Open Access
Add to BibSonomy
Abstract
Tomographic imaging of soft tissue such as skin has a potential role in cancer detection. The penetration of infrared wavelengths makes a confocal approach based on laser feedback interferometry feasible. We present a compact system using a semiconductor laser as both transmitter and receiver. Numerical and physical models based on the known optical properties of keratinocyte cancers were developed. We validated the technique on three phantoms containing macro-structural changes in optical properties. Experimental results were in agreement with numerical simulations and structural changes were evident which would permit discrimination of healthy tissue and tumour. Furthermore, cancer type discrimination was also able to be visualized using this imaging technique.
© 2017 Optical Society of America
1. Introduction
Tomography, the visualization of slices from the interior of an object, and in particular its applications in biomedical imaging have attracted significant interest over the past few decades [1, 2]. Optical tomography techniques such as diffuse optical tomography [3–5] and optical coherence tomography [6–8] use visible or near-infrared light to map the optical scattering from biological tissues. These techniques are safe and low-cost compared to other tomography techniques [7] and are able to image different functional and morphological features of biological tissues [9].
One relatively unexplored optical technique which can be exploited for cross-sectional imaging of biological tissues is confocal laser feedback tomography (LFT). This technique is based on laser feedback interferometry (LFI) [10–14] in which the emitted light from a laser will be partially reflected from an external target and re-enters the laser cavity where it interferes with the intra-cavity fields and creates an interferometric signal. LFI is an uncomplicated and compact configuration as it has one arm for both incident and detection beams. While there have been many reports of LFI for three-dimensional (3D) applications such as surface reconstruction of objects [15–17], in-depth cross-sectional imaging [18, 19], and 3D imaging of the interior of a microfluidic device [20], it has not been applied to biomedical 3D imaging and tomographic applications. For the first time, we propose a confocal LFT for 3D and cross-sectional imaging of biological tissues. This solution presents a low-cost and compact platform for tissue tomography with potential clinical roles.
Skin tissue phantoms with inhomogeneities emulating malignant changes provide an excellent test bed for the evaluation of the proposed technique in biomedical context. In this article we explore the performance of the proposed confocal LFT system on skin tissue phantoms designed to realistically model the optical properties of skin tissue with embedded keratinocyte carcinoma (KC) lesions (also referred to as nonmelanoma skin cancers). Three-layer silicone-based phantoms were used as models for different types of KCs. Models of cancer affected regions were incorporated into these phantoms as volumes of altered optical properties situated in the second layer [21]. Results show that confocal LFT can provide cross-sectional images of the tissue phantoms at depths of about 0.75 mm and is sensitive enough to detect just 10% reduction in the scattering coefficient (μs) of the second layer. To assess the technique’s criteria and parameters before conducting the experiments, we performed a Monte Carlo (MC) simulation. Numerical outcomes were used to optimize the experimental design. Then it was shown that experimental results conform well with the numerical results, validating the idea of using confocal LFT in such an application.
Skin cancer is the most common type of cancer among fair-skinned people [22]. Although melanoma is the most serious type of skin cancer, the incident rate of KC is much higher than melanoma and it can also be aggressive if not detected at early stages [22–24]. The proposed technique is equally applicable to different types of skin cancer, however we developed skin tissue phantoms which are specific to one type of cancer and precisely replicate the optical properties of that type. We chose KC as it is the most prevalent type of skin cancer, however a parallel study could be equally executed for melanoma. This work investigates the feasibility of a new optical technique which may have a potential role in skin cancer detection.
The remainder of this paper is structured as follows: section 2 discusses the methodology; configuration of confocal LFT and skin cancer tissue phantom preparation techniques. In section 3 a MC numerical model of the optical system and tissue phantoms are presented. Section 4 includes the experiments and results, followed by section 5 where more discussion is included. Finally, section 6 presents our conclusions and summary.
2. Methodology
2.1. Confocal laser feedback tomography
Confocal LFT is based on an LFI technique [10–14]. Sensing in this technique occurs when a small fraction of the laser beam reflects back from a distant target into the laser cavity and interferes with the strong intra-cavity fields, therefore modulating both the amplitude and frequency of the lasing field. This technique has a self-coherent nature which results in high sensitivity, as it only responds to the radiation of its own. The back-reflected beam contains information about the target which can be obtained by monitoring the laser’s output optical power or terminal voltage. The changes in the output optical power can be detected using a photodetector, while the fluctuations in the laser terminal voltage can be directly measured. LFI Signal depends strongly on the level of optical feedback and its characteristics can be categorized into different regimes based on the amount of light which reflects back from the external target [14]. Confocal LFT is self-aligned; we only need to focus the beam into the tissue and back reflection to the laser occurs through the same lensing system. We used a vertical cavity surface emitting laser (VCSEL) in this system. There is no need for a detection arm as detection takes place within the VCSEL’s cavity and signal can be obtained from the VCSEL’s terminal voltage or photocurrent of an in-built photodiode situated inside the VCSEL’s package. Although this phenomenon can be observed in different types of lasers, laser diodes are mainly used as the source/detector in such systems [11, 25]. LFI can be used in general remote sensing applications such as distance, displacement, vibration, and velocity measurements [11, 26–28], but two applications that show its potential functionality in biological tissue tomography are 3D imaging and microscopy applications. For instance, LFI has been used for 3D surface reconstruction of objects [15–17], and also in non-biological [29–32] and biological [33, 34] microscopy applications where it can easily achieve sub-micrometer lateral resolutions. In one of the first demonstrations, Juškaitis et al. represented a simple and compact confocal microscope using just a laser diode and a lensing system based on LFI technique [30].
