Non-invasive visual tools for diagnosis of oral cancer and dysplasia: A systematic review

Background Gold standard for the diagnosis of oral dysplasia (OD) oral squamous cell carcinoma (OSCC) and malignant lesions is the histological examination. Several adjunctive diagnostic techniques have been proposed in order to increase the sensitivity (SE) and specificity (SP) of conventional oral examination and to improve the diagnostic first level accuracy. The aim of this study is to perform a systematic review on non-invasive tools for diagnosis of OD and early OSCC. Material and Methods Medline, Scopus, Web of Knowledge databases were searched, using as entry terms “oral dysplasia AND diagnosis” / ”oral cancer AND diagnosis”. Data extracted from each study included number of lesions evaluated, histopathological diagnosis, SE, SP, positive and negative predictive values (PPV and NPV), diagnostic accuracy (DA) and the main conclusions. Results After title and abstract scanning of 11.080 records, we selected 35 articles for full text evaluation. Most evaluated tools were autofluorescence (AF), chemiluminescence (CL), toluidine blu (TL) and chemiluminescence associated with toluidine blue (CLTB). Conclusions There is a great inhomogeneity of the reported values and there is no significant evidence of superiority of one tool over the other. Further clinical trials with a higher level of evidence are necessary in order to assess the real usefulness visual diagnostic tools. Key words:Oral dysplasia, oral cancer, diagnosis, visual diagnostic tool, systematic review.


Introduction
Oral squamous cell carcinoma (OSCC) is the sixth most common malignant tumour, with an incidence of more than 500.000 cases per year (1). The most important prognostic factor influencing the disease-specific survival rate is the tumour stage at diagnosis. Patients with stage I tumours have a 5-year survival rate of 75%, which dramatically decreases in patients with tumours in stage III or IV, being 49% and 30%, respectively (1,2). The diagnostic pathway for oral suspicious lesions usually starts with the conventional objective examination (COE) based on inspection and palpation of the oral mucosa with the support of an incandescent light available on the dental chair. It is well known that COE mainly depends on a subjective interpretation, which is a consequence of the experience of the operator. Moreover, oral epithelial dysplasia (OED) and early OSCC may already be present within areas of oral mucosa macroscopically normal, as well as within the context of oral potentially malignant disorders such as leukoplakia, erythroplakia, submucous fibrosis and oral lichen planus (3). The gold standard for the diagnosis of oral dysplastic and neoplastic malignant lesions is the histological examination (4). Incisional or excisional biopsy techniques are the most reliable methods to collect a surgical specimen suitable for microscopic evaluation. However, despite the little invasivity of such techniques, they still have some disadvantages in terms of morbidity and possible artifacts induced by the method of collection. In a recent paper, Mehrotra et al. indicated that there are two approaches for detection of oral dysplasia and cancer: 1) oral cancer screening programs that identify asymptomatic patients with suspicious lesions and 2) specific diagnostic tools to identify dysplasia and early oral cancers in asymptomatic patients with an oral abnormality (5). Several visual diagnostic aids have been developed as adjunctive tools in order to increase the diagnostic accuracy (DA) and enhance the specificity (SP) and sensitivity (SE) of the conventional diagnostic pathway. However, results of studies on the usefulness of such tools show impressive discrepancies with regard to values such as the positive or negative predictive values (PPV, NPV), when the same tools is evaluated by different researchers. The aim of this study is to perform a systematic review on non-invasive tools for the diagnosis of OED and OSCC, taking into account factors as SE, SP, PPV, NPV and DA.

Material and Methods
The databases Medline, Scopus and Web of Knowledge were searched, using as entry terms "oral dysplasia AND diagnosis" / "oral cancer AND diagnosis". No time limits were specified in the present research.
Search flow is shown in figure 1. Papers with abstract unavailable were excluded for further evaluation. Titles and abstract were screened and the following exclusion criteria were applied: -papers not in English. -studies ex vivo or based on animal models. -typology of the study: case reports, case series with less than 10 patients, conference proceedings, personal communications, editorials, descriptive studies and reviews.
-studies that analyse salivary biomarkers.
-studies including also tumours of other head and neck regions (e.g. oropharynx). Papers with equivocal abstracts were included for fulltext evaluation. Further studies were excluded after full-text reading, if not pertinent with aim of the present review. Data extracted from each study included authors and publication year, typology of the study, diagnostic tool analysed, number of lesions evaluated, (if present) histopathological diagnosis, (if present) SE, SP, PPV, NPV, DA and the main conclusions of the study (Tables 1 and  1 continue,2). SE and SP measure the accuracy of a test without any relation to the disease or population, whereas PPV and NPV measure the proportion of people whose test results reflect their health status. DA is the proportion of true positive results (both true positive and true negative) in a selected population, with regard to a specific disease. The mean value of each variable analysed was calculated; range and standard deviation (SD) were indicated for samples having > 2 values. Level of evidence of each study was assessed according to the Oxford Evidence-based Medicine (OEBM) Levels for Diagnosis updated in March 2009.

