Open Access
Add to BibSonomy
Abstract
Quantifying the biomechanical properties of the tongue is significant for early diagnosis of tongue carcinoma. Therefore, an intraoral optical coherence elastography system based on a miniature probe was proposed here to evaluate the viscoelasticity of in vivo tongue for the first time. Results of experiments with Sprague-Dawley rats indicate that considerable elasticity diversity occurred between cancerous and normal tongues, and the corresponding ratio of their Young’s modulus was evaluated to be 3.74. It is also found that, viscosity in diseased tissue is smaller than that in normal tissue. Additionally, healthy, transitional and cancerous regions in the cancerous tongue can be distinguished easily by calculating viscoelasticity characteristics. Based on this preliminary attempt, our method with advantages of noninvasive, high-resolution, high-sensitivity and real-time detection and convenient operation may have good potential to become a useful tool for tongue carcinoma assessment after further optimization.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Oral cancer is a prevalent malignancy, including cancers of lips, tongue, gum, floor of mouth, palate, cheek mucosa, and a range of other parts of the oral cavity. Among them, the incidence of tongue cancer ranks first among neoplastic cancer cases due to frequent mechanical stimulation. [1] Early treatment of tongue cancer will lead to a reduction in mortality, particularly if the primary tumor can be eliminated before dissemination. The survival rate of stage I (T1N0) oral cancer is about 80 percent, at least at the five year-survival rate level (50%-60%), while stage IV carcinomas have a cure rate of a mere 20 percent, and thus early detection of tongue cancer is of the utmost importance. [2–4] Unfortunately, early diagnosis cases of tongue cancer stage I only account for 10.9%-25.4% nowadays. [5] Tissues biopsy followed by histopathological assessment remains the gold standard for differentiating and diagnosing different types of oral cancers at present. [6,7] However, there are still some limitations in biopsy techniques, such as invasiveness, high-cost and long-time operation. [8,9] Therefore, it’s urgent to introduce more beneficial tools to aid the early-detection efficiency of oral cancer.
During the past two decades, several optical imaging modalities and techniques have been used in detecting and monitoring oral tumors, comprising narrowband imaging, near-infrared fluorescence imaging, photoacoustic tomography, confocal laser Raman spectroscopy and optical coherence tomography (OCT). [10–14] In particular, OCT technique demonstrates excellent characteristics including noninvasive, high-resolution, real-time, and label-free tissue imaging, which has been proved to be an effective tool in various medical applications. [15,16] In recent years, OCT has been widely used for oral tissue imaging and regarded as an effective indicator for oral cancer detection. [17,18] Although it is possible to observe anatomical changes associated with tumor formation using OCT, it may still be insufficient for early diagnosis of oral carcinoma prior to occurrence of structural alterations.
It recently has been shown that biomechanical properties of the tongue change at the onset of malignancies due to the synthesis and remodeling of collagen fibers, which indicates that tongue biomechanics evaluation may contribute to early diagnosis of tongue cancer. [19,20] Optical coherence elastography (OCE), as a functional extension of OCT, has a superior micrometer scale resolution and is therefore suitable for imaging subtle mechanical changes in early stages of diseases. [21–23] There is evidence suggesting that OCE has the capability to identify tissue microarchitecture in benign and malignant human breast tissues. [24,25] Similar to breast cancers, tongue malignancies typically occur close to the surface, which allows OCE to characterize biomechanical properties of cancerous tongue.
In our recent work, a phase-resolved shaker-based OCE system was developed to evaluate elasticity of in vivo healthy beagle tongue, where the ability and effectiveness of the OCE system for tongue detection has been demonstrated. [26] However, the currently used shaker is not appropriate for clinical experiments because of its large size, poor operating performance and bad user experience. Besides, quantification of elastic wave velocity was usually based on group velocity, which ignores the dispersion effect of elastic wave and viscosity in samples. [27]
To overcome these issues, here we implemented an intraoral OCE system based on a miniature probe for viscoelasticity quantification of both normal and cancerous tongue tissues. More detailly, a miniature-ultrasonic transducer (MUT) was employed to induce tissue deformation, which was detected by a swept source (SS) OCT system. Furthermore, a phase velocity algorithm and a Rayleigh wave-based Kelvin-Voigt model were employed to evaluate tongue viscoelasticity. The feasibility and reliability of the proposed OCE system were first validated on a heterogeneous phantom model. Subsequently, in vivo measurements were performed on normal and cancerous tongues of Sprague-Dawley (SD) rats, where the corresponding Young’s modulus and viscosity were calculated, respectively.
