Predictive value of a non-invasive autofluorescence-based device (GOCCLES®) in the early diagnosis of oral potentially malignant disorders and oral cancer: a cross-sectional study

Abstract

Purpose

Oral cancer is the sixteenth most common malignancy globally and is characterized by a 5-year survival rate of less than 40%. In recent years, innovative methods have been developed to facilitate its early diagnosis, including the GOCCLES® device. This study aimed to evaluate the diagnostic value of GOCCLES® in detecting dysplasia in oral potentially malignant disorders (OPMDs) and diagnosing oral squamous cell carcinoma (OSCC). Additionally, this study sought to compare the device’s efficacy with toluidine blue (TB) staining and assess the value of combining both methods.

Methods

Patients clinically diagnosed with OPMDs or OSCC were included. Anamnesis, oral examination, inspection with GOCCLES®, and TB staining were performed. Finally, lesions were biopsied and the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of each technique were determined.

Results

Thirty-two patients were included in this study. The GOCCLES® device had a sensitivity of 33.3%, specificity of 40%, PPV of 25%, NPV of 50%, and accuracy of 37.5%. For TB staining, the sensitivity, specificity, PPV, NPV, and accuracy were 58.3%, 85%, 70%, 77.27%, and 75%, respectively. The respective values for the combined method were 44.4%, 77.78%, 66.67%, 58.33%, and 61.11%.

Conclusions

GOCCLES® and TB staining could help in the early diagnosis of OSCC and OPMDs, but a biopsy of the lesion remains essential for obtaining a definitive diagnosis.

Introduction

Oral cancer ranks as the sixteenth most common malignant neoplasm worldwide, with 90% of cases classified as oral squamous cell carcinoma (OSCC) of epithelial origin (Miranda-Filho and Bray 2020). In general, OSCC is more prevalent in men than in women, and its incidence increases with age, with the majority of cases occurring in individuals over 50 years old (Abati et al. 2020).

Despite advances in oncology, OSCC continues to have a poor prognosis, with 5-year survival rates below 40% (Romero et al. 2025). Early diagnosis, however, significantly improves outcomes, with survival rates exceeding 80% when identified at stages I or II (Warnakulasuriya and Kerr 2021), (Fernández-Martínez et al. 2023). Unfortunately, over half of oral cancers are diagnosed at advanced stages (III and IV) (Warnakulasuriya and Kerr 2021). The delayed diagnosis of OSCC is largely due to the asymptomatic nature of early-stage cancers, which are often mistaken for benign conditions. Consequently, patients frequently delay seeking medical attention, and the dentist may find it challenging to detect the pathology in its early stages (Abati et al. 2020), (Romero et al. 2025). A definitive diagnosis of OSCC requires a biopsy followed by histopathological analysis. However, biopsy is an invasive technique, and not all dental professionals are trained or inclined to perform it for preventive purposes. Studies indicate that only 15–21% of general dentists utilize biopsy as a diagnostic tool (Buenahora et al. 2021), (Schiavo-Di Flaviano et al. 2025).

Oral potentially malignant disorders (OPMDs) refer to a group of lesions or conditions that carry varying risk of malignant transformation into lip or oral cavity cancer (Warnakulasuriya et al. 2021). The report of a recent workshop organized by the World Health Organization indicates that OPMDs include leukoplakia, proliferative verrucous leukoplakia (PVL), erythroplakia, lupus erythematosus, oral submucosal fibrosis, oral lichen planus (OLP), actinic cheilitis, palatal lesions caused by inverted smoking, and dyskeratosis congenita. Additionally, there is sufficient evidence to warrant the inclusion of oral lichenoid lesions (OLLs) and oral lesions characteristic of graft-versus-host disease in this group (Warnakulasuriya et al. 2021).

The risk of malignant transformation in oral potentially malignant lesions varies widely, ranging from 1.4% to 49.5% (Iocca et al. 2020). Patients diagnosed with PVL and erythroplakia exhibit a higher frequency of malignant transformation, nearing 30–50%. In contrast, the risk in OLP patients is significantly lower, ranging between 1 and 2% (Warnakulasuriya et al. 2021). Predicting this risk remains a significant challenge in clinical practice. The presence and degree of dysplasia have proven useful in stratifying the risk of cancer development However, no pathognomonic clinical presentation exists for dysplasia. As a result, identifying the specific areas of a suspicious lesion that contains dysplasia is only possible through biopsy (Odell et al. 2021).

