Optical coherence tomography (OCT) has emerged as a novel noninvasive imaging device that allows for the real-time, in vivo, cross-sectional imaging of skin morphology. OCT has increased imaging depth and field of view compared with reflectance confocal microscopy, at the cost of decreased cellular resolution. Frequency domain OCT, dynamic OCT (D-OCT), and high-definition OCT (HD-OCT) are useful in the diagnosis, treatment planning, and treatment monitoring of nonmelanoma skin cancers. Research is currently underway to assess the utilization of these devices in distinguishing between malignant and benign melanocytic lesions based on vascular patterns on D-OCT and cellular information on HD-OCT.
Key points
- •
A review of the literature shows that optical coherence tomography (OCT) increases the overall sensitivity, specificity, and diagnostic accuracy compared with clinical and dermoscopy evaluation alone.
- •
Frequency Domain OCT (FD-OCT) has imaging depth of up to 2 mm, with enough cellular clarity to diagnose nonmelanoma skin cancers. Dynamic OCT (D-OCT) enables us to visualize vascular patterns in the skin, improving diagnostic accuracy. Finally, high-definition OCT (HD-OCT) has improved cellular resolution compared with FD-OCT and D-OCT, at the sacrifice of penetration depth and field of view. However, HD-OCT serves to fill the gap between reflectance confocal microscopy and conventional FD-OCT.
- •
OCT has also been shown to be useful in tumor margin delineation and is, thus, useful in preoperative treatment planning. In addition, OCT enables noninvasive treatment monitoring of skin cancers undergoing nonsurgical therapies.
Introduction
Over the past decade, optical coherence tomography (OCT) has emerged as a novel noninvasive imaging device that allows for the real-time, in vivo, cross-sectional imaging of skin morphology. The advantage of these noninvasive devices over histopathology is that they enable repeated imaging of the same unaltered skin sites to observe dynamic events and long-term changes over time. Therefore, OCT has been used in both clinical and research settings to aid in the diagnosis of clinical and subclinical lesions; delineate lesion margins; and, unique to OCT given its larger field of view (FOV) and increased depth, monitor lesions undergoing nonsurgical treatment.
Introduction
Over the past decade, optical coherence tomography (OCT) has emerged as a novel noninvasive imaging device that allows for the real-time, in vivo, cross-sectional imaging of skin morphology. The advantage of these noninvasive devices over histopathology is that they enable repeated imaging of the same unaltered skin sites to observe dynamic events and long-term changes over time. Therefore, OCT has been used in both clinical and research settings to aid in the diagnosis of clinical and subclinical lesions; delineate lesion margins; and, unique to OCT given its larger field of view (FOV) and increased depth, monitor lesions undergoing nonsurgical treatment.
Optical coherence tomography
OCT imaging is based on low-coherence interferometry to detect the intensity of backscattered infrared light from biological tissues by measuring the optical path length. With these imaging devices, there is an inverse relationship between cellular clarity and both FOV as well as depth. Basically, as imaging depth and lateral resolution increases, the cellular resolution decreases ( Table 1 ).
| Imaging Modality | Imaging Depth (mm) | Lateral Resolution (μm) | Axial Resolution (μm) | FOV (mm) | Probe Aperture Size |
|---|---|---|---|---|---|
| RCM | 0.2 | 0.5–1.0 | 3–5 | 0.5 × 0.5 | 3.16 cm |
| HD-OCT | 0.57 | 3 | 3 | 1.8 × 1.5 | 5 cm |
| FD-OCT, D-OCT/SV-OCT | 1.5–2.0 | 7.5 | 5 | 6.0 × 6.0 | 1.38 cm |
| HFUS (20 MHz) | 10 | 200 | 80 | 12 | 10–20 mm |
There are several different OCT imaging modalities that have been studied. The swept-source multi-beam Frequency Domain OCT (FD-OCT) (Vivosight, Michelson Diagnostics, Kent, United Kingdom) provides 2 real-time imaging modes: b-scan (vertical, cross-sectional), similar to histology, and en face modes (horizontal), similar to that of dermoscopy and reflectance confocal microscopy (RCM). The images have an optical resolution of less than 7.5 μm laterally and less than 5 μm axially, a penetration depth of up to 2 mm, and an FOV of 6.0 mm × 6.0 mm. A recent advancement, dynamic OCT (D-OCT) based on speckle variance OCT (SV-OCT), allows for visualization of skin microvasculature and the detection of blood flow. Angiogenesis is important in the growth and spread of cancers; thus, visualization of vessel morphology is helpful in improving diagnostic accuracy.
