Ultraviolet (UV) light serves as a standard and extensively employed diagnostic tool for evaluating vitiligo. Among the reviewed studies, UV light was utilized as a diagnostic technique in seven studies, with Wood's lamp being the preferred choice in six of them (46–51, 55). The theory behind UV light photography suggests that UV rays are selectively absorbed by melanin in the epidermis compared to visible light. This theory is primarily discussed in the context of using UV cameras. However, when it comes to diagnosing vitiligo specifically, most studies relied on standard cameras that capture images within the visible spectrum only. Therefore, the prominent visualization of vitiligo in photography mainly results from enhanced skin fluorescence due to the absence of superimposed pigment, rather than direct UV absorption. The most frequently used device to diagnose vitiligo using UV light is Wood's light, also called a Wood's lamp. It is a handheld device that emits long-wave UV light (wavelengths of 320–450 nm, peak 365 nm) and is equipped with a magnifying lens to enable close examination of the skin (47). Other devices for UV assessment have also been reported, including softboxes, camera flash, and high output flash. Uitentuis et al. discovered that varied UV set-ups produced significantly different quality images, with the high output flash technique producing the best characteristics for vitiligo assessment (47).

Furthermore, one study conducted by Kaliyadan et al. utilized a simple hand-held black-light source, specifically a rudimentary flashlight. This study revealed that the hand-held source was equally effective in detecting fluorescence or enhancing skin lesions, when compared to a standard Wood's lamp (55). Anbar et al. compared the accuracy of vitiligo lesion identification by dermatologists and patients with and without Wood's lamp (50). They observed that lesions outlined using Wood's lamp were significantly more prominent, leading to the detection of new clinically unseen lesions or the identification of expansions of clinically apparent lesions that were otherwise invisible under regular lighting conditions (50). Furthermore, the effectiveness of UV photography in monitoring disease stability has been demonstrated in multiple studies. For instance, Wang et al. conducted a study where Wood's lamp was utilized to assess vitiligo activity and disease stability, both crucial components for successful epidermal grafting surgery. The researchers observed that by examining amelanotic lesions with well-defined borders under Wood's lamp, which serves as an indicator of stability, dermatologists can identify suitable candidates for grafting. Moreover, the use of Wood's lamp facilitates the early detection of re-pigmentation, enabling accurate surveillance and evaluation of treatment outcomes (48).

Reflectance confocal microscopy (RCM) was used in eleven studies for vitiligo diagnosis and treatment monitoring (9–19). RCM performed in vivo is a non-invasive and repetitive imaging technique that generates real-time images with a resolution like histological images. The components of a reflectance confocal microscope include objective and condenser lenses, a detector, and a light source in the form of a near-infrared laser beam. The laser beam is directed to a particular spot on the skin. The microscope produces images from the stratum corneum to the upper dermis, with a maximum imaging depth of 250 μm and may be used as a non-invasive optical biopsy. RCM analysis facilitates recognizing typical characteristics of affected vitiligo skin and may be beneficial in distinguishing vitiligo in contrast to other hypopigmentary conditions such as post-inflammatory hypopigmentation, nevus depigmentosus, or nevus anemicus. The main RCM characteristics of vitiligo lesions observed throughout the studies included an apparent loss of melanin in the lesioned skin, an indistinct boundary separating the affected from the surrounding normal skin, loss of integrity of the bright dermal papillary rings generally seen at the dermo-epidermal junction level, and highly refractile inflammatory cell infiltration at the edge of the lesions. RCM has been shown useful to differentiate between active or stable vitiligo, proving to be useful in predicting stability in vitiligo. Pertaining to stable vitiligo, there is complete loss of melanin in affected skin, with no changes in the content of melanin nor the dermal papillary rings. In active vitiligo, there is also a loss of melanin alongside a disappearance of the dermal papillary rings (9).

Currently, multiple approaches have been investigated for computer-assisted imaging analysis of vitiligo. The literature search found twenty studies which used variations of computer-assisted imaging analysis for vitiligo diagnostics.