Discussion
The principles of functioning of non-invasive visual diagnostic tools for OSCC and dysplastic lesions are very different, being based on diverse specific cellular and tissue characteristics. Such a great diversity may partly explain the impressive discrepancy of results obtained in the studies analysed. Another reason which can give some reasons for the wide range of results, in terms of SE, SP and DA is the great variability both of the typology of the studied lesions and of the diagnostic criteria used for the clinical and histological assessment of such lesions. The difficulty to establish univocal and broadly-accepted criteria for the assessment of the OED has been widely reported, particularly, with regard to the inter-and intra-observer disagreement for the diagnosis. Moreover, SE and SP may well depend on the degree of development of a lesion, seeming quite reasonable that both these indicators increase with the progression of a lesion from normal, to dysplastic, early neoplastic and invasive and destructive lesion. Taking into account the abovementioned considerations, we report a short discussion for each tool analysed: -Auto fluorescence (AF) -Direct visual fluorescence examination (DVFE) Auto fluorescence (AF) uses natural fluochromes which are located within the epithelium and the submucosa and which are excited when irradiated with specific wavelengths. Using wavelengths between 375 and 440 nm, some fluochromes show fluorescence in the range of the green colour. Following such irradiation, normal, unaltered mucosa emits a pale green AF light when viewed through a selective, narrow-band filter. A proper filtration is crucial, due to the intense light used for excitation of the fluorochromes (13,15). Areas of reduced AF (dark areas) are suspicious for epithelial dysplasia or OSCC, whereas normal mucosa appears bright green (10). The VELscopeTM (LED Medical Diagnostics Inc., Barnaby, Canada) system consists of a non-invasive device designed to visualise early mucosal changes using the principles of tissue AF. According to such principles, dysplastic changes should be associated with a loss of stromal AF (29,32). It seems of paramount importance to highlight here that benign lesions, or those associated to inflammation, can also be characterized by a loss of stromal AF, which grossly limits the diagnostic specificity, especially in low-risk populations.  (21). Light emission from the laser source is guided to the oral mucosa through a 3 µm long bifurcated fiber optic probe that has a central fiber to deliver the excitation beam and 6 surrounding fibers (400 µm diameter each) to collect AF emissions. The red to green colour ratio is defined as the numerical color value (NCV). Two studies regarding this tool have been selected for this review. SE and SP values are reported in 1 study only and they are high (SE: 100%-95%; SP: 96%-86%, according to the histopathological diagnosis), but the OEBM level is low (4). Data from further studies with a higher OEBM level are necessary.
-5-aminolevulinic acid (ALA) induced protoporphyrin IX (PPIX) fluorescence Only one perspective study describing this technique was selected for this review (29). Topical or systemic administration of 5-ALA results in a selective accumulation of PPIX in neoplastic tissue, which is probably due to altered activity levels of the enzymes of the heme biosynthetic pathway within malignant transformed cells. In the protocol of Leuing et al., the patients performed a 15-minute continuous rinsing of the oral cavity using the 5-ALA solution. After an incubation period of 1 to 2.5 hours (maximum contrast after 1.5h), fluorescence investigation was performed. In 13.8% of the patients, additional findings like dysplasia, carcinoma in situ, OSCC were found through fluorescence in contrast to COE (27). An evaluation of the biopsy specimens resulted in a SP of 60% and a SE of 99% (29). ALA-induced fluorescence could represent a possible useful new diagnostic tool to detect early malignant lesions in the oral cavity. However, further studies seem to be necessary.
-Optical spectroscopy Optical spectroscopy is a non-invasive diagnostic method that has been investigated in many forms including fluorescence spectroscopy (FS), elastic or diffuse scattering spectroscopy (ESS), and Raman spectroscopy (RS). Spectroscopic measurements can detect biochemical and architectural alterations in tissue that are related to the carcinogenesis. These alterations may include changes in the concentrations of native fluorophores such as collagen, elastin, keratin, nicotinamide adenine dinucleotide (NADH), and flavin adenine dinucleotide (FAD); changes in hemoglobin concentration and oxygenation; increasing epithelial thickness; increasing nuclear size and nuclear/cytoplasmic ratio; change in vascularization (20,25,30,31). These principles are employed within experimental methods; SE and SP values seem to be high, but there is need for more data. Only one study for each optical spectroscopy method was identified and they had a low OEBM level (4, 3b and 2b, respectively).