2.1.

Canine Prostate Sample

A canine prostate (Labrador retriever, intact male, 10 to 12 years, 26 kg) was harvested at the Atlantic Veterinary College, University of PEI as per Animal Control and Biosafety protocol approved by the University. The urinary bladder and proximal urethra, extending the length of the pelvic cavity and including the prostate, were removed. The ureters were transected several centimeters from the trigone, and urine was expressed from the bladder. Collection of the prostate was done $\sim 4\mathrm{h}$ after euthanization. The largest dimension of the prostate was $\sim 40\mathrm{mm}$. The optical measurements began $\sim 17\mathrm{h}$ after collection. The prostate was never frozen before or between the measurements.

2.2.

Experimental Setup

A schematic of the experimental setup is shown in Fig. 1. The detailed description of the setup can be found in the earlier publications.^33,38,42 The prostate was held in a black plastic holder with internal diameter approximately matching the size of the prostate. To prevent the prostate from drying, distilled water was sprayed over the prostate and added to the urethra. A 20-W tungsten halogen white light source (HL-2000, Ocean Optics, Dunedin, Florida) provided tissue illumination through a fiber with a 2-mm spherical diffuser at the end. The total power output from the spherical diffuser was $\sim 18\mathrm{mW}$. In the absence of the well-defined rectum, the illuminating fiber was pressed against the side of the prostate with the diffuser tip positioned vertically approximately in the middle section of the prostate. Detection was performed with a custom-made 600-micron side-firing fiber (Pioneer Optics, Bloomfield, Connecticut) inserted to urethra or selected locations in the prostate and aligned vertically with the diffuser until the maximum signal was obtained. The typical insertion depth was $\sim 17\mathrm{mm}$. Both illuminating and detecting fibers were threaded through 15-gauge needles for mechanical stability, but only the protruding part of the detecting fiber was inserted inside the prostate to avoid self-scattering effects discussed previously.³⁸ The side-firing fiber was mounted on a computer-controlled rotation stage (PRM1Z8, Thorlabs, Newton, New Jersey). Radiance data were acquired by rotating the side-firing fiber over a 360-deg range with a 2-deg step. The side-firing fiber was connected to a computer-controlled spectrometer (USB 4000, Ocean Optics) that collected spectra at every angular step. It took $\sim 4\min$ to acquire a complete single angular profile containing 180 spectra.

Fig. 1

The schematic of the experimental setup for radiance measurements. Inset: prostate top view with locations of the measurement points.

Measurements were performed at four locations in the prostate as indicated in the inset of Fig. 1. The inset shows the top view of the prostate with the position of the illuminating fiber and multiple locations of the detecting fiber with corresponding distances to the illumination point. The distances were measured with a ruler. Aside from transurethral measurements, all other interstitial measurements required puncturing prostate with a 17-gauge needle to produce a hole to accommodate the protruding part of the detecting fiber.

2.3.

Gold Nanoparticles Inclusion

The inclusion was formed by filling a thin wall quartz capillary tube (1 mm diameter, $\sim 30\mathrm{mm}$ height) with the colloidal solution of gold nanorods (Nanopartz) with following characteristics: plasmon resonance: 808 nm (Fig. 2); individual rod size: 10 nm (diameter) and 41 nm (length); concentration: $6.2·{10}^{11}\mathrm{NPs}/\mathrm{mL}$ or $0.035\mathrm{mg}\mathrm{Au}/\mathrm{mL}$. Molar absorption and scattering at 808 nm were $9.19·{10}^8$ and $1.02·{10}^8{\mathrm{M}}^{-1}{\mathrm{cm}}^{-1}$, respectively, making the nanorods dominant absorbers. The inclusion contained $\sim 1.5·{10}^{10}$ NPs or 0.8 μg Au in 0.024 mL volume. The inclusion was inserted in the prostate at a 10-mm distance from the urethra at a 90-deg angle with respect to the light source.

Fig. 2

The extinction spectrum of gold nanorods used in the current work.

2.4.

Porcine Phantom Geometry

On a single occasion, we performed radiance measurements in the prostate-shaped porcine muscle phantom as in Ref. 33. The prostate size piece ($\sim 4\mathrm{cm}$ diameter, $\sim 3\mathrm{cm}$ height) was cut from the boneless pork loin and placed in the same plastic holder that accommodated the prostate (Fig. 1). Pork loin demonstrated a strong degree of heterogeneity manifested as a network of fatty inclusions between muscle fibers (see images in Ref. 33). Both detecting and illuminating fibers were inserted interstitially to the phantom with 15-mm separation keeping it in line with similar (transurethral) measurements in the prostate (position “15 mm” in the inset of Fig. 1).