2. Materials and methods
2.1 OCE setup
The experimental apparatus used in this paper contains an SS-OCT system and an excitation unit, which is equipped with a custom-made intraoral probe, as seen in Fig. 1. A swept laser with a central wavelength of 1310 nm, an A-line speed of 50 kHz, a bandwidth of 100 nm was employed in the OCT system. The output light from the laser was split into two beams through a 99:1 fiber coupler. In the reference arm, the beam passed through a circulator (CIR-1310-50-APC, Thorlabs), a collimator (F260APC-C, Thorlabs), an attenuator, a lens, and a mirror, respectively. Meanwhile, the light was transmitted to the samples through a circulator, a collimator, a dual-axis galvo mirror (GVS102, Thorlabs) and a tailor-made intraoral probe in the sample arm. Subsequently, interference signals were formed after two beams from reference and sample arms entered a 50:50 optical coupler, which was then collected by a photodetector (PDB480C-AC, Thorlabs). The lateral and axial resolution of our apparatus were estimated to be 20 µm and 6.7 µm in air, respectively. The signal-to-noise ratio and field of view (FOV) were 95 dB and 5 mm × 5 mm, respectively. Moreover, the spatial resolution of the OCE system was measured to be 38.5 µm, which enables precise detection of wave propagation and velocity changes within the 5 mm FOV.
Fig. 1. Experimental configuration. (a) Schematic of the OCE system. (b) Intraoral probe composed of a scan lens and a miniature ultrasound transducer.
An intraoral probe (diameter of about 21 mm) composed of a scan lens (OCTH-LK30, Thorlabs) and a MUT was built to better carry out the measurements. The MUT was provided by Guangzhou Doppler Electronic Technologies Co Ltd., Guangzhou, China. The center frequency, bandwidth and diameter of the MUT were separately devised to be 930 kHz, 150 kHz and 1.8 mm. Besides, the focal spot diameter is approximately 1.6 mm, as measured by the full-width half-maximum of the normalized pressure profile. Meanwhile, a kind of harmless, specialized hydrogel medium was evenly applied to the transducer surface to facilitate sound wave propagation. A sine wave signal from a function generator (AFG31102, Tektronic) was employed to drive the transducer when it was amplified by an amplifier (2100L-1328, ENI). The excitation location was positioned very close to the 5 mm FOV to ensure that the local excitation focal point is coaligned with the OCE sample beam focused on the sample surface. In this configuration, the ultrasound transducer generated the excitation force between the 101st and the 120th A-line scans, and a total of 20 cycles were used for each excitation. In addition, the probe does not make any contact with oral tissues, thus eliminating the risk of causing harm to the mouth. On the other hand, the hydrogel used in this study meets medical standards, ensuring the safety of the detection process.
2.2 Sample preparation
Our experiments have been approved by the Animal Experiment Ethics Committee of Nanchang Hangkong University (20190316/v1.0). To verify reliability of the proposed methods, a heterogeneous agar phantom with 1% and 0.5% concentration was prepared according to the reference, as described below. [28] (1) 1.0 g of agar powder was added into 100 mL of distilled water, the resulting mixture was heated to 96 °C and then cooled to 60 °C, and it was added with 0.6% (v/v) intralipid solution to increase light scattering for OCT imaging. (2) The solution was poured into a mold for solidification, then half of the solidified phantom was removed. The vacated area was subsequently filled with the 0.5% (w/v) agar solution which was fabricated by using the same method as in step (1). To induce tongue cancer, [29] the carcinogen 4-nitro-quinoline-1-oxide (4-NQO) stock solution (Sigma-Aldrich, America) was prepared in 1, 2-propyleneglycol at a concentration of 200 mg/ml. Subsequently, it was diluted to a final concentration of 100 µg/ml in the drinking water. The freshly prepared solution was replaced on a weekly basis, and the bottled drinking water was shielded with aluminum foil to prevent light exposure. Twelve male SD rats, aged eight weeks, were randomly and evenly divided into control and experimental groups (n = 6), which were housed at room temperature with a relative humidity of 30%-50%. After baseline data acquisition using OCE, the rats in the experimental group received drinking water containing 100 µg/ml 4NQO to 24 weeks, while the control group consistently drank normal water containing an equivalent concentration of 1, 2-propylene glycol. Throughout the model construction process, weekly observations and recordings were made on food intake, water intake, and body mass of the rats. Morphological changes in tongue mucosa were observed and recorded every two weeks. Subsequently, OCE examinations were conducted on both groups at week 24. All rats were anesthetized by intraperitoneal injection during examination operation, which were euthanized with carbon dioxide inhalation at the conclusion of the experiment. The tongues were then excised and fixed in neutral formalin 10% for 24 hours. Subsequently, the samples were embedded in paraffin, sectioned and then were stained with hematoxylin and eosin (H&E). All H&E staining slides undergo meticulous examination by two seasoned pathologists.
