Discussion

With the development of AK, it was reported that melanocytes were more active than normal situation due to proliferation and differentiation in the epidermis^[48][49], which caused the increased number of dendritic structures. But it cannot be recognized in H&E images except under specific staining. In our HGM results, we almost see dendritic-like cells in all cases except AK patient 02. It shows in Fig 5.3.1. Moreover, we list the distribution of THG dendritic-cell-like signals in 5 patients shown in Table 5.3.2.

Fig. 5.3.1 En face sectioned ex vivo and in vivo HGM images of AK patients in the stratum basale. (a) The stratum basale of AK patient 00. In the red circle, it shows dendritic-cell-like signals. (b) The stratum basale of AK patient

01. In the red circle, it is dendritic-cell-like signals. (c) The stratum basale of AK patient 03. In the red circle, it is dendritic-cell-like signals. (a) The stratum basale of AK patient 04. In the red circle, it is dendritic-cell-like signals.

Table 5.3.2 The basal distribution of THG-bright dendritic-cell-like signals in 5 AK patients. “+” represents that this feature can be recognized. “-” represents that this feature does not exist.AK grade represents the severity of AK.

AK grade I represents mild AK. AK grade II represents moderate AK. AK grade III represents severe AK.

From Table 5.3.2, the dendritic structures are related to the severity of AK. The AK patient 02 is the only one without dendritic structures and the patient is happened to be the only in mild AK (AK grade I). To prove the result, more patients should be collected in future research.

In current situation of non-invasive way for AK diagnosis, OCT and RCM have made progress.

For OCT, images obtained in 2014 using a “VivoSight” OCT system from Michelson Diagnostics, UK with a resolution of <5 μm axial and <7.5 μm lateral and a scan area of 6 × 6 mm. The penetration depth of system was 1-2 mm. The lateral resolution is 7.5 μm lateral and vertical resolution was 10 μm. Their images are shown in Fig. 5.3.3.^[12] From Fig. 5.3.3, OCT could not recognize the morphology and it distinguished AK from normal skin only dependent on the thickness of epidermis.

Therefore, the main work of OCT for AK was measuring the thickness of epidermis and calculating the sensitivity of OCT by kappa analysis.

Fig. 5.3.3 Example of OCT-images presented in the study set. (A) Normal skin located on the arm, showing a narrow, hyperreflective band corresponding to an entry signal (thick arrow). The epidermis is seen as a homogenous well-demarcated darker layer (marked by *). The dermis is seen as a lighter layer (marked by **) and the DEJ is seen as a clear transition between the layers (thin arrow). (C) AK lesion located on the scalp, showing thickening of the epidermis (marked by *) and purple streaks in the upper epidermis due to hyperkeratosis (thick arrow). The DEJ is disrupted beneath the thickened epidermis (marked by **). Copyright: J. Olsen, L. Themstrup, N. De Carvalho, M.

Mogensen, G. Pellacani, G. B. E. Jemec, “Diagnostic accuracy of optical coherence tomography in actinic keratosis and basal cell carcinoma”, Photodiagnosis and Photodynamic Therapy, 2016, 16 (44-49) with permission from Elsevier.

The result is that skilled OCT observers were able to diagnose AK lesions with a sensitivity of 59% to 97% (average 76%) and a specificity of 52% to 83% (average 68%). Skilled observers with at least one year of OCT-experience showed an overall higher diagnostic accuracy compared to inexperienced observers. The conclusion is that diagnostic accuracy of differentiating AK from healthy skin has improved obviously than earlier OCT technology. But in comparation with H&E, the features of AK in OCT to distinguish normal from AK are the thickness of epidermis and the location of dermis-epidermis junction, which is imprecise to compare AKs from normal skin without the morphology of basal cells. Moreover, OCT did not indicate the grade of AK and they also did not show treatment assessment of AK by OCT.

For RCM, the typical system is Vivascope 1000/1500 where the resolution is 0.5 to 1.0 μm in the lateral and 3 to 5 μm in the axial dimension. An examination depth is 350 μm, which corresponds to the papillary dermis. The FOV is 500 μm * 500 μm.

