Oral squamous cell carcinoma (OSCC) is a lethal disease, with rising incidence. There were 6767 new OSCC cases and 2056 deaths in the UK in 2011. Cancers are preceded by oral potentially malignant disorders (PMDs), recognizable mucosal diseases harbouring increased SCC risk, offering clinicians a ‘therapeutic window’ to intervene. Contemporary practice remains unable to predict lesion behaviour or quantify malignant transformation risk. No clear management guidelines exist and it is unclear from the literature whether early diagnosis and intervention prevents cancer. Between 1996 and 2014, 773 laser treatments were performed on 590 PMD patients in Newcastle maxillofacial surgery departments. The efficacy of the intervention was examined by review of the clinicopathological details and clinical outcomes of the patients (mean follow-up 7.3 years). Histopathology required up-grading in 36.1% on examining excision specimens. Seventy-five percent of patients were disease-free, mostly younger patients with low-grade dysplasia; 9% exhibited persistent disease and were generally older with proliferative verrucous leukoplakia. Disease-free status was less likely for erythroleukoplakia ( P = 0.022), ‘high-grade’ dysplasia ( P < 0.0001), and with lichenoid inflammation ( P = 0.028). Unexpected OSCC was identified in 12.0%, whilst 4.8% transformed to malignancy. Interventional laser surgery facilitates definitive diagnosis and treatment, allows early diagnosis of OSCC, identifies progressive disease, and defines outcome categories. Evidence is lacking that intervention halts carcinogenesis. Multicentre, prospective, randomized controlled trials are needed to confirm the efficacy of surgery.
Oral squamous cell carcinoma (OSCC), probably the most common head and neck malignancy, remains lethal and deforming due to local invasion, orofacial destruction, metastasis to cervical lymph nodes, and blood–borne tumour dissemination. In 2011 there were 6767 new cases and 2056 deaths in the UK, whilst globally OSCC shows a rising incidence and a prognosis significantly compromised by advanced disease and late presentation.
The ‘progression model’ for oral carcinogenesis proposes that, following irreversible genotypic mutation, phenotypic epithelial disorganization and dysmaturation changes occur, which, if allowed to progress, lead to invasive OSCC. Such features preceding cancer are identifiable microscopically and are collectively termed dysplasia. In clinical practice, assessment of the degree of dysplastic change is made following incision biopsy and classification into mild, moderate, or severe categories, according to the extent of atypical epithelium.
It has been recognized for many years that a spectrum of distinct mucosal abnormalities, now termed potentially malignant disorders (PMDs) and which accompany dysplasia, may be identified clinically, albeit non-specifically, during oral examination. There is, therefore, an opportunity for early diagnosis and therapeutic intervention during this ‘oral pre-cancer window’. PMD encompasses primarily localized lesions such as leukoplakia, erythroplakia, and erythroleukoplakia, together with progressive multifocal disorders such as proliferative verrucous leukoplakia (PVL).
The identification of a PMD in a patient does not mean that malignant transformation is inevitable. Many lesions do not progress, whilst others resolve spontaneously, but it remains impossible in clinical practice to predict behaviour for individual cases. Nonetheless, patients with PMDs remain at increased risk of OSCC.
‘Potential malignancy’ is difficult to define, so it is unsurprising that previous treatment interventions have been non-specific with ill-defined goals and end-points; the few clinical trials in the literature concentrate on medical therapies and are compromised by small numbers of patients and short study durations. No study has demonstrated long-term lesion resolution, reduction in disease incidence, or prevention of malignant transformation. Arduino et al. emphasized the need for long-term patient studies to improve understanding of the natural history of PMDs.
A lack of insight into the progression of PMDs has resulted in a variety of proposed treatments based upon clinician preferences and experience. PMDs are mucosal lesions only, so do not require the aggressive treatment necessary for removal or destruction of invasive cancer. It seems self-evident, therefore, to intervene early to excise dysplastic tissue at this ‘pre-invasive’ stage.
Interventional laser therapies evolved following the demonstrable failure of observational and medical treatments and limitations in conventional surgery to treat oral PMDs. The use of carbon dioxide (CO 2 ) laser facilitates surgical management, providing precise excision, full histopathological lesion assessment, minimal postoperative morbidity, and coordinated patient follow-up. Whilst excision is preferred, a limited role for ablation allows a defocused laser beam to destroy small or less dysplastic lesions, particularly at gingival/alveolar sites.
Previous studies on the treatment of PMDs are limited by small numbers of patients and varying follow-up periods. Detailed, long-term analysis of a large, well-defined PMD patient cohort undergoing coordinated treatment and follow-up was overdue.
