3 RESULTS
3.1 Smoking status and age between patients with unilateral and bilateral lesions
The number of cigarettes smoked per day didn’t appear significantly different between patients with unilateral and bilateral lesions (U = 56.9, p = .109 and though it seemed to be higher in the malignancies group, that was not statistically significant (U = 56, p = .064). The years of smoking were found to be similar between patients with unilateral and bilateral lesions (t = 0556, p = .956) and although patients who developed malignancies reported having smoked for more time, there was no significant difference when comparing their years of smoking with those of the patients with non-malignant lesions (t = -1.24, p = .227). It was age however that was found to be of true importance, since in our group malignancies were identified in older rather than younger subjects (t = -2.23, p = .034), but that was not observed as regards the bilateral presence of lesions (t = -0.98, p = .332). Table 2 presents the comparison of smoking status and age between patients with and without malignancies.
3.2 Smoking habits and age in different types of vascular classification
Moreover, a Kruskal-Wallis test showed no significant statistical difference in cigarettes smoked per day between lesions that had different types of intraepithelial papillary capillary loop classification, χ2 (5) = 9.93, p = .077 (Figure 1 ). No such difference was found in years smoking either, as determined by a one-way ANOVA that was also performed, F (5, 9.95) = 2.52, p = .100. Age however presented a significant effect regarding the vascular classification, F (5, 12.9) = 3.71, p = .027. Post-hoc comparisons using the Games-Howell test showed significant difference particularly between the age of patients in the group of type I lesions (M = 52.7, SD = 1.15) and the age in the group of type Vb lesions (MD = 63.7, SD = 8.50). Our results suggest that malignancies had significantly higher chances of developing in older individuals, when compared to the mildest type I lesions of vascular classification.
3.3 Smoking habits and age in different types of morphological classification
A Kruskal-Wallis test showed no significant statistical difference in cigarettes smoked per day between different types of morphological classification, χ2 (5) = 4.84, p = .089. In the same way a one-way ANOVA did not indicate statistical difference in the years smoking either F (2, 14) = 2.02, p = .170. Age, according to a one-way ANOVA, was significantly different among different types of morphological classification types, F (2, 35) = 7.72, p = .002. The significant difference in age was detected between the “flat and smooth” type of lesions (M = 51.4, SD = 6.33),p=.002 and the “rough” type of lesions (M = 61.5, SD = 6.83), p = .001 as indicated by post-hoc comparisons using the Tukey correction. The results are depicted in Table 3 .
3.4 Classification of leukoplakia and smoking habits
As mentioned above we have examined a total of 32 vocal cords with leukoplakia. By 22 of them a low-grade dysplasia has been found. The rest of them have been diagnosed with a high-grade dysplasia. A Kruskal-Wallis test showed no significant statistical difference in cigarettes smoked per day between different types of morphological classification, χ2 (5) = 4.64, p = .088. In the same way a one-way ANOVA did not indicate statistical difference in the years smoking either F (2, 08) = 2.02, p = .1640. The application of Kendal-tau criterion has shown the grade of dysplasia is correlated to age. By patients with low-grade dysplasia (n=18) this was estimated by τ=0.73. By those with high-grade dysplasia (n=14) the correlation was stronger (τ=0.82).
3.5 Contact Endoscopy in non-smoking and smoking subjects
We also tested if there was statistical significance between the vascularization type of the vocal cords of the healthy subjects (control group) and the vascularization type of patients suffering from leukoplakia and malignancy. Vocal cords with type I or type II vascularization pattern were more likely to belong to patients who have non-malignant contralateral vocal cords, whereas healthy cords of type III or IV were more likely to be associated with contralateral malignancies, according to a Fisher’s exact test that showed statistical significance, p = .002. In non-smoking subjects we have found that 18 of them had a type I vascularization. By the rest a type II vascularization on both sides has been documented.