Introduction

Fungal nail infection, onychomycosis, is a frequent disease responsible of 50% of all nail diseases [7]. It can cause pain, ulceration and stigmatization [7] and is associated with tinea pedis and cellulitis [5]. Furthermore, it is associated with lower Health-Related Quality of Life and a number of other diseases such as vascular disease, immunosuppression, psoriasis and diabetes [6, 10]. The prevalence in Europe and North America varies between 4.3% [95% confidence interval (CI): 1.9–6.8] in population-based studies and 8.9% (CI 4.3–13.6) in hospital-based cohorts [14]. Four clinical presentations exist of which all of them may lead to total dystrophy of the nail if not treated accordingly. Distal lateral subungual onychomycosis (DSLO) is the most frequent subtype. Other clinical presentations include superficial white onychomycosis (SWO), proximal subungual onychomycosis and least frequent, endonyx [10]. The latter implies that the fungal infection is restricted to the nail plate not affecting the nail bed or nail plate surface [13].

Diagnosing onychomycosis is essential before initiating antifungal treatment. This is because fungal organisms have different antifungal susceptibility patterns, e.g., terbinafine is the first-line drug for the dermatophyte Trichophyton (T.) rubrum, a prevalent species in toe nails, whereas fluconazole (or itraconazole) is used for treatment of the yeast Candida, which is more commonly isolated from finger nails [12]. Diagnostic methods include direct microscopy and histology of nail material, which are able to confirm the presence of fungal elements in the material and culturing and molecular-based methods such as polymerase chain reaction (PCR) analysis and matrix-assisted laser desorption-ionization time of flight (MALDI-TOF) which are able to genus/species-specific identification of the fungal pathogen. The diagnostic accuracy varies between the diagnostic tools, and also depends on obtaining sufficient and representative nail material. Microscopy is a rapid, low-cost procedure, though with a low sensitivity found to be 48% in a study by Wilsmann-Theis et al. [17]. Histopathology using periodic acid-Schiff (PAS) staining is able to produce results within 24 h with a high sensitivity of 82% [7]. Cultures are time consuming (2–4 weeks), but with the advantage of genus/species-specific identification. It has, however, a sensitivity of 46.7% for dermatophyte species [16]. Molecular diagnostic methods are fast but not available in many countries as a routine diagnostic tool. This means that in many countries microscopy of nail material is the only available fast diagnostic method for detection of fungal elements. The direct microscopy method does not allow differentiation between fungal colonization and infection of the nail. Histopathology is, therefore, used in some countries as it can confirm the growth of fungal elements within the nail. This method is unfortunately not genus/species specific, and it is also invasive with the risk of secondary bacterial infection. Development of a quick and reliable method to substantiate clinical diagnosis is, therefore, of interest.

Optical coherence tomography (OCT) is a non-invasive imaging tool enabling real-time imaging of skin and tissue. A number of diseases have been characterized with OCT, and as a diagnostic tool, it has proven useful in non-melanoma skin cancer, though the technique has been applied for a series of dermatological fields, e.g., inflammatory and bullous diseases [9]. OCT is able to visualize the nail plate, nail bed and nail matrix, with the nail plate appearing with a varying number of homogenous horizontal bands in healthy nails [8]. In 2012, Aydin et al. studied nail changes in psoriasis [2], and in 2010, Abuzahra et al. performed a pilot study on the potential of diagnosing onychomycosis using OCT, showing the technology’s ability to visualize fungal elements and good correlation with histology [1].

The aim of this study is to investigate the applicability of optical coherence tomography to detect and localize fungal elements in nails of patients with clinical signs of onychomycosis.

Materials and methods

Patient recruitment

This explorative prospective study was conducted at the dermatology outpatient clinic at Zealand University Hospital, Roskilde, Denmark from 2017 to 2019. Patients with clinically suspected onychomycosis had nail scrapings for microscopy, fungal cultures and dermatophyte-specific PCR performed as standard routine procedures. In addition to these, OCT scans of the clinical abnormal nails were performed prior to scrapings.

Ethics

The study was conducted in accordance with Helsinki Declaration and approved by the Zealand Regional ethics committee SJ-651. Patients provided written informed consent. Data collection was approved by the Danish Data Protection Agency REG-177–2017. Informed written consent was obtained from all participants.

