Non-Invasive Approaches for Oral Potential Malignant Disorders Surveillance: A Review

Oral potential malignant disorders (OPMDs), a series of local diseases with carcinogenic potential occurred in oral mucosa, have the possibility of developing into oral cancer. Effective diagnosis and surveillance at the early stage of OPMDs may be of help for interrupting the malignant transformation. In this review, we introduced the available non-invasive approaches of detecting the suspicious lesions, including vital staining, oral cytology, optical detection, saliva biomarkers detection, and image analysis, which can enhance the detection efficiency for the early diagnosis and surveillance of OPMDs. However, limitations which restrict the clinical application of these approaches still exist. How to improve the sensitivity and specificity of these techniques deserves further detailed study.


Introduction
Oral potential malignant disorders (OPMDs) are diseases with potential carcinogenesis in oral mucosa, which have evidently clinical and histological alternations, such as leukoplakia, erythroplakia, oral submucous fibrosis, lichen plaus and discoid lupus erythematosus. It was reported that approximately 16% -62% Advances in Bioscience and Biotechnology of epithelial dysplasia cases of OPMDs underwent malignant transformation and eventually developed into oral squamous cell carcinoma (OSCC) [1]. OSCC is the most common oral cancer, accounting for more than 90% of all oral cancers.
In the past few decades, the five-year survival rate of OSCC patients in the early stage is about 80%, while it rapidly decreased to 40% in the advanced stage. As early diagnosis and timely monitoring substantially reduce the cancer-specific morbidity and mortality in OSCC, the surveillance of OPMDs is of great significance for the prevention of OSCC [2].
Although the pathological examination is the gold standard for the diagnosis of OPMDs, it is invasive to obtain pathological tissue through surgical operation, which is hardly acceptable by patients. Furthermore, such surgical processes involve certain risks including postoperative scars and excessive bleeding, so that they are not suitable for multiple operations for the early diagnosis and long-term monitoring of OPMDs [3]. Currently, non-invasive detection methods for early diagnosis, such as vital biological dye staining, oral cytology, optical detection and image diagnosis have been widely used in clinical practice.
These methods seem to replace tissue biopsy because of their efficiency and accuracy, as well as the improved acceptance of patients [4]. The purpose of this review is to summarize the current non-invasive diagnostic techniques for the monitoring of the malignancy of OPMDs, which may effectively lower the occurrence of oral cancer.

Toluidine Blue Staining
Toluidine blue (TB), a common cationic metachromatic alkaline dye, has been used as an important stain to highlight oral potential malignant lesions since the early 1980s. Cells in dysplastic or malignant tissue are more disordered in arrangement, looser in connection and contain more nucleic acid than normal cells. Therefore, TB can enter the extracellular space in dysplastic or malignant tissue, and bind to the intracellular nucleic acid to make the tissue blue, which can be readily distinguished [5].
It was reported that the sensitivity and specificity of TB staining in detecting high-risk pre-cancer lesions were 94% and 45%, respectively, whereas for carcinoma, the data were 100% and 39%. Therefore, TB staining combined with clinical examination may help monitor the development of OPMDs and identify severe dysplasia or cancer [6]. Besides, the operation of TB staining is simple and its result is easy for observation. However, the drawback of TB staining is the possibility of false results. For instance, the space between epithelial cells in inflammatory lesions can be enlarged, and TB could also stain neutrophils and bacterial nuclear fragments, which lead to false positive results. On the contrary, there is no nucleus in the stratum corneum for the epithelium with hyperkeratosis, which can be hardly stained by TB and may result in false negative results [7].

Lugol's Iodine Staining
Lugol's iodine staining was firstly used to assist the diagnosis of early esophageal cancer in 1770s, and applied in the early screening of cervical, oral and some other cancers afterwards. The underlying mechanism of Lugol's iodine staining depends on its specific interaction with the glycogen in the cytoplasm. As the content of glycogen in the cytoplasm of normal oral mucosal tissue is higher than that in the dysplastic and malignant tissue, the color of the two kinds of tissue is different after staining [8]. Specifically, the normal mucosal tissue turns brown, while the dysplastic and malignant tissue becomes white or light yellow due to the active metabolism and significantly reduced glycogen. Thus, as a simple, fast, non-invasive and inexpensive method, Lugol's iodine solution staining can be used to indicate the lesion sites and better delineate their boundary.
The clinical accuracy of Lugol's iodine was reported to be 76% in detecting dysplasia of OPMDs from other benign mucosal lesions, indicating its value in the detection of dysplasia and malignant transformation of OPMDs. However, false results might be observed. On one hand, the content of glycogen in epithelial cells is positively correlated with the differentiation degree of tumor cells. The glycogen content in mild epithelial dysplasia and highly differentiated squamous cell carcinoma was higher than that in severe dysplasia and lowly differentiated cancer cells, resulting in false negative results in the former two when stained with Lugol's iodine. On the other hand, false positive results may occur in benign lesions with thick keratinizing layer because of the decreased glycogen [9].
As TB staining shows a high false positive rate while Lugol's iodine staining has a high false negative rate, it was recommended that a combination of them may improve the detection accuracy up to 90% [10].

