Ocular surface squamous neoplasia (OSSN) is the most common nonpigmented malignancy of the conjunctiva and the third most common of all ocular tumors in elderly patients.1,2 Von Graefe first reported these tumors in 1860, and the terminology used to identify them has evolved since that time. OSSN currently refers to a spectrum of epithelial abnormalities that ranges from dysplasia of the corneal and/or conjunctival epithelium to frankly invasive squamous cell carcinoma.1 Although most reports of OSSN in the United States are among elderly males of European origin with significant ultraviolet B exposure, the tumor occurs in both sexes and across all ages and regions of the world.3,4 The incidence has been estimated at 0.08 to 0.18 per 100,000 per year worldwide, and as high as 3.4 per 100,000 per year in countries such as Zimbabwe.1,5 Higher rates of OSSN are seen in patients with HIV infection and xeroderma pigmentosum.6–8 An association between OSSN and human papillomavirus has also been postulated.1,9–11
In the past, treatment of OSSN was primarily surgical with wide excision of the tumor and cryotherapy to tissues at the surgical margins with the intent of destroying adjacent but clinically unapparent malignant cells. In the last decade however, there has been a trend toward the use of topical chemotherapeutic agents to complement or replace surgical management of OSSN. Topical interferon alpha-2b (IFN-α2b) is of particular interest because of its apparent efficacy and low toxicity.12 The proposed advantages of IFN-α2b therapy include treatment of the entire ocular surface rather than solely tissue within the surgical field, less damage to fragile limbal stem cells, and the convenience of outpatient management.12 However, there has not been a randomized trial comparing IFN-α2b therapy with surgical management to confirm that these putative benefits are associated with comparable rates of success. In addition, efforts to indirectly compare the strategies have been limited by widely varying estimates for recurrence rates, ranging from 5% to 66% for surgical treatment13–15 and from 0% to 28% with IFN-α2b treatment,16–22 depending on the surgical technique, IFN-α2b regimen, and length of follow-up.
Most OSSN-related morbidity accrues from multiple rounds of surgical intervention, causing repeated damage to the ocular surface. This can result in corneal and conjunctival scarring, loss of limbal stem cells, and potentially exenteration in severe cases; therefore, resolution of the primary disease and prevention of recurrence are the principal goals of therapy. In the absence of randomized trials or other prospective comparative studies to direct clinical practice, we used a decision analysis informed by a systematic review of the literature to compare 4 strategies for treating OSSN, ranging from surgical excision alone to empiric topical IFN-α2b.
MATERIALS AND METHODS
We developed a decision-analytic model to evaluate the expected value (EV) of 4 treatment strategies aimed at minimizing the persistence/recurrence of disease in patients with OSSN: (1) traditional surgical excision with cryotherapy followed by repeat excision for cases with positive margins (excision-only), (2) surgical excision with cryotherapy followed by topical IFN-α2b for cases with positive margins (excision-directed IFN), (3) incisional biopsy followed by topical IFN-α2b for cases with positive biopsies (incisional biopsy–directed IFN), and (4) empiric treatment with topical IFN-α2b (empiric IFN) (Fig. 1).
The base-case population was a multiracial, generally immunocompetent group older than 50 years. Because the length of treatment was highly variable in our review of the literature, the model was evaluated over a single treatment course. Variables used in the model are listed in Table 1 and included test characteristics of excisional and incisional biopsies, rate of positive surgical margins after excision, recurrence rate for positive and negative margins, rate of successful treatment with IFN-α2b, and recurrence after clinical resolution after IFN-α2b treatment. The endpoints for each strategy were dichotomized as either complete clinical resolution, or disease persistence or recurrence.
For the excision-only strategy, a patient with a positive biopsy result and positive margins underwent a second intervention (reexcision) before facing the probability of recurrence. If the excisional biopsy result was positive but with negative margins, the patient faced a probability of recurrence associated with negative margins. We assumed that all patients entering the decision tree had OSSN; therefore, if the excision biopsy result was negative, the diagnosis was considered missed and the patient continued to have OSSN. The excision-directed IFN-α2b strategy is identical to the excision-only strategy except that a patient with a positive biopsy result and positive margins received IFN-α2b as the second intervention. Before this additional intervention, the appropriate proportion of individuals who would have not recurred despite positive margin status was removed. If IFN-α2b was successful, the patient went on to face a chance of recurrence after IFN-α2b treatment. If IFN-α2b was not successful, the patient assumed the risk of recurrence associated with positive margins. For the incisional biopsy–directed IFN-α2b strategy, a patient with a positive incisional biopsy received IFN-α2b. If the IFN-α2b was unsuccessful or if the incisional biopsy was negative, the patient continued to have OSSN. For the empiric IFN-α2b strategy, the patient was treated with IFN-α2b then followed the same path of probabilities as those with a positive biopsy result in the incisional biopsy–directed IFN strategy.
