Correlation of radiographic features and treatments with the frequency of recurrence in odontogenic keratocysts of the mandible

Odontogenic keratocysts (OKCs) have high recurrence potential. The aim of the study is to evaluate the correlation of radiographic features and treatments with the frequency of recurrence in OKCs in the mandible. This is a cohort retrospective study. The size of the lesions was measured on orthopantomography. Radiographic features (unilocular or multilocular), location, treatment type (enucleation or marsupialization), follow-up time, age, and sex were variable factors. Subjects were studied in 2 groups: a recurrence group and a nonrecurrence group. The radiographic size was considered to be a predictive factor, and the frequency of recurrence was an outcome of the study. An independent t-test and chi-squared test were used to compare variables between recurrence and nonrecurrence groups. Cox regression was applied to the time effect evaluation of the lesion's size and recurrence. A receiver operating characteristic test was used to find a cutoff point for recurrence in the size of radiographic views. Ninety-one subjects were studied. Radiographic features and sizes were different between the 2 groups. The frequency of recurrent lesions was significantly different in various sites of the mandible. Cox regression demonstrated a positive effect of follow-up time on the risk of recurrence. In a 1-year increase in the follow-up time, the recurrence rate increased 1.07 times. An analysis of the data showed a positive relationship between radiographic size and recurrence rate. A 1-cm increase in the lesion's size increased the recurrence rate 1.67 times. A receiver operating characteristic test demonstrated 5.25 cm as a cutoff point in radiographic size; the radiographic size sensitivity was 88.5% (69.8%-97.4%), and specificity was 93.8% (85%-98.3%) for recurrence lesions. Lesions that were 5.25 cm or more on radiographic views had a significant relationship with recurrence rate (P < 0.05). The positive predictive value was 85.2% (65.4%-95.1%), and the negative predictive value was 95.3% (86%-98.8%). The radiographic size may be a predictive factor for the frequency of recurrence lesions in OKCs. Primary multilocular lesions had a greater probability for recurrence than unilocular lesions.