In a reflectance confocal microscope, a fraction of the back-reflected beam coming from a minute spatial sensing volume around the focal point, which passes through a pinhole situated on its return path, is detected usually by photomultiplier tubes or avalanche photodiodes. The pinhole acts as a spatial filter to eliminate out of focus beams [35–37]. To achieve optical sectioning, we need to scan the focal point over the target at a constant depth. Confocal LFT using a semiconductor laser is a special embodiment of a confocal microscope where the VCSEL aperture (with a diameter normally in the range of a few micrometer) serves as a confocal pinhole. The main difference is that in a conventional confocal microscope, the sensed signal is the amount of back-reflected light detected by a photodiode. The sensed signal in a confocal LFT is as a result of interference in the laser cavity. Figure 1 illustrates the configuration of the confocal LFT used in this work and how VCSEL’s aperture acts as a confocal pinhole to eliminate out of focus beams. The VCSEL operates at 850 nm, within the biological window, to avoid high absorption by water and oxygenated and deoxygenated haemoglobin, as well as to benefit from maximum penetration into biological soft tissues [3]. Furthermore, the VCSEL operates just above the threshold current and incident optical power on the target is in the range of a few hundreds of micro-watts which makes it a safe and low-power device. Being safe, low-power, and compact, the confocal LFT has potential to be used as a portable tool for superficial cancer detection purposes. We have previously suggested confocal LFI for skin cancer detection purposes [38, 39] and its versatility is shown in a dual-modality imaging application [40].
Fig. 1 Confocal laser feedback tomography apparatus where the VCSEL’s aperture acts as a confocal pinhole to eliminate out of focus beam. VCSEL current is modulated at a low level around the operating average current to create two levels of LFI signal. Beam is focused on a silicone-based three-layer KC tissue phantom with a cylindrical cancerous region enclosed in the dermis layer.
2.2. Keratinocyte carcinoma tissue phantom
Human skin is a multi-layer tissue and mainly consists of three layers: epidermis, dermis, and subcutaneous tissue. Epidermis, the outermost layer, is composed of layers of keratinocytes and in the basal layer contains melanin, the main absorber of visible light, while dermis is composed of collagen and elastin and does not contain melanin and has a lower absorption coefficient (μa). Subcutaneous tissue is the innermost layer and consists of fat cells. Optical properties of human tissue have already been studied extensively [41]. However, universal hallmarks of malignant change such as angiogenesis [42], abnormal cell nucleus structure [43] (as the main scatterer of light), increased nuclear to cell volume [44], increased extra cellular fluid from leaky cancerous blood vessels [45], and other morphological and molecular changes will alter the optical properties of an affected skin tissue [21]. These micro- and macro-structural changes are considered as biomarkers and used in many optical systems for diagnostic and treatment purposes.
This work focuses on detecting regions at altered optical properties in three types of KCs through a series of examination on KC tissue phantoms. We took the values of the scattering coefficients, μss, of epidermis, dermis, and subcutaneous tissue phantom layers to be 23, 13, and 13 mm−1, respectively, and absorption coefficients, μas, of them to be 0.4, 0.1, and 0.1 mm−1, respectively [46]. We included the cancer affected regions as cylindrical regions situated at and enclosed by the second layer (dermis), with a diameter of 3 mm, and a thickness similar to dermis layer. Diameters of the cancerous models were in the range of typical KC precursors [47]. We considered reduction in the optical properties of the cancerous models with respect to dermis layer. In [21], optical properties of infiltrative basal cell carcinoma (infBCC), nodular basal cell carcinoma (nodBCC), and squamous cell carcinoma (SCC) samples were measured over an extended spectral range, and it was shown that at 850 nm the μss of infBCC, nodBCC, and SCC drop by factors of 0.9, 0.7, and 0.6, respectively, and μas of them drop by factors of 0.8, 0.2, and 0.4, again respectively, all with respect to dermis tissue optical properties [21]. We applied these levels of decrease in the optical properties of the cancerous regions in three types of KC tissue phantoms.
We chose to fabricate silicone-based tissue phantoms due to their stability and good performance in providing homogeneous samples. The base and curing agent for two-part polydimethylsiloxane (PDMS), which is a room-temperature vulcanizing silicone, were used (Sylgard® 184 Silicone Elastomer Kit). PDMS refractive index is about 1.4, in the range of near-infrared. This is close to its value for human skin tissue [48]. PDMS is a clear material with negligible μs and μa. Therefore, we used titanium dioxide (TiO2) and India ink to control its optical properties. TiO2 and India ink are among the most accepted scatterers and absorbers of light in this spectral range [49]. We purchased commercially available TiO2 (Sigma-Aldrich, Corp.) and India ink (Winsor & Newton, Co.). TiO2’s μa and India ink’s μs can be neglected as they are much smaller than their other optical properties. We used the Beer–Lambert technique [39] and measured the μs and μa of pure TiO2 and India ink, which were about 380 and 290 mm−1, respectively.
