FLEX October 2023

Risk and interval to malignancy of patients with laryngeal dysplasia 365

lished but no full study was found, it was agreed that the authors would be contacted, but no such instances were identified from the search. One further paper was excluded from the meta-analysis as it was a very small study, which reported a very high rate of transformation. This study had a high risk of sam pling bias. 3 Cases were only included if two biopsies had been performed at least 6 months apart. This may have excluded a significant number of dysplastic cases that only underwent single biopsy and then never progressed over the follow-up period, thereby introducing a sam pling bias of including cases more likely to progress. This small study had a MTR of 74% which is around five times higher than the pooled average and was therefore, deemed to be an outlier. The total list of abstracts obtained from the search was independently assessed by two reviewers, (MW, PN) to establish eligibility based on the above criteria. The two reviewers underwent training with the senior author (HM) before undertaking the research. If one reviewer considered the abstract relevant, then this was sufficient for the full article to be obtained and assessed. Outcome data were then extracted independently by the researchers (MW, PN) and checked by the senior author (HM). Data on the following primary outcome measures were extracted from the studies: the rate of malignant transfor mation to laryngeal cancer and the time interval to malignant transformation. Where available, data was sub classified by both histological grade of the lesion [mild, moderate, severe dysplasia, or carcinoma-in-situ (CIS)] and by management strategy (surgical intervention or radiotherapy). Data on the relative risks of malignant transformation for proven or potential clinical risk factors – gender, site of lesion, continuation of smoking, and alcohol intake – were also extracted if stated. There is currently no formally accepted method for quality assessment of observational studies. We have performed a quality assessment of the included studies using a method ological tool previously used to assess the quality of case series for a set of clinical guidelines issued by the National Institute for Clinical Excellence. 14 Three reviewers (PN, MW, and VP) independently scored each study using a Data extraction Outcome measures Quality and confounding assessment

to malignant transformation in patients with laryngeal dysplasia, as well as the effects that treatment modality, histological grade and risk factors (such as smoking and excessive alcohol intake) may have on this.

Materials and methods

Search strategy

A systematic review of the literature was performed in accordance with MOOSE guidelines. 13 The core search term was larynx coupled with dysplasia , leukoplakia , or erythroplakia . Each of these three combinations was then combined with each of the following terms: progression , follow - up , treatment , cohort , natural history , and recur rence . The databases searched were Medline (1966 to Jan uary 2010), EMBASE (1980 to January 2010), CINAHL (1981 to January 2010) and Cochrane databases (CEN TRAL, Cochrane Library, 1995 to January 2010). In addi tion, reference lists of obtained papers were examined for further potentially relevant papers. All the lead authors of the papers selected for inclusion were contacted for their original data by email or letter. This yielded responses from four of the authors, one of whom provided further data and three who could provide no further informa tion. The last search was performed on 29th January 2010. For inclusion in the meta-analysis, the studies had to fulfil the following criteria: a longitudinal cohort or case series, case-controlled study or as part of a randomised controlled trial of patients with a histologically confirmed diagnosis of laryngeal dysplasia, regardless of treatment modality; and at least one of the outcome measures and one intervention method or clinical risk factor must have been studied. The following exclusion criteria were applied: 1 Cross-sectional studies. 2 No histologically confirmed diagnosis of dysplasia. 3 Data from studies in which patients with dysplasia were not reported separately from those with other conditions, or could not be extracted from the provided data. 4 Any patients with synchronous cancer at or identified within 3 months of the time of diagnosis of dysplasia. 5 Studies with less than 12 months minimum follow-up. 6 Studies not translated into English. Where there were multiple publications using the same patient cohort, the data from the paper with the longest follow-up period was used. Where an abstract was pub Selection criteria

2010 Blackwell Publishing Ltd • Clinical Otolaryngology 35, 364–372