Convolutional Neural Networks (CNNs) have shown promising results in vitiligo diagnosis and were utilized in eight studies (20–25). CNNs analyze skin images and classify them as either vitiligo or non-vitiligo lesions using a model involving multiple convolutional layers that identify patterns and a fully connected layer that performs the classification. A large dataset of labeled images is utilized to optimize the network performance and to minimize the difference between its predictions and the true labels.

Compared to human raters, including practicing dermatologists, dermatology residents, and general practitioners, CNNs outperformed all groups consistently. This suggests the efficacy of machine learning in classifying vitiligo by case probability and the possible advantages of CNN techniques as a remote diagnosis tool for vitiligo in situations involving telemedicine or where a Wood's lamp is not accessible (21, 26).

Principal component analysis (PCA) and independent component analysis (ICA) were proposed as mechanisms for analyzing vitiligo lesion segmentation and progression in five different studies (20, 27–30). While CNN, PCA, and ICA all use machine learning for image analysis, CNNs are non-linear supervised learning techniques, while PCA and ICA are linear unsupervised learning techniques. Generally, PCA is used to segment RGB images into melanin and hemoglobin only images, followed by ICA-powered alignment of the two principal component axes. In a study by Nugroho et al., a set of 41 RGB images of vitiligo lesions from 18 patients used a combinational PCA and ICA. They concluded that at a 95% confidence interval, there was a high sensitivity (0.9105 ± 0.0161), specificity (0.9973 ± 0.0009), and accuracy values (0.9901 ± 0.0028) (29). This diagnostic accuracy was comparable to other studies using PCA and ICA (20, 27–30, 72).

Two studies used Fuzzy C-Means (FCM), a cluster-based algorithm, to segment vitiligo lesions (31, 32). The algorithm follows a two-step mechanism, whereby skin segmentation is followed by vitiligo-specific segmentation. Both studies demonstrated considerably fast processing, illustrating the algorithm's promise for use in clinical settings and applicability for teledermatology applications. As the algorithm accepts low-resolution images as input, picture acquisition can be readily carried out using smartphone cameras. Furthermore, Nugraha et al. implemented the FCM software to develop a mobile application called Vi-DA (Vitiligo Diagnostic Assistance), allowing vitiligo patients to self-assess at home (32).

One study used artificial neural networks (ANN) of multilayer perceptron (MLP) type to quantifiable measure skin depigmentation based on the pattern of light refraction (33). Based on the light refraction pattern, the MLP was taught to analyze each pixel in an image and classify it as either having healthy or affected skin. The MLP then produces a binary picture from the original image, where cells with 0 and 1 represent healthy and vitiligo-affected skin, respectively. Compared to ICA/PCA and FCM, the proposed method outperformed both methods in specificity and sensitivity over an 8-test span (33).

Several studies demonstrated that computer-assisted imaging analysis provided quantitative data on vitiligo lesion stability, namely, consistent coloration over time, with little variation in pigmentation intensity, relatively constant lesions sizes, showing minimal or no expansion or contraction, and well-defined, regular borders, suggesting a lack of active disease progression.

Four studies utilized Optical Coherence Tomography (OCT) for vitiligo assessment (38–41). OCT measures differences in optical path light, where one light path is directed to a tissue sample while the other is to a reference mirror (73). Through lateral scanning, OCT effectively constructs high-resolution two or three-dimensional cross-sectional images of microstructural morphology in biological tissue in situ (74, 75). A study by Su et al. demonstrated that OCT could effectively diagnose vitiligo in its early stages (40). This was accomplished by utilizing OCT for in vivo imaging to reconstruct a three-dimensional skin microstructure. This successfully identified any loss of pigment in the early stages of vitiligo, despite white patches of skin becoming prominent only in the later stages of the disease (40). Xie et al. came to a similar conclusion, supporting the efficacy of OCT in the early diagnosis of vitiligo (39).