To make the base silicone gel for each one of the layers, we added 2.5 g of PDMS curing agent to 25 g of base part (1:10 proportion). We added TiO2 and India ink to the base gel according to the calculated proportions to make realistic optical properties for each one of the layers. TiO2 was added in a weight-to-weight (w/w) proportions using a 0.1 mg precision balance, while India ink was added using a 0.2 μL precision pipette. After adding TiO2 and India ink we stirred the mixture mechanically for about 30 minutes to achieve a uniform distribution of scatterers and absorbers. In order to make the layers, we poured the mixture between acrylic glass sheets and used adhesive tape or other shims between the sheets to achieve specific thicknesses. Layers were then left for at least two days to cure. We chose cast acrylic sheet because it was easy to remove the layers after curing process. We built epidermis, dermis, and subcutaneous tissue layers with thicknesses of 80, 220, and 1100 μm, respectively, and the thicknesses of cancerous models are the same as epidermis thickness. We placed a 150 μm glass cover slip on top of the KC tissue phantoms as an anti-diffusive reflection layer. Schematic diagram of a KC tissue phantom can be seen in Fig. 1.
3. Simulations
3.1. Monte Carlo model
Monte Carlo simulation is a powerful technique for modeling light-tissue interactions and characterizing optical techniques [50]. We performed a MC simulation to assess the system criteria and examine the experimental parameters before conducting the experiments. We used an in-house object-oriented MC engine prepared in MATLAB (R2015a, academic license). To perform MC modeling, we numerically defined the optical system (laser and lenses), beam (a population of photons moving in 3D space), and KC models. Numerical models for infBCC, nodBCC, and SCC were defined with the exact geometrical shapes and optical properties of that for the KC phantoms, as explained in section 2.2. We took the refractive indexes of the models for cover slip and layers to be 1.35 for the purpose of index matching. A population of 0.2 million photons was defined with a spatial Gaussian distribution on the objective lens. Photons were launched toward the targets in a hyperboloid geometry [51]. A portion of the photons enters the KC models and interacts with scatterers and absorbers. We used a Henyey–Greenstein phase function with anisotropy factor of 0.85 to define scattering angles. A portion of the photons exits the KC models, passes through the lenses, and reenters the laser aperture which we considered as detected photons. Lenses were defined based on the physical characteristics of the ones we wanted to use in the experiments (C330TME-B and C240TME-B) released by Thorlabs incorporation. Numerical structure was precisely as in the setup shown in Fig. 1. Finally, numerical model was scanned over the KC models in a raster scanned style. A volume of 4.5×4.5×0.9 mm in 37×37×31 steps, at pitches of 125×125×30 μm was numerically scanned and at each point the number of detected photons was considered as the reflection signal level.
3.2. Results
By raster scanning the optical system over the tissue model, a 3D matrix of signal can be created from stacking two-dimensional (2D) scans at different depths. This 3D matrix is associated with the volume of the tissue model enclosing a tumorous model. We used isosurface of points at a constant signal level (CSL) within the 3D matrix of signal to visualize the boundaries of the tumorous area. In this context, an isosurface is a surface which renders the region at or below a given level of signal in the 3D space and a CSL is the level of signal which remains the same over the isosurface. Reduction in the optical properties of the affected area causes a decrease in the signal value and this implies that for an appropriate choice of CSL, the isosurface should display the boundary of the KC model. Figures 2(a) to 2(c) show (i) oblique, (ii) top, and (iii) side views of the isosurfaces for infBCC, nodBCC, and SCC numerical models, respectively.
Fig. 2 Numerical isosurfaces for (a) infBCC, (b) nodBCC, and (c) SCC, representing the boundaries of affected regions with relative changes in optical properties, when (i), (ii), and (iii) represent the oblique, top, and side views, respectively.
4. Experiments
4.1. Experimental setup
Experimental setup is as shown in Fig. 1. In this setup we used an 850 nm VCSEL (Litrax Technology Co., Ltd.) operating at 3.55 mA, just above the threshold current (at about 3.4 mA). VCSEL was temperature controlled at 35°C to have a better LFI signal [52]. Beam was collimated (using a C240TME-B, Thorlabs, Inc. lens) and focused (using a C330TME-B, Thorlabs, Inc. lens) on the target. Focused beam had a focal length and numerical aperture of 3.1 mm and 0.68, respectively, and a focal beam spot diameter and Rayleigh length of about 1.8 and 2.9 μm, respectively. Laser and lensing system were placed on a three axis translation stage (Zaber Technologies Inc.). The perpendicular beam was then raster scanned over the phantoms. The scanned area of each of the optical sections was 4.5×4.5 mm in 91×91 steps at 50 μm pitch (lateral resolution), and when complete, the focal point was moved deeper into the phantom by 20 μm (axial resolution) to the next optical section. After 46 steps an axial depth of 0.9 mm was scanned. Scans were started from the surface of the phantoms (surface of the cover slips) and a volume of 4.5×4.5×0.9 mm of the phantom was scanned (similar to the MC numerical model). VCSEL current was modulated at a low level around the operating current (as shown in Fig. 1) to create two levels of feedback which was then extracted as the LFI signal to quantify the reflection level [53]. Frequency of current modulation was 5 kHz. Signal was acquired from an in-built photodiode in the VCSEL package, using a data acquisition card at a rate of 100 kS/s. At each point 1000 samples of the signal were acquired and a fast Fourier transform was applied to the time series to convert the signal into frequency domain. LFI signal was then extracted at the frequency of modulation (5 kHz).
Signal should be depth compensated for attenuation as it penetrates the phantom in order to reconstruct the 3D images. In addition, we needed to consider proportional changes in the maximum range of the signal at each of the optical sections (the range between the maximum and minimum values of signal at an optical section). Therefore, we first normalized the signal levels for each of the optical sections relative to the maximum signal in the corresponding section. Then we measured the average signal levels at each of the optical sections and multiplied them by a reversed attenuation factor to compensate for the drop in their levels. Figure 3 shows the average signal levels, maximum range of the signal at optical sections, and depth compensated maximum range versus the scan depth from the phantom surface. Higher maximum ranges of signal in Fig. 3 occur at depths where cancerous models are incorporated and are indicators of strong changes in the optical properties.
Fig. 3 Average signal (red dotted line), maximum range of the signal at optical sections (green dashed line), and depth compensated maximum range (blue solid line) versus the scan depth from the top of the cover slip (all normalized to the maximum signal value across the depth of scan), for three types of KC tissue phantoms of (a) infBCC, (b) nodBCC, and (c) SCC. Shaded areas show the approximate situated depths of the KC models.
4.2. Results
Figure 4 shows the scan results at each one of the optical sections, starting at just above the top surface of the epidermis layer in the SCC phantom and ending at the depth of 400 μm, comprising 21 images. A 2D median filter has been applied to provide these images and they have been compensated for the drop in the level of signal and dynamic range (as discussed in section 4.1). They have also been normalized to the maximum level of signal at each one of the slices. The first five slices occur in the epidermis layer model (80 μm thick) where there is no cancerous region and low contrast circular drop in the level of signal is due to the cancerous model at deeper areas. Images at depths between 100 and 280 μm sliced the cancerous model (with lower optical properties) and are situated in the dermis layer (220 μm thick). Higher contrast can be seen in these slices. Other slices are situated in the subcutaneous tissue. A small high contrast area on the right side of the cancerous model in first images in Fig. 4, is due to an air gap between the epidermis and dermis layers where these layers are not completely adhered together.
Fig. 4 Images provided at different depths of SCC phantom, starting at just above the top surface of the epidermis layer and ending at the depth of 400 μm in subcutaneous tissue. Stronger contrast can be seen at depths where cancerous model is incorporated, enclosed in the dermis layer.
In a similar way as described in section 3.2, a 3D signal matrix can be derived from the experimental scans and isosurfaces representing the boundaries of the KC models within the tissue phantoms can be depicted. Figure 5(a) to 5(c) show (i) oblique, (ii) top, and (iii) side view of isosurfaces for infBCC, nodBCC, and SCC phantoms, respectively. The 3D volumes represented in this figure render the boundaries between the healthy and diseased regions in the tissue phantoms. Larger relative reduction in the optical properties of the cancer model results in higher contrast and consequently a more solid isosurface (the case in Fig. 5(c)), while smaller relative reduction results in lower contrast and a less solid isosurface (the case in Fig. 5(a), as infBCC has only 10 % reduction in its scattering coefficient and can also be considered as the sensitivity of the technique) where noise may drown out the signal area. Note that all the phantoms in Fig. 5 have identical geometry for the layers and enclosed affected volumes. It can be seen that experimental results are quantitatively similar to numerical results in Fig. 2 validating the outcome of the experiments. Figure 5(a) has a higher noise level compared to Fig. 2(a), which is expected as the experimental model is noisier as is its numerical counterpart.
Fig. 5 Isosurfaces for (a) infBCC at CSL of about 0.7, (b) nodBCC at CSL of about 0.5, and (c) SCC at CSL of about 0.3. (i), (ii), and (iii) represent the oblique, top, and side views, respectively.
In order to find an optimized CSL to create an isosurface separating normal and diseased regions, we plotted isosurfaces in the 3D signal matrix for CSLs from zero to one in 0.01 increments. Regions at lower signal values appeared first and when the CSL is optimized isosurface shows the boundary of the volume with reduced optical properties. Figure 6(a) shows the process of CSL optimization in the case of SCC phantom. The CSL increases from 0 to 0.3 in 0.01 increments and isosurfaces are drawn for each one of the CSLs in Fig. 6(b) (see Visualization 1). At the optimized CSL of 0.3, the corresponding isosurface clearly delineates the diseased volume from the healthy surrounding region. If we continue increasing the CSL, the boundary between the healthy and diseased tissue merges. Following this procedure, CSL is optimized at 0.7, 0.5, and 0.3 for Fig. 5(a), 5(b), and 5(c), respectively. This result matches the level of drop in μs for infBCC, nodBCC, and SCC which are by factors of 0.9, 0.7, and 0.6 relative to the healthy region, respectively (considering μs as the dominant optical property as it is much larger than μa). Part (c) of Fig. 6 illustrates a slice view (see Visualization 2) of the SCC phantom from the depth of 400 μm to the surface of the epidermis model, at increments of 20 μm. It helps to visualize the optical sections from within the tissue phantom in addition to the 3D visualization of the volume of the diseased region with relative changes in the optical properties.
Fig. 6 (a) Increasing constant signal level (CSL) (red plane) from 0 to 0.3, where it cuts a typical optical section from the squamous cell carcinoma (SCC) phantom (regions below the CSL will be enclosed in the isosurface view), (b) isosurfaces of SCC made at different CSLs from 0 to 0.3 in 0.01 increments (see Visualization 1), and (c) slice view of SCC phantom from the depth of 400 μm to the surface of the epidermis model in 20 μm increments (see Visualization 2).
5. Discussion
Dermoscopy, one of the main noninvasive techniques used in this application, is based on optical magnification and looks for morphological characteristics of skin lesion on or close to the surface of the tissue [54–56]. Therefore, dermoscopy is a 2D projection of a 3D pathology and you might get surrogate information as to the depth of the diseased tissue. Using a dermascope, you do not have objective measure of the depth which has been affected. Our proposed technique, provides a 3D visualization platform showing the volume of the tissue which has been affected and can augment to dermoscopy. 3D assessment gives precise information as to the depth to which the cancer is extended and in particular penetration through the basal membrane. Confocal LFT can be an effective tool to measure the invasiveness (penetration depth) of the lesion. The mortality of both melanoma and keratinocyte cancers are exquisitely proportional (or tightly coupled) to the invasiveness; therefore, a 3D detection as proposed by our technique can be of great value.
The other main noninvasive technique which has been used in this application is confocal microscopy [57–61]. It examines morphological features of the lesion in a cellular scale. As described in section 4.1, our proposed method has a confocal structure and a focal beam spot diameter of about 1.8 μm. Therefore, it has the capacity of being used as a laser scanning confocal microscope to provide microscopic images in a cellular scale. As a simple and low-cost embodiment of a confocal microscope, confocal LFT has the capability to probe cancerous changes in a skin tumorous tissue at both macroscopic and microscopic scales.
Angiogenesis which is one of the main bio-markers of cancerous activities occurs surprisingly early during the multistage development of invasive cancers [42]. The proposed technique also has the capability of quantifying the neovascularization level by measuring the laser Doppler perfusion signal at the same time as it measures the confocal reflectance signal [38]. It shows the versatility of confocal LFT to be used in such an application.
Human skin is a large organ and it is difficult to monitor its whole area at a microscopic scale for early skin cancer detection. A macroscopic approach usually has a much higher speed and it is more likely to be effectively applied to early skin cancer detection. Being able to probe tumorous changes in the optical properties of a tissue, we believe confocal LFT is capable of detecting tissue volumes suspected of being cancerous at early stages, and then provide microscopic images of limited areas within the tissue. Microscopic images then provide an expert dermatopathologist with a tool to recognize patterns of different types of skin cancers at a cellular scale. Moreover, this technique has the potential to be used for demarcation of cancerous boundaries which is of great importance in surgical management. The transition from tissue phantoms to biological tissue involves fast scanning which could be achieved by replacing the mechanical translation stage with fast scanning mirror which is a fairly straightforward procedure.
6. Conclusion
Confocal laser feedback tomography is proposed as a useful biomedical imaging application. This technique was applied to malignant tissue phantoms. Three-layer silicone based phantoms for infiltrative basal cell carcinoma, nodular basal cell carcinoma, and squamous cell carcinoma were built and models of cancer affected regions were incorporated in these phantoms as cylindrical volumes with altered optical properties. Cross-sectional imaging was performed and tomographic presentation of macro-structural changes in the optical properties of the phantoms were illustrated in 3D space. In addition to detecting macro-structural changes, it is discussed that confocal laser feedback tomography is also able to provide images at a cellular scale. To underpin the experimental results, the system was studied numerically using Monte Carlo simulation and it was shown that experimental results were consistent with the numerical results. Results of this work suggest the feasibility of a sensitive, low-cost, versatile, and compact device with applications in early skin cancer detection.
Funding
Australian Research Council (ARC) discovery projects (160 103910); Trevor and Judith, St Baker Family Foundation.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References and links
1. J. Frank, Electron tomography (Springer, Dordrecht, 2008).
2. J. Hsieh, Computed tomography: principles, design, artifacts, and recent advances (SPIE Bellingham, Washington, USA, 2009).
