2.1.

Raman System

An overview of the Raman system employed in this work is depicted in Fig. 1. The excitation source was a titanium-sapphire (Ti-saph) laser operating at 785 nm pumped by a Nd:YVO₄ solid-state laser (V5) with intracavity doubling operating at 532 nm. The output power of the Ti-Saph at the sample surface was 100 mW. After the incident light interacts with the sample, the remitted light from the surface of the sample enters the first portion of the collection optics, consisting of two autofocusing lenses (L) (Nikon 85 and 50 mm) and a notch filter (NF). The light is then focused onto a fiber optic bundle consisting of 43 fibers (200 μm, NA of 0.28, CeramOptec Industries, Incorporated, East Longmeadow, Massachusetts). The bundle is connected to an axial transmissive imaging spectrograph (H) (Holospec f/1.8i, Kaiser Optical Systems Incorporated, Ann Arbor, Michigan) with the fibers aligned along the entrance slit of the spectrograph. The grating had a spectral resolution of 12 cm⁻¹ when the 200-μm slit was used with 785-nm excitation. Light detection is made with a front illuminated, liquid nitrogen cooled charge-coupled device (CCD) (Photometrics Limited, Tucson, Arizona) coupled to the exit port of the spectrograph. All Raman spectra were collected with an integration time of 15 s, repeated for 15 cycles, for a total collection time of 225 s per spectrum, giving an acceptable signal-to-noise ratio. Two spectra were acquired for each sample, resulting in a dataset for each sample consisting of 60 spectra. The responsivity of the system was corrected according to the procedure outlined by Shim and Wilson,¹⁸ and the fluorescence background was subtracted using the adaptive minmax method recently developed by Cao and adapted appropriately for use with the collected raw data.¹⁹

Fig. 1

Schematic of the Raman system including autofocusing lenses (L), notch filter (NF), axial transmissive imaging spectrograph (H), and liquid nitrogen cooled charge-coupled device (CCD).

2.2.

Tissue Equivalent Albumen Phantoms

Albumen-based phantoms that coagulate on heating werefabricated based on a recipe published by Iizuka, Sherar, and Vitkin.²⁰ Albumen protein is comprised mainly of water (88.1%), globular proteins (10.2%), and lipids (0.05%), and is commercially available as dehydrated egg white powder. The phantom used consisted of chicken egg albumen (Crude, grade 2, Sigma, Saint Louis, Missouri), bacteriological agar (Oxoid, Hampshire, England), and Naphthol Green dye (Sigma, Saint Louis, Missouri). The solid phantoms were constructed by combining the agar-dye solution, heated to 85°C, and the albumen stock solution, heated to 40°C. This mixture was then poured into cylindrical molds with a diameter of 6 cm and height of 3 cm, and cooled to room temperature. Thermal coagulation was performed by submerging each phantom in a water bath maintained at 75°C until the target temperature was reached, as reported by a microthermocouple placed at the center of the phantom. The phantom was then immediately removed, placed into a cold water bath at approximately 22°C, and allowed to cool. Performing the coagulation in this manner mimics thermal therapy treatments, in that once a defined target temperature is reached, heating ceases and the tissue is allowed to cool. Each phantom was heated to a target temperature of 45, 55, 65, or 75°C. Each phantom was then cut in half and Raman spectra were acquired at the center of each sample, identified by the indentation in the phantom made by the microthermocouple. This ensured that Raman spectra were collected at the exact region where the temperature was recorded during heating.

2.3.

Laser Photocoagulation of Bovine Muscle

Bovine muscle was selected for laser photocoagulation based on previous experimental results and data available in the literature on the Raman spectra for muscle.^{3, 21} Photocoagulation was performed using an 805 nm diode laser (Diomed Incorporated, Andover, Massachusetts) coupled to a 400-μm core optical fiber with a 20-mm cylindrical diffusing tip. The laser fiber was placed at the center of a tissue slab. Microthermocouples were placed at 2, 5, and 10 mm from the center of the diffusing tip. A second tissue slab was placed on top, followed by a weight to ensure good tissue coupling. The input laser power was set to 7 W and the irradiation was terminated when a temperature of 90°C was recorded by the thermocouple positioned 2 mm from the source fiber. The tissue slab was then allowed to cool to room temperature (∼22°C).

