Selection of study specimens
Ethical approval was obtained from the Faculty of Dentistry, University of Otago Human Ethics committee. Specimens used for the study were all formalin-fixed, paraffin-embedded tissues that had been diagnosed with OSCC from the Oral Pathology Centre, Faculty of Dentistry, University of Otago. For this study, oral squamous cell carcinoma (OSCC) arising in the oral cavity and lip vermilion was used. Specimens from the skin of the lips and face were excluded. The most recent cases of OSCC were reviewed until 20 suitable cases had been collected for the study. Inclusion criteria were the presence of the ITF, intra-oral site, patient consent to the use of the biological material after diagnostic procedures were complete for teaching and/or research purposes and there being sufficient material in the block. For each chosen case, a haematoxylin and eosin (H&E) slide was histologically assessed, and a PCNA score was obtained after immunohistological staining.
Histological evaluation
Each H&E stained slide was assessed histologically according to the criteria outlined by Bryne (10) (Table 1). An IFG score was thus generated, where an increase in score is associated with a decrease in prognosis. Each H &E slide was initially evaluated by the primary author (EL) and later confirmed by AR, a consultant oral pathologist; this was repeated two weeks later to check the correlation. Where there was a discrepancy between examiners, the score from the most experienced examiner (AR) was taken. Each histological sample was then assigned an IFG score, as well as a conventional histological WHO grading (24).
Immunohistochemistry
A peroxidase-labelled streptavidin-biotin technique was performed, similar to previously described (20, 25). Briefly, 4mm thick tissue sections were deparaffinsed in xylene, rehydrated in graded absolute alcohols before washing in running water. The pre-treatment step involved treating the slides with proteinase K (DakoCytomation, Lot 00004344) for 10min, followed by immersion in citrate buffer and heating in a microwave at 90oC for 10min. The slides were then allowed to cool to room temperature before being washed under running water. All slides were then treated with 3% H2O2 (BDH Laboratory Supplies Poole, England, Lot K32950980) in methanol for 15min to block the endogenous peroxidase activity. To block nonspecific binding sites, 5% fetal calf serum (25ul frozen aliquot FCS: 5ml PBS) was added. This was followed by application of the monoclonal anti-PCNA antibody PC10 (DakoCytomation, Lot 00005087). Sections were then treated with the biotinylated link and subsequently with the streptavidin horseradish peroxidase (LSAB kit, DakoCytomation, Lot 00003534). In order to visualise the peroxidase activity, slides were saturated with the chromogen, diaminobenzidine hydrochloride (DAB, DakoCytomation, Lot 016067) in a ratio of 1ml DAB buffer: 1 drop DAB substrate. All of the previous steps were followed by a wash in PBS, unless otherwise stated. Finally, all slides were counterstained with Gills haematoxylin, washed in Scotts tap water before coverslips were applied.
All slides contained a section of normal epithelium to serve as a both positive and an internal control. A negative control in which the primary antibody (PC10) was omitted was included in each batch of slides stained.
Immunohistochemical evaluation: expression of PCNA
Each specimen was orientated by firstly identifying the epithelium and then locating the tumour front in the deepest invading area of the tumour. Counting was done systematically, whereby malignant keratinocytes in alternating high power fields at the invading front were counted until a total of 200 cells was reached. A PCNA positive cell was defined as one with clear, distinct brown nuclear staining. The number of positive cells was expressed as a percentage of the total number counted to generate a PCNA score (25). Intensity of PCNA expression was recorded with reference to a known positive control and was reviewed and verified by a specialist oral pathologist. Grade 1 denoted intense PCNA staining, grade 2 moderate PCNA staining and grade 3 indicated weak staining. The calibration process was carried out on a total of six slides where agreement was reached for each of the three grading intensities. Following calibration, each slide was assessed independently by two examiners. The final number of PCNA positive cells recorded for each slide was obtained from taking the average positive cell counts from the two examiners, EL and AR. An appropriate level of inter-examiner agreement was achieved across all slides (Cohen’s Kappa coefficient>0.6). To aid in the comparison between slides, images of the cells in the invading front were captured digitally for later reference.
Statistical analysis
Regression analysis was performed using Statistical Package for Social Studies software (SPSS version 24; IBM Corporation, Armonk, NY, USA) to compare variables. The level of significance was set at p<0.05).