Incorporation of differentiated dysplasia improves prediction of oral leukoplakia at increased risk of malignant progression

Abstract

Oral leukoplakia is the most common oral potentially malignant disorder with a malignant transformation rate into oral squamous cell carcinoma of 1-3% annually. The presence and grade of World Health Organization defined dysplasia is an important histological marker to assess the risk for malignant transformation, but is not sufficiently accurate to personalize treatment and surveillance. Differentiated dysplasia, known from differentiated vulvar intraepithelial neoplasia, is hitherto not used in oral dysplasia grading. We hypothesized that assessing differentiated dysplasia besides World Health Organization defined (classic) dysplasia will improve risk assessment of malignant transformation of oral leukoplakia. We investigated a retrospective cohort consisting of 84 oral leukoplakia patients. Biopsies were assessed for dysplasia presence and grade, and the expression of keratins 13 (CK13) and 17, known to be dysregulated in dysplastic vulvar mucosa. In dysplastic oral lesions, differentiated dysplasia is as common as classic dysplasia. In 25 out of 84 (30%) patients, squamous cell carcinoma of the upper aerodigestive tract developed during follow-up. Considering only classic dysplasia, 11 out of 56 (20%) patients with nondysplastic lesions progressed. With the incorporation of differentiated dysplasia, only 2 out of 30 (7%) patients with nondysplastic lesions progressed. The risk of progression increased from 3.26 (Hazard ratio, p = 0.002) when only classic dysplasia is considered to 7.43 (Hazard ratio, p = 0.001) when classic and differentiated dysplasia are combined. Loss of CK13, combined with presence of dysplasia, is associated with greater risk of malignant progression (p = 0.006). This study demonstrates that differentiated dysplasia should be recognized as a separate type of dysplasia in the oral mucosa and that its distinction from classic dysplasia is of pathological and clinical significance since it is a strong (co)prognostic histopathological marker for oral malignant transformation. In oral lesions without dysplasia and retained CK13 staining the risk for progression is very low.

Fig. 1. Examples of classic and differentiated dysplasia.

In classic dysplasia (a) the keratinocytes have enlarged and hyperchromatic nuclei, decreased nuclear–cytoplasmic ratio, mitoses in suprabasal layers and loss of differentiation towards the surface. Differentiated dysplasia (b) is characterized by a basal layer of small cells with hyperchromatic or open nuclei with small nucleoli with an abrupt transition to suprabasal large cells with abundant, eosinophilic cytoplasm with differences in eosinophilia, intercellular edema with clearly visible desmosomes, and large, open nuclei with prominent nucleoli.

Fig. 2. Overview of the distribution of progression and types of dysplasia for all oral leukoplakia patients.

The number between brackets is the number of patients within each group. As oral leukoplakia biopsies can contain both classic and differentiated dysplasia, the number of patients with any dysplasia is less than the total number of classic and differentiated dysplasia cases combined.

Fig. 3. Progression-free survival of oral leukoplakia patients stratified by classic dysplasia according to the WHO classification, only differentiated dysplasia and the combination of both classic and differentiated dysplasia.

a A Kaplan–Meier survival curve demonstrating the malignant progression-free survival of oral leukoplakia patients who presented with (blue line) or without (red line) classic dysplasia at the time of sampling. Note the number of patients who, despite the absence of classic dysplasia, nonetheless undergo malignant transformation. Patients presenting with classic dysplasia have a hazard ratio of 3.26 (95% CI 1.48–7.19, p = 0.002) for malignant progression. b A Kaplan–Meier survival curve demonstrating the malignant progression-free survival of oral leukoplakia patients that presented with (blue line) or without (red line) differentiated dysplasia at the time of sampling. Patients presenting with differentiated dysplasia have a hazard ratio of 5.48 (95% CI 1.18–25.35, p = 0.014) for malignant progression. c A Kaplan–Meier survival curve demonstrating the malignant progression-free survival of oral leukoplakia patients that presented with (blue line) or without (red line) any type of dysplasia, either classic or differentiated, at the time of sampling. Patients presenting with any type of dysplasia have a hazard ratio of 7.43 (95% CI 1.75–31.56, p = 0.001) for malignant progression.

Fig. 4. CK13 and CK17 staining of healthy oral mucosa and oral leukoplakia.

Example of a healthy oral mucosa biopsy specimen stained for CK13 (a) and CK17 (b) and an oral leukoplakia biopsy specimen stained for CK13 (c) and CK17 (d). While normal epithelium stains strongly for CK13 and not for CK17, disturbance of epithelial balance is often accompanied by loss of CK13 and gain of CK17 expression. The shown example is an oral leukoplakia sample with differentiated dysplasia, but no classic dysplasia. Although the keratin staining is in itself not pathognomonic for differentiated dysplasia, it is immensely helpful in assessing the range and heterogeneity of the lesion. CK13 keratin 13, CK17 keratin 17.

Fig. 5. Progression-free survival of oral leukoplakia patients stratified by presence of any type of dysplasia in combination with expression of CK13.

a Kaplan–Meier survival curve demonstrates the malignant progression-free survival of oral leukoplakia patients that presented without dysplasia and CK13 (blue line), without dysplasia and without CK13 (red line), with dysplasia and CK13 (yellow line) or with dysplasia without CK13 (green line). CK13 keratin 13.

Fig. 6. Proposed scheme for the surveillance of oral leukoplakia patients.

Based on our results patients with no dysplasia (classic and/or differentiated) and retained CK13 staining have a very low risk of malignant transformation and could be referred for less intensive surveillance, while other patients need to remain under regular surveillance. For our cohort this would mean that 22 out of 84 patients (26%) would be eligible for less intensive follow-up. CK13 keratin 13.