Although conventional FD-OCT has been shown to be useful in skin imaging, its limited resolution precludes visualization of the skin at the cellular level. High-definition OCT (HD-OCT) (Skintell device, Agfa Healthcare, Mortsel, Belgium) seems to bridge the gap between conventional FD-OCT imaging and RCM offering improved axial and lateral resolution of 3 μm, with the trade-off of a more limited penetration depth of about 750 μm and FOV of 1.8 mm × 1.5 mm.
High-frequency ultrasound is another imaging modality that has the largest penetration depth and FOV, however, lacks the cellular resolution necessary for skin visualization (<1 mm) and is, therefore, not frequently used in the diagnosis and management of skin.
When comparing the different imaging methods, it is important to be aware of the imaging mode. FD-OCT, D-OCT, and HD-OCT devices all provide both vertical and horizontal en face images, creating a 3-dimensional image. The vertical scans are helpful in that they mimic histology sections; the horizontal view, similar to RCM, helps bridge the gap of dermoscopy to histology.
Another parameter to consider is the probe aperture size to FOV ratio. Noninvasive devices have a probe that directly touches the skin and produces an image based on its FOV. Usually the FOV is much smaller than the probe itself. The smaller the aperture and the larger the FOV the less of a discrepancy between what the probe comes in contact with and what is actually being imaged. Thus, a smaller variation in the aperture to FOV ratio leads to more accurate probe placement and, therefore, a better correlation of what you are imaging and what you see clinically. This smaller variation is especially important at varying time points if, for example, you are monitoring a lesion undergoing treatment. Additionally, a small probe can be positioned to image in more cosmetically sensitive areas, such as the head and neck. Ultrasound has the best FOV to aperture size ratio followed by FD- and D-OCT as seen in Table 1 .
Frequency domain-optical coherence tomography
FD-OCT has mainly been used in the cross-sectional (vertical mode) similar to histology ( Table 2 ).
| First Author, Year, Country | Population Characteristics | Lesions, Number | Accuracy | Findings/Results | Limitations |
|---|---|---|---|---|---|
| Maher et al, 2016, Australia | 88 patients (47 M, 28 F), mean age 63 y | 88 equivocal amelanotic or hypomelanotic skin lesions (AHM = 13), mostly located on the trunk (n = 36), lower limb (n = 16), upper limb (n = 18), head/neck (n = 18) | NR | OCT features of icicles and dermal ovoid structures with dark borders both significant ( P <.05) for diagnosis of AHM compared with other study lesions DEJ disruption commonly seen in AHM on OCT (10 of 13 cases) but was also seen in other NMSC, thus, lacked specificity | Limited number of significant OCT features to help identify AHM because OCT only provides architectural, superficial view of a skin lesion but does not offer cellular resolution Weak inter-rater agreement |
| Markowitz et al, 2016, United States | 30 male patients, aged 67–93 y (mean 76 y), all being treated with ingenol mebutate gel 0.015%; 2 patients excluded | 336 lesions (168 clinical AKs, 168 perilesional) | OCT detected 100% (28 of 28) of clinical and 73% (16 of 22) of subclinical lesions at baseline | At day 60, OCT indicated 76% clinical lesion clearance rate (52 of 68) for ingenol mebutate treated areas vs 11% (6 of 55) for untreated areas Clearance rate for subclinical lesions with ingenol mebutate: 88% (21 of 24) vs 43% (6 of 14) |
|
| Cheng et al, 2016, Australia | 103 patients (63 M, 40F), aged 31–88 y (median 66 y) | 168 lesions: 52% sBCC, 26% other BCC variants, remainder were AK, SCCIS, other benign inflammatory processes and 2 other malignant tumors Lesions located on the trunk (55.4%), upper extremity (18.5%), head/neck (13.7%), lower extremity (12.5%) |