Model Endpoints and Analysis
To model the expected probability of persistent disease or recurrence, we assigned an outcome value of 0 for complete resolution with no recurrence or a value of 1 for persistence or recurrence. In the base-case analysis, intermediate probabilities were used for all chance nodes. We identified the strategies that minimized the expected probability of missed/persistent disease or recurrence. Sensitivity analyses were conducted using the range of reported values in the literature. An additional analysis was performed to determine the preferred strategy when avoidance of a missed diagnosis was considered the optimal outcome.
We performed a systematic review of the literature to obtain values and ranges for each variable used in the model. We searched MEDLINE (via Ovid), Web of Science, and the Cochrane Library, each from inception to January 2015. Known relevant articles were used to identify appropriate MeSH terms, which were combined with key words to maximize the sensitivity of the search. Combining the 3 databases produced 2049 results, of which 1162 duplicates and unrelated articles were excluded by initial screening. The remaining 887 abstracts were further screened, excluding those that were non-English, nonhuman, published before 1983, case reports or studies with uncertain methods, or focused on populations with HIV or xeroderma pigmentosum. The resulting 36 studies were then acquired and various data were extracted, including patient demographics (age, sex, and race), length of the disease course, treatment strategy implemented, accuracy of clinical suspicion, accuracies of incisional and excisional biopsies, rates of positive and negative margins, success rates, time to resolution, recurrence rates, and time to recurrence.13–48 For each variable used in the decision and sensitivity analyses, the number of contributing studies, the range of reported findings, the calculated weighted means used for the base case, and the ranges used can be found in Table 1. Ranges used in sensitivity analyses were extended beyond the most extreme values observed in the literature to ensure that potential threshold values were not overlooked.
The mean accuracies of clinical and pathologic diagnosis were 0.84 and 0.98, respectively.13,15,16,23–26 The risk of positive margins after excision was estimated at 0.42.13,15,24,25,27,28 The mean rates of recurrence with positive and negative margin status were 0.52 and 0.11, respectively.13–16,24,27,29–31 The rate of successful treatment with IFN-α2b was estimated to be 0.88 and subsequent risk of recurrence was 0.06.16–22,27,32,33,37–40,42–48 Incisional biopsy accuracy rate was difficult to estimate as the literature contained a mixture of minimally invasive testing including true incisional biopsy, impression cytology, optical coherence tomography, and confocal microscopy.
Based on EV analysis of the decision tree, excision-directed IFN-α2b emerged as the optimal strategy with the lowest EV of persisting or recurrent disease (EV 0.13). Other strategies, ordered from lowest to highest disease burden, were empiric IFN-α2b (EV 0.17), excision-only (EV 0.22), and incisional biopsy–directed IFN-α2b (EV 0.30).
Sensitivity analyses identified 3 variables that would cause a shift in the decision when certain threshold values within their estimated range were applied. Empiric IFN-α2b would become the optimal strategy in the following situations: the probability of successful treatment with IFN-α2b increases from 0.88 to 0.94, the probability of recurrence for excised lesions with negative surgical margins increases from 0.11 to 0.20, or the accuracy of excisional histopathologic diagnosis decreases from 0.98 to 0.92. When avoidance of a missed diagnosis was assigned as the primary endpoint, empiric treatment with topical IFN-α2b was superior to other strategies as all patients with suspected OSSN were treated.