3.1.

Tissue Equivalent Albumen Phantoms

Raman spectra, normalized to the band at 1000 cm⁻¹ for each phantom experiment, are shown in Fig. 2. When normalizing Raman spectra, it is essential to choose a band that is not sensitive to the conformational changes to ensure that any observed band shifts are identifiable. The 1000 cm⁻¹ band does not change in shape or position with heating, suggesting that it is not sensitive to any conformational changes. Raman peak intensity increases with final temperature achieved, most notably in the major bands representing albumen protein, such as phenylalanine C=C symmetric stretch (1002 cm⁻¹), CH₂,CH₃ deformation (1463 cm⁻¹), and the amide-1 β-sheet (1663 cm⁻¹).^{22, 23, 24, 25} Several features, such as the phenylalanine C=C symmetric stretch (1002 cm⁻¹) and CH₂/CH₃ deformation (1463 cm⁻¹), remain at the same frequency regardless of final temperature achieved. However, there is a shift in the amide-1 β-sheet from 1658 cm⁻¹ in the native state to 1663 cm⁻¹ in the 75°C coagulated state, which may be a result of changes in vibrational modes resulting from coagulation. This shift is in agreement with Ngarize, Adams, and Howell, who observed a 3 cm⁻¹ shift in the amide-1 band in egg albumen heated to 90°C.²³ The absence of shifts in other major bands (phenylalanine C=C at ∼1002 cm⁻¹ and CH₂/CH₃ deformation at ∼1463 cm⁻¹) may be indicative of the more stable secondary protein structure being less susceptible to the effects of coagulation. The denaturation temperature of two major components of albumen (ovalbumin and conalbumin) have been determined to be 84.5 and 61.0°C, respectively.²⁶ The shift in the amide-1 band may be an indication of the secondary structure of conalbumin being affected by heating to 65 and 75°C, in addition to tertiary structure alterations due to thermal coagulation. These results suggest that the increased signal is primarily due to increased optical scattering associated with protein coagulation. It has been reported that the scattering coefficient (μ_s) of albumen-based phantoms can increase by a factor of ∼17 after 40 min of heating in a 70°C water bath.²⁷ These results are consistent with other reports of coagulation-induced increases in optical scattering and diffuse reflectance in tissues.^{28, 29} It is believed that these increases in scattering result from denaturation and hyalinization (a glassy appearance) of collagen and other proteins. The coagulation-induced increase in scattering can increase the photon flux from the sample surface, similar to the increase in diffuse reflectance that occurs with increased scattering.²⁹

Fig. 2

Raman spectra of albumen-based phantoms at room temperature, and heated to temperatures of 45, 55, 65, and 75°C. Spectra are individually normalized to the intensity at 1000 cm⁻¹.

3.2.