OCT scan acquisition

For this study, we used the commercially available Vivosight Dx (Michelson Diagnostics, Kent, UK) with a central wavelength of 1305 nm; an optical resolution of < 7.5 µm lateral and < 5 µm axial; a scan area of 6 × 6 mm and a penetration depth around 1 mm. The settings for all scans were 120 frames and 6 mm width. As this was an exploratory study, the scan pattern evolved until the authors could review the first scans. It was then decided that longitudinal scans in the nail’s direction of growth seemed to provide the best information on the morphological features of the nails. Furthermore, it was concluded that most nails would need two scans to cover the entire length of the nail. The guidelines provided by Ulrich et al. for performing OCT scans were used with regards to optimal position for the OCT image [15].

Analysis of scans

Image analysis was performed by the authors JO and PLA in unison, initially evaluating morphological features described in former studies on onychomycosis, previously defined as horizontal homogenous bands of varying intensity and thickness [3]. The authors, however, also recorded other deviations from healthy nail morphology based on known literature [8] and Fig. 1. The previously described morphologic criteria for onychomycosis were: hyperreflective lines seen as lengthy structures inside the nail; hyperreflective dots seen as round aggregated structures inside the nail and irregular surface seen as an irregular intense entrance signal as defined by previous studies [1, 11]. After the first assessment of all scans, each scan was reassessed with our added features in mind (see “Results”). Observers were blinded when reviewing scans with regards to clinical and mycological diagnosis. Agreement of the presence of each morphological feature was reached for each scan. For image analysis, we used ImageJ 1.52i (Wayne Rasband, NIH, USA).

Fig. 1
figure 1

Cross-sectional OCT images of a healthy nail. a Distal, b median and c proximal parts of a healthy nail from the hallux, demonstrating of a narrow entry signal and homogenous nail plate with partly visible nail bed. In a, the terminal nail plate shows a hyperreflective line and increased signal density at the hyponychium. In c, the lunula is shown as a slightly hyperreflective band

Full size image

Results

Patient inclusion and characteristics: a total of 32 patients referred with clinical signs of onychomycosis were included in the study. Out of these, 29 patients had positive microscopy, fungal cultures and/or PCR, while two patients had negative samples. None of the included patients had a history of nail psoriasis. One patient was excluded due to complete removal of the affected nail prior to scanning. The included patients had a mean age of 57.7 ± 19.7 SD (standard deviation) years and a sex distribution male/female of 18/11. The clinical presentations of onychomycosis were 17 cases of DLSO, 7 cases of total dystrophy, 2 cases of endonyx and proximal subungual onychomycosis respectively and 1 case of SWO. The fungal agent (with clinical subtyping) was in 19 cases T. rubrum (11 DLSO, 5 total dystrophy, 2 proximal subungual onychomycosis and 1 SWO), 2 cases Aspergillus terreus (both DLSO) and 1 of each of Fusarium solani (DLSO), T. interdigitale (total dystrophy) and Acremonium spp. (DLSO). In five cases, direct microscopy was positive, but the culture and dermatophyte-specific PCR was negative.

The OCT scans were reviewed based on the suggested morphological features including refinements shown in Figs. 1 and 2 with a healthy nail in Fig. 1 and onychomycosis in Fig. 2. The suggested added features were:

  • Dark bands Dark elongated structures not due to artefacts caused by hyperreflective lines and

  • Disturbed architecture Disruption of the normal homogenous layering of healthy nail.

Fig. 2
figure 2

Cross-sectional OCT images of nails with onychomycosis. a Sharply demarcated hyperreflective lines. b Clustered hyperreflective dots, proximal nail. c Dark band and a mildly irregular surface, proximal nail. d Diffusely demarcated hyperreflective lines and clustered hyperreflective dots, distal nail. e Singular hyperreflective dots in and sharply demarcated hyperreflective lines, distal nail. f Disturbed architecture and sharply demarcated hyperreflective lines, distal nail. g Mildly irregular surface in proximal nail. h Moderately irregular surface, proximal nail. i Severely irregular surface, distal nail

Full size image

The refinements of the original features were:

  • Hyperreflective lines Divided into sharply and diffusely demarcated lines.

