measurements. Spectroscopic Measurements Diffuse reflectance spectra were measured using a Skinskan spectrofluorometer (JY Horiba, Edison, NJ). The instrument consists of a 150 W xenon lamp, double grating excitation and emission monochromators having fixed bandpasses of 5 nm for both excitation and emission, and a photomultiplier tube (PMT) set at 950 V. The fiber optic probe used for illumination and collection consisted of a central collection core with a diameter of 1.52 mm surrounded by an illumination ring with an outer diameter of 2.18 mm. The illumination ring and collection core were

3 922 MILLON ET AL. made up of 31 individual fibers, each with a core/cladding diameter of 200/245 mm. The numerical apertures (NA) of the illumination and collection fibers were and 0.12, respectively. The diffuse reflectance spectra were measured between the wavelength range of nm with a 0.1 second integration time at each wavelength and at a wavelength increment of 5 nm. These measurements were completed in synchronous scan mode, where the excitation and emission gratings are moved simultaneously. The diffuse reflectance spectrum measured from each sample was background subtracted (diffuse reflectance measured in distilled water) and divided by the diffuse reflectance spectrum measured from a 99% reflectance Spectralon puck (SRS , Labsphere, Inc., North Sutton, NH) to correct for variations in throughput, and the wavelength response of the instrument and probe. Data Analysis A Monte Carlo based inverse model developed by our group [18] was employed to extract the absorption and scattering properties of the hamster cheek pouch from the measured diffuse reflectance spectra. In the forward model, a set of absorbers are presumed to be present in the medium, and the scatterer is assumed to be singlesized, spherically shaped and uniformly distributed. The wavelength dependent absorption coefficients of the medium are calculated from the concentration of each absorber and the corresponding wavelength dependent extinction coefficients. The wavelength dependent scattering coefficients and anisotropy factor are calculated from scatterer size, density and the refractive index of the scatterer and surrounding medium using Mie theory for spherical particles. The absorption and scattering coefficients are then input into a scalable Monte Carlo model of light transport to obtain a modeled diffuse reflectance spectrum. In the inverse model, inversion is achieved by adaptively fitting the modeled diffuse reflectance to the measured tissue diffuse reflectance using a Gauss-Newton non-linear least squares optimization algorithm. When the sum of squares error between the modeled and measured diffuse reflectance are minimized, the concentrations of absorber, the scatter size and density are thereby extracted. A detailed description of this model is provided elsewhere [18]. The free parameters in the inversion process were the scatterer size, the scatterer density and the concentrations of the absorbers. The scatterer size was varied over a range between 0.35 and 1.5 mm diameter [17,19]. The fixed parameters were the refractive index mismatch between the scatterers and the surrounding medium, the type of absorbers and the extinction coefficient of the absorbers. The refractive index mismatch was assumed to be 1.4 for the scatterers and 1.36 for the surrounding medium. The effect of different values of refractive index mismatch between the scatterer and surrounding medium (extending beyond that which would be expected in tissue) was evaluated in the previous study [18]. For a wide range of refractive index mismatches, (refractive index mismatch ratio range of ) there was a negligible degradation in the accuracy with which the optical properties were extracted using the model. Thus, keeping the refractive index mismatch as a fixed value within a range that would be expected in tissue was a reasonable assumption. The absorbers in this study were assumed to be oxygenated (oxy) hemoglobin and deoxygenated (deoxy) hemoglobin whose extinction coefficients (for the hamster) are available for the nm wavelength range [20]. Thus, the wavelength range used in the analysis was dictated by the available extinction coefficients of hamster hemoglobin. A comparison between hamster hemoglobin and human hemoglobin extinction coefficient spectra was made to determine if it would be reasonable to use human hemoglobin extinction coefficients which are available over a much wider wavelength range. The human and hamster hemoglobin extinction coefficient values in the literature are distinctly different in magnitude. Human hemoglobin extinction coefficients are cm/m for deoxy hemoglobin and cm/m for oxy hemoglobin over nm. Hamster hemoglobin extinction coefficients are cm/m for deoxy hemoglobin and cm/m for oxy hemoglobin over nm. In addition, the line shape of human and hamster hemoglobin extinction coefficients are different. Extinction coefficients can be reported per hemoglobin, which is the common nomenclature or per heme molecule. This could potentially be a reason for the differences in magnitude between human and hamster hemoglobin extinction coefficients. However, the reference from which the extinction coefficients of hamster hemoglobin was retrieved does not indicate if the extinction coefficients are per hemoglobin or per heme molecule [20]. The accuracy of the model used has been previously verified using phantom studies [18]. However these experiments were conducted over a larger wavelength range than that used in the current study. Therefore, the effect of using the more limited wavelength range ( nm) was evaluated using the same set of phantom experiments described in a previous publication [18]. Briefly, the phantoms consisted of polystyrene spheres ( , Polysciences, Inc., Warrington, PA) and human hemoglobin (H0267, Sigma Co., St. Louis, MO) dissolved in water. Five phantoms were prepared, having reduced scattering coefficients ranging from 10.9 to 16.4 cm 1 and absorption coefficients ranging from 0 to 17.5 cm 1. The optical properties were extracted from each of these phantoms using the model described above, and compared to known values to determine the error in extracting the optical properties. The model was used to extract the optical properties from the diffuse reflectance spectrum of each phantom using every other phantom as the reference measurement (i.e., all possible combinations of targetreference phantoms were considered). The root mean square (RMS) error for extracting the optical properties of the phantoms, averaged across wavelengths and phantoms was then calculated to evaluate the accuracy of this model. It was found that the errors for the limited wavelength

4 EFFECT OF OPTICAL CLEARING AGENTS IN VIVO 923 range were comparable to those of the full wavelength range (see Table 1). The Monte Carlo model assumes that the medium is semi-infinite. A series of experiments were conducted to ensure that the hamster cheek pouch, which has been measured and found to be approximately 3 mm thick, could be approximated as being semi-infinite over the nm range for the particular probe geometry used. Diffuse reflectance measurements were made from the cheek pouch before and after application of the agent for two different cases. In the first case, aluminum was placed beneath the cheek pouch. In the second case, black felt was placed beneath the cheek pouch. It is expected that aluminum will reflect any transmitted light, while the felt will absorb it. The percent difference between the integrated diffuse reflectance over the nm wavelength range measured with a felt base and aluminum base was at most 7% when either OCA was used (n ¼ 2 hamster cheek pouches). Therefore, the tissue was assumed semi-infinite for the purposes of Monte Carlo modeling. Prior to fitting to the inverse Monte Carlo model, each tissue diffuse reflectance spectrum was divided point by point by the diffuse reflectance spectrum of a reference tissue phantom with known optical properties (absorption coefficient of cm 1 and reduced scattering coefficient of cm 1 over wavelength range of nm). Each modeled diffuse reflectance spectrum was calibrated in a similar manner, that is, by dividing the modeled diffuse reflectance spectrum by that of a modeled reference tissue phantom with the same pre-defined optical properties to correct for any differences in the magnitude of the modeled and measured diffuse reflectance spectra. This calibration also accounts for the wavelength dependence of the instrument and probe, which in this case is redundant since the Spectralon puck achieves the same purpose. This allowed direct comparison of the modeled and measured diffuse reflectance spectrum during the inversion process. The outputs from the inverse Monte Carlo model were the wavelength dependent absorption and reduced scattering coefficients, the concentration of oxy and deoxy hemoglobin, and the scatterer size and density. The parameters used for further analysis in this study were the intermediate results of the fit, that is, the absorption and reduced scattering coefficients. The reduced scattering coefficient was used in place of the scatterer size and density because different values of scatterer diameter and TABLE 1. The Accuracy With Which the Inverse Monte Carlo Model Extracts the Phantom Absorption (l a ) and Reduced Scattering Coefficients (l 0 s ) From Diffuse Reflectance Spectra Over the Wavelength Range Used in This Study ( nm) and the Wider Wavelength Range ( nm) Previously Used to Validate This Model Wavelength range (nm) Mean RMS error in 0 s (%) Mean RMS error in a (%) density can yield similar values for the scattering coefficient using Mie theory. Thus, the reduced scattering coefficient (which is what directly affects the diffuse reflectance spectra) was used. The concentrations of oxy and deoxy hemoglobin are not reported in this publication due to the uncertainty as to whether the wavelength dependent extinction coefficients of hemoglobin are reported per hemoglobin or per heme molecule. A paired Wilcoxon sign-rank test was employed to determine if there are statistically significant differences in the absorption and scattering properties of the pre-treated and posttreated cheeks within the same cheek pouch in all animals. RESULTS Table 1 shows the accuracy with which the inverse Monte Carlo model extracts the phantom absorption ( a ) and reduced scattering coefficients ( 0 s ) from diffuse reflectance spectra over the wavelength range used in this study ( nm) and the wider wavelength range ( nm) previously used to validate this model [18]. As was previously reported, the percent RMS error of the absorption and reduced scattering coefficient over the tested wavelength range were determined and then averaged over all target-reference phantom combinations to produce the average percent RMS error. It was found that the errors for the limited wavelength range were comparable to those of the full wavelength range. Figure 1 shows the measured diffuse reflectance spectra (calibrated to that of a reference tissue phantom) and the best fit to the diffuse reflectance spectra (using the Monte Carlo model) before and after application of (A) PBS, (B) DMSO, and (C) glycerol. It can be seen that there is minimal deviation between the measured and fitted spectra, indicating that the model is able to effectively describe these measured diffuse reflectance spectra measured from the control and treated cheek pouches (for all three agents). The quality of these fits is representative of remaining tissue spectra. The fits to the diffuse reflectance spectrum measured from a glycerol treated cheek had the smallest residual error. From the inverse Monte Carlo model the following parameters were calculated: the reduced scattering and the absorption coefficients over the wavelength range between 400 and 460 nm. Table 2 contains the mean and standard deviation of the reduced scattering coefficient ( 0 s ) at every 20 nm between 400 and 460 nm of all cheek pouches before and after the application of each agent. The fifth column of Table 2 shows that the pre-glycerol scattering is uniformly greater than that of the pre-pbs (column 1) and pre-dmso (column 3) scattering. It should be noted that an outlier exists in this data (mean reduced scattering coefficient value of cm 1 over the wavelengths, 400, 420, 440, and 460 nm). These values were discarded prior to calculating the mean and standard deviation in the seventh and eighth columns of Table 2 (i.e., the mean and standard deviations do not include data from the outlier). It was found that removing the outlier data from the statistical analysis did not affect the conclusions or statistical significance of any of

5 924 MILLON ET AL. treatment with each agent. The percent change for each treated cheek was calculated relative to that of its corresponding untreated cheek within an animal (paired comparison) and the percent change was averaged across all animals treated with the same agent. These results indicate that DMSO causes a statistically significant decrease in the reduced scattering coefficient (P<0.01) at all wavelengths, while glycerol does not cause a significant change at any of the wavelengths. Table 3 shows the absorption coefficient ( a ) at every 20 nm between 400 and 460 nm of all cheek pouches before and after the application of each agent. Figure 3 shows the percent change of the absorption coefficient ( a ) at each wavelength upon treatment with each agent. The percent change for each treated cheek was calculated relative to that of its corresponding untreated cheek within an animal (paired comparison) and the percent change was averaged across all animals treated with the same agent. Statistically significant differences were observed in the absorption coefficients of both DMSO and glycerol (P<0.05). DMSO causes a statistically significant decrease, while glycerol causes a significant increase in the absorption coefficient at all wavelengths. Fig. 1. Measured diffuse reflectance spectra (calibrated to that of a reference tissue phantom) and the fit to the diffuse reflectance spectra (using the Monte Carlo model) before and after application of (A) PBS, (B) DMSO, and (C) glycerol. the results. Thus, the outlier data remains as part of the succeeding tables and figures. Figure 2 shows the percent change of the reduced scattering coefficient ( 0 s ) at each wavelength upon DISCUSSION Our study has found that DMSO has a more dramatic effect on the absorption and scattering properties of tissues as compared to glycerol, and thus may prove to be a more effective OCA. DMSO significantly decreases the reduced scattering coefficient and the absorption coefficient. By reducing both the scattering and absorption of the tissue it clears the turbid tissue. Treating a tissue with glycerol decreases the reduced scattering coefficient, but this was not found to be statistically significant for the nm wavelength range. However, it does significantly increase the absorption coefficient. The increase in absorption clearly indicates that glycerol is not as effective as DMSO in optically clearing tissues. These findings agree with a previously published study which reported that glycerol is not as effective of an OCA as DMSO [2]. The application of glycerol increased the absorption coefficient, thereby reflecting increased hemoglobin absorption (primary absorber in soft tissues). Vargas et al. [7] showed that glycerol slows or ceases the flow of blood within vessels possibly causing an inflammatory response. Galanzha et al. [15] found a significant decrease in blood flow and increase in the diameter of micro vessels within the rat mesentery with the application of glycerol. Higher blood content is consistent with the increased absorption coefficient observed with the application of glycerol in our study. DMSO decreased the absorption coefficient thereby reflecting decreased hemoglobin absorption. DMSO is used as an anti-inflammatory and anti-coagulant agent [16] and has the effect of restricting blood flow to the tissue. This may explain the reduction in absorption, and thus, total hemoglobin absorption in the DMSO treated tissue. These findings illustrate the importance of in vivo testing of OCAs. The physiological response of the tissue to each OCA

6 TABLE 2. Mean and Standard Deviation of the Reduced Scattering Coefficient ( 0 s ) at Every 20 nm Between 400 and 460 nm of All Cheek Pouches Before and After the Application of Each Agent Wavelength (nm) EFFECT OF OPTICAL CLEARING AGENTS IN VIVO 925 Scattering coefficient versus wavelength 0 s values (cm 1 ) PBS DMSO Glycerol Glycerol* Before After Before After Before After Before After The asterisk denotes the mean and standard deviation after the removal of the one outlier from the cheek pouches treated with glycerol. differs greatly and can have a significant effect on the optical properties, particularly as related to the blood content of the tissue, which would not be seen with just in vitro testing. DMSO showed a statistically significant decrease in the reduced scattering coefficient. This is consistent with previously published ex vivo studies and is consistent with the expected effect of OCAs [1,4]. Glycerol did not cause a statistically significant decrease in the reduced scattering coefficient. This result is inconsistent with the previously published ex vivo studies [1,4], and the cause of this discrepancy is not clear at this time. It is possible that the effectiveness of glycerol as an OCA may be dependent on the concentration used. Fig. 2. Percent change of the reduced scattering coefficient ( 0 s ) at each wavelength upon treatment with each agent. The percent change for each treated cheek was calculated relative to that of its corresponding untreated cheek within an animal (paired comparison) and the percent change was averaged across all cheek pouches of all animals treated with the same agent. An asterisk indicates that statistically significant differences in that variable were observed between the pre- and post-treated cheek pouch.

7 926 MILLON ET AL. TABLE 3. Mean and Standard Deviation of the Absorption Coefficient (l a ) at Every 20 nm Between 400 and 460 nm of All Cheek Pouches Before and After the Application of Each Agent Absorption coefficient versus wavelength a values (cm 1 ) Wavelength (nm) Before PBS After PBS Before DMSO After DMSO Before glycerol After glycerol Our study shows that DMSO is a much more effective agent than glycerol for optical clearing of tissues. However, DMSO is not necessarily a benign agent, though it can be used in some cases as a transport drug to allow the passage of a second functional drug to penetrate membranes otherwise impermeable, and it may also prove, in lower concentrations, to be a safe and effective OCA [16]. Glycerol is a more benign agent, but may not adequately clear the tissue. In addition, concentrations of 100% are not ideal, and thus, in future studies lower doses of both OCAs will need to be tested to assess their effect on in vivo tissue optical properties. It may well be that the optical properties and hence, the optical clearing potential of these agents will be strongly influenced by the concentrations at which they are used. This study established that it is feasible to provide a quantitative assessment of the effect of OCAs on tissue optical properties in vivo. The ability to monitor the scattering and absorption properties of tissue upon treatment with these agents could help assess the effectiveness of the agent for optically clearing tissues. This study showed that the two agents, glycerol and DMSO have distinctly different effects on the absorption and scattering properties of hamster epithelial tissues. Future studies should focus on individual agents or combinations of agents and examine quantitatively, their effect on the optical Fig. 3. Percent change within a cheek pouch (averaged over all cheek pouches in all animals treated with the same agent) of the absorption coefficient ( a ) at each wavelength upon treatment with each agent. The percent change for each treated cheek was calculated relative to that of its corresponding untreated cheek within an animal (paired comparison) and the percent change was averaged across all cheek pouches of all animals treated with the same agent. An asterisk indicates that statistically significant differences in that variable were observed between the pre- and post-treated cheek pouch.

