This study evaluates the accuracy of the results of liquid-based oral brush cytology and compares it to the histology and/or the clinical follow-ups of the respective patients. A total of 1352 exfoliated specimens were collected with an Orcellex brush from an identical number of oral lesions, then cytological diagnoses were made using liquid-based cytology. The final diagnoses in the study were 105 histologically proven squamous cell carcinomas (SCCs), 744 potentially malignant lesions and 503 cases of traumatic, inflammatory or benign hyperplastic oral lesions. The sensitivity and specificity of the liquid-based brush biopsy were 95.6% (95% CI 94.5–96.7%) and 84.9% (95% CI 83.0–86.8%), respectively. This led to the conclusion that brush biopsy is potentially a highly sensitive and reliable method to make cytological diagnoses of oral neoplasia. The main advantage of a brush biopsy over a scalpel biopsy is that it is less invasive and is more tolerated by the patients. Therefore, more lesions can be screened and more cancers can be detected at an early stage.

Background This study compares two different cell collectors, the Orcellex Brush (rigid brush) and the Cytobrush GT (nylon brush), using liquid-based cytology. A comparison of their obtainment procedures was also considered. The aim was to determine the diagnostic accuracy for detection of malignancy in oral brush biopsies. PICO-Statement: In this consecutive and retrospective study we had as population of interests, patients with oral lesions, the intervention was the brush biopsy with two different cell collectors and the control was healthy oral mucosa. The outcome of the study was to compare both cell collectors. Methods From 2009 to 2018, 2018 patients with oral lesions were studied using the nylon brush (666 cases) and rigid brush (1352 cases). In the first cohort five smears per patient were taken with the nylon brush, while each patient received one smear with the rigid brush in the second cohort. These were further processed in a liquid-based procedure. Cytological evaluations were categorised into ‘negative’, which were considered as negative, whereas ‘doubtful’, ‘suspicious’ and ‘positive’ cytological results were overall considered as positive for malignancy in comparison to the final histological diagnoses. Additionally, the clinical expenditure for each collector was estimated. Results 2018 clinically and histologically proven diagnoses were established, including 181 cases of squamous cell carcinomas, 524 lichen, 454 leukoplakias, 34 erythroplakias and 825 other benign lesions. The sensitivity and specificity of the nylon brush was 93.8% (95% CI 91.6–95.5%) and 94.2% (95% CI 91.8–95.5%) respectively, whereas it was 95.6% (95% CI 94.4–96.6%) and 84.9% (95% CI 83.8–87.5%) for the rigid brush. The temporal advantage using the plastic brushes was 4×  higher in comparison to the nylon brush. The risk suffering from a malignant oral lesion when the result of the brushes was positive, suspicious, or doubtful was significantly high for both tests (nylon brush OR: 246.3; rigid brush OR: 121.5). Conclusions Both systems have a similar sensitivity, although only the rigid brush achieved a satisfactory specificity. Additional methods, such as DNA image cytometry, should also be considered to improve the specificity. Furthermore, the rigid brush proved to be more effective at taking a sufficient number of cells, whilst also being quicker and presenting less stress for the patient.