Study design and sample
We designed and implemented a cross-sectional study using the oral exfoliative cytology results of patients who had been referred to the Kagawa Prefectural Central Hospital (Takamatsu, Japan) for diagnosis, treatment, and examination of oral lesions. During the period from April 2010 to March 2019, a total of 1234 specimens were obtained from the cytology specimens collected by the Department of Oral and Maxillofacial Surgery and diagnosed by the Department of Pathology. Of these specimens, nine were insufficient and excluded. Of the remaining 1225 specimens, 251 specimens that underwent histological diagnosis ranging from benign or malignant oral lesions for biopsy and/or surgical resection were included in this study. All specimens were collected, processed, and diagnosed in a single general hospital.
Cytology specimens were processed by conventional cytology from April 2010 to March 2015 and by the LBC method from April 2015 to March 2019. Cells were harvested by scraping with a cotton cytobrush device in all cases. In the conventional method, the collected cells were smeared onto a glass slide to prepare a sample, immersed in 95% ethanol, fixed, and stained with Papanicolaou stain. The LBC method involves dipping a cotton brush containing the sample directly into the transport medium, which is an alcohol-based preservative (BD CytoRich blue preservative, BD Japan, Tokyo, Japan). The liquid-based cellular material in the vial was processed according to the manufacturer’s protocol (BD Japan). The processing steps included vortexing of the sample, density reagent centrifugation, decantation and resuspension of cell pellets followed by gravity sedimentation on poly-l-lysine coated slides and subsequent staining with the Papanicolaou stain.
Procedure of cytological diagnosis
The specimens were reviewed by raters who had passed the board examination for cytology of the Japanese Society of Clinical Cytology. Cytology diagnostic experts confirmed that the sample was appropriate for cytology diagnosis. According to the criteria for specimen adequacy, we identified non-diagnosable specimens as inappropriate due to the presence of hypocellular or air-drying artifacts. The specimens were evaluated independently by at least two raters, and a representative cytology result of each case was determined by a majority vote. Cytological diagnoses were made based on the Bethesda system according to the Japanese society of clinical cytology (JSCC) diagnostic guideline and were classified into negative for intraepithelial lesion or malignancy (NILM), oral low-grade squamous intraepithelial lesion (OLSIL), oral high-grade squamous intraepithelial lesion (OHSIL), SCC, and indefinite for neoplasia [8]
Procedure of histological diagnosis
A histological diagnosis was provided by pathologists, and then, the number of biopsy samples was determined at the investigator’s discretion. These histological slides were subjected to hematoxylin and eosin staining, and their histological findings were divided into two categories as negative group and positive group. Negative was defined as non-malignant lesions, including inflammatory ones and mild, or moderate dysplasia. Positive was defined as severe dysplasia, carcinoma in situ (CIS), SCC, and other malignancies. Histological diagnosis was based on the WHO criteria [9]. CIS was in accordance with the general rules for clinical and pathological studies on oral cancer [10].
The design and methodology of this study have been approved by the Ethics Committee of Kagawa Prefectural Central Hospital (Approval No. 946).
Statistical analysis
Data were entered into a database using Microsoft Excel (Microsoft Inc., Redmond, WA, USA). The database was transferred to JMP version 14.2 for Macintosh computers (SAS Institute Inc., Cary, NC, USA) for statistical analysis. To compare between cytological and histological diagnoses, the histological diagnoses were classified into negative and positive, and the cytological diagnoses were also classified into negative (NILM and OLSIL) and positive (OHSIL, SCC, and other malignancies). The diagnostic performance metrics was examined by comparing the cytological diagnosis against the histological diagnosis, for which the sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were calculated, followed.
TP and TN indicate true positive and true negative classifications, respectively; FP and FN indicate false-positive and false-negative classifications, respectively.