2.3 Data acquisition and processing
The structural and viscoelastic information were obtained by M-B protocol. A total of 500 A-lines were involved in one M scan at each transverse position, which takes about 10 ms. Then, OCT beam was moved to the next position using the dual-axis galvo mirror. One B scan could be acquired after M scans at 750 locations have been collected, with a time of 7.5 s.
Subsequently, the achieved raw data was processed by phase-resolved Doppler OCT algorithm [30], and axial phase shift Δφ of two adjacent A-lines was obtained using the following Eq. (1):
2.4 Rayleigh wave-based viscoelasticity quantification
The tongue is assumed as a Kelvin-Voigt viscoelastic, homogeneous and nearly incompressible material, as a result, dynamic particle motion of the tissue can be described by the following Navier’s equation [31]:
Meanwhile, considering the status of free boundary in tongue tissue, Rayleigh wave could be suitable wave mode in our experiments. The solution of Rayleigh wave from Eq. (3) can be estimated using cylindrical coordinates, as follows [32]:
where kr and ks separately denote the wave number of Rayleigh wave and shear wave, and ${\beta _r} = \sqrt {k_r^2 - k_s^2} $. According to the Poisson’s ratio $\upsilon = 0.5$ in incompressible soft tissue, Eq. (5) can be further simplified to be ${c_r} \approx 0.95{c_s}$.In the meantime, the process of Rayleigh wave propagation is accompanied by energy attenuation, which is closely related to viscosity. According to equation $\omega = c(\omega )k$ and complex wavenumber $\kappa (\omega )= k(\omega )- i\alpha (\omega )$, the following equation can be obtained:
Consequently, attenuation coefficient α(ω) and frequency-dependent wave velocity cr(ω) can be calculated by the following equations:
The energy distribution map during the propagation process of Rayleigh wave can be acquired by applying 1D-FFT to the spatiotemporal displacement map. Then, the normalized energy distribution curve can be obtained using the energy around the strongest energy. The coefficient α(ω) can be estimated through exponential function curve fitting due to the assumption that energy decays exponentially. Meanwhile, the phase velocity cr(ω) calculation is repeated and then averaged depth-wise (over 0.2 ∼ 0.5 mm depending on sample optical properties), and then the shear modulus and viscous modulus are estimated at the main frequency, whose determination has been discussed in detal in our previous study. [28] Ulteriorly, Young’s modulus of the tongue can be achieved using the equation: $E = 2\mu ({1 + \nu } )$.
Besides, it’s thought that elastic wave propagation shows different characterization in soft and stiff areas when external incentives are applied to the organization, as described in Fig. 2. Therefore, the abnormal components involved in tissues can be recognized based on this hypothesis.
Fig. 2. Wave propagation characterization in soft and stiff regions. (a) Wave propagation at the beginning moment. (b) High and low speed propagation in two regions with time lapse.
3. Results
3.1 Heterogeneous phantom imaging
The tissue-mimicking phantom was used to verify the feasibility of the system and algorithm. Figure 3(a) shows the OCT B-scan image of the heterogeneous phantom, where the left part (1%) is darker than the right part (0.5%), resulting in a clear dividing line that is marked by an orange dotted line. Rayleigh wave propagation in the heterogeneous phantom at different time points is described in Figs. 3(b)-(e). It can be found that the excitation location was applied at the center of the heterogeneous materials, as shown in Fig. 3(b). The phase change caused by the MUT can be observed in Fig. 3(b) (indicated by a white arrow), where an elastic wave was induced and then propagated to both sides of the sample (indicated by white arrows in Fig. 3(c)). Then, spatiotemporal images of Rayleigh wave propagation can be achieved by reslicing the Doppler OCT images at all the data acquisition times, as depicted in Fig. 3(f). The slope difference between left and right sides of the demarcation line is relatively significant, indicating the ability of our system to distinguish tissues with different hardness.