2.5.

Data Analysis and Optical Parameters Extraction

Optical radiance ($I$) is a function of the source-detector separation ($r$), the angle between the directions of light propagation and scattering ($\theta$), and the wavelength of light ($\lambda$). Individual spectral profiles of radiance measured every 2 deg over 360 deg in the prostate were combined into the prostate spectro-angular matrix, $I_{\text{prostate}}(r,\theta ,\lambda )$, with columns as wavelengths from 600 to 950 nm and rows as angles from $-178$ to 178 deg. In a similar way, the water matrix, $I_{\text{water}0}(r,\theta ,\lambda )$, was assembled in which the same spectrum of radiance measured at 0 deg in water was replicated for all rows (angles) with the subscript 0 emphasizing it. Constructing a ratio of these two matrices eliminated spectral responses from the white light source, illuminating and detecting fibers in addition to referencing prostate measurements to those in water. The ratio $I_{\text{prostate}}(r,\theta ,\lambda )/I_{\text{water}0}(r,\theta ,\lambda )$ focuses on optical properties of the prostate, has a clear physical definition, and corresponds to relative radiance that can be, in principle, converted to the absolute value with proper normalization as was shown previously.³⁸ This ratio is presented via surface or contour plots demonstrating intensity and spectral variations versus angle. Such plots map the photon distribution in the angular domain and are defined as spectro-angular maps.

For detecting the presence of the localized optical inhomogeneity (inclusion) in the prostate, another ratio was constructed. As was shown earlier,^33,43 constructing the radiance extinction ratio (RER) from two radiance data sets (one for the phantom or organ with the localized inclusion and the other for the phantom or organ without the inclusion) allows isolating signatures of the inclusion from the dominant background (the prostate, in the current work) contribution, identifying it spectroscopically and locating it in the angular domain. For the inclusion of Au NPs in the prostate, the RER was defined as $\mathrm{RER}=I_{\text{prostate}}(r,\theta ,\lambda )/I_{\text{prostate}+\mathrm{Au}}(r,\theta ,\lambda )$. Hence, RER focuses on optical extinction properties of the inclusion and is presented as the contour spectro-angular plot.

Complete details of the approach for extracting optical properties can be found in Ref. 41. Only final formulas are presented here. The effective attenuation coefficient ${\mu }_{\mathrm{eff}}(\lambda )$ is determined from a simple algebraic expression constructed from a ratio of two radiance measurements at two different source-detector separations ($r_0$ and $r$) and the same 90-deg angle.

Then, with the knowledge of ${\mu }_{\mathrm{eff}}(\lambda )$, the diffusion coefficient $D(\lambda )$ is determined from another ratio constructed from two radiance measurements at two angles (0 and 180 deg) and the same source-detector separation.

Once ${\mu }_{\mathrm{eff}}(\lambda )$ and $D(\lambda )$ are obtained, the absorption coefficient ${\mu }_a(\lambda )$ and the reduced scattering coefficient ${\mu }_s^{\prime }(\lambda )$ can be determined from them as follows:

The accuracy of the method depends on the accuracy of the source-detector separation measurements, which appears in the difference between two different source-detector separations ($r_0$ and $r$). For distance measurements with a ruler, the typical accuracy is $\pm 0.5\mathrm{mm}$. In the worst case scenario of systematic errors, if two errors add up resulting either in the increased ($\pm 1\mathrm{mm}$ offset) or decreased ($-1\mathrm{mm}$ offset) difference, the uncertainties would be $\sim 18\%$ for ${\mu }_{\mathrm{eff}}(\lambda )$ and ${\mu }_a(\lambda )$ and $\sim 10\%$ for $D(\lambda )$ and ${\mu }_s^{\prime }(\lambda )$. However, in practice, it is possible to keep the error for ${\mu }_{\mathrm{eff}}(\lambda )$ and ${\mu }_a(\lambda )$ $<10\%$.

4.1.

Analysis of Results

Because of a large source-detector separation used in the work, photons sample a substantial portion of the prostate on a way to the detector such that their absorption and scattering are affected by interactions with the tissue in distant regions. Hence, the spectro-angular map as in Fig. 3 represents a characteristic snap-shot of bulk optical propagation in prostate tissue encoded in spectral and angular domains. Such spectro-angular mapping is obtained in a desirable minimally invasive way with the detector in the urethra and illumination near a position of the rectum. It is important to understand what information can be obtained from such measurements because this approach to canine prostate has yet to be reported.