Rose Bengal Staining
Rose Bengal (RB) is a derivative of luciferin. The main staining mechanism of RB is its special affinity to DNA polymerase. RB dye binds DNA polymerase together in the form of ionic bond to stain the nucleus. RB also can strongly bind to the nuclear histones and other organelles [11]. The contents of nucleic acids and DNA polymerases are abnormally increased in dysplastic and malignant cells due to its rapid proliferation, resulting in different absorptions of RB in patients with OPMDs and OSCC. Therefore, RB staining was considered to assist the early diagnosis of precancerous malignant change and OSCC. It was found that the intensity of RB staining was correlated with dysplasia grades, indicating its application in identifying normal tissue, simple hyperplasia, mild, moderate and severe dysplasia. Nevertheless, false positive results might be caused by the inflammatory lesions, which are easier to be stained by RB [12].

Cytological Techniques
The process of cytological techniques is divided into two steps. The first step is cell collection from the oral mucosal surface, including traditional exfoliative cell

Traditional Exfoliative Cell Collection Technique
In the past few decades, the most commonly used traditional tools for collecting oral mucosa cells were metal spatula, wooden tongue depressor and cotton swab.
The truth is that the sample cells collected with these traditional tools are insufficient for further detection and analysis, which may cause a high false negative rate. Furthermore, metal spatula could induce trauma and infection, which is hardly acceptable to patients [13].

Cell Morphology Detection
The collected cells can be made into smears for cytological diagnosis. Oral liquid-based cytology (OLBC) is characterized by displaying the cell morphology and staining for the instant fixation of cells while removing unwanted substances [16]. High feasibility and effectiveness of OLBC with oral brush samples have been reported. The detection specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV) and accuracy rate of OLBC with brush biopsy of Orcellex ® were found to be 76%, 75%, 76%, 75% and 75%, respectively [17]. It was also reported that OLBC with clinical oral brush biopsy was simpler and more efficient than conventional cytology. The sensitivity, specificity, PPV and NPV of the cytological diagnosis were 97.5%, 68.8%, 88.76% and 91.7%, respectively, which may improve the diagnostic accuracy of oral brush biopsy in the future [18].

DNA Content/Ploidy Detection
DNA is the basis of cell growth, differentiation and proliferation. Genetic alterations caused by various carcinogenic factors lead to DNA content variation which was mainly detected by DNA flow or DNA-image cytometry on stained nucleus of cells obtained from clinical brush samples from OPMDs [19].
Aneuploidy refers to the number abnormality of DNA, which is represented with DNA index (DI), i.e. the ratio of the DNA content in the sample cells to that in the normal diploid cells. Aneuploidy with a presence of 20% -92% in the patients with oral epithelial dysplasia is often taken as a specific indicator of malignant transformation and carcinogenesis [20]. Also, it was found that the DI of the normal diploid cells was about 0.85 -1.15 and increased significantly in dysplastic and malignant cells [21].

Copy Number Aberrations Detection
Copy number aberrations (CNAs) refer to the abnormal deletions and duplications in the genome, which are associated with malignant transformation of OPMDs. CNAs can be analyzed by the array comparative genomic hybridization (aCGH), single-nucleotide polymorphism technique (SNP), and next generation sequencing (NSP) from extracted DNA [25].
Some specific CNAs have shown a high expression frequency of 40% -60% in OPMDs. It was reported that the average of CNAs was 1.9 in non-dysplastic mucosa and 6.5 in OPMDs. Besides, the gains of 20q13.31 -q13.33 and 5p13.33-pter, and the loss of 9p21.3 occurred more frequently in patients with OPMDs than the control, at frequencies of 62.5%, 50% and 50%, respectively [26]. Some studies manifested that the gains in 3q, 5p, 7p, 8q, 11q and 20q, and the losses in 3p, 8p, 9p and 18q, were the most commonly observed chromosome aberration regions [27].

