Background: Oral leukoplakia may develop into squamous-cell carcinoma, whic
h has a poor prognosis. Risk factors for oral carcinoma have been identifie
d, but there are no reliable predictors of the outcome in individual patien
ts with oral leukoplakia.
Methods: We identified 150 patients with oral leukoplakia that was classifi
ed as epithelial dysplasia and measured the nuclear DNA content (ploidy) of
the lesions to determine whether DNA ploidy could be used to predict the c
linical outcome. Biopsy specimens obtained at annual follow-up visits were
graded histologically and classified with respect to DNA content in a blind
ed fashion. Disease-free survival was assessed in relation to DNA ploidy an
d the histologic grade. The mean duration of follow-up was 103 months (rang
e, 4 to 165).
Results: Among 150 patients with verified epithelial dysplasia, a carcinoma
developed in 36 (24 percent). Of the 150 patients, 105 (70 percent) had di
ploid (normal) lesions, 20 (13 percent) had tetraploid (intermediate) lesio
ns, and 25 (17 percent) had aneuploid (abnormal) lesions at the time of the
initial diagnosis. A carcinoma developed in 3 of the 105 patients with dip
loid lesions (3 percent), as compared with 21 of the 25 patients with aneup
loid lesions (84 percent), yielding a negative predictive value of 97 perce
nt with respect to the diploid lesions and a positive predictive value of 8
4 percent with respect to the aneuploid lesions. Carcinoma developed in 12
of 20 patients with tetraploid lesions (60 percent). The mean time from the
initial assessment of the DNA content to the development of a carcinoma wa
s 35 months (range, 4 to 57) in the group with aneuploid lesions and 49 mon
ths (range, 8 to 78) in the group with tetraploid lesions (P=0.02). The cum
ulative disease-free survival rate was 97 percent among the group with dipl
oid lesions, 40 percent among the group with tetraploid lesions, and 16 per
cent among the group with aneuploid lesions (P<0.001).
Conclusions: The DNA content in cells of oral leukoplakia can be used to pr
edict the risk of oral carcinoma. (N Engl J Med 2001;344:1270-8.) Copyright
(C) 2001 Massachusetts Medical Society.