Department of Oral and Maxillofacial Surgery and Oral Pathology, VU University Medical Center (VUmc)/Academic Centre for Dentistry Amsterdam (ACTA), Amsterdam, The Netherlands
Correspondence: Prof. Isaäc van der Waal, Department of Oral and Maxi-llofacial Surgery and Oral Pathology, VU University Medical Center, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands. Tel: +31 20 4444039, Fax: +31 20 4444046, E-mail: firstname.lastname@example.org
Oral leukoplakia is a potentially malignant disorder of the oral mucosa. The aim of this retrospective study was to identify the factors that possibly predict malignant transformation in a well-defined cohort of patients with a long-term follow-up. All leukoplakias were staged according to a clinicopathological classification and staging system. Furthermore, a certainty factor has been used with which the diagnosis has been established.
Material and methods
The group consisted of 144 patients. The size, presence and degree of epithelial dysplasia were incorporated into a clinicopathological classification and staging system. Initial management consisted of surgical excision, CO2 laser vaporisation or observation only. The mean follow-up period was 51.2 months (s.d. = 39.33, range 12–179 months).
In 16 of 144 patients (11%), malignant transformation occurred between 20 and 94 months (mean 57.0 months) after the first visit, the annual malignant transformation rate being approximately 2.6%. A large size of the lesion (≥ 4 cm) showed to be the only statistically significant predictor of malignant transformation (P =0.034).
A size of ≥ 4 cm showed to be the only significant predicting factor of malignant transformation in oral leukoplakia. No other epidemiological, aetiological, clinical or histopathological parameters were of statistical significance.
Oral leukoplakia is a potentially malignant disorder which means that in this morphologically altered tissue, squamous cell carcinoma is more likely to occur than in its apparently normal counterpart (Warnakulasuriya et al, 2007). The reported annual malignant transformation of oral leukoplakia into oral squamous cell carcinoma (OSCC) is approximately 1–2% (van der Waal, 2009). Several factors have been suggested to predict an increased risk of malignant transformation of oral leukoplakia, such as age, gender, tobacco habits, homogeneity and size of the lesion, oral subsite, degree of epithelial dysplasia, if present, loss of heterozygosity, survivin, matrix metalloproteinase 9 and DNA content (Schepman et al, 1998; Dietrich et al, 2004; Holmstrup et al, 2006; Fillies et al, 2007; Smith et al, 2009; van der Waal, 2009; Bremmer et al, 2011; Brouns et al, 2012b). The possible role of human papilloma virus infection with regard to malignant transformation of oral leukoplakia is yet unclear (Syrjanen et al, 2011; Feller and Lemmer, 2012). Management of oral leukoplakia consists of surgical excision, laser surgery, CO2 laser vaporisation and observation (Mehanna et al, 2009; Ribeiro et al, 2010).
The aim of the present retrospective study was to identify the factors that possibly predict malignant transformation in a well-defined cohort of patients with a long-term follow-up. All leukoplakias were staged according to a clinicopathological classification and staging system. Furthermore, a certainty factor has been used with which the diagnosis has been established.
Material and methods
For the purpose of this retrospective study, 144 consecutive patients were included who were referred to the Department of Oral and Maxillofacial Surgery and Oral Pathology at VU Medical Center/ACTA, Amsterdam, the Netherlands, between 1997 and 2012 and in whom a definitive clinicopathological diagnosis of oral leukoplakia was established. In case of possible aetiological factors, a period of maximum 3 months was allowed to evaluate the results of elimination of these factors. As a result, only patients with a diagnostic certainty factor 3 and 4 have been included (Table 1) (van der Waal and Axell, 2002; Brouns et al, 2012c). The follow-up period lasted at least 12 months. Patients who developed a malignancy within a period of 12 months after the biopsy (n =2) have been excluded beforehand. In these two patients, the initial biopsy may not have been representative.
Table 1. Certainty (C) factor of a diagnosis of oral leukoplakia
C1 Evidence from a single visit, applying inspection and palpation as the only diagnosis means (provisional clinical diagnosis), including a clinical picture of the lesion.
C2 Evidence obtained by a negative result of elimination of suspected aetiological factors, for example, mechanical irritation, during a follow-up period of 6 weeks (definitive clinical diagnosis)
C3 As C2, but complemented by pretreatment incisional biopsy in which, histopathologically, no definable lesion is observed (histopathologically supported diagnosis)
C4 Evidence following surgery and pathological examination of the resected specimen
The group of 144 patients consisted of 44 men and 100 women, with a mean age of 58.7 years (s.d. = 14.11, range 26–98 years) (Figure 1). The use of tobacco and alcohol was registered in a simplified manner as being user, non-user or unknown (Table 2).