Fig. 3. Results of heterogeneous phantom imaging. (a) The 2D OCT image. (b)-(e) Maps of Rayleigh wave propagation at different time points. (f) Spatiotemporal image of Rayleigh wave propagation. (g) Wavenumber-frequency distribution image (left) and phase velocity dispersion curve (right). (h) Power distribution map (left) and normalized spatial distribution curve of the peak intensity (right).
Subsequently, the Rayleigh wave velocity can be determined according to the phase velocity algorithm discussed in our previous studies. [28,33] The phase data in Fig. 3(f) was converted to displacement data followed by 2D Fourier transform. The resulted wavenumber-frequency distribution image and corresponding dispersion curves of phase velocity are demonstrated in Fig. 3(g), where Rayleigh wave speed in 1% phantom and 0.5% phantom were determined to be 4.37 ± 0.48 m/s at main frequency of 730 Hz (red curve) and 2.85 ± 0.24 m/s at main frequency of 708 Hz (blue curve), respectively. The results show that Rayleigh wave travels faster in the 1% phantom than in the 0.5% phantom. Meanwhile, a power distribution map can be acquired by applying 1D FFT transformation on Fig. 3(f), and then corresponding frequency-dependent α(ω) was estimated, as shown in Fig. 3(h). It’s suggested that Rayleigh wave energy dissipates more slowly in phantom with 1% concentration by comparing two intensity curves. Finally, Young’s modulus of 1% and 0.5% phantoms was evaluated to be 65.39 ± 1.78 kPa and 24.87 ± 1.95 kPa, respectively, and their corresponding viscous modulus was separately estimated to be 0.53 ± 0.09 Pa·s and 0.98 ± 0.06 Pa·s.
3.2 In vivo tongue imaging
After system verification, in vivo experiments on the control and experimental groups were conducted to investigate differences in viscoelastic properties between normal and cancerous tongues. Figures 4(a) and (b) respectively show the 2D OCT structure of a healthy tongue in the control group and a diseased tongue in the experimental group. To compare the viscoelastic differences, spatiotemporal images corresponded to Figs. 4(a)-(b) were provided, as depicted in Figs. 4(d)-(e), respectively. It can be found that the slope of Fig. 4(d) is steeper than that of Fig. 4(e), suggesting a comparatively lower stiffness of the normal tongue in comparison with the cancerous tongue.
Fig. 4. Results of in vivo tongue imaging. (a)-(c) The 2D OCT image of normal tongue (a), and cancerous tongue (b and c). (d)-(f) Spatiotemporal images of Rayleigh wave propagation.
Post-processing of the obtained spatiotemporal images was carried out to make a quantitative analysis of the difference, as shown in Fig. 5. The wavenumber-frequency distribution image (Fig. 5(a)) was first achieved by 2D-FFT applied on the spatial-temporal image. Subsequently, the dispersion curves of phase velocity for the normal tongue (blue curve) and cancerous tongue (red curve) were obtained (Fig. 5(b)), where the Rayleigh wave velocity is 2.54 ± 0.16 m/s (main frequency: 701 Hz) and 4.61 ± 0.68 m/s (main frequency: 707 Hz), respectively. At the same time, the power distribution image (Fig. 5(c)) and the normalized distribution profile of the wave energy intensity (Fig. 5(d)) were obtained to provide the viscous information. According to the exponential decay fitting, attenuation of the Rayleigh wave in the normal tongue was found to be over twice than that in the cancerous tongue. Finally, the Young’s modulus E (19.84 ± 0.92 kPa/normal tongue and 74.29 ± 1.63 kPa/cancerous tongue) and viscous modulus η (1.21 ± 0.09 Pa·s/normal tongue and 0.67 ± 0.02 Pa·s/cancerous tongue) were determined based on the Kelvin-Voigt viscoelastic model, as depicted in Fig. 5(e).
Fig. 5. (a) Wavenumber-frequency distribution image. (b) Phase velocity dispersion curves of normal and cancerous tongue. (c) Power distribution map. (d) The normalized spatial distribution curves of the peak intensity and the exponential fitting of normal and cancerous tongue. (e) The Young’s modulus and viscosity of normal and cancerous tongue.
