1 Introduction

Noticeable progress is realized during recent years in the multidisciplinary field of biophotonics to develop the noninvasive cancer diagnostic techniques. The current diagnostic methods are mainly based on the painful and sluggish biopsy leading to the pathologist report. Despite their widespread success, these methods require expert interpretation of the microscopic information. Nowadays, there is a significant medical demand for the optical biopsy as the minimally invasive alternative in the diagnostic oncology (Majumder et al. 2005). Therefore, the optical spectroscopic instruments are utilized to modify the traditional pathology. Among various techniques, the fluorescence spectroscopy exhibits an efficient and promising method. Despite that the poor autofluorescence signals limit the clinical applications of the endogenous fluorophore, however, the laser-induced fluorescence (LIF) has shown the competence for the efficient fluorescence spectroscopy/imaging in oncology by virtue of exogenous agents. The redshift of the fluorescence emissions from the Rd6G-infiltrated tissue is suggested to discriminate the healthy and the cancerous tissues (Ghasemi et al. 2015, 2017). Recently, the observation of the random laser (RL) emission in biomaterials has been reported leading to the demonstration of a single-cell biological laser (dos Santos et al. 2014).

Furthermore, random laser is known as the coherent emission in highly scattering gain media, where the optical feedback comes from the light scattering instead of the regular mirrors. Resonant cavities are randomly closed-loop paths of the light which is self-formed due to multiple scatterings (Choi and Kim 2014). Laser oscillation occurs when the amplification along such an optical loop exceeds the loss, and the interference effect gives rise to the resonant modes at certain frequencies (Lü et al. 2014; Cao et al. 1999). The overall scattering properties depend strongly on the size and the refractive index of the scatterer media (dos Santos et al. 2014; Lü et al. 2014). The laser emission spectra comprise valuable information regarding the optical scattering events from the biomaterials (Polson and Vardeny 2004).

Biological tissues are known as the turbid optical media, in which light transport is dominated by elastic scattering. The primary scattering centers are most likely due to the collagen fiber network of the extracellular matrix, the mitochondria, and other intracellular substructures, and all exhibit dimensions smaller than optical wavelengths. The larger structures, such as cell nuclei with typically 5–15 μm in diameter, are also light scatterers based on the Mie scattering. In highly scattering tissues, the scatterers are randomly distributed to the photons usually encounter the multiple scattering events. The diffusive light which is not absorbed in the tissue eventually returns back to the surface, passing once more through the epithelium, where it is again subject to the scattering from the cell nuclei. Thus, the emerging light consists of a large diffusive background plus the component of the forward and the backscattered photons from the nuclei in the epithelial layer (Perelman et al. 1988; Mourant et al. 1995).

When penetrating biological tissues, dye solutions act as a gain media and emit coherent photons. Examples of such include the dye-infiltrated colon tissues for diagnostic purposes (Polson and Vardeny 2004, 2010) and dye-infused bone structures (Song et al. 2010). In the presence of the scatterers, the fluorescence spectra of the dye vary as a function of scatterer density (Ahmed et al. 1994). The resonant modes are typically Stokes shifted with respect to the absorbance. A different spatial distribution of the same scatterers may generate a perturbation that results in a sensible spectral shift (Song et al. 2010). It is particularly important in the medical point of view, since the spectral distribution may relate to the medical conditions to be diagnosed.

In this work, the random laser emission spectra of the normal and cancerous Rd6G-infiltrated thyroid tissues are investigated at different excitation energies to distinguish the healthy and the malignant ones. We have also carried out the Fourier transform (FT) and power Fourier transform (PFT) of the laser emission spectra to estimate the effective length of the randomly formed resonators. Moreover, the averaged PFT of the emission spectra attests the evidence of some distinct features, useful for the efficient diagnosis of the healthy and cancerous tissues.

2 Experimental Methods

Samples of human thyroid tissues, previously fixed and well diagnosed by a pathologist as either nominally healthy or cancerous, were immersed in the solution (0.1 mM) of rhodamine 6G dye/ethanol for 45 min. Rd6G (C28H31N2O3Cl), with molecular weight 479.02 g/mol as well as ethanol alcohol (C2H5OH) with spectroscopic grade purity, was supplied by Sigma-Aldrich Chemie GmbH and Merck Chemical companies, respectively. The solution of dye was prepared by dissolving the certain amounts in ethanol solvent at various concentrations. The dimension of the specimen was chosen to be ~ 1 cm with the thickness of ~ 0.3–0.5 cm.

Experimental setup consists of coherent sources, conducting optics, laser pulse diagnostics, spectrometer, and microprocessor. A second harmonic Q-switched Nd:YAG laser model Brilliant made by Quantel system was employed as the coherent excitation source with 10-ns duration and 1–10 Hz repetition rate at 532 nm to generate random laser in tissues. A CW Nd:YAG laser at 532 nm was also utilized to irradiate the dye solution in order to obtain the fluorescence emissions, as well. The light emission from the sample was collected by a fiber bundle, UV 600 (core)/660 (cladding) type with SMA-905 fiber connector having NA 0.22 and 1 m length. The fiber output was coupled to the entrance slit of a compact wide-range spectrometer (200–1100) model S150 Solar Laser System TM (50 mm focal length, transmission diffraction grating with 200 grooves/with 0.02 × 3.0 of entrance slit and 0.5 nm spectral resolutions). A charge-coupled device array detector model Toshiba TCD 130AP with 3648 number of pixels was used to detect the emission signals. The pump beam is focused vertically on to the surface of the sample below the ablation threshold.