Table 2. Distribution of tobacco and alcohol habits among 144 patients with oral leukoplakia
Total n =144
Malignant transformation n =16
In this study, the definition of leukoplakia, as proposed by the WHO in 2005, has been used: ‘A predominantly white plaque of questionable risk having excluded (other) known diseases or disorders that carry no increased risk for cancer’(Warnakulasuriya et al, 2007). Clinically, a distinction was made between homogeneous (flat, thin or wrinkled, uniform white in colour) and non-homogeneous (white or white-and-red, either speckled, nodular or verrucous) leukoplakia (Warnakulasuriya et al, 2007). There were 65 patients with a homogeneous leukoplakia and 79 with a non-homogeneous type. Although many patients had multiple or widespread leukoplakias, not a single one fully qualified to be characterised as proliferative verrucous leukoplakia as being described in the literature (Hansen et al, 1985; Cerero-Lapiedra et al, 2010).
The location of leukoplakia was specified according to eight subsites: (i) tongue, (ii) floor of mouth (FOM), (iii) lower lip, (iv) hard palate, (v) buccal mucosa, (vi) upper alveolus and gingiva, (vii) lower alveolus and gingiva, and (viii) multiple sites (Table 3). In all patients, clinical pictures of the oral leukoplakia were taken at the first visit and in most cases also during follow-up, particularly in case of recurrence or malignant transformation. The clinical picture taken at the first visit was used for this study. The size, presence and degree of epithelial dysplasia were incorporated into the OL-classification and OL-staging system (Table 4) (van der Waal, 2009). There were 95 incisional and 49 excisional biopsies. In eleven patients, there was an additional excision specimen available for histopathological grading. In five of the eleven patients, the final histopathological examination resulted in a different P-classification; the highest pathological score was used in these cases. All histological sections were revised by the same pathologists (EB/IW); in a limited number of cases that were graded differently, a consensus was arrived (Table 5). In 53 patients, PAS-D sections were available for assessment of the presence of Candida albicans. In 14 of 53 (26%) sections, Candida albicans was present.
Table 3. Site distribution of 144 patients with oral leukoplakia
Total n =144
Malignant transformation n = 16
Tongue (dorsal and lateral surfaces)
Floor of mouth
Upper alveolus and gingiva
Lower alveolus and gingiva
Table 4. Classification and staging system for oral leukoplakias (OL-system)
General rules of the OL-staging system: (i) If there is doubt concerning the correct L category to which a particular case should be allotted, then the lower (i.e. less advanced) category should be chosen. This will also be reflected in the stage grouping; (ii) in case of multiple biopsies of single leukoplakia or biopsies taken from multiple leukoplakias, the highest pathological score of the various biopsies should be used; (iii) carcinoma in situ has been excluded from this classification; (iv) for reporting purposes, the oral subsite according to the ICD-DA should be mentioned (World Health Organisation, International Classification of Diseases. Tenth Revision. Application to Dentistry and Stomatology, ICD-DA, Geneva, 1992).
L (size of the leukoplakia)
Size of single or multiple leukoplakias together < 2 cm
Size of single or multiple leukoplakias together 2–4 cm
Size of single or multiple leukoplakias together > 4 cm
Size not specified
No epithelial dysplasia (includes ‘no or perhaps mild epithelial dysplasia’)
Mild or moderate epithelial dysplasia
Severe epithelial dysplasia
Absence or presence of epithelial dysplasia not specified in the pathology report
L3P0 or L1L2P1
L3P1 or any L P2
Table 5. Size, histopathology and stage in 144 patients with a definitive clinicopathological diagnosis of oral leukoplakia
Total n =144
Malignant transformation n =16
Management and treatment
The initial treatment was performed within 6 months after the first visit to the clinic. In 71 of 144 patients (49.3%), initial treatment consisted of surgical excision (n =49) or CO2 laser vaporisation (n =22). CO2 laser vaporisation was mainly used in non-dysplastic leukoplakias and leukoplakias at sites that were less amenable for surgical excision, such as floor of the mouth and mucobuccal folds. In most cases, a margin of a few millimetres clinically normal-looking mucosa was observed. In 73 of 144 (50.7%) patients, initial management consisted of observation. The treatment modalities during follow-up of these patients consisted of surgical excision (n =11) and CO2 laser vaporisation (n =13). Forty-nine patients never underwent any type of treatment.
The follow-up period ranged from 12 to 179 months with a mean of 51.2 months (s.d. = 39.33). Follow-up visits were scheduled at 3- or 6-month intervals. The follow-up period started at the first visit of the patient to the clinic. It ended in case of lost to follow-up, death, development of OSCC at the site of the leukoplakia or elsewhere in the oral cavity or in the head and neck region.