Furthermore, detecting the stratum basale and dermal papillae is vital to diagnosing skin pigmentation disorders. To effectively identify such epidermal structures, OCT analyzes the low scattering property of the dermal papillae, contrasting it with the high scattering property of the pigmented basal layer (39). OCT's ability to autodetect such papillae structures are accomplished by scanning large areas of skin and assessing the decrease or absence of the scattering contrasting between the epidermal structures. Gao et al. studied OCT's ability to accurately quantify the optical path length, measuring in vivo tissue's refractive index (41). Their research demonstrated the ability of OCT to evaluate a lower refractive index in skin tissue with vitiligo, showcasing the contrast in scattering coefficients of skin with vitiligo versus without. This change in scattering coefficients was associated with a decrease in melanin content, thus confirming OCT's effectiveness in diagnosing vitiligo. Furthermore, OCT has shown promise in evaluating vitiligo stability. By analyzing the epidermal microstructure, OCT can identify well-defined lesion borders, preserved structural integrity, and relatively constant scattering coefficients in stable vitiligo cases. In contrast, unstable vitiligo was shown to exhibit irregular borders, disrupted epidermal architecture, and changes in scattering coefficients (41).

Two studies investigated the use of visible light and digital photography for assessing vitiligo (53, 54). As vitiligo is primarily a clinical diagnosis, visible light and digital photography are essential tools for clinicians to use for educational, clinical or telemedicine applications. To better aid physicians in their assessments, digital photography's capture and storage help create a visual database of clinical presentations. The construction of robust visual databases is the foundation for advanced computational analysis to improve early diagnosis, monitor high-risk patients, and diagnose atypical lesions (53). For telemedicine, when imagining quality is consistent and adequate, it can substitute for inpatient physical examinations in up to 83% of cases (53). A recent article from Geel et al. highlights the importance of standardization in photographing vitiligo to improve documentation and comparison among different sites to produce a more efficient and reliable interpretation of results (54). Under visible light, distinguishing between hypomelanosis and amelanosis in vitiligo patients, particularly those with very fair skin (Type I or II) or children, has been shown to be challenging (76).

Three more imaging techniques employed in vitiligo diagnosis include infrared photography, multiphoton microscopy and third-harmonic generation microscopy. Table 1 provides a summary of the detailed characteristics.

Dermoscopy, also known as epiluminescence microscopy, is a real-time, dynamic, nondiagnostic tool that enables in vivo examination of skin lesions. Clinically, dermoscopy is widely used in vitiligo diagnostics due to its ability to enhance visualization and detect subtle changes in pigmentation and skin lesions that may be difficult to discern with the naked eye. The technique involves using a handheld instrument equipped with a polarized light source and magnifying lens to visualize the epidermal layer's microstructures and the skin's superficial dermis. Dermoscopy can provide valuable information about the structure and appearance of affected vitiligo skin lesions, allowing for improved diagnostic accuracy.

Dermoscopy was utilized in eleven different studies (46, 49, 55–63), of which two studies used ultraviolet light (46, 49), and one used high dynamic range conversion of images for diagnostic and monitoring purposes (55). Perifollicular changes and interfollicular pigmentation constituted the most common dermoscopic observations seen by Al-Refu et al. and Jha et al. (56, 57). Two studies showed that employing UV-dermoscopy can identify various distinctive microscopic characteristics of vitiligo not seen in polarized dermoscopy, namely, an enhanced perifollicular border and a more distinguishable depigmented junctional zone (46, 49). Kaliyadan et al. observed that prominent pigmentary features were perceived to show significant enrichments after HDR conversion (55). Furthermore, a novel study by Scarfi et al. proposed utilizing fluorescence-advanced videodermatoscopy which enables dynamic examination of superficial skin structures with a cellular resolution and holds potential for improved disease monitoring, prognosis, and treatment outcomes in vitiligo patients (58). Dermoscopy also offers valuable insights into assessing disease stability in vitiligo by identifying key features such as indistinct boundaries, perifollicular depigmentation, satellite lesions, and the micro-Koebner phenomenon. These characteristics are significantly associated with active vitiligo. Conversely, the presence of perifollicular repigmentation serves as a promising indicator of vitiligo stabilization. These insights can assist in evaluating the activity of vitiligo lesions and provide useful guidance for patient counseling regarding disease prognosis.