| Clefting, hyporeflective ovoid structure, and absence of a fully encompassing ovoid structure highly predictive of sBCC Good diagnostic accuracy with OCT for diagnosing sBCC and measuring depth in tumors <0.4 mm Potential to reduce the need for biopsy in clinically suspected sBCCs with OCT use; careful follow-up required as there is a small risk (5%) of misdiagnosis; a potential 76% biopsy reduction rate of biopsy associated with a 5% error rate Good interobserver agreement between experienced and inexperienced observers | Potential pitfall: identified case of amelanotic melanoma that was diagnosed as sBCC clinically and on OCT |
| Meekings et al, 2016 | NR | 40 BCCs (13 nodular, 22 superficial, 5 morpheaform) | NR | Diagnostic criteria for BCC: hyporeflective ovoid structures (40 of 40), dark halo boundaries (38 of 40), epidermal thinning (28 of 40), and collagen compression (14 of 40) |
|
| Olsen et al, 2016, Denmark | 162 patients (90 M, 73 F), mean age 69.4 y | 142 lesions (41 BCCs, 30 AKs, 71 normal skin) located mainly on the head (90 of 142), followed by the trunk (32 of 142) and extremities (20 of 142) |
| Note: skilled observers: 1–10 y of OCT research and 3–28 y clinical experience in dermatology; unskilled observers: no experience in OCT and 1–15 y of clinical experience in dermatology Skilled observers better at interpreting OCT images compared with unskilled observers with a significantly higher sensitivity and specificity in diagnosing BCC and healthy skin; thus, potential increase in diagnostic accuracy with intensified training in OCT For AK: no significant difference in sensitivity or specificity between groups |
|
| Wahrlich et al, 2015, Germany | 130 patients (58 F, 72 M), average age 68.1–61.3 y | 98 BCCs, located on the head, trunk, extremities, 29 other skin diseases |
| 88% of all diagnoses correctly classified & confirmed by histopathology | Invasive SCC most frequent false-positive diagnosis |
| Maier et al, 2015, Germany | 4 patients (3 F, 1 M) | 14 AK lesions located on the upper extremity undergoing topical ingenol mebutate | NR | OCT features of AK pretreatment: crusts or scaling, epidermal broadening and thickening, ill-defined dermoepidermal boarder, hyperkeratosis During treatment: subepidermal blistering, dermal edema Noninvasive imaging superior to clinical evaluation to detect nonresponding lesions | Case series |
| Ulrich et al, 2015, Germany | 156 patients, aged 33–90 y (median 70 y) | Total: 235 nonpigmented pink lesions suspicious for BCC Histology identified 141 of 235 (60%) as BCCs (44 nBCC, 59 sBCC, 19 sclerosing BCC, 19 other BCC), located mostly on head (41.0%) and upper body (48.8%) | Increase in sensitivity from 90.0% (clinical examination) to 95.7% (clinical + dermoscopy + OCT) ( P = .099) Increase in specificity from 28.6% by clinical assessment to 54.3% using dermoscopy and to 75.3% with the addition of OCT ( P <.001) a PPV: 66.0% (clinical), 75.0% (dermoscopy), 85.2% (OCT) NPV: 65% (clinical), 79.4% (dermoscopy), 92.1% (OCT) | Diagnostic accuracy: increase from 65.8% (clinical) 76.2% (clinical + dermoscopy) to 87.4% (clinical + dermoscopy + OCT) | Caution for rare cases of amelanotic melanoma, which can present as a pink patch, plaque, or nodule |
| Blumetti et al, 2015, Brazil | NR | 39 lesions: 19 melanomas (10 in situ, 9 invasive, with Breslow thickness <1 mm), 15 compound nevi, 5 junctional nevi | NR |
| Small population |