The design of this study adheres to the code for proper secondary use of human tissue of the Dutch Federation of Biomedical Scientific Societies (http://www.federa.org) (Oosterhuis et al, 2003).
Minimum, maximum and mean values of continuous variables were calculated. Statistically significant relations were tested with the chi-square test. The results were statistically significant if the P-value was < 0.05. The limited number of patients did not allow to perform a multivariate analysis. For malignant transformation, a survival curve was plotted according to the Kaplan–Meier method. For all statistical analyses, SPSS 17.0 for Windows (SPSS Inc, Chicago, IL, USA) was used.
In 16 of 144 (11%) patients, four men and twelve women, malignant transformation occurred during follow-up (Figure 2). There was no statistically significant relation between malignant transformation and gender (P =0.609). The age of these sixteen patients ranged from 26 to 81 years (s.d. = 14.11, mean 58.2 years). The age of the patients with malignant transformation was categorised into 60 years and older (n =10) and < 60 years (n =6); age was not related to malignant transformation (P =0.593). Five patients had tobacco habits, and six patients had regular use of alcohol. Also, tobacco habits and regular alcohol use were not associated with malignant transformation (P =0.894 and P =0.953, respectively). Malignant transformation occurred between 20 and 94 months (mean 57.0 months) after the first visit, the annual malignant transformation rate being approximately 2.6%.
Four oral leukoplakias were homogeneous and 12 non-homogeneous. No statistically significant relation was found between the clinical appearance and malignant transformation (P =0.086). The oral leukoplakias were located at the tongue (n =8), FOM (n =1), lower lip (n =1), hard palate (n =2), buccal mucosa (n =2) and multiple sites (n =2). No statistically significant relation was found between the oral subsite and malignant transformation (P =0.144), even when the subsites were subdivided into high risk (floor of mouth and tongue) versus low risk (remaining oral subsites) (P =0.408). All carcinomas arose at the same subsite of the oral leukoplakia.
Four oral leukoplakias were L1 in size, four were L2 and eight lesions were L3. A large size of the lesion (L3 ≥ 4 cm in largest diameter) showed to be a predictor for malignant transformation (P =0.034).
The initial biopsy showed no dysplasia (P0) in 10 patients, while in two patients mild/moderate dysplasia (P1) and in two other cases severe epithelial dysplasia (P2) were observed. No statistically significant relation was found between the presence and degree of epithelial dysplasia and malignant transformation (P =0.604); this was also true when using a binary system, that is, no dysplasia (P0) and presence of dysplasia (P1 and P2) (P =0.334). The patients were also staged according to the OL-classification and OL-staging system: Stage 1 (n =1), Stage 2 (n =3), Stage 3 (n =7) and Stage 4 (n =5). Stage was not a statistically significant factor (P =0.187). Of the 16 patients with a malignant transformation, PAS-D sections were available in eight cases of which four were positive for Candida albicans (P =0.101).
Management and treatment
The initial treatment of the 16 patients with a malignant transformation consisted of surgical excision (n =3), CO2 laser vaporisation (n =3) and observation (n =10). Treatment during follow-up of the patients with initial observation consisted of surgical excision (n =2) and CO2 laser vaporisation treatment (n =2). Six patients (37.5%) never underwent any type of active treatment.
Of the 60 patients who underwent surgical excision as the initial treatment (n =49) or during follow-up (n =11), 25 (42%) patients had a recurrence in a mean follow-up period of 48.8 months. There was no statistically significant relation between recurrence after surgical excision and malignant transformation (P =0.134). Of the 35 patients who underwent CO2 laser vaporisation as the initial treatment (n =22) or during follow-up (n =13), 14 (40%) patients had a recurrence in a mean follow-up period of 61.9 months (Brouns et al, 2012a). There was no statistically significant relation between recurrence after CO2 laser vaporisation and malignant transformation (P =0.324) (Brouns et al, 2012a).
Patients who underwent active treatment (surgical excision or CO2 laser vaporisation) (n =95) did not have a statistically significant lower risk of malignant transformation than patients managed by only observation (n =49) (P =0.756).
In the present study, 144 patients with a histopathological diagnosis of oral leukoplakia and a minimum follow-up of 12 months were included. Moreover, a classification and staging system has been used (Tables 1 and 4) (van der Waal and Axell, 2002; van der Waal, 2009; Brouns et al, 2012c). There were 16 of 144 (11%) patients in whom a malignant transformation occurred during the follow-up, resulting in an annual malignant transformation rate of 2.6%. Ninety-five of 144 (66%) patients underwent any type of treatment at the initial stage or during follow-up, while in a previous study from our institute, only 87 of 166 (52%) underwent any type of active treatment (Schepman et al, 1998). In the latter study, the annual malignant transformation rate amounted 2.9% (Schepman et al, 1998). Apparently, active or passive management policy is not related to the risk of malignant transformation. This observation has also been made in various other studies (Schepman et al, 1998; Lodi et al, 2006; Arduino et al, 2009). It is actually unknown whether the width of the margin, either in surgical removal or CO2 laser vaporisation, would be of relevance in this respect. At present, no molecular markers are available to determine the optimal width of the margin.