In recent years, various non-invasive techniques and devices have been developed to aid in the early detection of dysplastic changes (Mazur et al. 2021). One such method is autofluorescence (AF), which exploits the fluorescence of biofluorophores, present in tissues, when excited using a specific wavelength of light. AF of the oral mucosa primarily derives from oxidized flavin adenine dinucleotide (FAD) and other fluorophores. When excited using blue violet or ultraviolet (UV) light, healthy tissues emit light at 515 nm (green spectrum). Conversely, tissues exhibiting cellular metabolic disorders, such as dysplasia, display altered fluorescence and appear as darker regions under illumination (Lajolo et al. 2022), (Moro et al. 2015).

GOCCLES® (Glasses for Oral Cancer Curing Light Exposed Screening) is a medical device developed by Pierrel and approved by the Food and Drug Administration (FDA) in 2015. It incorporates optical filters composed of a three-layer lamellar structure, allowing the isolation of the fluorescent component emitted by FAD (515 nm) (Lajolo et al. 2022) when oral tissue is illuminated using a dental curing lamp (Moro et al. 2015). Since the approval of GOCCLES in 2015, only two studies have evaluated the device as an auxiliary tool in the diagnosis of OPMDs and oral cancer.

The primary objective of this study was to analyze the utility of the GOCCLES® device as an adjunctive diagnostic method in routine clinical practice. Specifically, we evaluated its diagnostic performance in detecting oral potentially malignant disorders (OPMDs) and oral squamous cell carcinoma (OSCC) and compared its efficacy with toluidine blue (TB) staining, a commonly used diagnostic technique in oral medicine. We hypothesized that the GOCCLES® device could represent an effective tool for the early diagnosis of OPMDs and OSCC, either alone or in combination with TB.

Materials and methods

Study design

This study was a cross-sectional, prospective clinical investigation that commenced in January 2022 and concluded in May 2024. It was conducted in compliance with the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines.

Patient selection

Patients visiting the Oral Medicine Unit of the University Clinic Foundation at Rey Juan Carlos University with lesions compatible with OPMDs or OSCC and aged > 18 years were recruited for this study. The exclusion criteria were patients who did not consent to participate, those with systemic conditions contraindicating biopsy or those who had recently undergone a biopsy of the oral cavity. Figure 1 shows the process of participant selection.

Fig. 1.

Flowchart showing the process of determining the final analysis participants

Study development

A visual examination of each lesion was conducted, and descriptive data were recorded, including the lesion’s exact location in the oral cavity, size, shape, consistency, color, and type (white, red, mixed, exophytic, or ulcerative). Lesions were photographically documented using a Canon EFS 60 mm camera and a Canon Macro Ring Lite MR-14EX® ring flash (Canon® Japan).

Subsequently, scanning was performed using the GOCCLES® device (Pierrel S.p.A-Italy) to detect areas of fluorescence loss indicative of possible dysplasia. For this procedure, the tissue was irradiated with a UV light source from an Sdi Radii Xpert polymerizing lamp (SDI, Australia) with a wavelength of 440–480 nm and an intensity of 1500 mW/cm2 applied at a distance of 20 cm from the lesion. The irradiated tissue was observed through the GOCCLES® device and photographs were taken using the mobile camera filter included in the device kit with a Xiaomi Redmi Note 9S smartphone (Xiaomi China). Lesions showing fluorescence loss, visible as brown or dark green areas through the device, were classified as GOCCLES® positive. Lesions without visible dark areas, where all tissue appeared within the green spectrum, were classified as GOCCLES® negative.

TB staining was performed following the technique described by Mashberg (1983). Lesions with intense blue staining were classified as positive for TB, while those with no staining or with only slight or diffuse staining were considered negative. Photographs were captured using the previously mentioned camera.