In Fig. 3, the signal measured at 0 deg (the detector facing the light source) is composed of forward-scattered photons and can be considered as the transmission of the medium. When the detector is facing opposite to the light source (180 deg), the measured signal is composed of backscattered photons in the backreflection geometry, while all intermediate angles provide information about other angular-sensitive pathways of photon propagation in the prostate. The maximum signal is marked with red and occurs in the forward-scattering direction (along 0 deg), which is typical for most biological tissues. The highest transmission is observed for wavelengths around 810 nm with the secondary peak of lower intensity at $\sim 730\mathrm{nm}$. Such shape is usually caused by overlapping transmission spectra of oxygenated and deoxygenated hemoglobin that are present in biological tissues.³³ Plots from Fig. 3 identify the spectral range of $\sim 790$ to 850 nm as the biomedical transparency window where light propagates longest distances in the excised prostate. While the concept of a generic biomedical transparency window in biological tissues is well known, possible differences in chromophore concentrations due to both morphological diversity and pathology (as applied to prostate) would require confirming it for each sample to avoid false assignments.

A degree of ellipticity of contour lines in the horizontal direction from contour plots serves as an indicator of angular light confinement, which is broadened by multiple scattering. More elongated contour lines (consistent with higher optical scattering) also tend to bring higher values to the signal measured in the backscattered direction (at 180 deg). A simple estimation from data in Fig. 3 indicates that relative radiance in the backscattered direction drops down to $\sim 50\%$ of the value measured along 0 deg for the most transmitted wavelength in the canine prostate. This is slightly below the $\sim 80\%$ that was measured in Intralipid-1% under similar conditions.³⁸ It indicates qualitatively that light scattering in canine prostate is relatively strong and only marginally smaller than that in Intralipid-1%. On the other hand, multiple scattering spreads photons over the larger volume, which should facilitate detectability of localized inhomogeneities in distant regions of the sample, as was the case for detecting Au NP inclusions in Intralipid.³⁹

Spatial distribution of light inside turbid media is a wavelength-dependent quantity based on absorption and scattering of individual chromophores. Even though the major chromophores in most biological tissues are the same, i.e., oxygenated (${\mathrm{HbO}}_2$) and deoxygenated (Hb) hemoglobin, water, lipids, etc., their concentrations differ from tissue to tissue making their combined effect on optical properties very unique for the particular tissue. To demonstrate it, Figs. 3(c) and 3(d) present the spectro-angular map in the contour and surface formats obtained for the prostate-shaped porcine muscle phantom under identical illumination/detection conditions. Other than the larger signal’s intensity, the shape of the pattern is distinctly different. As expected, the maximum of intensity is detected along 0 deg. Two peaks at $\sim 810$ and $\sim 730\mathrm{nm}$ that are well defined for the prostate have broadened, indicating a different composition of the porcine phantom. The contour lines are elongated strongly in the vertical direction with the signal measured at 180 deg down to $\sim 30\%$ of the signal measured along 0-deg angle. Altogether, it indicates that porcine muscle tissues exhibit less scattering than the canine prostate and light is (relatively) better confined in them. With this knowledge, we expect that localized inclusions of Au NPs in the canine prostate can be detected at larger distances (from the detector in urethra) than those reported for the porcine phantom.³³

4.2.

Implications of the Work in Light of Other Studies

The effect of prostate heterogeneity on optical properties manifested in up to 70% spot-to-spot variation was reported in a number of publications.^12,13 It is interesting to note that in our experiments the set involving transurethral measurements produced the highest absorption and scattering coefficients. While it may indicate different optical properties of tissues composing the urethra, more systematic measurements on a large set of samples are required to provide conclusive results. It should be noted that this radiance method does not require absolute measurements, detector calibration, or a fitting procedure for parameter extractions,^46,47 which have certain advantages.

Optical properties of the canine prostate have been the topic of a number of publications^10–19 with average values and standard deviations summarized in Table 1.

Table 1

Optical properties of canine prostate from literature data.