In the absence of national PMD disease statistics or a robust evidence base for treatment, the aims of this study were to profile patient demography and clinicopathological data for a patient cohort presenting to a specialist PMD service in the north-east of England, to determine clinical outcomes following interventional CO 2 laser surgery and to assess treatment efficacy and its influence on the progress of oral carcinogenesis.
Caldicott approval was obtained from the Newcastle University/Newcastle upon Tyne Hospitals NHS Foundation Trust to facilitate anonymized, retrospective data collection from the medical records, operating logs, and pathology reports of patients with PMDs treated at Newcastle Dental Hospital and the Royal Victoria Infirmary. Inclusion criteria required a new, untreated, single-site PMD, confirmed by incision biopsy. Patients with a previous or multifocal PMD and those with a history of OSCC or head and neck radiotherapy were excluded.
Patient and treatment details
The following demographic and clinicopathological details were collected for patients who had undergone CO 2 laser surgery between August 1996 and December 2014: date of treatment, age and sex, clinical appearance and anatomical site of the oral lesion, histopathology diagnoses for incision and post-laser excision biopsies, follow-up data, and any further laser treatment. The clinical outcome was determined at the study census date (31 December 2014), unless malignancy supervened when this became the study exit point.
Incision biopsies were performed under the direction of the first author (PJT) and excision specimens were obtained following interventional laser treatment, as detailed previously. Laser surgery was performed by the first author (PJT), or by colleagues working under direct supervision; this was done within 6–12 weeks following incision biopsy to avoid disease progression. Formalin-fixed tissue specimens were assessed via standardized histopathology examination by oral pathologists at the Royal Victoria Infirmary working to agreed diagnostic criteria with peer review and consensus grading. Using the World Health Organization classification, specimens were graded as mild, moderate, or severe dysplasia, carcinoma in situ, or OSCC. In addition, the presence of hyperkeratosis and lichenoid inflammation (LI) was recorded, as well as diagnoses of PVL and chronic hyperplastic candidosis.
Clinical outcome categories
By reviewing the case records, the patients were assigned to one of the following clinical outcome categories: disease-free (the absence of PMD), further/disease-free (further disease achieving disease-free status following additional intervention), further/persistent (further disease persisting despite intervention), and malignant transformation (if OSCC was confirmed by histopathological examination).
Descriptive statistics were used to summarize details of patient demography, clinical features, and pathological diagnoses, together with documentation of treatment interventions, clinical outcome, and follow-up data. Histopathology diagnoses were treated as categorical variables, and testing of agreement between incision biopsy and post-laser excision biopsy diagnoses was performed using weighted kappa statistics; a coefficient of 1 represented perfect agreement.
Clinical outcomes were stratified as disease-free or persistent disease (further PMD or malignant transformation). Success rates and 95% confidence intervals (CI) were calculated. Multivariate logistic regression was used to analyze factors potentially prognostic of a disease-free outcome (age at first treatment, sex, lesion appearance, anatomical site, and histopathology data). Relationships between factors were explored using χ 2 tests for categorical variables or logistic regression to check for collinearity. Factors were first explored through univariate analysis and then a multivariate model was built using a stepwise procedure until all variables were significant at the 10% level. Candidate variables for the initial model were those significant at the 20% level on univariate analysis. If collinearity was suspected, only the most significant univariate relationships were included in the model. Malignant transformation was analyzed using univariate Cox regression to investigate clinicopathological variables that might influence time to transformation. All statistical analyses were performed using SAS/STAT 9.3 software (SAS Institute Inc., Cary, NC, USA).
Table 1 summarizes the study results. Five hundred and ninety patients with a PMD were treated during the 19-year period; 347 were male (58.8%) and 243 were female (41.2%). Mean age at presentation was 59.7 years. Five hundred and thirteen (86.9%) were current or ex-smokers and 496 (84.1%) regularly consumed alcohol.