Laser Photocoagulation of Bovine Muscle

A thermal lesion created in bovine muscle is shown in Fig. 3. The locations of the laser fiber and three thermocouples are clearly visible by indentations in the tissue. Fig. 4 shows the corresponding Raman spectra at 2, 5, and 10 mm, where temperatures reached 90.0, 57.0, and 46.6°C, respectively. The relative intensities of all spectra have been normalized to the band at 1000 cm⁻¹. Several Raman bands are easily identifiable, specifically those near 935 cm⁻¹ [ν(CC) skeletal α-helix], 952 cm⁻¹ (CH₂ bending), 1000 cm⁻¹ [Phenylalanine ν(C=C)], 1074 cm⁻¹ and 1123 cm⁻¹ [ν(CN)], 1317 cm⁻¹ [ν(Cα-H)], 1334 cm⁻¹ (tryptophan), 1448 cm⁻¹ [δ(CH₂/CH₃)], and 1655 cm⁻¹ (amide-1 α-helix).^{3, 22, 24, 25} The Raman spectra acquired at 2 mm from the laser fiber (90°C) show a number of notable band shifts. Specifically, there are band shifts in the [ν(CN)] group from 1121 to 1123 cm⁻¹, the [ν(Cα-H)] group from 1312 to 1317 cm⁻¹, and the [δ(CH₂/CH₃)] group from 1446 to 1448 cm⁻¹. The largest shift was observed in the amide-1 α-helix, which shifted from 1648 cm⁻¹ in the native state to 1655 cm⁻¹ in the 90°C coagulated state, similar to the results observed in the albumen phantoms. The bands at 952 and 1000 cm⁻¹ do not undergo a Raman shift at temperatures up to 90°C, which may indicate that the secondary and tertiary structures of these particular components of the collagen protein are stable and less susceptible to effects of heating. Alternatively, the absence of shifts in these bands may indicate only that they arise from vibrations that are not interested in the structural variations by heating. This result is consistent with Dong, who noted that the Raman band near 1000 cm⁻¹ [phenylalanine ν(C=C)] did not undergo a shift with heating to 80°C.²⁴

Fig. 3

Thermal lesion created in bovine muscle samples heated with an 805-nm laser operating at 7 W. Indentations by the laser fiber and three thermocouples are indicated.

Fig. 4

Raman spectra of bovine muscle in the native state and heated by laser photocoagulation to maximum temperatures of 90.0, 57.0, and 46.6°C, measured at 2, 5, and 10 mm from the laser fiber, respectively. Spectra are individually normalized to the intensity at 1000 cm⁻¹.

Unlike Raman measurements on the phantoms, the relative intensities of the Raman bands in muscle varied with temperature, which affected the overall spectral shape. A change is observed in the slope of the shoulder leading up to the band at (Cα-H) 1317 cm⁻¹. In the native and 46.6°C states, this shoulder gradually rises toward the peak; however, in the 57.0°C state, the shoulder becomes steeper, and in the 90.0°C spectrum, the shoulder rises sharply from ∼1230 to ∼1270 cm⁻¹ and then plateaus. A second sharp rise can be seen at ∼1290 cm⁻¹ toward the peak at 1313 cm⁻¹. Another point of note is the significant reduction in relative intensity of the tryptophan band at 1334 cm⁻¹ on heating to 90°C. As the neighboring bands increase in intensity, this band nearly disappears from the Raman spectrum. This intensity decrease may be explained by the degradation of tryptophan's large indole side chain due to heating. Protein-bound tryptophan residues degrade on heating to temperatures over 100°C.^{30, 31} Further, ∼46 to 55% of tryptophan residues in a lactose-tryptophan mixture were lost with heating at 155°C for 2 min.³² Therefore, heating the bovine muscle sample to 90°C may cause a reduction in the protein-bound tryptophan residues thus reducing the number of Raman interactions that can cause the tryptophan band at 1334 cm⁻¹ to be reduced in intensity. Reduction in the intensity of the amide-1 α-helix at 1655 cm⁻¹ may be due to dehydration of this component as water is removed from the protein via heating, causing a reduction in the number of Raman interactions. Further, significant changes in the shoulders leading up to both bands {1317 cm⁻¹ [ν(Cα-H)] and 1655 cm⁻¹ (amide-1 α-helix} are observed. Disruption of secondary structures is typically detected through changes of the band shape, which in turn leads to a frequency shift of the vibrational mode. It may be this phenomenon that results in the reduction in intensity on heating to 57.0°C. However, on heating to 90.0°C, the intensity in both bands increases. This increase is interesting and may represent a transition in the tissue response to heating at higher temperatures. For example, tissue browning is observed near the 2-mm thermocouple (Fig. 3). It is also reported that the biological effects of heating on tissues begins as enzyme inactivation and mitochondrial injury and transitions at approximately 60°C to protein denaturation, membrane rupture, pyknosis, and hyperchromasia.³³ Moreover, such a transition from decreasing intensity to increasing intensity may prove to be a relevant metric for thermal therapy monitoring.