The gender distribution of the entire patient group showed a female preponderance with a male/female ratio of approximately 1:2,3 (Figure 1). Of the patients with malignant transformation, the male/female ratio is even 1:3,7. In most long-term follow-up studies of oral leukoplakia, there is a preference for males in the entire patient group (Banoczy, 1977; Lumerman et al, 1995; Saito et al, 1999; Liu et al, 2011). However, other studies have shown that female gender was a higher risk of malignant transformation (Silverman et al, 1984; Schepman et al, 1998). Despite the high female preponderance in the present study, no statistically significant difference was found between gender and malignant transformation (P =0.609).
A large size of the lesion (L3 ≥ 4 cm) showed to be the only significant predictor of malignant transformation (P =0.034). This relationship has also been shown in another study (Holmstrup et al, 2006). As mentioned by Saito et al (Saito et al, 1999), widespread multiple leukoplakias have a higher potential for the development of cancer than the localised ones. These authors also mentioned that their multiple leukoplakias probably represented examples of proliferative verrucous leukoplakia. The latter term has been the subject of discussion related to its definition for several decades. Anyhow, every leukoplakia, how small and how benign histopathologically looking, may turn into malignancy and widespread multiple leukoplakias apparently carry a higher risk of malignant transformation than localised small ones.
Of the patients with a malignant transformation, eight carcinomas arose at the tongue and not one at the upper or lower alveolus and gingiva (Table 3). There was no statistically significant relation found between the oral subsite and malignant transformation (P =0.144) and also not when the subsites were subdivided into high risk (floor of mouth and tongue) versus low risk (remaining oral subsites) (P =0.408). We also did not identify ‘alveolar ridge keratosis’ as a separate entity, and we did not exclude cases of ‘frictional keratosis of the attached gingiva’ that did not disappear after changing the brushing habits (Chi et al, 2007; Natarajan and Woo, 2008; Mignogna et al, 2011).
Other reported predicting factors of malignant transformation such as age, the absence of tobacco habits, heterogeneity of the lesion, presence of C. albicans and presence and degree of epithelial dysplasia were not found in the present study (Schepman et al, 1998; Holmstrup et al, 2006; Liu et al, 2010; Bremmer et al, 2011; Warnakulasuriya et al, 2011; Ho et al, 2012).
In 11 patients, surgical excision was performed after an incisional biopsy. The advantage of additional surgical excision after an incisional biopsy is the availability of a surgical specimen for additional histopathological examination. It has been shown that the histopathological findings in an incisional biopsy from leukoplakia are not always representative of the entire lesion (Vedtofte et al, 1987; Holmstrup et al, 2007).
No reliable comparison can be made in our study between the treatment results of CO2 laser vaporisation and cold-knife surgery, because of different indications for both treatment modalities in the present study. In the literature, no randomised controlled trials are available to compare both treatment modalities with regard to the local recurrence rate and the risk of malignant transformation. For both treatment modalities, high recurrence rates have been reported in the literature (Vedtofte et al, 1987; Brouns et al, 2012a; Kuribayashi et al, 2012). The present study showed for both surgical excision and CO2 laser vaporisation treatment a recurrence rate of approximately 40% during a mean follow-up period of 48.8 months and 61.9 months, respectively.
The annual malignant transformation rate (2.6%) in this study may be difficult to compare with that of other studies, due to possible differences in study populations, different tobacco and alcohol habits, different diagnostic criteria used for oral leukoplakia and allied white lesions, different treatment strategies and perhaps also different follow-up policies. A size of ≥ 4 cm showed to be the only significant predicting factor of malignant transformation in oral leukoplakia. No other epidemiological, aetiological, clinical or histopathological parameters were shown to be of statistical significance. Although the efficacy of removal of oral leukoplakia with regard to the risk of malignant transformation is uncertain, such removal should be considered in each patient with such lesion, particularly also in the extensive ones.
The study design came from Dr. I. van der Waal. Drs. J. A. Baart and K. H. Karagozoglu treated a large number of the patients. Dr. E.R.E.A. Brouns analysed all data and wrote the manuscript. The statistical analysis was carried out by Mrs. I.H.A. Aartman and Dr. E. R. E. A. Brouns. Drs. E. Bloemena and I. van der Waal revised the histopathological slides.