Currently, many options exist for the treatment of OSSN. Among the 4 strategies we explored, our analyses demonstrate that the optimal approach for minimizing the likelihood of disease persistence or recurrence is surgical excision followed by topical IFN-α2b for cases with positive surgical margins. Presumably, this combination of both surgical and chemotherapeutic strategies produces the lowest risk of recurrence because it combines the strengths of both arms; however, it also confers the risks and expenses associated with both modalities, which are factors not considered in this analysis. Similar conclusions regarding the benefits of topical IFN-α2b after surgical excision were found in a recent retrospective analysis of 389 patients with OSSN.49 The success rates of other modes of treatment were only slightly less effective in this model and remain reasonable alternatives given the individual needs and challenges of particular patients. A recent case–control study of 98 patients with OSSN treated with either surgery or topical IFN-α2b found no significant difference in recurrence rates.27 With avoidance of a missed diagnosis as the primary endpoint, empiric treatment with IFN-α2b was naturally superior because it presumes disease in all patients with suspected OSSN. In this regard, it is important to reiterate that our model assumes that all patients entering the decision have OSSN. Although our review of the literature estimates the accuracy of OSSN diagnosis on clinical grounds alone to be approximately 84%, the analysis does not take into account that some cases will be misdiagnosed as pterygia, pinguecula, papillomata, dyskeratoses, and nevi.11,13,14,21,41 Surgical approaches have higher diagnostic power, but will produce some missed diagnoses because of false-negative biopsies. Additionally, the decision to proceed with surgical excision depends on a preoperative clinical diagnosis and subsequently conveys all the risks associated with the excision mentioned earlier. Patients treated empirically with IFN-α2b without biopsy will have a number of false-positive empiric diagnoses but without the same surgical risks.
Sensitivity analyses to detect variables exerting the most influence on our outcome identified that the optimal decision was sensitive to 3 variables: success rate of IFN-α2b therapy, recurrence rates of excised tumors with negative surgical margins, and accuracy of excisional biopsy histopathology. Although the accuracy of excisional biopsy histopathology was well above 92% in all the studies examined, the other variables show considerable variation among reports, some of which readily meet the threshold values required for a change in the decision. The identification of key variables in the sensitivity analysis has important clinical implications. First, it points clinicians' attention to factors that may determine success or failure of different treatment strategies in their practices. Furthermore, the paucity of prospective randomized trials in the OSSN literature highlights the need for such trials and identifies these variables as important areas of investigation in such a study.
Our excision-only strategy is somewhat problematic because of challenges in the literature, as the management of cases with positive margins is uncertain in many reports. Most of these reports are retrospective pathology reviews in which a correlation with the initial treatment and pathologic findings are correlated with recurrence. The details of the management of positive margins are not well described and can involve an amalgam of simple excisional techniques, additional cryotherapy, and adjunct radiation or chemotherapeutic therapies; moreover, the efficacies of these interventions are highly variable and difficult to estimate. Whether clinicians just observed these patients hoping that marginal tumor had been cleared by cryotherapy or whether additional surgery, cryotherapy, or topical therapy was performed is difficult to extract from most of these reports. The rate of recurrence after surgery with negative margins may also be driven by variation in clinical practice. Because topical IFN-α2b therapy requires months of outpatient administration of eye drops, patient variation in compliance with these medications over long periods of time plays a considerable role in the selection of the optimal treatment strategy. Furthermore, many clinicians do not use IFN-α2b monotherapy for large thick tumors. Inclusion or exclusion of such patients in the case series may also explain the variation in the success rates with interferon. Finally, reports varied widely in their treatment regimen, definitions of success and failure, and length of follow-up.
Several limitations of our study should be noted. The heterogenous nature of the reports from which the data were extracted creates challenges in making comparisons among different studies. Articles varied significantly in IFN-α2b dosing, duration of treatment, and length of follow-up. In addition, we did not include other topical treatments such as mitomycin C (MMC) and 5-flurouricil (5-FU). We chose to focus on IFN-α2b for several reasons. First, IFN-α2b treatment has the least toxicity of these different modalities, and despite the lack of a randomized prospective trial, IFN-α2b has been the focus of most reports in recent years. A survey of cornea specialists conducted in 2012 reported that 56% preferred IFN-α2b as a first-line treatment over MMC and 5-FU. This represented a 5-fold increase in the use as monotherapy for lesions <2 mm and a 10-fold increase for lesions >8 mm compared with surveys from 2003.50 Furthermore, although efficacy and treatment regimens of MMC and 5-FU differ from those of IFN-α2b, analogous variables would be expected to influence their efficacy in comparison with surgical approaches. Like IFN-α2b, there is also a similar need for prospective randomized data to provide reliable data regarding the comparative efficacy and side effects of these medications in the treatment of OSSN to better inform their use in relation to surgical management.
In summary, our study identified surgical excision followed by topical IFN-α2b for cases with positive surgical margins as the most effective strategy for minimizing persistence or recurrence of OSSN. We also identified the efficacy of topical IFN-α2b, recurrence risk with negative surgical margins and IFN-α2b treatment, and accuracy of pathologic diagnosis as important drivers of clinical outcomes. Until additional data from a prospective trial are available, ophthalmologists hoping to improve the outcomes of patients with OSSN should strive to find ways to optimize these variables within their practices.