| Markowitz et al, 2015, United States | 100 patients, >18 y old | 115 clinically challenging BCCs located on the head or neck | Diagnostic accuracy: 57.4% (clinical), 69.6% (dermoscopy), 87.8% (OCT) Increase in specificity from 48.9% by clinical assessment to 55.6% using dermoscopy and to 80% with the addition of OCT Increase in sensitivity from 62.9% by clinical assessment to 78.6% using dermoscopy and to 92.9% with the addition of OCT PPV: 65.7% (clinical), 74.3% (dermoscopy), 87.8% (OCT) NPV: 45.8% (clinical), 62.5% (dermoscopy), 87.8% (OCT) | Significant improvement in sensitivity and specificity over clinical or dermoscopic evaluation alone with OCT Certainty of diagnosis increased by OCT, with a positive effect on accuracy: increase in accuracy of diagnosis with OCT from 88% to 96% when diagnostic certainty was accounted for | Observational study involving only clinically difficult lesions and, thus, likely underestimates the specificity for obvious lesions |
| Mogensen et al, 2009, Denmark | 104 patients, mean age 69.3 y | 64 BCCs, 1 baso-squamous carcinoma, 39 AKs, 2 malignant melanomas, 9 benign lesions Locations NR | Sensitivity 79%–94%, specificity 85%–96% for all NMSCs depending on experience | Error rate of 50%–52% with discrimination of AK from BCC, much higher than any other study, likely due to older OCT | |
| Gambichler et al, 2007, Germany | 75 patients (42 M, 33 F), aged 15–95 y (mean 54.4) | 92 melanocytic lesions (52 BN, 40 MM), mostly located on the trunk, followed by limbs and head. | NR | OCT of MM: marked architectural disarray and DEJ; large, vertical, icicle-shaped structures | Conventional OCT: not enough clear-cut differences demonstrated between MM and BN to be used |
Basal Cell Carcinoma
There have been several studies investigating the accuracy of FD-OCT in diagnosis of basal cell carcinoma (BCC). The literature indicates that FD-OCT increased sensitivity, specificity, and diagnostic accuracy compared with clinical and dermoscopy assessment alone. There are few studies showing that FD-OCT is able to distinguish between the different BCC subtypes. However, some studies think FD-OCT lacks the cellular clarity to make this distinction.
The major diagnostic criteria on FD-OCT for BCC are alteration of the dermoepidermal junction (DEJ) and dark ovoid basal cell islands in the dermis, which are typically surrounded by a darker, hyporeflective peripheral border. Often these ovoid structures are surrounded by a bright fibrous tumor stroma. Additional features described include absence of normal hair follicles and glands as well as prominent dilated vessels in the superficial dermis directed toward the basaloid cell islands. Some lesions also have small, well-circumscribed, black/hyporeflective areas inside the tumor nests, representing tumor necrosis ( Fig. 1 ).
Squamous Cell Carcinoma/Actinic Keratosis
Actinic keratosis (AK) is considered to be the initial lesion in a continuum that progresses to invasive squamous cell carcinoma (SCC). While clinically it can be difficult to distinguish between AK and SCC, the distinction is critical in determining appropriate treatment. The use of noninvasive imaging can both improve diagnostic accuracy as well as increase the detection of early subclinical disease states.
On FD-OCT, AKs appear as white streaks in the upper epidermis, indicated by hyperechogenic areas that correspond to hyperkeratotic scale. There is thickening of the epidermis with a haphazard pattern in the upper portion, disruption of the normal layered skin architecture as well as ill-defined borders at the DEJ. Compared with AKs, SCCs tend to have more epidermal thickening and haphazard patterning and appear more broadly throughout the FOV ( Figs. 2 and 3 ).
Related posts:
Clinical Diagnosis of Skin Cancer
Enhancing Skin Cancer Diagnosis with Dermoscopy
Using Raman Spectroscopy to Detect and Diagnose Skin Cancer In Vivo
Proteomic Mass Spectrometry Imaging for Skin Cancer Diagnosis
Smartphone-Based Applications for Skin Monitoring and Melanoma Detection
Teledermatology Applications in Skin Cancer Diagnosis
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