The colorimeter is a non-invasive instrument used to quantify skin color and has also been utilized to determine the capacity for pigmentation. Colorimetry allows for objective quantification of epidermal changes associated with vitiligo and offers a standardized approach to evaluating disease stability. By precisely measuring parameters such as luminance value and melanin index, colorimetry provides valuable insights into the progression and response to treatment in vitiligo. Two articles illustrate the potential role and efficacy of the colorimeter as a non-invasive diagnostic tool for diagnosing vitiligo and its severity (64, 65). Brazzelli et al. used a portable colorimeter to evaluate the progressive development of vitiligo patches in a Caucasian male. The colorimeter evaluated an area of 8 mm² in diameter of skin over four months, and they found the luminance value significantly increased, indicating increased relative lightness and depigmentation.

Tawfik et al. utilized colorimetry to evaluate the treatment response of vitiligo to narrow-band UVB phototherapy over the course of six months. The results were significant for regimentation in 90% of patients, as shown by an increase in melanin index, indicating improved stability of vitiligo lesions. An important finding was that the colorimeter could assess regimentation a month before it was apparent clinically, implicating its potential use as a prognostic tool (27). Compared to other measurement assessment techniques, such as the point counting method, colorimetry was found to be less time intensive and more standardized across all anatomic regions (27).

Six studies investigated the use of spectrometry in vitiligo assessment (66–71). These non-invasive tools can be used to assess disease activity in vitiligo patients. Spectrometers are portable tools used to investigate the physiological and morphological properties of skin tissue by utilizing optical signals with a high spectral resolution (72). The spectrometry mechanism of action involves exposing the epidermal surface to white light produced by an incandescent source. Melanin, one of the skin's chromophores, absorbs the majority of incoming light but exhibits a monotonic rise in intensity as wavelengths get shorter (absorption is almost wholly attenuated for wavelengths longer than 700 nm) (68). Spectrometers then detect backscattered photons emerging from various layers of skin tissue, resulting in a tissue surface emission profile. De Bruyne et al. used spectrometry on the perilesional skin of 70 vitiligo patients in different anatomic regions and noted a classification model generating a correct prediction in 82.9% of the cases (66). Poojary et al. analyzed using a portable fluorescence spectrometer for diagnosing vitiligo in 260 patients and recorded a similar sensitivity and specificity of 74.6% and 73%, respectively (67). They also observed a critical point (cut-off) of 975.995 nm to differentiate vitiligo from other hypopigmented with an increased sensitivity of 93.1% and specificity of 86.4%. Hegyi et al. utilized diffuse reflectance spectroscopy with an experimental spectrophotometer to measure skin pigmentation in patients with vitiligo undergoing PUVA therapy. They saw an increase in pigmentation compared to pre-treatment as demonstrated by a statistically significant difference in pre- and post-treatment melanin quantification (70). Park et al. utilized a narrow-band reflectance spectrophotometer to measure melanin indexes (Mis), relative melanin indexes (RMIs), and erythema to differentiate between two similar hypopigmented disorders, vitiligo and nevus depigmentosus (ND) (71). The study found that the mean Mis and RMIs are statistically different between patients with vitiligo and ND, with mean RMI scores of 50% and 74% for vitiligo and ND patients, respectively. Spectrometry techniques also offer a non-invasive means to assess vitiligo disease activity. Unlike invasive blood markers, spectrometry offers immediate and real-time information about the specific conditions of the skin and its biochemical alterations. Stability evaluation relies on monitoring the absence of significant changes in biochemical parameters and presents with distinctive benefits, including rapid analysis (<10 s), portability, and compactness, making it suitable for use in a dermatologist's office during consultations (66).