  • Hyperreflective dots Divided into clustered and singular dots

  • Irregular nail surface Divided into an ordinal scale of mild, moderate or severe irregularity

In Fig. 3, the overall frequencies of each OCT morphology for both proximal and distal nail scans are reported. In Table 1, the occurrence of each OCT morphology is reported for each clinical subtype. Irregular surface was the most predominant feature with a frequency of 80.9–83.4% and hyperreflective lines were the second most predominant feature with a frequency of 71.4–83.4%. Hyperreflective dots, dark bands and disturbed architecture were overall present in 23.8–50.0%, 52.4–66.7% and 42.9–45.8% of scans, respectively.

Fig. 3
figure 3

Frequency chart, displaying the frequency of each morphology proximally and distally. Plotted using Microsoft Office 2013 Excel

Full size image
Table 1 Frequencies of OCT morphologies for each clinical subtype
Full size table

Discussion

In this study, we explored the morphological heterogeneity of onychomycosis with OCT. The suggested morphological features are an addition to known features described in previous studies [1, 11]. We confirmed the presence of hyperreflective lines, hyperreflective dots and irregular surface as previously described [1, 3]. We furthermore refined the established features and added new qualitative features.

Histology of the highly scattering parts of mycotic nails in OCT has previously shown to consist of conglomerates of hyphae, while low scattering represent lacunae of hyperkeratotic nail plate [1]. Both hyperreflective lines and hyperreflective dots may be considered highly scattering areas, and a high number of cases were shown to have especially hyperreflective lines. This suggests that the finding of highly scattering elements in OCT scans may consist of fungal material. As shown in Fig. 2, the depth of these features can be easily determined. Thus, we hypothesize that OCT may be used as guidance for nail scraping by applying OCT to guide the clinician to sample from an area where the fungal elements are present thereby contributing to increased detection of fungal elements and a higher sensitivity of nail scrapings. Dark bands was described as low scattering areas, thus we hypothesize that this morphology represents lacunae of hyperkeratotic nail plate and does not necessarily contain fungal hyphae [1]. Dark bands should, therefore, not be considered a target for nail scrapings. It should be emphasized that even though OCT adds rapid and non-invasive detection of nail tissue likely containing fungal elements, the technique is not genus nor species specific and cannot render the use of cultures and molecular-based methods unnecessary. As the resolution of the OCT device is approximately 5 µm, it is not able to directly detect fungal hyphae.

The frequencies of each morphology subdivided into clinical subtypes are presented in Table 1. Here hyperreflective lines were divided equally between sharply and diffusely demarcated, except endonyx, which only presented diffusely demarcated lines. Furthermore, endonyx is the only clinical subtype with only one case of irregular surface (mild), suggesting that the nail in this endonyx case may be affected to a severe extent trending towards total dystrophy. The lack of surface irregularity and solely diffusely demarcated hyperreflective lines could be pathognomonic for endonyx, but this needs to be validated in a larger sample. DLSO accounts for 9 of 12 cases distal cases of hyperreflective dots. Interestingly, the only clinical subtype with clustered dots was DLSO, which may contribute to the understanding of the pathophysiology for DLSO. We hypothesize that clustered dots represent a proximal spread of infection, i.e., hyphae that are not yet dense enough to form hyperreflective lines. Disturbed architecture was found in 52.4–66.7% of cases and was only present in DLSO and total dystrophy. Disturbed architecture is described as a disturbance in the normal layering of the nail (Fig. 2f), a finding which has also been described in psoriasis nails [2], which may indicate the degree of nail dystrophy. The OCT morphology results (Fig. 3) suggest a slightly higher occurrence rate of morphological features distally than proximally. As 17/29 cases were DLSO, the overall distribution of morphologies may be a result of the many cases of DSLO. The average number of morphologies is highest in the distal part of the nail in cases with ‘total dystrophy’ (4.2) followed by DLSO (3.4), which is a clinical subtype defined by distal nail involvement. More surprising, DLSO has just about the same number of morphologies proximally (2.8) as ‘total dystrophy’ (2.6), where irregular surface is more frequent in total dystrophy and hyperreflective dots is more frequent in DLSO. As total dystrophy is considered, a continuum of the four other clinical subtypes, especially developing from long-standing DLSO or proximal subungual onychomycosis [10], and because DLSO is by far the most common, it is warranted to expect the two to have somewhat of the same morphological features.