Their images are shown in Fig. 5.3.4 to Fig. 5.3.7.^[13][14]

Fig. 5.3.4 Horizontal sections from depths in epidermis. Left column is conventional histopathology of AK, center is RCM of AK, and right column is RCM of adjacent normal skin. (A) Stratum corneum. (B) Stratum granulosum.

(C) Stratum spinosum. (D) Stratum basale. Copyright: D. Aghassi, R. Anderson, S. González, “Confocal laser microscopic imaging of actinic keratoses in vivo: A preliminary report.” Journal of the American Academy of Dermatology, 2000, 43 (42-48)with permission from Elsevier.

Fig. 5.3.5 The stratum corneum (A–C) RCM images; (D–F) representative histologic images. Copyright: M. Mlrich, A. Maltusch, F.-D. Rius, J. Röwert-Huber, S. González, W. Sterry, E. Stockfleth, S. Astner, “Clinical applicability of in vivo reflectance confocal microscopy for the diagnosis of actinic keratoses.” Dermatologic Surgery, 2008, 34:5 (610-619)with permission from Elsevier.

Fig. 5.3.6 Stratum granulosum and stratum spinosum(A–C) RCM images; (D–F) representative histologic images.

Copyright: M. Mlrich, A. Maltusch, F.-D. Rius, J. Röwert-Huber, S. González, W. Sterry, E. Stockfleth, S. Astner,

“Clinical applicability of in vivo reflectance confocal microscopy for the diagnosis of actinic keratoses.”

Dermatologic Surgery, 2008, 34:5 (610-619) with permission from Elsevier.

Fig. 5.3.7 Dermis (A, B) RCM images; (C, D) representative histologic images. Copyright: M. Mlrich, A. Maltusch, F.-D. Rius, J. Röwert-Huber, S. González, W. Sterry, E. Stockfleth, S. Astner, “Clinical applicability of in vivo reflectance confocal microscopy for the diagnosis of actinic keratoses.” Dermatologic Surgery, 2008, 34:5 (610-619)

with permission from Elsevier.

In Fig. 5.3.4, upon evaluation of the stratum corneum, superficial disruption can be found but the parakeratosis is not easily identified. The morphology of the stratum basale is not easily recognized in comparison with the morphology of the stratum spinosum in Fig. 5.3.4 C and D. In Fig. 5.3.5 to Fig. 5.3.7, they did not show the images in the stratum basale in this paper^[14]. However, the morphology of the stratum basale is the most important for AK diagnosis. Because pathological changes of the mild AK are only in the stratum basale. In the fact, spinosum cells are bigger than basal cells.

The morphology of the stratum spinosum of RCM is more legible than the morphology of the stratum basale of RCM due to the resolution. So, RCM prefers to show the images of the stratum spinosum instead of the images of the stratum basale. This is not appropriate for diagnosis of doctors. Moreover, in Fig. 5.3.4 the images are in low signal to background ratio in comparison with HGM images.^[50] Low signal to background ratio makes the morphology of cells hard to be recognized. In Fig. 5.3.7, they declared that RCM could distinguish the solar elastosis because the signal intensity would be enhanced by elastin materials which is shown in the bright area of the dermis (arrows in Fig. 5.3.7 A and B) in RCM images. But collagen fibers also can cause the bright area of the dermis (Situations are the same as in arrows in Fig. 5.3.7 A and B.). RCM cannot distinguish elastin fibers fromcollagen fibers. So, it is unpersuasive that RCM can distinguish solar elastosis. The same as OCT, the main work of RCM was also calculating the sensitivity of AK by kappa analysis. Although RCM could show the morphology of all epidermis, they did not show the comparison results of H&E and RCM images in each diagnosis standard. With a total of 44 AKs included in the final

analysis and following blinded evaluation by two independent well-trained investigators, 97.7% of all skin samples were identified as AK using RCM. 2.3% were incorrectly identified as normal skin by RCM, while routine histology showed features consistent with AK (RCM and H&E in each patient). While if the observers were not trained, the result was 20 percent drop to 75.6% of all skin samples to be identified as AK correctly.Moreover, RCM did not indicate the grade of AK and they also did not show treatment assessment of AK. We list the summary of HGM, RCM and OCT in Table 5.3.8. Details of diagnosis criterions of HGM, RCM and OCT are shown in Table 5.3.9.