Figure 4(c) visualizes more detailed structural information of the cancerous tongue, where the normal region (indicated by a white arrow), transitional region (recognized by a rectangular dotted box) and cancerous region (indicated by a red arrow) can be clearly distinguished. Moreover, there is an obvious difference in tongue thickness among these three regions. A further in-depth study concerning viscoelasticity of the cancerous tongue was also carried out since it is very significant to establish a cutoff value for tumor tissue. [34] It can be observed in the spatiotemporal image (Fig. 4(f)) that there exists an obvious discrepancy in the slope, which reflects the elastic differences among the normal, transitional and cancerous tissues in Fig. 4(c).
Subsequently, detailed viscoelastic properties were calculated using the Rayleigh wave-based model, whose results are summarized in Table 1. Velocity of the three regions was measured to be 2.57 ± 0.32 m/s, 3.44 ± 0.56 m/s, and 4.61 ± 0.68 m/s, respectively, and then their viscoelasticity was separately calculated as well. It’s revealed that Young’s modulus and viscosity values of the transitional tissue are within the middle range between the normal and cancerous tissues, and maximum Young’s modulus and minimum viscosity exist in the cancerous tissue. Besides, Rayleigh wave velocity of the baseline data was calculated to be 2.49 ± 0.46 m/s (main frequency: 719 Hz), and then corresponding Young’s modulus and viscosity were separately estimated to be 21.06 ± 0.74 kPa and 1.02 ± 0.03 Pa·s. The result is nearly identical to that in the normal region.
Table 1. The experimental data of in vivo rats’ tongue.
To offer more information about morphological alteration in the cancerous tongue, a 3D image was reconstructed (Fig. 6(a)). Besides, Fig. 6(b) demonstrates the thickness of the normal, transitional, and cancerous regions, which was measured to be 1.11 ± 0.04 mm, 1.34 ± 0.13 mm, 1.48 ± 0.11 mm, respectively. Moreover, there is little variance between the normal tongue of the control group (1.12 ± 0.07 mm) and the normal region of the cancerous tongue in tongue thickness. After OCE experiments, all rats were euthanized with carbon dioxide inhalation and the tongues were then excised and prepared for histological analysis. The corresponding H&E staining result is shown in Fig. 7(b), which indicated that the tongue cancer model was successfully established.
Fig. 6. The morphological results. (a) 3D image of the cancerous tongue. (b) Different thickness of the tongue in control and experimental groups.
Fig. 7. The results of histological study. (a) Photography of the cancerous tongue (the cancerous region is labeled with a white box). (b) Histopathological image of the cancerous tongue (HE × 400).
4. Discussion
The purpose of this preliminary study is to investigate viscoelastic properties of cancerous tongue using an OCE system based on an intraoral probe, and therefore to afford possible evidence for early diagnosis of tongue carcinoma. As far as we know, it’s the first time that OCE measurements with high resolution, high sensitivity, and non-invasive imaging have been applied in cancerous tongue research. While intraoral OCT, acoustic radiation force excitation, and the Kelvin-Voigt model are indeed established techniques, their integration for intraoral OCE aimed specifically at assessing the mechanical properties of the tongue represents a novel application not previously explored. This unique combination allows us to non-invasively probe and quantify viscoelastic properties of the tongue with high precision, a significant advancement over existing methods that have not yet been applied to the early detection of oral cancer. Furthermore, the mechanical index and the spatial peak pulse average intensity of the MUT were measured to be 1.06 and ∼10 W/cm2, respectively, which are all within the FDA standardization. The system was first validated by a heterogeneous phantom model, where the wave velocity variation due to concentration change was observed. As a result, corresponding difference in elastic property was revealed, in which the ratio of Young’s modulus in the 1% phantom to that in the 0.5% phantom was calculated to be 2.63. In contrast, it was found that when the concentration was increased, the changing trend of their viscous modulus showed in opposite direction.
