FLEX October 2023

Risk and interval to malignancy of patients with laryngeal dysplasia 369

with 317 patients and scored highly on the quality assessment.

Discussion Widely varying transformation rates of between 2% 2 and 74% 3 have been reported by different studies. In an attempt to ascertain the true MTR of laryngeal dysplastic lesions, we performed a systematic review and meta-analy sis. This demonstrates an overall MTR for laryngeal dyspla sia of 14% (95% CI – 8%–22%) in 940 patients. It appears that the MTR is highly correlated with severity of grade of dysplasia, with the risk increasing threefold between mild ⁄ moderate dysplasia (10.6%) and severe dysplasia ⁄ CIS (30.4%). This is a statistically significant difference. Although the meta-analysis does not include large numbers of patients, it is more likely to accurately reflect the true MTR because it utilised strict criteria to include only the studies of higher quality. It also excluded small studies. The transformation rate found by the meta-analy sis was similar to that of one of the highest quality stud ies, which is also the largest single study of 317 patients. Finally, this rate is comparable to that found for oral dys plasia in a separate meta-analysis, which showed a rate of malignant transformation of 12%. 55 Indeed, a sensitivity analysis was performed excluding the one small outlying study with a high transformation rate. 9 This gave an overall transformation rate for the 899 patients included of 12% (95% CI – 8%, 18%). The studies reported a wide range of times to malignant transformation, with the longest reported time being over 14 years. It also appears that the time to transformation is not dependent on the grade of dysplasia, although the quality of this data was almost universally poor. Due to this, there is no evidence available to support the practice of early discharge in patients with mild or moderate dysplasia, which some clinicians follow. The data shows lesions treated by surgical excision have a lower MTR than those that have not. This was not statis tically significant, however, and therefore, it is not possi ble to make strong recommendations regarding the role of surgery in this condition. However, this provides a good basis for further research, enabling sample size calculations to be performed. Time to transformation Effect of treatment modality

Sensitivity analysis

One study with only 41 patients appeared as an outlier with a high transformation rate. 9 An analysis was, there fore, performed with this study excluded. This analysis of 899 patients gave a MTR of 12% (95% CI – 8%, 18%).

Mean time to transformation

Seven studies reported mean time to transformation, however, only six gave results subdivided by grade, (Table 2). The overall mean time to transformation was 5.8 years (range 1.8–14.4 years). Sub analysis by grade showed mean time to transformation of 4.9 years for mild dysplasia (range 2.4–14.4 years), 3.4 years for mod erate dysplasia (range 1.8–4.7 years) and 4.1 years for severe dysplasia (range 2.1–11.0 years). When analysing the effect of histological grade on the rate of malignant transformation, histological grades were combined into two groups – mild and moderate or severe and CIS. The MTR was 10.6% (95% CI – 5.1–20.7) for the mild-moderate group and 30.4% (95% CI – 16.1– 49.9) for the severe-CIS group. This difference is statisti cally significant ( P < 0.0002). The variety of interventions reported and the lack of clar ity regarding interventions and follow-up regimes made it difficult to extract data on intervention type. The studies were subdivided into two groups, namely those where for mal surgical excision had been performed in an attempt to excise the lesion, (using either cold steel or LASER) and a ‘non-excision’ group where only a biopsy had been performed or the modality of treatment was not made clear. When these two groups were analysed, the MTR was 15% (12%, 18%) for the surgical group (752 patients) and 21% (CI – 16%, 27%) for the ‘non-excision’ group (188 patients). This did not reach statistical signifi cance even after adjustment for grade ( P = 0.12). Malignant transformation rate and histological grade Malignant transformation and intervention type

Risk factors for malignant transformation

Limitations

There was insufficient data in the studies to comment on differences according to gender or risk factors such as smoking or alcohol consumption.

Despite high number of studies on initial search, most were excluded because they were not specific to the

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