As this is a cross-sectional study, a treatment response for onychomycosis was not assessed. It is difficult clinically to determine the mycotic cure in nails, as successful eradication may still leave abnormal nails [13]. It is possible that OCT may aid the assessment of whether a mycotic cure has been reached, based on the presence of morphologies indicating possible presence of fungal hyphae, i.e., hyperreflective lines/dots. The purpose for OCT scans of onychomycosis may, therefore, not only be used as a diagnostic tool, but could also have its merits in treatment assessment, where species/genus has already been determined, considering the technique is more rapid and less invasive than direct microscopy and histology.

The suggested morphologies for onychomycosis may not be specific for this diagnosis. OCT morphologies of other nail diseases should be compared with those of onychomycosis to assess the potential for differential diagnostics. As with onychomycosis, nail psoriasis has also been described with hyperreflective spots, hyperreflective lines and irregular surface in OCT [3]. Since there are no specific morphological criteria for onychomycosis in OCT, our suggested terminology may contribute in differentiating onychomycosis from other nail diseases. An observer-blinded study on the differentiation between nail psoriasis and onychomycosis may be of value. Since nail psoriasis is a common nail disease and increases the risk of onychomycosis [6], it is important to include mycological diagnostics in study designs of OCT morphology of nail psoriasis, which lacks in current published studies [2, 3].

The diagnostic value of OCT in onychomycosis has been evaluated by Rothmund et al. OCT and confocal laser scanning microscopy (CLSM) were compared to direct microscopy, histology and cultures in a study of 60 patients. The study showed a sensitivity of 92.3% and specificity of 42.9% for OCT and sensitivity of 81.0% and specificity of 79.5% for CLSM [11]. This study did not define specific diagnostic criteria for diagnosing onychomycosis with OCT, thus further warranting specific criteria to raise specificity. As there are very few studies on the morphology of onychomycosis in OCT, we did not investigate the diagnostic accuracy of OCT in our study, even though future studies on the diagnostic accuracy of the refined criteria may be warranted.

Limitation

As this is an explorative study, the method for describing the morphological features of onychomycosis in OCT has not been validated, and especially the newly added features of ‘dark bands’ and ‘disturbed architecture’ need to be further assessed to verify them.

In case of surface irregularities, deep infection cannot be excluded. Compared to other imaging modalities like high-frequency ultrasound, OCT relies on back-scattering of light, thus it is known to be unable to detect morphological change underneath highly scattering tissue such as crusting [4], exemplified in Fig. 2h, showing moderately irregular surface. It is, therefore, possible that deeper parts of the nail and nail bed affected by onychomycosis are not assessable and, therefore, overlooked with OCT. Even though OCT can be applied for rapid diagnosing, it is not able to determine the fungal agent (just as microscopy). Therefore, cultures and/or molecular-based methods are still needed in determining the pathogen and initiate appropriate treatment. It has been demonstrated that confocal microscopy (CLSM) is able to visualize fungal hyphae directly [11] due to the higher resolution, achieving a higher specificity than OCT. On the other hand, OCT has been shown to have a slightly higher sensitivity than CLSM [11], which may be due to the higher penetration depth, especially in endonyx, where the surface is less affected. However, the lack of clear diagnostic criteria in the study may warrant a new comparison of the techniques. Our sample mainly consisted of patients with DLSO reflecting the clinical distribution of onychomycosis subtypes. Thus, the overall frequency distribution may be skewed by the relative overrepresentation of this subtype.

In summary, this study presents the findings in OCT morphology in patients with onychomycosis. The suggested terminology is a refinement of previous literature and may contribute in diagnostic accuracy and differential diagnostics of onychomycosis. Though OCT cannot replace genus or species specific identification, it may be a useful assistance tool in performing nail scrapings to obtain representative nail material and may support control of infection eradication. Validation of the suggested terminology and OCT’s usefulness in specimen sampling and treatment assessment is needed.