Micronucleus Detection
Micronucleus is a centerless chromatid, or chromosome fragment, formed after genetic damage caused by toxic substances contained in tobacco and betel nut.
The newly formed cells of oral epithelium replace the exfoliative cells as they migrate to the epithelium surface. Thus, the genetic damage of the basal cells can be reflected by detecting the material of nucleus after staining the smear of exfo- liative cells [28]. It was found that micronuclei count increased significantly during the transformation of normal mucosa to OPMDs, which indicated that micronucleus can be used as a marker for the diagnosis of OPMDs. Besides, micronucleus detection of exfoliated epithelial cells was found to be a good biomarker for detecting chromosome loss or mitotic spindle dysfunction [29].

Optical Detection Techniques
The optical detection system has special light sources, which can be absorbed and reflected by abnormal and normal tissue of oral mucosa with different degrees. Currently, chemiluminescence techniques, such as ViziLite and ViziLite Plus emit chemiluminescent light to visually identify lesions, and autofluorescence techniques, including Visually Enhanced Lesion scope (VEL scope) and Identafi induce tissue fluorescence to detect suspicious tissue, both of which are commonly used in optical detection of OPMDS and oral cancer ( Figure 1).

ViziLite
As  As reported, the ViziLite showed a diagnosis accuracy of 100% sensitivity, 30% specificity, 26% PPV and 100% NPV for the low specificity malignancies.
Vizilite Plus identified both serious and benign abnormalities in the OPMDs and inflammation cases with sensitivity, specificity, PPV and NPV of 80%, 97.5%, 89% and 95%, respectively [31]. In the comparison of sensitivity and specificity between Vizilite and toluidine blue, the former was found to be more sensitive but less specific than the latter. Compared with Vizilite Plus with a sensitivity of 100%, specificity of 97.3%, PPV of 100% and NPV of 75%, Vizilite is more effective but less specific for detecting abnormal hyperplasia [32]. What should be mentioned is that the color difference of oral mucosa caused by Vizilite and/or Vizilite Plus may be related to epithelial thickness, which may lead to misdiagnosis. In addition, Vizilite was found to improve the visibility and clarity of white lesions, especially for oral leukoplakia, which can effectively distinguish white spots from normal tissue, but for the observation of red spots, the distinguishing ability dropped to 50%. Therefore, the effect of this system in detecting red lesions remains to be further studied [33].

Identafi
Angiogenesis is a known process in the progression of OPMDs and oral cancer.
The vascular activity in the dysplasia and malignancy tissue is more active than that in the normal tissue, which can be detected by Identafi device (DentalEZ

Saliva Detection Techniques
It is evident that molecular expression alternations are closely related to carci- The expression level of such molecules in saliva is usually 100 to 400 times lower than that in plasma, but they can also be detected through techniques with high detection sensitivity such as DNA microarray, miRNA-microarray, mass spectrometry and nanoscale biosensor, which makes saliva to be a diagnostic tool (Table 1). It is hoped that analysis of the genome, transcriptome, proteome and cellular metabolites will provide further insight into what happens at the subcellular level before clinically or histopathologically observable changes in OPMD.
Salivary diagnosis is minimally invasive compared with oral tissue biopsy. However, the detection of a single marker is far insufficient to achieve a reliable result, which means the combination of some other biomarkers or detection tools is required for accurate diagnosis [41].

Genomic Detection
Carcinogenesis is usually accompanied with genetic alterations, such as loss of  9p, 13q and 17p was highly correlated with carcinogenesis in oral cancer [42]. It was reported that LOH occurred at 3p and 9p appeared a 3.8-fold risk of malignant transformation than the normal. Furthermore, individuals with LOH in 4q, 8p, 11q, or 17p had a 33-fold increased carcinogenesis risk than the normal. Besides, the malignant risk degree of LOH occurred in 3p or 9p showed a 22.6-fold increase compared with the normal [43].
TP53 gene is a well-known tumor suppressor gene which inhibits cell cycle and initiates apoptosis in DNA damage. TP53 mutation is one of the most common genetic events occurred in the early stage of OPMDs malignant transformation [44]. and miRNA -3065-5p appeared a downregulation trend [48]- [54]. Such miRNAs can be used as biomarkers for predicting malignancy of OPMDs to some extent.
Once more data are obtained to analyze the miRNA expression of patients and a reference threshold of miRNA expression is developed, miRNA may be used as a reference index for OPMDs detection.