In contrast to the present study, available literature data combine in vitro, ex vivo, and in vivo results and refer to different but fixed wavelengths used for monitoring PDT. It is understood that these data are dependent on the sample preparation technique, prostate storage, the measurement method, and optical properties extraction algorithms in addition to instrumental errors and noise. Another important factor is prostate heterogeneity reflected in intrasample and intersample variations that can easily outweigh all other factors. A difference between in vivo and ex vivo measurements is also expected. It was suggested that postmortem tissue storage produces a decrease in ${\mu }_a(\lambda )$ and an increase in ${\mu }_s^{\prime }(\lambda )$ accompanied by losses in hemoglobin.⁴⁸ Since all experimental methods have their merits, existing data cannot be ignored. With all the factors in mind, the agreement between our ex vivo data from the single prostate and reported discrete average values of ${\mu }_{\mathrm{eff}}(\lambda )$ appears to be quite good [Fig. 7(a)]. Literature values of ${\mu }_a(\lambda )$ [Fig. 7(b)] and, to a greater extent, ${\mu }_s^{\prime }(\lambda )$ [Fig. 7(c)] show a larger spread. A possible explanation is that in fluence measurements, ${\mu }_{\mathrm{eff}}(\lambda )$ is determined from the slope, and thus, the values can be quite accurate. Then, since both ${\mu }_a(\lambda )$ and ${\mu }_s^{\prime }(\lambda )$ contribute to ${\mu }_{\mathrm{eff}}(\lambda )$, various fitting routines are used to separate them, bringing more uncertainties at this stage. In general, when comparing a discrete measurement with a continuous waveform, it might be difficult to state whether the entire curve shifted or the discrepancy occurred only at a particular wavelength.

Fig. 7

Comparison of our experimental data with published values for (a) the effective attenuation coefficient, (b) absorption coefficient, and (c) reduced scattering coefficient for canine prostate.

Spectroscopic measurements provide indispensible tools for tissue characterization and diagnostics. The absorption spectrum can be linked to contributions of individual chromophores quantifying their presence in the tissue. Figure 5 shows that the spectrum of the averaged absorption coefficient of canine prostate can be matched well by the synthetic absorption spectrum composed of $9.0(\pm 2)\mu \text{M}$ of ${\mathrm{HbO}}_2$, $29.6(\pm 0.5)\mu \text{M}$ of Hb, and $0.47(\pm 0.07)$ fractional volume of ${\mathrm{H}}_2\mathrm{O}$, which produces 38.6 μM for the total hemoglobin concentration (THC) and 23% tissue oxygen saturation (${\mathrm{StO}}_2$). Contributions from Hb and ${\mathrm{H}}_2\mathrm{O}$ are associated with very distinct spectral features, while the one from ${\mathrm{HbO}}_2$ is the lowest and provides almost constant values in the entire spectral range. Thus, the concentration of ${\mathrm{HbO}}_2$ has the largest uncertainty ($\sim 22\%$). Unfortunately, such data do not exist for ex vivo canine prostate and are very scarce for in vivo canine prostate. Solonenko et al.¹⁶ obtained $51\pm 11\mu \text{M}$ and $50\pm 16\%$ for THC and ${\mathrm{StO}}_2$ correspondingly for canine prostate in vivo, while Jiang et al.⁴⁹ reported a wider range for THC of 110 to 150 μM also in vivo. While our value for THC is close to the lower end of the range reported by Solonenko et al.,¹⁶ it is expected that ex vivo canine prostate would exhibit lower THC and oxygenation values as do most postmortem tissues.⁴⁸ Unfortunately, there are no studies on variation of THC and oxygenation values with time after excision for canine or human prostates. We could not find any published data on the water content in the canine prostate. However, for the human prostate, the fractional volume of water was reported to be 0.7 from in vivo measurements.^50–52 Water content in excised organs will depend on a number of factors, including time elapsed after extraction and organ handling, but overall, it is expected to be lower than the one obtained from in vivo measurements.

There are a limited number of spectroscopic studies on the water content in normal and neoplastic tissues in the prostate. This is unfortunate because this interesting chromophore may bring some promising developments in prostate cancer diagnostics. It appears from in vitro measurements that prostate with cancer may have lower water content than a normal prostate.⁵³ However, a number of ex vivo and in vivo measurements on breast^54–57 and lung⁵⁸ tumors demonstrated an increase in water content in neoplastic tissues. Thus, monitoring spectrally resolved optical properties, concentrations of major chromophores, ${\mathrm{StO}}_2$ and water content, could offer a comprehensive approach to a systematic study of the canine prostate ex vivo and potentially in vivo. As compared with discrete wavelength measurements, the use of full spectroscopy data collection improves the accuracy of calculated values.^16,59 While THC depends on total absorption, determination of ${\mathrm{StO}}_2$ requires detecting more subtle spectroscopic changes than are possible to derive reliably at few wavelengths.⁶⁰

