1.

Introduction

There has been an increasing incidence of carcinomas of the gastroesophageal junction in the Western industrialized countries in recent years. Radical esophageal resection is the only potentially curative therapeutic option currently available.^{1, 2} Despite constant progress in surgical management, radical resection is still associated with a high morbidity and mortality rate.³ An alternative therapeutic strategy is photodynamic therapy (PDT),^{4, 5, 6, 7, 8, 9} a nonthermal technique in which the interaction between a photosensitizer and light of a specific wavelength leads to phototoxic cell destruction in tissue.^{10, 11} Besides the concentration of the photosensitizer, the diffusion of light in tissue is responsible for the extent of tumor destruction.¹² To prevent complications such as perforations and stenoses, several factors must be taken into account during clinical application. Apart from the concentration of the photosensitizer and the choice of the time interval between the application and the irradiation, other important parameters for the successful use of PDT are the duration of exposure and the light dose. Knowledge about the behavior of light in the specific tissue is imperative, since the therapeutic effect depends primarily on the amount of light absorbed.^{13, 14} In this connection, the penetration depth of light is dependent on the optical tissue parameters (absorption coefficient ${\mu }_a$ , scattering coefficient ${\mu }_s$ , anisotropy factor $g$ , and penetration depth $d$ ). The clinical application of PDT has thus far been based on purely empirical values for the light dose and exposure time. Knowledge about the behavior of light in healthy and dysplastic tissue is of great interest for careful scheduling of irradiation. The current literature offers little information about the optical tissue properties. Data are available for stomach or carcinoma of the esophagus, but not especially about the gastroesophageal junction. Thus, the aim of this study was to determine the optical properties of healthy and carcinomatous human tissue of the gastroesophageal junction to provide reproducible parameters for optimal dosimetry in PDT. In this context, we examined a wavelength range of $300\text{to}1400\mathrm{nm}$ to determine the differences between healthy and tumorous tissue and those in the effective area of currently used photosensitizers [hypericin, 5-aminolaevulinic acid (5-ALA) and porfimer sodium].

2.

Materials and Methods

3.

Results

3.1.

Optical Parameters of Healthy Stomach and Adenocarcinoma Tissue

Evaluation of the data showed marked variations between the parameters. The values of the absorption coefficient ${\mu }_a$ for healthy tissue were between 0.01 and $13{\mathrm{mm}}^{-1}$ in the survey spectrum. The results for carcinomatous adenocarcinoma tissue were significantly lower at values between 0.02 and $3.84{\mathrm{mm}}^{-1}$ (Fig. 1 ).

Fig. 1

Absorption coefficient ${\mu }_a$ in the survey spectrum for stomach and adenocarcinoma tissue.

After reaching a maximum value of $23.13{\mathrm{mm}}^{-1}$ at $300\mathrm{nm}$ , the scattering coefficient ${\mu }_s$ for stomach tissue dropped continuously to a minimum value of $7.56{\mathrm{mm}}^{-1}$ at $1140\mathrm{nm}$ . Scattering coefficient values for adenocarcinomas were between 23.08 and $9.26{\mathrm{mm}}^{-1}$ (Fig. 2 ).

Fig. 2

Scattering coefficient ${\mu }_s$ in the survey spectrum for stomach and adenocarcinoma tissue.

Both healthy stomach and adenocarcinoma tissue showed strong forward scattering over the entire light spectrum, i.e., the anisotropy factor $g$ ranged from 0.7 to 0.9.

The optical penetration depth $d$ was highest from $0.072\text{to}7.42\mathrm{mm}$ in healthy stomach tissue and 0.11 and $6.49\mathrm{mm}$ in adenocarcinomas, with a maximum value of $7.42\mathrm{mm}$ at $1060\mathrm{nm}$ (healthy stomach tissue) and $6.49\mathrm{mm}$ at $1090\mathrm{nm}$ (adenocarcinoma) (Fig. 3 ). The optical penetration depth in carcinoma tissue was higher, up to $1000\mathrm{nm}$ , and lower, from $1010\mathrm{nm}$ , than in healthy tissue.