Table 5.3.8 The comparison of OCT, RCM and HGM.

Table 5.3.9 Diagnosis criterions of HGM, RCM and OCT. “+” represents that this feature can be recognized. “-”

represents that this feature is uncertain to recognize.

In our work, the resolution of images is better than RCM and way much better than OCT. FOV is less than them and penetration depth is lowest on 250 μm, due to the limitation on the working distance. But 250 μm already corresponds to the papillary dermis and it can satisfy almost cases of AK diagnosis (especially AK I and AK II) if the epidermis is not thickened significantly. However, for AK III it suggests better penetration depth. Because the thickness of epidermis is thickened significantly. In AK patient 00, the thickness of epidermis was ~ 250 μm. Moreover, our resolution and signal to background ratio are better than the RCM and OCT and it means better morphology. In Table 5.3.9, it shows the details of diagnosis criterions of HGM, RCM and OCT. OCT only distinguishes AK from normal skin by the thickness of epidermis and it is completely unpersuasive for the pathologist. HGM and RCM can show the morphology of all diagnosis standards in epidermis. But the morphology of HGM is more legible than the morphology of RCM.

To show the HGM morphology of different layers of AKs, we collected 5 patients including 1 ex vivo and 4 in vivo shown in chapter 5.2. Moreover, we also tracked 1 patient to show the treatment assessment of AK.

In our study, the main work is the comparison of morphology of HGM and H&E images especially in the stratum basale which is the most important standard for diagnosis. We also tracked the location of multiple AK lesions for patients and showed the treatment assessment.

From the above, the diagnosis and treatment of AK are important because these lesions represent precursors to SCC that can be cured before malignant degeneration.

Real-time HGM offers a non-invasive opportunity to differentiate AKs from benign and malignant lesions without biopsy. By this study, pathologic features of AKs can be recognized legibly by HGM compared to existing technology including OCT and RCM.

Moreover, HGM can provide some unique findings.

1. HGM provides more legible morphology due to resolution and the much-improved signal to background ratio.

2. In the dermis, HGM can distinguish solar elastosis accurately by quantified signal intensity.

3. Especially, HGM can find dendritic structures in the stratum basale without specific staining.

Unfortunately, depth of penetration imposes a limitation in this HGM system in the diagnosis of AKs (AK III) especially for hyperkeratotic lesions. But generally speaking, the wavelength of laser source of RCM is 800 to 900 nm. The wavelength of the laser source of OCT is 1320 nm and the wavelength of the laser source of HGM is 1260 nm. So, considering the light diffusion and absorption of skin, the penetration depth of HGM is the best. In practice, two reasons are obvious. One is the limitation by working distance of our adopted objective (UAPON340/40X/NA=1.15, Olympus, Tokyo, Japan). Because our HGM system needs the good morphology (resolution) which means high NA. So, it is hard to be long in work distance for objective with high NA in the meanwhile. The other reason is that AKs are mostly surmounted by significant hyperkeratosis (AK III), which impedes the penetration of laser beam into the epidermis.

In addition, another limitation is that those AK patients are usually too old to keep still. So, the images are easy to be fuzzy by shaking of patients. More seriously, some layers were disappeared in image stacks due to the violent shaking in our clinical trials.

In this study, the 5 patients were all over 70 years old. Therefore, we developed the movable system for clinical trials to reduce the impact of shaking by patients.

Most people get more than one AK lesion, and patients who have multiple AKs continue to get new AKs for life. Clinical diagnosis and treatment follow-up for multiple AKs mainly depend on visual inspection and direct counting of all visible lesions. Problems in the direct counting approach may reside in difficulty in dealing with small or almost contiguous lesions. In addition, it is difficult to track the same AK lesion according to the location on the face and its size progression if there are multiple lesions scattered on the whole face. With our 3D facial images system, HGM is potential to track the multiple locations of AKs and it can record all locational information accurately. We have tested one patient to show the tracking of an AK lesion.