We have investigated the normal and cancerous thyroid specimens of the two patients lead to the similar results.

3 Results and Discussion

Rhodamine 6G solution experiences a strong absorbance and fluorescence properties over the green and red spectral ranges, respectively. Thus, dye solutions of various concentrations were irradiated at 10 mW pump energy of a CW Nd:YAG laser at 532 nm to find the optimum concentration to be ~ 0.1 mM. The typical emission spectra of the Rd6G/ethanol at various concentrations are shown in Fig. 1.

Fig. 1
figure 1

Fluorescence emission spectrum of rhodamine 6G dissolved in ethanol at various concentrations: a 10, b 1, c 0.1, and d 0.01 mM at 10 mW excitation energy

Full size image

The typical microscopic structure of the thyroid tissues is shown in Fig. 2 for the healthy tissue (2b) and the malignant one (2d). The tissues are immersed in the Rd6G solution at optimum concentration of 0.1 mM. The Rd6G-infiltrated samples were irradiated at two distinct points by the second harmonic Nd:YAG laser at the same excitation energies. The light emitted by the dye gain medium is scattered in the tissues providing field feedback for lasing. The typical random laser emission spectra from the healthy and malignant samples are shown in Fig. 2a–c, respectively.

Fig. 2
figure 2

Typical random laser emission spectra of a healthy and c malignant Rd6G- infiltrated samples. The microscopic structure of: b the healthy and d malignant thyroid tissues

Full size image

The overall emission spectra taken by the random laser are much narrower than the fluorescence emission of the dye by itself at the same excitation intensity.

Laser action requires the formation of some sort of the optical cavity. When the loss rate becomes smaller than the spontaneous emission rate, laser oscillation takes place in the certain modes corresponding to the microcavities. As the excitation intensity exceeds above the threshold, the emission spectra narrowing occurs versus the broad spontaneous emission leading to the formation of multiple sharp peaks. This addresses the wavelength of the modes from a specific resonator formed within the tissue. The typical emission spectra of the Rd6G-infiltrated tissues at various excitation energies are shown in Fig. 3.

Fig. 3
figure 3

Typical random emission spectra of a healthy and b malignant Rd6G-infiltrated tissues at various excitation energies of: 216 mJ/pulse (dark blue), 169 mJ/pulse (red), 156 mJ/pulse (green), 126 mJ/pulse (violet), 112 mJ/pulse (blue), and 86 mJ/pulse (maroon)

Full size image

As expected, the signal magnitude lucidly rises with the pump energy. Figure 4 depicts the maximum emission signals in terms of the pump energies.

Fig. 4
figure 4

Maximum emission intensity versus the excitation laser energy for: the healthy (open circle blue) and malignant (asterisk red) samples

Full size image

The fluorescence emission and reabsorption rates are significantly increased inside the malignant tissues as a scattering media due to the higher degree of irregularity. Accordingly, a significant redshift took place in laser emission of the cancerous tissues compared to the healthy ones. This is due to the different random walk paths of the light. Longer random walk of the scattered photons leads to the higher reabsorption rates, and eventually, the sensible spectral shift occurs. Figure 5 illustrates the redshift of the emission spectra from a typical cancerous tissue against the normal one at 216 mJ/pulse excitation energy.

Fig. 5
figure 5

The typical spectra which shows the red shift of the emission: the cancerous samples (red) respect to the normal one (blue)

Full size image

To be in agreement with the previous reports studying fluorescence emission of breast and Hodgkin’s tissues (Ghasemi et al. 2015, 2017), the redshift assists to discriminate the healthy from cancerous tissues as well. This shift was also observed at the various excitation energies for the different samples of interest. In fact, by elevating the pumping energy, the dye molecules located in the deeper regions are excited due to longer light penetration depth. Hence, the luminescence coming from the deeper regions is strongly reabsorbed before exit resulting in the redshifted emission spectra. Since the spatial distribution of tissues, as the scatterers, is heterogeneous, the numerous resonators are formed with plenty of central wavelengths and FWHM. Figure 6 depicts the significant redshift of the central wavelength emissions versus the pumping energy of the cancerous samples against the normal one.

Fig. 6
figure 6

Spectral shift of central wavelength of the emission versus the pumping energy of: the normal (open circle blue), and malignant (asterisk red) samples

Full size image

The FT and PFT of the lasing spectra are also utilized to elucidate that real randomly formed resonators are involved in the random laser. Each FT component is related to the specific resonators in the particular illuminated area of the sample that correspond to the certain microresonator lengths (Polson and Vardeny 2010). The Fourier transform of the optimum dye solution emission (spectrum c in Fig. 1) is shown in Fig. 7. The FT profile is a smooth spectrum with no distinct harmonic modes. No lasing spike has been observed, and the spectral shape merely follows the spontaneous emission proportional to the gain profile.