To systematically evaluate viscoelastic properties of tongue, in vivo imaging was performed on the normal and cancerous tongues that respectively belong to the control and experimental groups. The 2D OCT structural images showed a greater thickness and more ambiguous boundary among different layers in the cancerous tongue, which can be attributable to the hyperplasia and increased irregularity of cancer cells. [35] In addition to the structural change, the results revealed a distinguished contrast in biomechanical properties between the two kinds of tongues. The mean Young’s modulus of the cancerous tongue is more than three times bigger than that of the normal tongue, which may be associated with changes in collagen concentration and alignment of diseased tissue. [36] It is notable that a similar conclusion was obtained in breast cancer study by using OCE, [25,37] supporting the efficacy and reliability of this technique in tongue cancer investigation. Meanwhile, the viscoelastic differences within the cancerous tongue were carefully researched, where the normal, transitional and cancerous regions can be distinguished by values of the Young’s modulus and viscosity. The information may be very helpful to establish a cutoff value for cancerous tissue. Furthermore, elasticity of the normal, transitional, and cancerous tissues increased gradually, which is in accordance with previous findings reported by ultrasound elastography. [38,39] It is also worth noting that the viscous information of the cancerous tongue was first achieved by using the OCE system, which may be an effective indicator for diagnosis and prognostic evaluations of tongue carcinoma according to the above discussion.
Although our method shows relatively good application ability, there are still some limitations need to be addressed. Firstly, the process of pathological changes has not been monitored yet in this preliminary attempt and thus experiments based on a larger sample size should be carried out to establish more improved viscoelastic models of different cancer stages. Based on the results, we will determine the range of changes in the viscoelastic values of tongue tissues corresponding to each stage of tongue cancer, thus providing a reliable basis for the assessment of tongue cancer. Secondly, the imaging depth, spatial resolution and scanning range of the proposed system should be enhanced to achieve better measurement results, which can be resolved well by system modification based on the OCT principle. [40,41] Thirdly, the frequency used for curve fitting in the KV model should be selected more comprehensively to ensure accuracy and relevance in estimating the viscoelastic properties of the tissue. Fourthly, the image reconstruction algorithm must be further optimized to achieve real-time two-dimensional mapping of tissue viscoelasticity. This could enable doctors to directly read the viscoelastic values of suspected areas in real-time images and thus make immediate decisions. Last but not least, the intraoral probe designed in this paper should be further optimized to improve its utility properties. On the one hand, a MEMS mirror can be integrated with the probe to achieve a hand-held exterior structure. On the other hand, the MUT can be substituted by an air-coupled ultrasonic transducer.
5. Conclusion
In conclusion, a novel method that combined an intraoral OCE system, a probe design and a Rayleigh wave-based Kelvin-Voigt model was proposed to evaluate biomechanical properties of tongue tissues. After the verification process was completed, it was applied successfully in rat experiments, where distinct viscoelasticity differences of normal and cancerous tongues were obtained. Furthermore, differences in viscoelastic properties within a cancerous tongue were achieved, which allows for easy identification of normal, transitional and cancerous regions. To our knowledge, this paper presented a new application of OCE, which might provide an ideal means for study of tongue cancer.
Funding
National Natural Science Foundation of China (12164028, 51863016,12064029).
Acknowledgments
Financial support from the National Natural Science Foundation of China (12164028, 51863016, 12064029).
Ethics Policies
The protocols and handling of the animals had been approved by [Nanchang Hangkong University (20190316/v1.0)].