Despite these limitations, our case study indicates that HGM may become an alternative to biopsy in the diagnosis of AKs and also help doctors to improve treatment efficiency.

REFERENCES

[1] K. Duncan, A. Sadanand, L. Davachi, “Memory’s penumbra: episodic memory decisions induce lingering mnemonic biases.” Science, 2012, 337:6093 (485-487).

[2] H. Robert, M. Lawrence, “Solar keratosis: an evolving squamous cell carcinoma, benign or malignant?” Dermatologic Surgery, 1995, 21:2 (184).

[3] M.-H. Gold, M.-S. Nestor, “Current treatments of actinic keratosis.” Journal of Drugs in Dermatology, 2006, 5:2 (17-25).

[4] E.-E. Uhlenhake, “Optimal treatment of actinic keratoses.” Clinical Interventions in Aging, 2013, 8 (29-35).

[5] R. Marks, G. Renni, T.-S. Selwood, “Malignant transformation of solar keratoses to squamous cell carcinoma.” Lancet, 1988, 331:8589 (795-797).

[6] M.-J. Dodson, J.-D. Spain, J.-E. Hewett, “Malignant transformation of actinic keratosis and the controversy over treatment: a patient-oriented perspective.” Archives of Dermatology, 1991, 127:7 (1029-1031).

[7] F. Aaron, M. Ellen, “The kinetics of skin cancer: progression of actinic keratosis to squamous cell carcinoma.” Dermatologic Surgery, 2007, 33:9 (1099-1101).

[8] W. Fu, C.-J. Cockerell, “The actinic (solar) keratosis: a 21st-century perspective.” Archives of Dermatology, 2003, 139:1 (66-70).

[9] K.-O. Duncan, J.-K. Geisse, D.-J. Leffell, “Epithelial precancerous lesions.” 2012, (261-492).

[10] Y.-H. Liao, K.-H. Chen, M.-P. Tseng, C.-C. Sun, “Pattern of skin diseases in a geriatric patient group in Taiwan: A 7-year survey from the outpatient clinic of a university medical center.”

Dermatology, 2001, 203:4 (308-313).

[11] L. Themstrup, G. Pellacani, J. Welzel, J. Holmes, G.-B.-E. Jemec, M. Ulrich, “In vivo microvascular imaging of cutaneous actinic keratosis, Bowen's disease and squamous cell carcinoma using dynamic optical coherence tomography.” The Journal of the European Academy of Dermatology and Venereology, 2017 ,31:10 (1655-1662).

[12] J. Olsen, L. Themstrup, N. Carvalho, M. Mogensen, G. Pellacani, G.-B.-E. Jemec, “Diagnostic accuracy of optical coherence tomography in actinic keratosis and basal cell carcinoma.”

Photodiagnosis and Photodynamic Therapy, 2016, 16 (44-49).

[13] M. Ulrich, A. Maltusch, J. Röwert-Huber, S. González, W. Sterry, E. Stockfleth, S. Astner,

“Actinic keratoses: non-invasive diagnosis for field cancerization.” The British Journal of Dermatology, 2007, 156:3 (13-17).

[14] M. Mlrich, A. Maltusch, F.-D. Rius, J. Röwert-Huber, S. González, W. Sterry, E. Stockfleth, S.

Astner, “Clinical applicability of in vivo reflectance confocal microscopy for the diagnosis of actinic keratoses.” Dermatologic Surgery, 2008, 34:5 (610-619).

[15] M.-R. Tsai, S.-Y. Chen, D.-B. Shieh, P.-J. Lou, C.-K. Sun, “In vivo optical virtual biopsy of human oral mucosa with harmonic generation microscopy.” Biomedical Optics Express, 2011, 2:8 (2317-2328).

