FLEX October 2023

368 M.D. Weller et al.

The spread of MTRs is wide, showing a degree of het erogeneity.

Overall malignant transformation rate

The untransformed study MTRs are shown in Table 3 and with confidence intervals in Fig. 3. The overall pooled MTR estimated from a random effect logistic regression model taking into account the different study sizes and the observed variation is 14% [95% confi dence interval (CI) – 8%, 22%]. This is also shown in Fig. 3. This overall pooled rate is similar to that dem onstrated by Sllamniku, which was the largest study

Fig. 2. Funnel plot: malignant transformation rate (arcsine trans formed) against study size.

Table 3. Malignant transformation rates (MTR) for all studies

Total dysplasia Mild Moderate Severe CIS

Malignant transformation Mild Moderate Severe CIS MTR

Study

Ricci 52 Plch 51

111

46 42 63 25 204 23 32 26 56 28 23 65

23

10 2 4 6 0 44 NR 13 NR 27 NR 4

5 0

3 0 4 1 6

9%

99

7

3

3 6.1% 0 13.8%

Sllamniku 53

317

90

15

25

Minni 2 Gallo 49

58

NR NR 1

0 4

NR NR 1.7% 3 NR 11.2% 113 NR 23.9% 2 NR 15.4%

116 113

32 25

Jeannon 50

1 13

Uno 54

26 59 41

10

9

7 9

2 3

0 5

Blackwell 48 Stenersen 9

26 15 NR NR

9 13

4

1 22%

41

19

NR NR

19

46.3%

Total

940

137

16% (pooled MTR 13.6%)

NR, not recorded; CIS, carcinoma-in-situ.

Pooled MTR = 14% (95% CI 8%–

22%)

Ricci et al. 2003 ( N = 111)

Sllamniku et al. 1989 ( N = 317)

Plch et al. 1998 ( N = 99)

Jeanon et al. 2004 ( N = 113)

Minni et al. 2008 ( N = 59)

Gallo et al. 2001 ( N = 116)

Uno et al. 1997 ( N = 26)

Stenersen et al. 1987 ( N = 41)

Blackwell et al. 1995 ( N = 59)

Pooled overall mean

0

0.1

0.2

0.3 0.4 0.5 0.6 0.7

0.8

0.9

1

Transformation rate

Fig. 3. Exact binomial confidence intervals of malignant transformation rate of laryngeal dysplastic lesions for studies included in the meta-analysis.

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