Protein Detection
Although numerous proteins in saliva have been reported as potential biomarkers, no one can be used as a direct and evident diagnostic indicator for OPMDs and OSCC. Thus, mass spectrometry and enzyme-linked immune sorbent assay (ELISA) technique is commonly used for proteinic quantitative analysis to provide auxiliary diagnostic approach for OPMDs.
Recently, it was found that the expression levels of complement factor H (CFH), fibrinogen alpha chain (FGA) and alpha-1-antitrypsin (SERPINAL1) were of great clinical significance to the conversion of OPMDs to OSCC [55].

Sialic Acid Detection
Sialic acid, also known as N-acetylneuraminic acid, is a derivative of 9-carbon monosaccharide, serving as an important component of receptors on the surface of cell membrane. It could react with diphenylamine and its content can be quantitatively determined by spectrophotometer [61]. In OPMDs and OSCC, the level of sialic acid is found to increase due to the secretion, transformation and shedding of malignant cells at the early stage of carcinogenesis. It was reported that the sialic acid content in both OPMDs and OSCC was significantly higher than that in normal control. Moreover, the sialic acid content in OSCC was almost nine times more than that in OPMDs. It was also found that the average levels of salivary sialic acid in the control, OPMDs and oral cancer were 40.373, 57.562 and 80.422 mg/dl, respectively, demonstrating a significant increase of sialic acid in OPMDs and oral cancer [62].

Reactive Oxygen Species and Reactive Nitrogen Species Detection
Reactive oxygen species (ROS) and reactive nitrogen species (RNS) are reactive Nitric oxide (NO), as a representative substance of RNS, is an important biological signal molecule. It was found that the salivary NO concentration in control group (median = 4.21 ug/ml) was significantly lower than that in OPMDs group (median = 12.91 ug/ml, P < 0.001) [64]. Another study reported that ROS in oral lichen plaus (OLP) group was significantly increased compared with control group. Moreover, the levels of NO and nitrite in saliva of OLP patients were significantly higher than that in control group [65]. Thus, it seems evident that RNS could be a biomarker for the early diagnosis of OPMDs.

Optical Coherence Tomography: OCT
Optical coherence tomography (OCT) is a non-invasive, high-resolution optical Advances in Bioscience and Biotechnology imaging technique that captures images by measuring the intensity of light reflected by a target tissue. With a penetration depth of 1 -3 mm, OCT detects dysplasia by the fluctuation of light scattering due to cellular changes in dysplastic tissue in comparison to normal mucosa. This method is more accurate in detecting pathological tissue and can be used to roughly determine the edges and contours of OPMDs and OSCC [66]. It was believed that OCT can distinguish the mucosa with mild and moderate dysplasia by imaging the structures of the oral epithelial layer and lamina propria. The sensitivity, specificity, PPV and NPV of OCT for OPMDs and OSCC were 85%, 78%, 86.5% and 77.5%, respectively. Therefore, OCT can be used as a simple auxiliary technique in the detection of OPMDs and OSCC [67].

Reflectance Confocal Microscopy: RCM
Reflective confocal microscopy (RCM) has been widely used in dermatology, and gradually applied to the observation of dental hard and soft tissue. As the refractive index of the detected components is different from that of the surrounding media, the microscopic details of the tissue and high-quality images for detection and diagnosis could be obtained by tomographic scanning, possibly with other techniques [68]. It works not only in detecting benign and malignant lesions, but also in identifying inflammatory reaction. Experimental results showed that RCM in vivo had a high consistency with histological detection. It has the potential as a non-invasive approach for the early detection of OPMDs and OSCC [69].

Conclusion
With the potential of malignant transformation, OPMDs may even develop into OSCC with poor prognosis, indicating that the surveillance of malignant progression for OPMDs may be the most effective way of oral cancer prevention.
Since traditional approaches for the early diagnosis of OPMDs like tissue biopsy These techniques help to undertake oral mucosal screening, but additional levels of education including the level of knowledge and understanding of oral mucosal lesions do influence the effectiveness of the techniques. Besides, as extensively discussed above, each currently-available non-invasive approach shows its own advantages and drawbacks. It seems that a combination of such methods may overcome their limitations and improve detection accuracy for long-term disease surveillance of OPMDs.