Fig. 3

Optical penetration depth $d$ in the survey spectrum for stomach and adenocarcinoma tissue.

3.2.

Optical Parameters of Healthy Esophagus and Squamous Cell Carcinoma Tissue

The absorption coefficient ${\mu }_a$ ranged from $0.02\text{to}11.19{\mathrm{mm}}^{-1}$ for human esophagus tissue, and from $0.01\text{to}5.44{\mathrm{mm}}^{-1}$ for squamous cell carcinoma (Fig. 4 ). Here, the values were significantly lower for squamous cell carcinoma than for healthy esophagus tissue.

Fig. 4

Absorption coefficient ${\mu }_a$ in the survey spectrum for esophageal and squamous cell carcinoma tissue.

The scatter coefficient ${\mu }_s$ dropped continuously from $22.32{\mathrm{mm}}^{-1}$ at 300 nm to $5.94{\mathrm{mm}}^{-1}$ at $1140\mathrm{nm}$ (healthy esophagus tissue). The scatter coefficient for squamous cell carcinomas ranged from $23.08{\mathrm{mm}}^{-1}$ at 300 nm to $4.07{\mathrm{mm}}^{-1}$ at $1140\mathrm{nm}$ (Fig. 5 ).

Fig. 5

Scattering coefficient ${\mu }_s$ in the survey spectrum for esophageal and squamous cell carcinoma tissue.

In healthy esophagus tissue and squamous cell carcinomas, the anisotropy factor likewise ranged between 0.7 and 0.9, corresponding to strong forward scattering.

The optical penetration depth $d$ ranged from $0.05\text{to}9.63\mathrm{mm}$ in healthy esophagus tissue (maximum of $9.63\mathrm{mm}$ at $1130\mathrm{nm}$ ), and from $0.09\text{to}8.29\mathrm{mm}$ in squamous cell carcinoma (maximum of $8.29\mathrm{mm}$ at 1110) (Fig. 6 ). Here, the optical penetration depth was higher in squamous cell carcinoma than in healthy tissue at a wavelength of up to $780\mathrm{nm}$ and usually lower above $780\mathrm{nm}$ .

Fig. 6

Optical penetration depth $d$ in the survey spectrum for esophageal and squamous cell carcinoma tissue.

Table 2

OP of healthy esophagus versus squamous cell carcinoma at 330, 630, and 650nm .  * signifies p<0.05 versus esophagus at current wavelength; # signifies p<0.05 versus esophagus at 330nm ; and + signifies p<0.05 versus squamous cell carcinoma at 330nm .

	330nm		630nm		650nm
	Esophagus	Squamouscell carcinoma	Esophagus	Squamouscell carcinoma	Esophagus	Squamouscell carcinoma
Absorptioncoefficient $\left[{\mathrm{mm}}^{-1}\right]$	$2.47\pm 1.11$	$1.48\pm {0.63}^{*}$	$0.21\pm 0.36\#$	$0.13\pm {0.04}^{*}+$	$0.18\pm 0.3\#$	$0.11\pm {0.03}^{*}+$
Scatteringcoefficient $\left[{\mathrm{mm}}^{-1}\right]$	$20.34\pm 2.03$	$19.44\pm {13.58}^{*}$	$12.56\pm 2.23\#$	$9.35\pm {8.9}^{*}$	$12.22\pm 2.23\#$	$8.98\pm {9.97}^{*}$
Anisotropy factor[ $-1$ to $+1$ ]	$0.85\pm 0.02$	$0.86\pm 0.04$	$0.94\pm 0.01\#$	$0.92\pm 0.02+$	$0.94\pm 0.01\#$	$0.92\pm 0.02+$
Penetration depth[mm]	$0.17\pm 0.06$	$0.23\pm {0.05}^{*}$	$1.47\pm 0.95\#$	$1.71\pm {0.31}^{*}+$	$1.65\pm 1.11\#$	$1.84\pm {0.31}^{*}+$

4.

Discussion

Photodynamic therapy is an established alternative procedure for management of pT1 tumor stages and early esophageal and gastric carcinomas, as well as for palliative therapy of extensive esophageal carcinoma or inoperable cases.^{17, 18, 19} However, the clinical approach of PDT has thus far been based solely on empirical data. There are only a few reports in the literature on the optical tissue properties. Thus, knowledge about the behavior of light in healthy and dysplastic tissue is of great interest for careful irradiation scheduling. The aim of this study was to initially determine all optical parameters (absorption, scattering, anisotropy factor, and penetration depth) in healthy and carcinomatous human tissue of the gastroesophageal junction to obtain reproducible parameters for optimal dosimetry in PDT application. We examined the question of whether differences exist between healthy and tumorous tissue and in the effective range of currently used photosensitizers [hypericin, 5-aminolaevulinic acid (5-ALA), and porfimer sodium]. The measurements were performed by a new integrating sphere spectrometer in an indirect measurement procedure.¹⁵ The advantages of the new spectrometer over a double Ulbricht’s sphere system were the automated measurement and the resultant simplified, faster handling. Measurement errors in spectrometer application occurred because a certain light component was lost by lateral diffusion due to multiple scattering from the sample holder, and was thus not registered by the light detector. The detectors also had higher absorption than the spectralon coating of the sphere’s inside surface. This loss of light caused a slight increase in the absorption coefficient.²⁰ Our analysis took into account this error, which is estimated to be 1 to 2% in the literature. An in-vitro study design was selected to determine the optical properties as accurately as possible. The technique of tissue sampling is considered a potential source of error. The most unadulterated, i.e., freshest, sample without essential changes in the tissue properties is the quality criterion. The freshly removed samples should not lose water or hemoglobin, nor should they be exposed to nonphysiological oxidation. Air bubbles enclosed in the tissue, inhomogeneity, and uneven surface structures cause changes in the layer thickness and thus alter the tissue properties.²¹ To ensure transferability to the in-vivo situation, the tissue was snap frozen and homogenized immediately after removal from the surgical biopsy specimen. It was subsequently placed in a cell of a given size without the inclusion of air bubbles. According to the literature, the optical properties of native and cryohomogenized tissue diverge by a maximum of 3%.^{22, 23} In a previous study, we demonstrated that the absorption level was not significantly influenced after cryohomogenization (increase of 5.9% in relation to the reference preparation). Also, the calculated reduced scattering factor ${\mu }_s^{\prime }$ decreased not significantly by a mean value of $-3.4\%$ . Hence, the changes in ${\mu }_s$ and $g$ compensated each other, and no cryohomogenization effect is expected from these data if the reduced set of optical properties is used.²² Accordingly, Peters found a maximum deviation of 3% in comparison to the nonhomogenized tissue sections.²³

As a rule, biological tissue consists mainly of water. The absorption properties of water are therefore of decisive importance for light-tissue interactions. In addition, there are various pigments. The most important example of these so-called chromophores is hemoglobin.

Hemoglobin and cytochrome oxidases are generally known to be the most important absorbers in the short infrared wavelength range. Thus, they exert a decisive influence on the optical parameters, especially on absorption. In this study, blood was not removed from the samples. However, the optical parameters of blood and water have to be considered for in-vivo irradiation planning. So, further experiments should be performed to study the influence of blood and water residuals on the absorption coefficients.

The dependence of the optical parameters on the wavelength must also be mentioned as another important factor. Except for the anisotropy factor, the optical properties vary widely, especially in the range of visible, ultraviolet, and infrared light.^{24, 25} Earlier publications on tissue properties were limited to individual wavelength values.^{3, 7, 20, 23, 25} These were mainly in the laser application $\left(1064\mathrm{nm}\right)$ and PDT $\left(630\mathrm{nm}\right)$ range. Apart from measuring the tumor diameter, Maier determined the extinction coefficient and the penetration depth of esophageal carcinoma tissue at $630\mathrm{nm}$ in vivo.¹⁸ There are no reports in the literature on the optical properties of stomach tissue. Our study assessed all optical parameters in healthy and cancerous esophageal and gastric tissue in the wavelength range of $300\text{to}1140\mathrm{nm}$ .

