The ocular surface is composed of the conjunctiva, the epithelial surface of the cornea and the junctional limbal epithelium. The term ocular surface squamous neoplasia (OSSN) encompasses the range of conjunctival and corneal intraepithelial neoplasia through to invasive squamous cell carcinoma of the ocular surface. This uncommon disorder is related to prolonged exposure to solar ultraviolet light and has been proposed as an AIDS associated tumor. HPV infection has a possible etiologic role. A study of OSSN is of particular relevance to Queensland, which has a subtropical to tropical climate and a majority of the population with British, Irish, or other European ancestry. The standard cytology textbooks contain very little information regarding the cytology of OSSN and until recently little information was available in the scientific literature.
The impression cytology (IC) from 267 patient episodes with subsequent histopathology was used to give a detailed description of the cytomorphology of intraepithelial and invasive OSSN. Within the intraepithelial group, keratinized dysplastic cells, which were often accompanied by hyperkeratosis, syncytial-like groupings and nonkeratinized dysplastic squamous cells were described. Eightyseven percent of clinically detectable intraepithelial OSSN cases were high-grade lesions on histopathology. Within the invasive group, cases with significant keratinization, and a further group of cases with little keratinization and sometimes also prominent macro-nucleoli were described. Keratinized cases were the most numerous in both the intraepithelial and invasive groups. A description was also given for the first time of a group of cases that in the absence of a clinically detectable lesion, had cytology and also subsequent histopathology indicating the presence of intraepithelial OSSN.
To assess the accuracy of IC in the detection of OSSN, multiple impressions were collected from 300 consecutive patients with a range of ocular surface conditions. Impressions were examined in a masked fashion. For 194 eyes, histopathology was evaluated within the following six months, this examination being masked to the cytology results. It was determined that IC could detect OSSN with a sensitivity of 78% when IC with indeterminate results was excluded. The specificity imder these conditions was 100%. The study found IC to have a higher sensitivity of 80% for the detection of histologically intraepithelial OSSN compared to 70% for histologically invasive or microinvasive disease. While the positive predictive value of the test as performed in this study was 100%, the negative predictive value was only 42% due in part to the low proportion of histologically confirmed negative cases.
A teaching program incorporating a manual, slide sets and an evaluation test was developed. The aim was to teach experienced cervical smear screeners to evaluate ocular surface specimens collected by IC with a minimum of individual tuition. The use of the manual was well accepted and three of the six candidates were able to adequately master the new skill within the eight hours allowed for the course. It was considered that the differences between the cytology of OSSN and the equivalent lesions of the uterine cervix are sufficient to prevent some experienced cervical smear screeners from rapidly acquiring these skills. The program could be readily modified to a workshop or computer assisted tutorial.
Impression cytology was also used to investigate apparent intraepithelial OSSN (subclinical OSSN) that did not display clinical symptoms, but could be identified by IC and histopathology. A randomly selected group from a normal south-east Queensland population; a group judged to be at high risk of OSSN as a result of the increased UVR exposure resulting from their occupational or recreational activities; and the contralateral eye of a group that already had developed clinical OSSN were examined. The prevalence of subclinical OSSN within these groups was calculated to be 3.5%, 20% and 14.5% respectively. Multifocal subclinical OSSN and bilateral OSSN were also identified. While increased solar exposure appeared to be associated with an increased prevalence of subclinical OSSN, further studies are needed to establish a definite link between subclinical OSSN and clinical disease.
The scientific literature related to the cytology of Ocular Surface Squamous Neoplasia has been considerably expanded by the work presented in this thesis.