[16] M.-R. Tsai, D.-B. Shieh, P.-J. Lou, C.-F. Lin, C.-K. Sun, “Characterization of oral squamous cell carcinoma based on higher‐harmonic generation microscope.” Journal of Biophotonics, 2012, 5:5-6 (415-424).

[17] M.-R. Tsai, Y.-H. Cheng, J.-S. Chen, Y.-S. Sheen, Y.-H. Liao, C.-K. Sun, “Differential

diagnosis of nonmelanoma pigmented skin lesions based on harmonic generation microscopy.”

Journal of Biomedical Optics, 2014, 19:3 (036001).

[18] M.-R. Tsai, C.-Y. Lin, Y.-H. Liao, C.-K. Sun, “Applying tattoo dye as a third-harmonic generation contrast agent for in vivo optical virtual biopsy of human skin.” Journal of Biomedical Optics, 2013, 18:2 (026012).

[19] G. J. Brakenhoff, J. Squier, T. Norris, A. C. Bliton, M. H. Wade, B. Athey, “Real-time two-photon confocal microscopy using a femtosecond, amplified Ti:sapphire system.” Journal of Microscopy, 1996, 181:3 (253-259).

[20] Y.-C. Guo, P.-P. Ho, H. Savage, D. Harris, P. Sacks, S. Schantz, L. Feng, N. Zhadin, R. R.

Alfano, “Second-harmonic tomography of tissues.” Optics Letters, 1997, 22:17, (1323-1325).

[21] C. G. Omar, A. C. Millard, C. B. Schaffer, J. A. Au, P.-S. Tsai, J. A. Squier, D. Kleinfeld,

“Spectroscopy of third-harmonic generation: evidence for resonances in model compounds and ligated hemoglobin.” Journal of the Optical Society of America B, 2006, 23:5 (932-950).

[22] J. Squier, “High resolution nonlinear microscopy: A review of sources and methods for achieving optimal imaging.” Review of Scientific Instruments, 2001,72:7 (2855-2867).

[23] S.-Y. Chen, Y.-H. Wu, S.-K. Sun, “In vivo harmonic generation biopsy of human skin.” Journal of Biomedical Optics, 2009, 14:6 (060505).

[24] H.-Y. Liao, S.-Y. Chen, S.-Y. Chou, P.-H. Wang, M.-R. Tsai, C.-K. Sun, “Determination of chronological aging parameters in epidermal keratinocytes by in vivo harmonic generation microscopy.” Biomed Optics Express, 2013, 4:1 (77-88).

[25] Y.-H. Liao, W.-C. Kuo, S.-Y. Chou, C.-S. Tsai, G.-L. Lin, M.-R. Tsai, Y.-T. Shih, G.-G. Lee, C.-K. Sun, “Quantitative analysis of intrinsic skin aging in dermal papillae by in vivo harmonic generation microscopy.” Biomed Optics Express, 2014, 5:9 (3266-3279).

[26] S.-Y. Chen, S.-U. Chen, H.-Y. Wu, W.-J. Lee, Y.-H. Liao, C.-K. Sun, “In vivo virtual biopsy of human skin by using noninvasive higher harmonic generation microscopy.” IEEE Jourmal of Selected Topics in Quantum Electronics, 2010, 16:3 (478-492).

[27] C. J. Cockerell, “Histopathology of incipient intraepidermal squamous cell carcinoma.” The American Journal of Dermatopathology, 2000, 42:1 (11-17).

[28] https://commons.wikimedia.org/wiki/File:Skinlayers.png

[29] https://www.skincancer.org/skin-cancer-information/actinic-keratosis

[30] G. M. James, M. Jeffery, “Looking bill and mark principles of dermatology.” 2006 4th edition (40).

[31] R. E. Jones, “Questions to the editorial board and other authorities.” The American Journal of Dermatopathology, 1984, 4:1 (91-95).

[32] D. Weedon, “Weedon's Skin Pathology” 2010 3rd edition (56-57).

[33] A. Michele, D. Mittelbronn, D. L. Mullins, R.-C. A. Francisco, F. P. Flowers, “Frequency of pre‐existing actinic keratosis in cutaneous squamous cell carcinoma.” International Journal of Dermatology, 1998, 37:9 (677-681).