3. D. A. Boas, D. H. Brooks, E. L. Miller, C. A. DiMarzio, M. Kilmer, R. J. Gaudette, and Q. Zhang, “Imaging the body with diffuse optical tomography,” IEEE Signal Process. Mag. 18, 57–75 (2001). [CrossRef]
4. A. Gibson, J. Hebden, and S. R. Arridge, “Recent advances in diffuse optical imaging,” Phys. Med. Biol. 50, R1 (2005). [CrossRef] [PubMed]
5. T. Durduran, R. Choe, W. Baker, and A. Yodh, “Diffuse optics for tissue monitoring and tomography,” Rep. Prog. Phys. 73, 076701 (2010). [CrossRef] [PubMed]
6. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. Fujimoto, “Optical coherence tomography,” Science 254, 1178 (1991). [CrossRef] [PubMed]
7. A. F. Fercher, W. Drexler, C. K. Hitzenberger, and T. Lasser, “Optical coherence tomography-principles and applications,” Rep. Prog. Phys. 66, 239 (2003). [CrossRef]
8. W. Drexler, M. Liu, A. Kumar, T. Kamali, A. Unterhuber, and R. A. Leitgeb, “Optical coherence tomography today: speed, contrast, and multimodality,” J. Biomed. Opt. 19, 071412 (2014). [CrossRef] [PubMed]
9. Y. Zhao, Z. Chen, C. Saxer, S. Xiang, J. F. de Boer, and J. S. Nelson, “Phase-resolved optical coherence tomography and optical Doppler tomography for imaging blood flow in human skin with fast scanning speed and high velocity sensitivity,” Opt. Lett. 25, 114–116 (2000). [CrossRef]
10. T. Bosch, N. Servagent, and S. Donati, “Optical feedback interferometry for sensing application,” Opt. Eng. 40, 20–27 (2001). [CrossRef]
11. G. Giuliani, M. Norgia, S. Donati, and T. Bosch, “Laser diode self-mixing technique for sensing applications,” J. Opt. A: Pure Appl. Op. 4, S283 (2002). [CrossRef]
12. S. Donati, “Developing self-mixing interferometry for instrumentation and measurements,” Laser Photon. Rev. 6, 393–417 (2012). [CrossRef]
13. S. Donati and M. Norgia, “Self-mixing interferometry for biomedical signals sensing,” IEEE J. Sel. Topics Quantum Electron. 20, 104–111 (2014). [CrossRef]
14. T. Taimre, M. Nikolić, K. Bertling, Y. L. Lim, T. Bosch, and A. D. Rakić, “Laser feedback interferometry: A tutorial on the self-mixing effect for coherent sensing,” Adv. Opt. Photonics 7, 570–631 (2015). [CrossRef]
15. T. Bosch, N. Servagent, R. Chellali, and M. Lescure, “Three-dimensional object construction using a self-mixing type scanning laser range finder,” IEEE Trans. Instrum. Meas. 47, 1326–1329 (1998). [CrossRef]
16. E. Gagnon and J.-F. Rivest, “Laser range imaging using the self-mixing effect in a laser diode,” IEEE Trans. Instrum. Meas. 48, 693–699 (1999). [CrossRef]
17. J. Keeley, P. Dean, A. Valavanis, K. Bertling, Y. Lim, R. Alhathlool, T. Taimre, L. Li, D. Indjin, A. Rakić, et al., “Three-dimensional terahertz imaging using swept-frequency feedback interferometry with a quantum cascade laser,” Opt. Lett. 40, 994–997 (2015). [CrossRef] [PubMed]
18. E. Lacot, R. Day, and F. Stoeckel, “Laser optical feedback tomography,” Opt. Lett. 24, 744–746 (1999). [CrossRef]
19. Y. Tan, W. Wang, C. Xu, and S. Zhang, “Laser confocal feedback tomography and nano-step height measurement,” Sci. Rep. 3, 2971 (2013). [CrossRef] [PubMed]
20. C. Xu, Y. Tan, S. Zhang, and S. Zhao, “The structure measurement of micro-electro-mechanical system devices by the optical feedback tomography technology,” Appl. Phys. Lett. 102, 221902 (2013). [CrossRef]
21. E. Salomatina, B. Jiang, J. Novak, and A. N. Yaroslavsky, “Optical properties of normal and cancerous human skin in the visible and near-infrared spectral range,” J. Biomed. Opt. 11, 064026 (2006). [CrossRef]
22. H. P. Soyer, D. Rigel, and E. M. Wurm, Dermatology, 3rd ed. ( Elsevier, London2012), chap. Actinic Keratosis, Basal Cell Carcinoma and Squamous Cell Carcinoma, pp. 1773–1794.