The most suitable area for biopsy was subsequently determined, always prioritizing the judgment of the investigator. Incisional or excisional biopsies were performed under local anesthesia, depending on lesion characteristics such as size or suspected malignancy. Each sample was placed in 10% formalin and sent to a specialized pathologist. All samples were analyzed by the same pathologist to ensure consistency.

All biopsy procedures and the processes described above were performed by the same professional with specialized training in oral medicine. The entire procedure was completed for each patient during a single appointment. Patients were scheduled for a follow-up one week after the intervention for review, suture removal, and delivery of results.

Ethical aspects

This study received approval from the Rey Juan Carlos University Ethics Committee (internal registration number 3011202123121) in December 2021 and was registered in ClinicalTrials.gov under the code NCT05562375. Informed consent was obtained from all patients prior to their participation. To ensure anonymity, each patient was assigned a numerical code for identification throughout this study.

Statistical analysis

The sample size was estimated a priori based on the expected diagnostic accuracy of the GOCCLES® device, using data reported in previous studies (Lajolo et al. 2022), (Moro et al. 2015). Assuming an expected proportion of 90%, a confidence level of 90% (corresponding to a z-score of 1.645), and accepting a maximum allowable error of 9% relative to the expected proportion, the minimum sample size required was determined to be 30 participants.

The clinical and demographic characteristics of the samples were described using descriptive statistics. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of GOCCLES®, TB, and the combination of both methods were calculated. The combined diagnostic performance of the two techniques was evaluated in 18 cases, where the results of GOCCLES® and TB were concordant (positive or negative). Data analyses were conducted using Microsoft Excel and SPSS® Statistics V22.0.

Results

The final sample included 32 patients (14 men and 18 women) aged 43–81 years. In total, 65.6% of the participants had no harmful habits, with such habits being more frequent among men. After biopsy sample analysis, the anatomopathological diagnosis of each lesion was confirmed. The most frequent diagnosis was lichenoid disease, including OLP or OLLs (46.9%), followed by leukoplakia (37.5%). Three patients were diagnosed with traumatic irritative pathology, one with pemphigus or pemphigoid and one with well-differentiated OSCC. The most common lesion location was the gingiva (50%), followed by the buccal mucosa (25%) and tongue (15.6%). Twelve patients (37.5%) exhibited anatomopathological changes indicative of epithelial dysplasia or malignant transformation, including 8 with low-grade dysplasia, 1 with moderate-grade dysplasia, 2 with high-grade dysplasia, and 1 with OSCC. These data are summarized in Tables 1 and 2.

Table 1.

Demographic characteristics of the studied cases (n = 32)

Demographic characteristics	Men (N = 14)	Women (N = 18)	N = 32	%
Age (years)	43–73	52–81	Mean age 64.66
No harmful habits	7	14		65.6
Tobacco	4	1		15.6
Alcohol	1	3		12.5
Tobacco and alcohol	1	0		3.1
Tobacco and hashish	1	0		3.1

Table 2.

Location of lesions and diagnosis confirmed with pathological analysis of biopsy specimens

	OLLs/LPO	Leucoplakia	Traumatic Irritative Pathology	Pemphigus/pemphigoid	OSCC	N	%
Tongue	3	2	0	0	0	5	15.63
Hard palate	1	1	0	0	0	2	6.25
Buccal mucosa	4	2	1	0	1	8	25.00
Keratinized gingiva	7	6	2	1	0	16	50.00
Labial mucosa	0	1	0	0	0	1	3.13
N	15	12	3	1	1
%	46.88	37.5	9.38	3.13	3.13

OLLs oral lichenoid lesions, OLP oral lichen planus, OSCC oral squamous cell carcinoma

GOCCLES®

Out of 32 cases, 16 exhibited fluorescence loss during the GOCCLES® scan and were classified as positive. However, after confirmation through anatomopathological analysis, 4 of these 16 lesions were confirmed as true positives (TPs), while the remaining 12 were false positives (FPs). The sensitivity and specificity of GOCCLES® for detecting dysplasia or OSCC were 33% and 40%, respectively. The PPV, NPV, and accuracy were 25%, 50%, and 37.5%, respectively.