[34] R. A. Schwartz, “The actinic keratosis: a perspective and update.” Dermatologic Surgery, 1997, 23:11 (1009-1019).

[35] L. Julien, C. Chen, G. Goldenberg, “Actinic keratosis as a marker of field cancerization in excision specimens of cutaneous malignancies.” Cutis, 2016, 97:6 (415-420).

[36] https://commons.wikimedia.org/wiki/File:Bowenoid_actinic_keratosis_-_high_mag.jpg [37] T. Day, S. M. Holland, J. P. Scurry, “Normal vulvar histology: variation by site.” Journal of

Lower Genital Tract Disease, 2016, 20:1 (64-69).

[38] R. R. Lubritz, S. A. Smolewski, “Cryosurgery cure rate of actinic keratoses.” Journal of the American Academy of Dermatology, 1982, 7:5 (631-632).

[39] Y.-C. Guo, R.-R. Alfano, “Optical harmonic generation from animal tissues by the use of picosecond and femtosecond laser pulses.” Applied Optics, 1996, 35:34 (6810-6813).

[40] “The 10 golden rules of laser safety” 2007, University of Nottingham.

[41] “Laser safety manual” 1998, California Institute of Technology.

[42] S.-Y. Chen, H.-Y. Wu, C.-K. Sun, “In vivo harmonic generation biopsy of human skin.” Journal of Biomed Optics 2009, 14:6 (060505).

[43] A. Böhling, S. Bielfeldt, A. Himmelmann, M. Keskin, K. P. Wilhelm, “Comparison of the stratum corneum thickness measured in vivo with confocal Raman spectroscopy and confocal reflectance microscopy.” Skin Research and Technology, 2014, 20:1 (50-57).

[44] Y.-X. Zhen, S. Takaki, H. Tagami, “Number of cell layers of the stratum corneum in normal skin –relationship to the anatomical location on the body, age, sex and physical parameters.”

Archives for Dermatological Research, 1999, 291:10 (555-559).

[45] https://www.histology.leeds.ac.uk/skin/epidermis_layers.php

[46] C.-H. Yu, C.-K. Sun, “In vivo and ex vivo imaging of intra-tissue elastic fibers using third-harmonic-generation microscopy.” Optics Express, 2007, 15:18 (11167-11177).

[47] L. Schmitz, P. Kahl, M. Majores, E. Bierhoff, E. Stockfleth, T. Dirschka, “Actinic keratosis:

correlation between clinical and histological classification systems.” The Journal of the European Academy of Dermatology, 2016, 30:8 (1303-1307).

[48] J. V. Schmitt, H. A. Miot, “Actinic keratosis: a clinical and epidemiological revision.” Brazilian Society of Dermatology, 2012, 87:3 (425-434).

[49] W.-H. Weng, C.-K. Sun, “Differentiating intratumoral melanocytes from Langerhans cells in nonmelanocytic pigmented skin tumors in vivo by label-free third harmonic generation microscopy.” Journal of Biomedical Optics, 2016, 21:7 (076009).

[50] Y.-F. Qi, C. Xu, “Multi-color background-free coherent antistokes Raman scattering microscopy using a time-lens source.” Optics Express, 2018, 26:26 (34474-34483).

COPYRIGHT

1. Fig. 2.2.1 is modified by Serephine in 16 November 2006. These images are in the public domain. The licensing is shown in website:

https://commons.wikimedia.org/wiki/File:Skinlayers.png

2. Fig. 2.3.3.1(a) is modified in Feb 2010 by Nephron. It is own work in Wikimedia commons.

The licensing is from website:

https://commons.wikimedia.org/wiki/File:Bowenoid_actinic_keratosis_-_high_mag.jpg

3. The license of Fig. 2.3.3.1(b):

4. The license of Fig. 5.3.3:

5. The license of Fig. 5.3.4:

6. The license of Fig. 5.3.5, Fig. 5.3.6 and Fig. 5.3.7: