This systematic review was undertaken to determine the overall and detailed recurrence rate of keratocystic odontogenic tumour in relation to specific treatment methods. Online electronic databases were searched to identify articles published in English language from 1956 to 2010. Articles were independently appraised by two reviewers in three separate rounds. Any disagreement was settled by discussion with a third judge. Of 1568 potentially relevant articles, 168 articles related to the treatment of keratocystic odontogenic tumour/parakeratinised type of odontogenic keratocyst and its recurrence rate entered the second round for evaluation. Fourteen papers entered the third round for critical appraisal. Two retrospective reviews entered the final analysis. One hundred eight lesions were found in the material analysed. Six treatment modalities were identified. The recurrence rates were 0% for resection, 0% for enucleation with peripheral ostectomy and Carnoy’s solution, 18.18% for enucleation with peripheral ostectomy, 26.09% for enucleation alone, 40% for marsupialisation, and 50% for enucleation with Carnoy’s solution. The overall recurrence rate was 23.15%. The present review discusses the methodological weaknesses of many of the studies analysed. No high quality evidence was obtained to evaluate recurrence rates related to treatment modalities of keratocystic odontogenic tumour.
A keratocystic odontogenic tumour (KCOT) is a benign neoplasm of odontogenic origin with potential for aggressive and infiltrative behaviour. The lesion, as a distinct entity, was first described in the literature by Philipsen in 1956; he used the designation ‘odontogenic keratocyst’ (OKC) to describe jaw cysts that exhibit keratinisation of their epithelial lining. Further studies have discussed the difference between the more common and clinically aggressive parakeratinised type of OKC (P-OKC) in contrast to the rarer and clinically benign orthokeratinised variant. In addition to the aggressive clinical behaviour of P-OKC, the increased mitotic activity in the cystic epithelium, the budding potential of the basal layer and the presence of daughter cysts attached to the side walls of the pericystic cavity all led to isolating this type of OKC from the spectrum of odontogenic cysts. More advanced morphogenetic studies discovering chromosomal abnormalities and genetic alterations, such as the mutation of the PATCHED gene and the over-expression and amplification of genes located in 12q, appeared to confirm this concept. In 2005, the World Health Organization Working Group acknowledged P-OKC to be a cystic neoplasm and employed the term ‘keratocystic odontogenic tumour’. However, cystic jaw lesions that are lined by orthokeratinising epithelium are not included in the spectrum of KCOT.
Epidemiological studies on the prevalence of KCOTs are not available. Before 2005 when the lesion was assigned to a group of odontogenic cysts, OKCs were the third most common jaw cyst and constituted 10.2% of all odontogenic cysts. The average annual incidence rate rose to 11.58. KCOT is one of the most common odontogenic neoplasms of the jaw. In a recent analysis, Gaitán-Cepeda et al. point out that reclassification of P-OKC into a KCOT significantly changed the frequency distribution of odontogenic tumours making KCOTs the most frequent odontogenic tumour. The issue of appropriate treatment of KCOTs remains a subject of controversy. The spectrum of treatment modalities contains conservative methods such as simple enucleation, decompression or marsupialisation and more aggressive techniques using cryosurgery or chemical destruction and radical surgical tactics with bone resection. Although it is generally accepted that more radical treatment results in lower recurrence rates, the literature does not give adequate evidence for determining which treatment modality is the most effective in either lowering morbidity or preventing recurrence.
Most of the published material comes from the period before 2005, and authors generally do not consider P-OKC as a separate entity in their analyses; instead, they pool the data together with the orthokeratinised variant. Thus, in the majority of published material, it is impossible to extract the detailed data on a specific type of OKC. Many studies do not have an adequate follow-up period making the reported recurrence rate untrustworthy. Some of the studies do not characterise the treatment used in detail and use vague terms such as conservative or radical surgery. Current medical databases contain numerous papers on the treatment and recurrence rate of OKCs and KCOTs, but their scientific credibility is questionable for the aforementioned reasons. Therefore, it is of vital importance to appraise the published material critically and organize it according to its reliability.