Toluidine Blue (TB)

Of the 32 cases, 10 stained intensely (dark royal blue) with TB and were classified as positive lesions. Of these 10 cases, 7 were TP and 3 were FP. The remaining 22 cases either showed no staining or exhibited light and diffuse staining (pale royal blue) and were therefore classified as negatives. This group included five false negatives (FNs) and 17 true negatives (TNs). The sensitivity of TB staining was 58.33%, the specificity was 85%, the PPV was 70%, the NPV was 77.27%, and the accuracy was 75%.

Combination of both techniques

GOCCLES® and TB staining yielded concordant results (both positive and negative) in 18 of the 32 cases. Six cases tested positive for both techniques, of which four were TP and two were FP. Twelve cases were negative for both techniques, comprising five FNs and seven TNs. The sensitivity of the combination of both techniques was 44.4%, the specificity was 77.78%, the PPV was 66.67%, the NPV was 58.33%, and the accuracy was 61.11%.

The data are presented in Tables 3 and 4.

Table 3.

Ratio of true positive (TP), true negative (TN), false negative (FN), and false positive (FP) cases with the GOCCLES® device, toluidine blue staining, and the combination of both techniques

	TP	TN	FN	FP
GOCCLES®	4	8	8	12
Toluidine blue	7	17	5	3
Combined results of GOCCLES® + Stain	4	7	5	2

Table 4.

Test results for sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy for GOCCLES®, toluidine blue, and combination of both techniques

	GOCCLES®	Toluidine Blue	Combination
Sensitivity	33.33%	58.33%	44.44%
Specificity	40.00%	85.00%	77.78%
PPV	25.00%	70.00%	66.67%
NPV	50.00%	77.27%	58.33%
Accuracy	37.5%	75.00%	61.11%

Discussion

Based on the foundational work on AF by researchers such as Kluftinger et al. (1992), Betz et al. (1999), Onizawa et al. (1999), and Svistun et al. (2004) and Lane et al. (2006) developed the prototype for what eventually became the GOCCLES® device. In their study, the authors analyzed 50 biopsy samples of oral cavity lesions from patients with a history of oral dysplasia or OSCC and they reported a sensitivity of 98% and specificity of 100% in distinguishing benign oral lesions from high-risk OPMDs and OSCC lesions.

Moro et al. (2010) utilized a device derived from the prototype created by Lane et al. (2006) to investigate 32 patients. Their study yielded promising results, demonstrating a sensitivity of 100%, specificity of 95%, PPV of 92%, and NPV of 100%. However, the study had some methodological limitations: it was conducted in a high-risk population and classified as TN all patients who showed no clinical signs of visible relapse or suspicious lesions during oral examination and subsequent follow-up. Due to these limitations, along with the small sample size, the authors emphasized the need for further research involving larger populations and broader demographic groups.

In 2015, the GOCCLES® device received FDA approval. In the same year, Moro et al. (2015) introduced the current iteration of the device through a multicenter study analyzing 64 lesions. While the sensitivity, specificity, and predictive values were not calculated in their study, the authors highlighted a significant risk of FP, which occurred in more than 50% of cases. Consequently, they concluded that GOCCLES® should be employed strictly as an adjunct to oral examination. Additionally, they stressed the importance of proper training for dental practitioners intending to use the device (Moro et al. 2015).

Lajolo et al. (2022) conducted a retrospective study analyzing 41 samples of 25 patients. Their study closely aligned with the present investigation, as they evaluated the diagnostic value of GOCCLES® and compared its performance with TB. Their results indicated a sensitivity of 66% and specificity of 34% for GOCCLES®. The PPV, NPV, and accuracy for GOCCLES® were 34%, 77%, and 53%, respectively. TB demonstrated superior diagnostic performance, with a sensitivity of 91%, specificity of 68%, PPV of 55%, NPV of 95%, and accuracy of 75%. These findings are consistent with the results obtained in our current investigation. In our study, a significant proportion of FPs (n = 12, 37,5%) and FNs (n = 8, 25%) was observed with GOCCLES®. As previously noted in earlier studies, FN cases can occur, particularly in hyperkeratotic lesions, where keratin—a key endogenous fluorophore—can increase fluorescence and obscure dysplastic or neoplastic areas (Lajolo et al. 2022). This limitation is also highlighted in the manufacturer's instructions, which state that GOCCLES® “is not suitable for the study of hyperkeratotic lesions, but may help to identify their margins, which should appear as a dark halo around the main lesion”. In white lesions, we observed that distending the lesions to identify possible fissures could enhance detection. These fissures often revealed areas of fluorescence loss within the lesion, as demonstrated in Fig. 2 (a). Conversely, FP cases were frequently observed in red or mixed lesions, where erythematous regions appear darker than adjacent healthy or white tissues during both oral examination and GOCCLES® evaluation, as illustrated in Fig. 2 (c) and (d). This phenomenon is also noted in the device’s instructions, which explain that lesions with an inflammatory component may appear darker due to an increased blood component. Similarly, lesions located in highly vascularized areas, such as the labial mucosa or the lateral and ventral tongue, may produce similar effects (Lajolo et al. 2022).