|Mean (SD)||59.7 (12.6)|
|Median (range)||60.0 (23–94)|
|Sex, n (%)|
|Lesion, n (%)|
|Site, n (%)|
|Floor of mouth||172 (29.2%)|
|Lateral tongue||130 (22.0%)|
|Buccal mucosa||59 (10.0%)|
|Ventral tongue||56 (9.5%)|
|Labial commissure||29 (4.9%)|
|Fauces/retromolar region||26 (4.4%)|
|Labial mucosa||11 (1.9%)|
|Dorsum of tongue||11 (1.9%)|
|Histopathological diagnosis (most significant), n (%)|
|Hyperkeratosis + LI||28 (4.7%)|
|Chronic hyperplastic candidosis||17 (2.9%)|
|Mild dysplasia||118 (20.0%)|
|Mild dysplasia + LI||28 (4.7%)|
|Mild dysplasia + PVL||40 (6.8%)|
|Moderate dysplasia||105 (17.8%)|
|Moderate dysplasia + LI||24 (4.1%)|
|Moderate dysplasia + PVL||15 (2.5%)|
|Severe dysplasia||99 (16.8%)|
|Severe dysplasia + LI||8 (1.4%)|
|Severe dysplasia + PVL||4 (0.7%)|
|Carcinoma in situ||81 (13.7%)|
|Duration of follow-up a , years|
|Mean (SD)||7.3 (5.1)|
|Median (range)||6 (0–19)|
|Treatment interventions per patient, n|
|Mean (SD)||1.2 (0.6)|
|Median (range)||1 (1–8)|
|Clinical outcome, n (%)|
|Further/persistent disease||53 (9.0%)|
|Malignant transformation||99 (16.8%)|
With regard to clinical features, 468 lesions were leukoplakia; erythroleukoplakia ( n = 99) and erythroplakia ( n = 23) were less common. Floor of mouth and ventrolateral tongue sites were involved in 358 cases.
Histopathological grading based upon the most significant diagnosis obtained from either incision or laser excision biopsy showed that 522 lesions (88.5%) exhibited dysplasia or carcinoma in situ. Incision and post-laser excision biopsies were available for direct comparison in 609 cases. Diagnoses agreed in 307 cases (50.4%); however, in 220 cases (36.1%), the excision specimens were ‘up-graded’ due to more severe dysplasia ( n = 121, 19.9%) or OSCC diagnosis ( n = 99, 16.2%). In 82 cases (13.5%), excision specimens were less severe than incision samples. Whilst overall kappa agreement was deemed moderate at 0.45 (95% CI 0.40–0.51), specific agreement regarding OSCC diagnosis between incision and excision biopsies was only 0.25 (95% CI 0.15–0.35).
The initial treatment modality was laser excision in 526 patients (89.2%) and ablation in 64 (10.8%). In total, 773 treatments were performed on the 590 patients: 513 (86.9%) underwent one laser surgery, whilst 77 (13.1%) required repeat intervention. The mean number of treatments per patient was 1.2 (standard deviation 0.6), with a median of 1 (range 1–8).
The mean duration of follow-up was 7.3 years (standard deviation 5.1 years), with a median of 6 years (range 0–19 years). Four hundred and thirty-eight patients (74.2%, 95% CI 70.7–77.8%) were disease-free at the study census date. Four hundred and four patients achieved this immediately following laser surgery, whilst an additional 34 required further laser intervention. For these latter patients, 77 repeat treatments were provided, with a mean number of 2.26 (range 2–4) to achieve disease-free status over a mean time of 32.88 months (range 4–130.9 months). Out of 87 patients who required repeat treatment, 53 exhibited persistent PMD at census. Malignant transformation was seen in 99 patients. In 71 of these patients (12.0%), a diagnosis of OSCC was made unexpectedly following histopathological examination of the laser excision specimens, whilst the remaining 28 patients (4.8%) underwent malignant transformation during follow-up; the median time to malignancy was 87.3 months (95% CI 59.9–149.2 months). Overall, 80 OSCCs were identified at the same site as their PMD precursor, whilst 19 arose at new, distinct sites. Erythroleukoplakia was most likely to exhibit malignancy ( P = 0.0019; Fisher’s exact test).
The univariate logistic regression analysis of clinicopathological features potentially influencing disease-free status is summarized in Table 2 . PVL did not meet the 20% criterion for inclusion in the multivariate model. Anatomical site was sparse and highly correlated to a number of other factors and was, therefore, also excluded. The final multivariate model contained lesion appearance ( P = 0.022), dysplasia ( P < 0.0001), and the presence of LI ( P = 0.028) ( Table 3 ). Erythroleukoplakia lesions had significantly lower odds of achieving disease-free status compared with leukoplakia (odds ratio (OR) 0.51, 95% CI 0.31–0.83), whilst the odds of disease-free status were significantly increased for lesions exhibiting mild compared with severe dysplasia (OR 2.25, 95% CI 1.27–3.98). The likelihood of disease-free status was reduced for lesions exhibiting LI features (OR 0.49, 95% CI 0.26–0.93). Twenty-eight patients underwent malignant transformation, with the median time to malignant transformation in the cohort of 87.3 months (95% CI 59.9–149.2 months), Fig. 1 . Univariate Cox regression analysis showed no significant influence of clinicopathological variables on malignant transformation time, but this could be due to the small number of events in the analysis.