8 EFFECT OF OPTICAL CLEARING AGENTS IN VIVO 927 properties of tissues as well as tissue morphology (via microscopy) in vivo in order to fully assess the effectiveness of the agent. OCAs have the potential to be very helpful for making tissues less turbid, but using them prior to fully understanding their effects on the tissue may have unintended consequences. REFERENCES 1. Vargas G, Chan EK, Barton JK, Rylander HG III, Welch AJ. Use of an agent to reduce scattering in skin. Lasers Surg Med 1999;24(2): Vargas G, Chan KF, Thomsen SL, Welch AJ. Use of osmotically active agents to alter optical properties of tissue: Effects on the detected fluorescence signal measured through skin. Lasers Surg Med 2001;29(3): Khan MH, Chess S, Choi B, Kelly KM, Nelson JS. Can topically applied optical clearing agents increase the epidermal damage threshold and enhance therapeutic efficacy? Lasers Surg Med 2004;35(2): Choi B, Tsu L, Chen E, Ishak TS, Iskandar SM, Chess S, Nelson JS. Determination of chemical agent optical clearing potential using in vitro human skin. Lasers Surg Med 2005;36(2): Khan MH, Choi B, Chess S, K MK, McCullough J, Nelson JS. Optical clearing of in vivo human skin: Implications for lightbased diagnostic imaging and therapeutics. Lasers Surg Med 2004;34(2): Tuchin VV, Maksimova IL, Zimnyakov DA, Kon IL, Mavlutov AH, Mishin A. Light propagation in tissues with controlled optical properties. J Biomed Opt 1997;2(4): Vargas G, Readinger A, Dozier SS, Welch AJ. Morphological changes in blood vessels produced by hyperosmotic agents and measured by optical coherence tomography. Photochem Photobiol 2003;77(5): Xu X, Wang RK. Synergistic effect of hyperosmotic agents of dimethyl sulfoxide and glycerol on optical clearing of gastric tissue studied with near infrared spectroscopy. Phys Med Biol 2004;49(3): Yeh AT, Choi B, Nelson JS, Tromberg BJ. Reversible dissociation of collagen in tissues. J Invest Dermatol 2003;121(6): Cicchi R, Pavone F, Massi D, Sampson D. Contrast and depth enhancement in two-photon microscopy of human skin ex vivo by use of optical clearing agents. Optics Express 2005; 13(7): Wang RK, Xu X, Tuchin VV, Elder JB. Concurrent enhancement of imaging depth and contrast for optical coherence tomography by hyperosmotic agents. J Opt Soc Am 2001; 18(7): Tuchin VV. Optical clearing of tissues and blood using the immersion method. J Phys D: Appl Phys 2005;38: Tuchin VV. Optical clearing of tissues and blood. Bellingham, WA: SPIE Press; p. 14. Tuchin VV. tissue Optics: Light scattering methods and instruments for medical diagnosis. Bellingham, WA: SPIE Press; Galanzha EI, Tuchin VV, Solovieva AV, Stepanova TV, Luo Q, Cheng H. Skin backreflectance and microvascular system functioning at the action of osmotic agents. J Phys D: Appl Phys 2003;36: Brayton CF. Dimethyl sulfoxide (DMSO): A review. Cornell Vet 1986;76(1): Skala MC, Palmer GM, Zhu C, Liu Q, Vrotsos KM, Marshek- Stone CL, Gendron-Fitzpatrick A, Ramanujam N. Investigation of fiber-optic probe designs for optical spectroscopic diagnosis of epithelial pre-cancers. Lasers Surg Med 2004; 34(1): Palmer GM, Ramanujam N. A Monte Carlo based inverse model for calculating tissue optical properties, part I: Theory and validation on synthetic phantoms. Appl Opt 2006;45(5): Palmer GM, Ramanujam N, Zhu C, Breslin TM, Gilchrist KW. A Monte Carlo based inverse model for calculating tissue optical properties, part II: Application to breast cancer diagnosis. Appl Opt 2006;45(5): Ellsworth ML, Pittman RN, Ellis CG. Measurement of hemoglobin oxygen saturation in capillaries. Am J Physiol 1987;252(5 Pt 2):H1031 H1040.