Being able to detect Au NPs in the prostate is important for their diagnostic and therapeutic applications. To the best of the authors’ knowledge, detection of Au NPs in the prostate has not been reported. Figure 6 shows our data on detection of the localized inclusion made from Au NPs ($1.5·{10}^{10}$ NPs or 0.8 μg of Au) in the canine prostate. Au NPs with the plasmon resonance at 808 nm were selected to fit the biomedical transparency window of the prostate ($\sim 790$ to 850 nm). The position of the inclusion was chosen to match the most distant location of the inclusion that could be detected under similar conditions in the prostate-shaped porcine muscle phantom.³³ However, the diameter of the capillary was reduced from 2 to 1 mm to detect a smaller number of Au NPs, i.e., $1.5·{10}^{10}$ versus $4·{10}^{10}$. Detecting $\sim {10}^{10}$ Au NPs or $\sim 0.8\mu \text{g}$ of Au in 10 to 15 g prostate with a clinical CT can be quite challenging.³³ However, this number of Au NPs is very important for early detection of cancer in the prostate when targeting PSMA receptors with Au NPs.

Receptors of PSMA are expressed in primary and metastatic prostate cancer tumors (but not in healthy vasculature)⁶¹ at $\sim {10}^6$ sites per cell.⁶² Since most receptors are expressed at levels of only ${10}^3$ to ${10}^4$ copies per cell, it makes PSMA an attractive biomarker of prostate cancer that can be used for diagnostic and therapeutic applications. A targeted (via PSMA) delivery of gold NPs to prostate cancer cells was explored using various targeting agents: antibodies, aptamers, and small molecules.^63–66 Binding Au NPs to 1 to 10% of total PSMA receptor sites would produce $\sim {10}^4$ to ${10}^5$ Au NPs per cell. Early cancer detection requires detecting $\le {10}^5$ malignant cells (in the background of 10 to 15 g prostate), which corresponds to a transition from diffusion limited nutrition to neovascularization (labeled as angiogenic switch).⁶⁷ Hence, detecting ${10}^5$ malignant cells with Au NPs attached to 1 to 10% of receptor sites is equivalent to detecting ${10}^9$ to ${10}^{10}$ Au NPs in the prostate.

The reported value of detected Au NPs, $1.5·{10}^{10}$, is approaching the targeted range of ${10}^9$ to ${10}^{10}$. It should be stressed that these numbers refer to gold nanorods used in the current work. Optical detection of Au NPs is based on the ability of NPs to absorb light that is expressed via the absorption coefficient proportional to the product of the NP concentration and the individual absorption cross-section. Here a motivation for developing new types of gold NPs for diagnostic purposes becomes very specific: NPs with absorption cross-sections larger than those of Au nanorods would require fewer total numbers in targeted prostate cancer cells in order to be detected. One of the promising candidates for this role can be gold nanocages.⁶⁸ Hence, improved sensitivity offered by optical detection may stimulate further applications of theranostic Au NPs when, for example, functionalized gold NPs targeting PSMA can be detected in the prostate and then used in NP-assisted laser thermal therapy ablating cancerous tissues.

4.3.

Methodological Limitations

For in vivo applications, the experimental setup described in the current work should be modified because it was originally designed for working in liquid phantoms. Radiance detector should be placed in an external housing such that no rotating parts would be in a direct contact with tissues. Illuminating fiber with a 2-cm-long cylindrical diffuser, which is a typical fiber for PDT applications, might be a better choice than the fiber currently used. Having the cylindrical diffuser would eliminate the need of vertically aligning the detecting fiber to the illuminating fiber and provide quasi-three-dimensional sectioning capabilities by incrementally translating the detecting fiber in the vertical direction. To avoid unnecessary loss of light from the diffuser to the rectum, the cylindrical diffuser can accommodate a simplistic reflector that would direct most of the light toward the prostate.

4.4.

Future Research

As one of the directions of future work, it would be extremely interesting to perform in vivo systematic radiance measurements with transurethral detection and transrectal illumination on multiple human prostates and correlate spectral and angular features with a particular state of the prostate health obtained from established methods.^69,70 Transurethral measurements with a side-firing fiber should be familiar to urologists because in the existing practice of laser prostatectomy, a side-firing fiber is already used for laser energy delivery from prostatic urethra to the adenoma for vaporization.²⁵

We will also explore approaches that enable the extraction of both absorption and scattering properties of the prostate from a single spectro-angular map obtained via transurethral measurements.

Our technique offers the framework for a systematic ex vivo study of major chromophores in the canine and, potentially, human prostate that can be used as comprehensive diagnostic markers in determining the state of the prostate health. Hence, the technique is also well suited for elaborating the effect of postmortem prostate aging on THC and water content. Developing further theoretical aspects of the approach⁴⁰ would enable researchers to extract additional information about localized inclusions, such as concentration of Au NPs and the distance to the inclusion.