Fig. 2.

a Areas with loss of fluorescence are observed during scanning with the GOCCLES® device when distending the lesion; (b) Toluidine blue staining of a leukoplakia lesion with high-grade dysplasia; (c) and (d) erythematous areas appear darker than the adjacent mucosa during oral exploration and the same occurs when scanning with GOCCLES®; (e) histopathology of the lesion corresponding to images (a) and (b) where epithelial hyperplasia with hyperkeratosis, chronic mucositis, and high-grade epithelial dysplasia can be seen; (f) histopathology of the lesion corresponding to images (c) and (d) where chronic mucositis with hyperkeratosis, epithelial atrophy, and ulceration without epithelial dysplasia can be seen

It is important to highlight that the results reported by Moro et al. in 2010 and 2015 are significantly more promising than those obtained by Lajolo et al. in 2022 and in the present study. The studies by Moro et al. were performed in a high-risk population with a prior history of oral potentially malignant disorders (OPMDs) or oral squamous cell carcinoma (OSCC), suggesting that the GOCCLES® device may be more suitable as a monitoring and follow-up tool for at-risk populations rather than as an adjunctive diagnostic method in routine clinical practice. For TB, a lower proportion of FPs and FNs was observed compared to that for GOCCLES®. Specifically, three FP cases (9.37%) and five FN cases (15.6%) were detected. It is worth noting that interpreting staining results is simpler than analyzing mucosal images observed with GOCCLES®. However, inter-examiner variability in defining a lesion as positive or negative after staining has also been reported in the literature. This variability arises from the absence of an objective scale to quantify the intensity of the blue staining. To ensure accurate interpretation, the guidelines described by Mashberg (1980) were followed in the present investigation, as illustrated in Fig. 2 (b).

Kim et al. (2021) conducted a systematic review and meta-analysis evaluating TB as a diagnostic tool. They reported a sensitivity of 72.9%, specificity of 69.6%, NPV of 71.6%, and accuracy of 76.6%, which are comparable to the results obtained in our study.

Regarding the diagnostic value of AF, it is also important to compare our findings with those of VelsCope® (Visually Enhanced Light Scope) (LED Dental Inc., Canada), one of the most widely recognized and frequently used AF devices. A recent systematic by Yeladandi et al. (2023) revealed that the sensitivity of this device varied across studies, ranging from 22 to 97% while specificity ranged from 8.4% to 95.8%, depending on the study reviewed.

An alternative AF device currently available is the Identafi® 3000 (StarDental-DentalEZ, Lancaster, PA, USA). This device features three types of lights: white LED light, green light, and violet light. In 2021, Sharma et al. conducted a study with 49 patients with lesions compatible with OPMDs and obtained a sensitivity of 73%, specificity of 46.2%, PPV of 57.6%, NPV of 63.11%, and accuracy of 61.9% for violet light. For the green/amber light, they obtained a sensitivity of 78.4%, specificity of 15.4%, PPV of 56.86%, NPV of 33.3%, and accuracy of 52.38% (Sharma et al. 2021).