23. L. J. Loescher, M. Janda, H. P. Soyer, K. Shea, and C. Curiel-Lewandrowski, “Advances in skin cancer early detection and diagnosis,” in “Seminars in oncology nursing,”, vol. 29 (Elsevier, 2013), vol. 29, pp. 170–181. [CrossRef]
24. H. W. Walling, S. W. Fosko, P. A. Geraminejad, D. C. Whitaker, and C. J. Arpey, “Aggressive basal cell carcinoma: presentation, pathogenesis, and management,” Cancer Metastasis Rev. 23, 389–402 (2004). [CrossRef] [PubMed]
25. S. Shinohara, A. Mochizuki, H. Yoshida, and M. Sumi, “Laser Doppler velocimeter using the self-mixing effect of a semiconductor laser diode,” Appl. Opt. 25, 1417–1419 (1986). [CrossRef] [PubMed]
26. Y. L. Lim, T. Taimre, K. Bertling, P. Dean, D. Indjin, A. Valavanis, S. P. Khanna, M. Lachab, H. Schaider, T. W. Prow, et al., “High-contrast coherent terahertz imaging of porcine tissue via swept-frequency feedback interferometry,” Biomed. Opt. Express 5, 3981–3989 (2014). [CrossRef] [PubMed]
27. A. D. Rakić, T. Taimre, K. Bertling, Y. L. Lim, S. J. Wilson, M. Nikolić, A. Valavanis, D. Indjin, E. H. Linfield, A. G. Davies, et al., “Thz qcl self-mixing interferometry for biomedical applications,” in “SPIE Optical Engineering+ Applications,” (International Society for Optics and Photonics, 2014), pp. 91990M.
28. A. D. Rakić, Y. L. Lim, T. Taimre, G. Agnew, X. Qi, K. Bertling, S. Han, S. J. Wilson, A. Grier, Z. Ikonić, et al., “Optical feedback effects on terahertz quantum cascade lasers: modelling and applications,” in “SPIE/COS Photonics Asia,” (International Society for Optics and Photonics, 2016), pp. 1003016.
29. A. Bearden, M. P. O’Neill, L. C. Osborne, and T. L. Wong, “Imaging and vibrational analysis with laser-feedback interferometry,” Opt. Lett. 18, 238–240 (1993). [CrossRef] [PubMed]
30. R. Juškaitis, N. Rea, and T. Wilson, “Semiconductor laser confocal microscopy,” Appl. Opt. 33, 578–584 (1994). [CrossRef]
31. C.-H. Lu, J. Wang, and K.-L. Deng, “Imaging and profiling surface microstructures with noninterferometric confocal laser feedback,” Appl. Phys. Lett. 66, 2022–2024 (1995). [CrossRef]
32. M. Wang and G. Lai, “Self-mixing microscopic interferometer for the measurement of microprofile,” Opt. Commun. 238, 237–244 (2004). [CrossRef]
33. O. Hugon, I. Paun, C. Ricard, B. Van der Sanden, E. Lacot, O. Jacquin, and A. Witomski, “Cell imaging by coherent backscattering microscopy using frequency-shifted optical feedback in a microchip laser,” Ultramicroscopy 108, 523–528 (2008). [CrossRef]
34. O. Hugon, F. Joud, E. Lacot, O. Jacquin, and H. G. de Chatellus, “Coherent microscopy by laser optical feedback imaging (lofi) technique,” Ultramicroscopy 111, 1557–1563 (2011). [CrossRef] [PubMed]
35. J. White, W. Amos, and M. Fordham, “An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy,” J. Cell Biol. 105, 41–48 (1987). [CrossRef] [PubMed]
36. M. Minsky, “Memoir on inventing the confocal scanning microscope,” Scanning 10, 128–138 (1988). [CrossRef]
37. J. White, “Reflecting on confocal microscopy: a personal perspective,” Confocal Microscopy: Methods and Protocols 10751–7 (2014).
38. A. Mowla, T. Taimre, Y. L. Lim, K. Bertling, S. J. Wilson, T. W. Prow, H. P. Soyer, and A. D. Rakić, “Concurrent reflectance confocal microscopy and laser Doppler flowmetry to improve skin cancer imaging: A Monte Carlo model and experimental validation,” Sensors 16, 1411 (2016). [CrossRef]
39. A. Mowla, T. Taimre, Y. L. Lim, K. Bertling, S. J. Wilson, T. W. Prow, H. P. Soyer, and A. D. Rakić, “Diffuse reflectance imaging for non-melanoma skin cancer detection using laser feedback interferometry,” in “SPIE Photonics Europe,” (International Society for Optics and Photonics, 2016), pp. 98870T.