Given the questionable conclusions of the various studies on the surgical management of KCOT with respect to the recurrence rates, the rationale behind this study was to identify and critically appraise the evidence currently available in the literature. This systematic review attempted to determine the recurrence rate of keratocystic odontogenic tumour following treatment.
Materials and methods
The search covered Embase (1 January 1956 to 30 September 2010), Cochrane Library (1956 to September 2010) and MEDLINE using PubMed (1 January 1956 to 30 September 2010), EIFL (January 1956 to September 2010), Ovid (1956 to September 2010) and ISI (1956 to September 2010). Search keywords included: (keratocystic AND odontogenic AND tumour) OR (odontogenic AND keratocyst) OR (primordial AND cyst) AND (recur OR relapse OR recurrence). The search was in all fields and in English. The reference list from the retrieved articles was reviewed for additional papers. Databases searches and hand-searches were last conducted on 30 September 2010.
The search, evaluation of the relevant articles, and their critical appraisal were performed by two independent judges (T.K. and I.M.) blind to each other. Any discrepancies between the results in any round were settled by discussion. If an agreement could not be reached, advice was sought from a third party (J.S.). The flow chart of the systematic article selection is presented in Fig. 1 .
The primary outcome measure was the overall recurrence rate of KCOT irrespective of treatment modality. The secondary outcome measure was the recurrence rate of KCOT for specific treatment modality.
In the first round – SEARCH , abstracts were reviewed and articles that were relevant to the recurrence rate and related treatment were selected. Papers with no abstracts were also considered. The full text of all the articles in the first round was obtained for detailed inspection. To avoid risk of selection bias, articles regarding patients with the peripheral (extraosseous) variant of KCOT/OKC were excluded. Series including patients with nevoid basal cell carcinoma syndrome (Gorlin-Goltz syndrome) were excluded as multiple tumours in this syndrome are often not synchronous and it is difficult to distinguish between recurrent tumours and new ones arising in contiguous sites.
In the second round – EVALUATION , using the standardised evaluation charts, the same two judges independently assessed the full text for each of the four criteria established by Lau and Samman ( Fig. 1 ). Only articles that met all four criteria were marked as accepted and included in the third round.
In the third round – CRITICAL APPRAISAL , to evaluate the methodological quality, each article was critically appraised using 7 standards established by Lau and Samman. The authors developed their own criteria for assessment of risk bias; each of the articles was ranked into one of the three categories: low risk of bias if all 7 standards were met; moderate risk of bias if one of the 7 standards was not met; high risk of bias if two or more of the 7 standards were not met.
Only articles with a low risk of bias were included in the final analysis. Using the standardised data extraction sheets, data were extracted from the selected articles for final analysis. Sample size, demographic data, type of treatment modality, follow-up period and recurrence rate were all recorded. 95% confidence intervals (CI) for overall recurrence rate as well as for specific treatment modalities were estimated.
One thousand five hundred sixty-eight hits were recorded for the electronic search. MEDLINE yielded 1415 hits (EIFL 444; Ovid 118; PubMed 319; ISI 534). The Embase database yielded 152 hits and Cochrane yielded one.
Of the 1568 hits, 137 potentially relevant articles related to the treatment of KCOTs or P-OKCs and the recurrence rate were selected. Manual search of the reference lists for these 137 selected articles yielded 31 more relevant citations. Thus, a total of 168 papers were subjected to the evaluation.
Of the 168 articles selected for the second round, 154 articles did not fulfil one or more criteria and were excluded. The remaining 14 papers were determined relevant and entered into the third round for critical appraisal. The publication dates ranged from 1988 to 2010.
The 14 articles were thoroughly analysed and critically appraised according to the above-mentioned 7 standards. Eight articles were excluded due to possible high risk of bias and another four due to possible moderate risk of bias ( Table 1 ). Two publications entered the final review ( Table 2 ). The well-defined and clear evaluation process allowed for the selection of the best available credible data for analysis. A literature search flow diagram is presented in Fig. 1 .