Considering the results obtained in this study, as well as those from the previously mentioned works, we believe it is important to highlight the research conducted by Marumo et al. (2023), in which the use of free FAD as the target molecule for AF devices is questioned. These authors propose that the molecule these devices should target is flavinated succinate dehydrogenase subunit A (SDHA), a mitochondrial protein that is covalently bound to FAD. According to these authors, flavinated SDHA would be responsible for 90% of the AF traditionally attributed to free FAD. The researchers point out that current AF devices are not equipped to detect changes in this molecule and suggest that, in order to improve their diagnostic efficacy, their excitation range should be adjusted to 360–400 nm and their emission range to 575–650 nm. GOCCLES® device is used with curing lamps that have excitation ranges of 440–490 nm and an emission of 515 nm, corresponding to FAD, and it is also designed to block other wavelengths.

Although the study by Marumo et al. is recent and its results should be interpreted with caution, it may be of interest to propose new lines of research aimed at improving current AF devices or developing new ones that adjust the wavelengths to focus on the study of the fluorescence of flavinated SDHA.

We consider it important to highlight the limitations of the present study. First, similar to the investigations by Moro et al. (2010, 2015) and Lajolo et al. (2022), this study shares the limitation of a small sample size, underscoring the need for further research with larger sample sizes conducted across diverse populations.

Another limitation is the lack of examiner blinding, as all procedures including biopsy site selection were performed by the same professional. While this introduces a risk of incorporation bias, it reflects routine clinical practice where a single clinician integrates multiple sources of diagnostic information to make decisions. We consider that conducting further studies incorporating the judgment of multiple professionals could be valuable to address this limitation.

Conclusions

GOCCLES® is a non-invasive auxiliary diagnostic tool designed to assist in the early detection of OPMDs and OSCC. Its purpose is to facilitate the identification of these lesions by dental professionals and to determine the most optimal sites for biopsy. However, in this study, the GOCCLES® device did not demonstrate sufficient sensitivity, specificity, PPV and NPV to be considered a reliable diagnostic test. Although the results were not as expected, we have observed that combining the device with toluidine blue may improve its diagnostic efficacy, and recent research indicates that AF techniques and current devices have room for improvement and updating. Furthermore, to date, few studies have been published on the GOCCLES® device, highlighting the importance of conducting further research aimed at analyzing and addressing its current limitations.

Nevertheless, although AF techniques show promising results, to date, no auxiliary diagnostic method can replace the oral examination performed by a trained professional. Furthermore, biopsy remains the only way to obtain a definitive diagnosis and determine the presence and degree of dysplasia.

Abbreviations

AF

Autofluorescence

CI

Confidence interval

ED

Epithelial dysplasia

FAD

Flavin adenine dinucleotide

FDA

Food and Drug Administration

FN

False negative

FP

False positive

GOCCLES®

Glasses for Oral Cancer Curing Light Exposed Screening

IARC

International Agency for Research on Cancer

NPV

Negative predictive value

OPMDs

Oral potentially malignant disorders

OSCC

Oral squamous cell carcinoma

OLLs

Oral lichenoid lesions

OLP

Oral lichen planus

PPV

Positive predictive value

PVL

Proliferative verrucous leukoplakia

TN

True negative

TP

True positive

UV

Ultraviolet

Author’s contributions

Conceptualization, Rocío Martín-Muñoz, Jesús Rodríguez-Molinero, Blanca Migueláñez-Medrán and Antonio López-Sánchez; Formal analysis, Rocío Martín-Muñoz; Investigation, Rocío Martín-Muñoz; Methodology, Rocío Martín-Muñoz, Jesús Rodríguez-Molinero, Blanca Migueláñez-Medrán and Antonio López-Sánchez; Resources, Juan Ruiz Roca, Pía Lopez-Jornet and Rosario Ramírez-Puerta; Supervision, Antonio López-Sánchez; Validation, Juan Ruiz Roca, Pía Lopez-Jornet and Rosario Ramírez-Puerta; Writing – original draft, Rocío Martín-Muñoz and Blanca Migueláñez-Medrán; Writing – review & editing, Rocío Martín-Muñoz, Jesús Rodríguez-Molinero, Pía Lopez-Jornet and Antonio López-Sánchez.

Funding

This research received no external funding.

Data availability

The data presented in this paper are available upon reasonable request from the corresponding author.

Declarations

Ethics statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the Rey Juan Carlos University Ethics Committee (protocol code: 3011202123121; December 2021).

Conflicts of interest

The authors declare no conflicts of interest.