The absorption coefficient ${\mu }_a$ showed marked differences for both tissues over the course of the measured wavelength spectrum. The highest values were found in the lower wavelength range. The maximum was $13{\mathrm{mm}}^{-1}$ for the stomach and $11.19{\mathrm{mm}}^{-1}$ for the esophagus at a wavelength of $430\mathrm{nm}$ . After lowering the values, there was a second peak at $550\mathrm{nm}$ with $1.76{\mathrm{mm}}^{-1}$ for the stomach and $1.26{\mathrm{mm}}^{-1}$ for the esophagus. All values were under $0.1{\mathrm{mm}}^{-1}$ at wavelengths above $750\mathrm{nm}$ . There was another increase in the absorption coefficient from $1140\mathrm{nm}$ , which can be explained by the absorption of water in this range. The maximums at 430 and $550\mathrm{nm}$ are attributed to the absorption of light by deoxyhemoglobin.²⁴ Yaroslavsky found markedly lower values for ${\mu }_a$ when measuring optical parameters in human brain tissue after washing out the hemoglobin. They thus obtained an absorption coefficient in the range of $0.7\text{to}0.4{\mathrm{mm}}^{-1}$ .²⁶ Anderson, Parrish, and Parrish described absorption of $0.26{\mathrm{mm}}^{-1}$ at $456\mathrm{nm}$ for skin tissue.²⁷ This very low value can clearly be attributed to the lower amount of hemoglobin. In agreement with this, Gottschalk found absorption coefficients between 1.4 and $3{\mathrm{mm}}^{-1}$ at a wavelength of $450\mathrm{nm}$ for healthy gastric and esophageal tissue.²¹ All absorption coefficient data were under $0.1{\mathrm{mm}}^{-1}$ between 740 and $1140\mathrm{nm}$ . In 1981, Parrish defined this phenomenon as the “optical window of biological tissue.”²⁸

Over the entire spectrum, the scattering coefficient ${\mu }_s$ showed a continuous decrease from $22.32{\mathrm{mm}}^{-1}$ (esophagus) and $23.13{\mathrm{mm}}^{-1}$ (stomach) to 7.56 and $5.49{\mathrm{mm}}^{-1}$ at the beginning of the measurement range. Simpson speak of a decreasing scattering coefficient with increasing wavelengths for all biological tissue.²⁶ In agreement with the present study, scattering was described in various tissues as fluctuating only slightly compared to the absorption coefficient. At a wavelength of $456\mathrm{nm}$ , Gottschalk found a scattering coefficient of $10.9\text{to}26{\mathrm{mm}}^{-1}$ for stomach tissue. This was $15.6{\mathrm{mm}}^{-1}$ in our study.

The anisotropy factor $g$ for the two tissues showed hardly any variation over the entire spectrum of measured tissue. Nearly all results were in the range of 0.9. A value of 1.0 corresponds to complete forward scattering. Thus, the scattering direction for the measured tissue is almost forward. In Yaroslavsky, the anisotropy factor for gray and white matter was between 0.81 and 0.96.²⁶ At a wavelength of $630\mathrm{nm}$ , he determined a mean anisotropy factor between 0.9 and 0.95. The value of 0.94 measured in our study showed clear agreement.

The impact of the absorption coefficients on the parameters can be determined from the course of the optical penetration depth $d$ over the measurement range. Due to the high absorption in the lower wavelength range, the penetration depth is less than $1\mathrm{mm}$ up to $600\mathrm{nm}$ . The two negative peaks of the curve result from the high light absorption caused by hemoglobin in the tissue. The values are almost the reversed image of the decrease in the absorption coefficients between 700 and $1140\mathrm{nm}$ . The highest penetration depth was $9.63\mathrm{mm}$ at $1110\mathrm{nm}$ for the esophagus, and $7.42\mathrm{mm}$ at $1060\mathrm{nm}$ for the stomach. Germer obtained almost identical data for the penetration depth in human liver tissue ( $7.49\mathrm{mm}$ at $1070\mathrm{nm}$ ).²⁹

The optical properties of carcinoma tissue showed clear parallels to those of healthy tissue in the survey spectrum. In the qualitative comparison, the absorption coefficient also demonstrated a decreasing course above $580\mathrm{nm}$ , just like healthy tissue. The values of the scatter coefficient and anisotropy factor hardly changed in either tissue.