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. M. Tang, W. Dai, H. Wu, et al., “Transcriptome analysis of tongue cancer based on high-throughput sequencing,” Oncol. Rep. 43(6), 2004–2016 (2020). [CrossRef]
2. I. A. El-Naaj, Y. Leiser, M. Shveis, et al., “Incidence of oral cancer occult metastasis and survival of T1-T2N0 oral cancer patients,” Journal of oral and maxillofacial surgery 69(10), 2674–2679 (2011). [CrossRef]
3. I. Van der Waal, R. de Bree, R. Brakenhoff, et al., “Early diagnosis in primary oral cancer: is it possible?” Medicina oral, patologia oral y cirugia bucal 16(3), e300–e305 (2011). [CrossRef]
4. S. Abati, C. Bramati, S. Bondi, et al., “Oral cancer and precancer: a narrative review on the relevance of early diagnosis,” Int. J. Environ. Res. Public Health 17(24), 9160 (2020). [CrossRef]
5. M. F. A. Ribeiro, M. C. M. Oliveira, A. C. Leite, et al., “Assessment of screening programs as a strategy for early detection of oral cancer: A systematic review,” Oral Oncol. 130, 105936 (2022). [CrossRef]
6. C. Mannelli, “Tissue vs liquid biopsies for cancer detection: ethical issues,” Journal of bioethical inquiry 16(4), 551–557 (2019). [CrossRef]
7. A. Cykowska, L. Marano, A. D’Ignazio, et al., “New technologies in breast cancer sentinel lymph node biopsy; from the current gold standard to artificial intelligence,” Surgical Oncology 34, 324–335 (2020). [CrossRef]
8. V. M. L. de Sousa and L. Carvalho, “Heterogeneity in lung cancer,” Pathobiology 85(1-2), 96–107 (2018). [CrossRef]
9. T. Hirahata, R. ul Quraish, A. U. Quraish, et al., “Liquid biopsy: a distinctive approach to the diagnosis and prognosis of cancer Cancer,” Cancer Inf. 21, 117693512210760 (2022). [CrossRef]
10. D. H. Kim, S. W. Kim, J. Lee, et al., “Narrow-band imaging for screening of oral premalignant or cancerous lesions: A systematic review and meta-analysis,” Clinical Otolaryngology 46(3), 501–507 (2021). [CrossRef]
11. J. Wan, M. L. Oblak, A. Ram, et al., “Determining agreement between preoperative computed tomography lymphography and indocyanine green near infrared fluorescence intraoperative imaging for sentinel lymph node mapping in dogs with oral tumours,” Vet. Comp. Oncol. 19(2), 295–303 (2021). [CrossRef]
12. J. Xiong, J. Feng, L. Qiu, et al., “SDF-1-loaded PLGA nanoparticles for the targeted photoacoustic imaging and photothermal therapy of metastatic lymph nodes in tongue squamous cell carcinoma,” Int. J. Pharm. 554, 93–104 (2019). [CrossRef]
13. C. van Lanschot, T. B. Schut, E. Barroso, et al., “Raman spectroscopy to discriminate laryngeal squamous cell carcinoma from non-cancerous surrounding tissue,” Lasers Med. Sci. 38(1), 193 (2023). [CrossRef]
14. S. P. Sunny, S. Agarwal, B. L. James, et al., “Intra-operative point-of-procedure delineation of oral cancer margins using optical coherence tomography,” Oral Oncol. 92, 12–19 (2019). [CrossRef]
15. E. Auksorius, D. Borycki, P. Stremplewski, et al., “In vivo imaging of the human cornea with high-speed and high-resolution Fourier-domain full-field optical coherence tomography,” Biomed. Opt. Express 11(5), 2849–2865 (2020). [CrossRef]
16. M. Arya, O. Sorour, J. Chaudhri, et al., “Distinguishing intraretinal microvascular abnormalities from retinal neovascularization using optical coherence tomography angiography,” Retina 40(9), 1686–1695 (2020). [CrossRef]
17. W. Yuan, L. Cheng, J. Yang, et al., “Noninvasive oral cancer screening based on local residual adaptation network using optical coherence tomography,” Med. Biol. Eng. Comput. 60(5), 1363–1375 (2022). [CrossRef]
18. V. Panzarella, F. Buttacavoli, A. Gambino, et al., “Site-Coded Oral Squamous Cell Carcinoma Evaluation by Optical Coherence Tomography (OCT): A Descriptive Pilot Study,” Cancers 14(23), 5916 (2022). [CrossRef]
19. S. Kawashiri, A. Tanaka, N. Noguchi, et al., “Significance of stromal desmoplasia and myofibroblast appearance at the invasive front in squamous cell carcinoma of the oral cavity,” Head & Neck: Journal for the Sciences and Specialties of the Head and Neck 31(10), 1346–1353 (2009). [CrossRef]
20. M. Shibata, A. Ishikawa, J. Ishii, et al., “Stiffness of tongue squamous cell carcinoma measured using strain elastography correlates with the amount of collagen fibers in the tumor,” Oral Radiol. 38(2), 278–287 (2022). [CrossRef]
21. G. Lan, M. Singh, K. V. Larin, et al., “Common-path phase-sensitive optical coherence tomography provides enhanced phase stability and detection sensitivity for dynamic elastography,” Biomed. Opt. Express 8(11), 5253–5266 (2017). [CrossRef]