Fig. 7
figure 7

FT of the Rd6G fluorescence emission (spectrum c in Fig. 1) with no distinct harmonic modes

Full size image

The ensemble averaging of the power Fourier transforms of the laser emission spectra is also carried out either over the sample area or over the exposure time. The averaged PFT spectrum contains a sharp, well-resolved Fourier component and its harmonics characterizing a well-defined laser resonator. The PFT of the emission spectrum from such laser cavity is characterized by several peaks of the Fourier components. The location of the dominant harmonics (dm) is equal to mLn/π where m is an integer denoting the FT harmonics, L is the cavity path length, and n is the refraction index of the gain medium. If the scatterer media contain cavities with similar sizes, then distinct components appear in the average FT, since a particular cavity becomes dominant upon averaging (Polson and Vardeny 2004).

Conversely, if the scatterer contains cavities with different sizes, then average will not form strong distinctions. Thus, the FT components of random lasing emission of such a disordered medium become faint. In a healthy tissue, the cell sizes and spacing are rather regular, and the average PFT contains a number of well-resolved harmonics. On the contrary, in the cancerous tissues, there are wide ranges of disordered cavities with different sizes leading to a rather smooth average PFT. Therefore, cancerous tissues demonstrate a single line in the gain spectrum and no harmonic features of the average PFT may take place (Cao et al. 1999).

Figure 8a, b displays the Fourier and power Fourier transforms for the emission spectra shown in Fig. 2. The resulting components in the normal samples are larger and stronger than those in cancerous ones.

Fig. 8
figure 8

Fourier transforms (a) and power Fourier transforms (b) of the emission spectra in Fig. 2 for the malignant (red) samples with more laser lines with respect to the healthy one (blue)

Full size image

Here, the PFT averaging technique is applied to estimate the typical random resonator size within the tissue. This procedure is carried out by changing the stripe on the samples to irradiate many different positions with spacing each ~ 0.2 mm. Figure 9 indicates the average power Fourier transform of the 30 different spectra of the healthy and malignant tissues at certain pump energy of 186 mJ/pulse.

Fig. 9
figure 9

Average power Fourier transforms of laser emission spectra of: the healthy (blue) and malignant (red) thyroid tissue

Full size image

The average PFT components in the healthy tissue samples are larger and stronger than those obtained in malignant ones, corresponding to the regular structures with specific cell sizes at certain wavelengths. Conversely, typical malignant tissue exhibits irregular structures having plenty of microcavities forming with no definite size. This resembles each spectrum to contain a series of wavelengths with a smooth average PFT.

In addition to thyroid tissues, other organs samples were also examined. Tests administered on various healthy and cancerous colon and breast tissues from different patients have led to similar results as well. Consequently, the comparison between the average PFT random resonator emission spectra from healthy and malignant tissues suggests a simple diagnostic technique for the malignant ones.

Moreover, the sizes of the underlying microresonators are estimated from the PFT. The location of the dominant harmonic of the averaged PFT emission spectra for samples of interest is given to be 17 μm. Let the refractive index of thyroid tissue be 1.5. Thus, the cavity path length of dominant microcavity is deduced to be 35 μm. This path length is comparable to the average size of the typical thyroid cell (~ 20 μm).

4 Conclusion

Here, the random laser of the dye-infiltrated human thyroid samples is employed to discriminate the normal and cancerous tissues. Inherent tissue disorders provide the optical scattering events necessary for the formation of different resonators and subsequent laser emission. The random laser emissions are investigated in terms of various effective parameters such as the dye concentration, the intensity of excitation, and the angle between the incident light and the detector probe. The spectrum is also taken from different positions by moving the sample to consider the disordered structure of tissues as the scatterers.

A notable redshift takes place for the random laser emission of cancerous tissues due to the longer random walk path that leads to the strong reabsorption and reemission events. The redshift of the random laser emissions can well discriminate the healthy and cancerous tissues. A similar redshift was previously observed in the laser-induced fluorescence emission of breast and Hodgkin’s tissues. Moreover, it is shown that the signal magnitude as well as the redshift increases by elevating the pumping energy due to longer light penetration depth.

Furthermore, the distinct difference in the averaged PFT of cancerous and normal tissues of the random laser spectra is taken into account for the rapid identification and diagnosis. The average process over different PFT spectra of the healthy tissues demonstrates the existence of several peaks indicating a number of dominant microcavities, comparable to the mean cell sizes. In comparison, the corresponding cavities due to the malignant tissues do not exhibit a certain peak leading to a smooth average PFT. Eventually, the length of dominant microcavity is estimated to be ~ 35 μm, which is comparable to the average size of the thyroid cell.

It is shown that optical spectroscopy based on the random lasing detection is useful for diagnostic purposes. It can be suggested that the drug molecule with fluorescent properties and high quantum efficiency will open up new horizons for research of in vivo application.