The optical penetration depth increased significantly with longer wavelengths. There were no significant quantitative differences between the optical penetration depth of healthy and cancerous tissue.

The qualitative change in the optical parameters can be explained by the hemoglobin component in the tissue. Like Gottschalk and Roggan, we can also confirm an effect on the parameters above $900\mathrm{nm}$ due to the water content of each tissue.

Significant differences were found in the quantitative comparison of absolute values. The values for the absorption and scattering coefficients were 52.3 and 9.4% higher in stomach tissue than in adenocarcinoma tissue. The optical properties of the squamous cell carcinoma had a significantly lower absorption and scatter coefficient in the quantitative comparison with healthy esophagus. This may be due to the lower content of chromophores, cell nuclei and membranes in tumor tissue.³⁰

In contrast, the recorded data showed no quantitative difference between tumorous and healthy tissue for the anisotropy factor.

The optical penetration depth $d$ was significantly higher in adenocarcinoma tissue than in stomach tissue up to $1000\mathrm{nm}$ . The difference from healthy tissue was $0.36\mathrm{mm}$ using a $550\text{-}\mathrm{nm}$ wavelength. The differences of 0.44 and 0.40 at 630 and $650\mathrm{nm}$ were also significant. This corresponds to a 45.9% times higher mean penetration depth in carcinoma tissue. Using $300\mathrm{nm}$ , the penetration depth was $0.059\mathrm{mm}$ higher in squamous cell carcinoma tissue than in healthy tissue. The difference in the wavelengths at 630 and $650\mathrm{nm}$ was 0.24 and $0.196\mathrm{mm}$ . This corresponds to a 20.67% times higher mean penetration depth in carcinoma tissue than in the esophagus.

Several authors have described a higher penetration depth in tumor tissue.^{20, 24, 25} Maier reported a penetration depth of $2.68\text{to}3.21\mathrm{mm}$ for $630\mathrm{nm}$ , depending on the tumor thickness. However, no distinction was made here between adenocarcinoma and squamous cell carcinoma.¹⁸ The interindividual differences in the tumor tissue may be explained by tissue structures varying according to the differentiation grade. Histological grading of the tumors was not included in our study. Denaturated proteins within the tumor necroses may also cause changes in the optical properties.

Further studies should be performed using the available data to examine the effect of the photosensitizers on the optical tissue properties. It would be of great interest in this connection to determine the optical parameters of tissues in the presence of photosensitizers. Such investigations would have to be performed in animal models. The present study deals with in-vitro measurement of samples from surgical preparations, but a photosensitizer cannot be previously applied in patients for ethical reasons.

Up to now, there have been hardly any reports on the tissue properties of the human esophagus and stomach. Measurements over a broad light spectrum have not been previously published for esophageal tissue. This is the first study to determine the penetration depth, absorption, and scattering for a given tissue.

Our data can be used as the basis for exact dosimetry to ensure effective and safe photodynamic therapy. In summary, it may be said that the measuring procedure with the spectrometer described in our study is an efficient and fast in-vitro method for determining optical parameters of healthy and dysplastic tissue. In view of the high penetration depth at the end of the light spectrum, PDT should be carried out in this range whenever possible. Thus, a photosensitizer with an optimal effect in the high wavelength range would be desirable for clinical use to ensure adequate tumor destruction, ideally by a single application. The sensitizers frequently used today have an effective range between 300 and $650\mathrm{nm}$ . In this range, light does not penetrate deeper than a few millimeters, achieving only superficial destruction of the tumor tissue and requiring high energy doses or repeated ablation treatments.