22. J. Zhu, X. He, and Z. Chen, “Acoustic radiation force optical coherence elastography for elasticity assessment of soft tissues,” Appl. Spectrosc. Rev. 54(6), 457–481 (2019). [CrossRef]
23. G. Lan, S. R. Aglyamov, K. V. Larin, et al., “In-vivo human corneal shear-wave optical coherence elastography,” Optometry and Vision Science 98(1), 58–63 (2021). [CrossRef]
24. L. Chin, B. Latham, C. M. Saunders, et al., “Simplifying the assessment of human breast cancer by mapping a micro-scale heterogeneity index in optical coherence elastography,” J. Biophotonics 10(5), 690–700 (2017). [CrossRef]
25. E. V. Gubarkova, A. A. Sovetsky, D. A. Vorontsov, et al., “Compression optical coherence elastography versus strain ultrasound elastography for breast cancer detection and differentiation: pilot study,” Biomed. Opt. Express 13(5), 2859–2880 (2022). [CrossRef]
26. Y. Zhang, J. Luo, X. Han, et al., “Quantitative Evaluation of Biomechanical Properties of the Tongue Using a Shaker-Based Optical Coherence Elastography System,” Photonics 10(10), 1143 (2023). [CrossRef]
27. A. Nair, C. H. Liu, M. Singh, et al., “Assessing colitis ex vivo using optical coherence elastography in a murine model,” Quant. Imaging Med. Surg. 9(8), 1429–1440 (2019). [CrossRef]
28. G. Liu, Y. Zhang, Y. Wang, et al., “Point-to-point optical coherence elastography using a novel phase velocity method,” Optics and Lasers in Engineering 163, 107489 (2023). [CrossRef]
29. K. Danial, H. Ahmed, and T. Hamid, “Exploring animal models in oral cancer research and clinical intervention: A critical review,” Vet. Med. Sci. 9(4), 1833–1847 (2023). [CrossRef]
30. Y. Li, S. Moon, J. J. Chen, et al., “Ultrahigh-sensitive optical coherence elastography,” Light: Sci. Appl. 9(1), 58 (2020). [CrossRef]
31. S. P. Kearney, A. Khan, Z. Dai, et al., “Dynamic viscoelastic models of human skin using optical elastography,” Phys. Med. Biol. 60(17), 6975–6990 (2015). [CrossRef]
32. A. C. Saavedra, F. Zvietcovich, R. J. Lavarello, et al., “Measurement of surface acoustic waves in high-frequency ultrasound: Preliminary results,” 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), IEEE, 3000–3003 (2017).
33. Y. Wang, Y. Zhang, G. Shi, et al., “Evaluation of Residual Corneal Stromal Bed Elasticity by Optical Coherence Elastography Based on Acoustic Radiation Force,” Photonics 10(3), 266 (2023). [CrossRef]
34. D. Aktar, H. Tomioka, H. Hirai, et al., “Value of real-time tissue elastography for squamous cell carcinoma of the tongue,” Dent Oral Craniofac Res 2(6), 371–375 (2016). [CrossRef]
35. E. Epstein, D. Fischerova, L. Valentin, et al., “Ultrasound characteristics of endometrial cancer as defined by International Endometrial Tumor Analysis (IETA) consensus nomenclature: prospective multicenter study,” Ultrasound in Obstetrics & Gynecology 51(6), 818–828 (2018). [CrossRef]
36. S. M. Zunder, H. Gelderblom, R. A. Tollenaar, et al., “The significance of stromal collagen organization in cancer tissue: An in-depth discussion of literature,” Critical Reviews in Oncology/Hematology 151, 102907 (2020). [CrossRef]
37. E. V. Gubarkova, A. A. Sovetsky, V. Y. Zaitsev, et al., “OCT-elastography-based optical biopsy for breast cancer delineation and express assessment of morphological/molecular subtypes,” Biomed. Opt. Express 10(5), 2244–2263 (2019). [CrossRef]
38. M. Shibata, A. Ishikawa, J. Ishii, et al., “Stiffness as measured with strain elastography is a prognostic factor for pT1/T2 tongue squamous cell carcinoma with muscle-layer invasion,” Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology 135(4), 558–566 (2023). [CrossRef]
39. I. Ogura, Y. Sasaki, M. Sue, et al., “Strain elastography of tongue carcinoma using intraoral ultrasonography: a preliminary study to characterize normal tissues and lesions,” Imaging science in dentistry 48(1), 45–49 (2018). [CrossRef]
40. S. Ishida and N. Nishizawa, “Quantitative comparison of contrast and imaging depth of ultrahigh-resolution optical coherence tomography images in 800–1700 nm wavelength region,” Biomed. Opt. Express 3(2), 282–294 (2012). [CrossRef]
41. T. Hirano, S. Kakihara, Y. Toriyama, et al., “Wide-field en face swept-source optical coherence tomography angiography using extended field imaging in diabetic retinopathy,” Br. J. Ophthalmol. 102(9), 1199–1203 (2018). [CrossRef]