40. A. Mowla, T. Taimre, Y. L. Lim, K. Bertling, S. J. Wilson, T. W. Prow, and A. D. Rakić, “A compact laser imaging system for concurrent reflectance confocal microscopy and laser Doppler flowmetry,” IEEE Photon. J. 8, 1–9 (2016). [CrossRef]
41. A. Bashkatov, E. Genina, V. Kochubey, and V. Tuchin, “Optical properties of human skin, subcutaneous and mucous tissues in the wavelength range from 400 to 2000 nm,” J. Phys. D: Appl. Phys. 38, 2543 (2005). [CrossRef]
42. D. Hanahan and R. A. Weinberg, “Hallmarks of cancer: the next generation,” Cell 144, 646–674 (2011). [CrossRef] [PubMed]
43. D. Zink, A. H. Fischer, and J. A. Nickerson, “Nuclear structure in cancer cells,” Nat. Rev. Cancer 4, 677–687 (2004). [CrossRef] [PubMed]
44. P. Jorgensen, N. P. Edgington, B. L. Schneider, I. Rupeš, M. Tyers, and B. Futcher, “The size of the nucleus increases as yeast cells grow,” Mol. Biol. Cell 18, 3523–3532 (2007). [CrossRef] [PubMed]
45. R. K. Jain, “Taming vessels to treat cancer,” Scientific American 18, 64–71 (2008). [CrossRef]
46. I. Fredriksson, M. Larsson, and T. Strömberg, “Optical microcirculatory skin model: assessed by Monte Carlo simulations paired with in vivo laser Doppler flowmetry,” J. Biomed. Opt. 13, 014015 (2008). [CrossRef] [PubMed]
47. R. Schwartz, T. Bridges, A. Butani, and A. Ehrlich, “Actinic keratosis: an occupational and environmental disorder,” J. Eur. Acad. Dermatol. Venereol. 22, 606–615 (2008). [CrossRef] [PubMed]
48. F. Ayers, A. Grant, D. Kuo, D. J. Cuccia, and A. J. Durkin, “Fabrication and characterization of silicone-based tissue phantoms with tunable optical properties in the visible and near infrared domain,” in “Biomedical Optics (BiOS),” (International Society for Optics and Photonics, 2008), Proc. of SPIE Vol. 6870, 687007.
49. B. W. Pogue and M. S. Patterson, “Review of tissue simulating phantoms for optical spectroscopy, imaging and dosimetry,” J. Biomed. Opt. 11, 041102 (2006). [CrossRef] [PubMed]
50. V. Tuchin, Tissue optics: light scattering methods and instruments for medical diagnosis (SPIE press, Bellingham, Washington USA, 2007). [CrossRef]
51. A. Tycho, T. M. Jørgensen, H. T. Yura, and P. E. Andersen, “Derivation of a Monte Carlo method for modeling heterodyne detection in optical coherence tomography systems,” Appl. Opt. 41, 6676–6691 (2002). [CrossRef] [PubMed]
52. R. S. Matharu, J. Perchoux, R. Kliese, Y. L. Lim, and A. D. Rakić, “Maintaining maximum signal-to-noise ratio in uncooled vertical-cavity surface-emitting laser-based self-mixing sensors,” Opt. Lett. 36, 3690–3692 (2011). [CrossRef] [PubMed]
53. K. Bertling, T. Taimre, G. Agnew, Y. L. Lim, P. Dean, D. Indjin, S. Hoefling, R. Weih, M. Kamp, M. von Edlinger, J. Koeth, and A. D. Rakić, “Simple electrical modulation scheme for laser feedback imaging,” IEEE Sens. J. 16, 1937–1942 (2016). [CrossRef]
54. G. Argenziano and H. P. Soyer, “Dermoscopy of pigmented skin lesions–a valuable tool for early,” Lancet Oncol. 2, 443–449 (2001). [CrossRef]
55. G. Argenziano, H. P. Soyer, S. Chimenti, R. Talamini, R. Corona, F. Sera, M. Binder, L. Cerroni, G. De Rosa, G. Ferrara, et al., “Dermoscopy of pigmented skin lesions: results of a consensus meeting via the internet,” J. Am. Acad. Dermatol. 48, 679–693 (2003). [CrossRef] [PubMed]
56. R. Hofmann-Wellenhof, A. Blum, I. H. Wolf, D. Piccolo, H. Kerl, C. Garbe, and H. P. Soyer, “Dermoscopic classification of atypical melanocytic nevi (clark nevi),” Arch. Dermatol. 137, 1575–1580 (2001). [CrossRef] [PubMed]
57. R. Hofmann-Wellenhof, G. Pellacani, J. Malvehy, and H. P. Soyer, Reflectance Confocal Microscopy for Skin Diseases (Springer Science & Business Media, 2012). [CrossRef]
58. C. Longo, F. Farnetani, S. Ciardo, A. Cesinaro, E. Moscarella, G. Ponti, I. Zalaudek, G. Argenziano, and G. Pellacani, “Is confocal microscopy a valuable tool in diagnosing nodular lesions? a study of 140 cases,” Br. J. Dermatol. 169, 58–67 (2013). [CrossRef] [PubMed]
59. A. Scope, U. Mahmood, D. Gareau, M. Kenkre, J. Lieb, K. Nehal, and M. Rajadhyaksha, “In vivo reflectance confocal microscopy of shave biopsy wounds: feasibility of intraoperative mapping of cancer margins,” Br. J. Dermatol. 163, 1218–1228 (2010). [CrossRef] [PubMed]
60. A. Rishpon, N. Kim, A. Scope, L. Porges, M. C. Oliviero, R. P. Braun, A. A. Marghoob, C. A. Fox, and H. S. Rabinovitz, “Reflectance confocal microscopy criteria for squamous cell carcinomas and actinic keratoses,” Arch. Dermatol. 145, 766–772 (2009). [CrossRef] [PubMed]
61. S. Segura, S. Puig, C. Carrera, J. Palou, and J. Malvehy, “Development of a two-step method for the diagnosis of melanoma by reflectance confocal microscopy,” J. Am. Acad. Dermatol. 61, 216–229 (2009). [CrossRef] [PubMed]