1. Lee GA, Hirst LW. Ocular surface squamous neoplasia
. Surv Ophthalmol. 1995;39:429–450.
2. Shields CL, Demirci H, Karatza E, et al.. Clinical survey of 1643 melanocytic and nonmelanocytic conjunctival tumors. Ophthalmology. 2004;111:1747–1754.
3. Sun EC, Fears TR, Goedert JJ. Epidemiology of squamous cell conjunctival cancer. Cancer Epidemiol Biomarkers Prev. 1997;6:73–77.
4. Tulvatana W, Bhattarakosol P, Sansopha L, et al.. Risk factors for conjunctival squamous cell neoplasia: a matched case-control study. Br J Ophthalmol. 2003;87:396–398.
5. Gichuhi S, Sagoo MS, Weiss HA, et al.. Epidemiology of ocular surface squamous neoplasia
in Africa. Trop Med Int Health. 2013;18:1424–1443.
6. Porges Y, Groisman GM. Prevalence of HIV with conjunctival squamous cell neoplasia in an African provincial hospital. Cornea. 2003;22:1–4.
7. Waddell KM, Lewallen S, Lucas SB, et al.. Carcinoma of the conjunctiva and HIV infection in Uganda and Malawi. Br J Ophthalmol. 1996;80:503–508.
8. Napora C, Cohen EJ, Genvert GI, et al.. Factors associated with conjunctival intraepithelial neoplasia: a case control study. Ophthalmic Surg. 1990;21:27–30.
9. Nakamura Y, Mashima Y, Kameyama K, et al.. Detection of human papillomavirus infection in squamous tumours of the conjunctiva and lacrimal sac by immunohistochemistry, in situ hybridisation, and polymerase chain reaction. Br J Ophthalmol. 1997;81:308–313.
10. Sjo NC, von Buchwald C, Cassonnet P, et al.. Human papillomavirus in normal conjunctival tissue and in conjunctival papilloma: types and frequencies in a large series. Br J Ophthalmol. 2007;91:1014–1015.
11. Verma V, Shen D, Sieving PC, et al.. The role of infectious agents in the etiology of ocular adnexal neoplasia. Surv Ophthalmol. 2008;53:312–331.
12. Nanji AA, Sayyad FE, Karp CL. Topical chemotherapy for ocular surface squamous neoplasia
. Curr Opin Ophthalmol. 2013;24:336–342.
13. Erie JC, Campbell RJ, Liesegang TJ. Conjunctival and corneal intraepithelial and invasive neoplasia. Ophthalmology. 1986;93:176–183.
14. Tunc M, Char DH, Crawford B, et al.. Intraepithelial and invasive squamous cell carcinoma of the conjunctiva: analysis of 60 cases. Br J Ophthalmol. 1999;83:98–103.
15. McKelvie PA, Daniell M, McNab A, et al.. Squamous cell carcinoma of the conjunctiva: a series of 26 cases. Br J Ophthalmol. 2002;86:168–173.
16. Sturges A, Butt AL, Lai JE, et al.. Topical interferon or surgical excision for the management of primary ocular surface squamous neoplasia
. Ophthalmology. 2008;115:1297–1302.
17. de Keizer RJ, de Wolff-Rouendaal D. Topical alpha-interferon in recurrent conjunctival papilloma. Acta Ophthalmol Scand. 2003;81:193–196.
18. Esquenazi S, Fry CL, Holley E. Treatment of biopsy proved conjunctival intraepithelial neoplasia with topical interferon alfa-2b. Br J Ophthalmol. 2005;89:1221.
19. Holcombe DJ, Lee GA. Topical interferon alfa-2b for the treatment of recalcitrant ocular surface squamous neoplasia
. Am J Ophthalmol. 2006;142:568–571.
20. Nemet AY, Sharma V, Benger R. Interferon alpha 2b treatment for residual ocular surface squamous neoplasia
unresponsive to excision, cryotherapy and mitomycin-C. Clin Exp Ophthalmol. 2006;34:375–377.
21. Schechter BA, Schrier A, Nagler RS, et al.. Regression of presumed primary conjunctival and corneal intraepithelial neoplasia with topical interferon alpha-2b
. Cornea. 2002;21:6–11.
22. Boehm MD, Huang AJ. Treatment of recurrent corneal and conjunctival intraepithelial neoplasia with topical interferon alfa 2b. Ophthalmology. 2004;111:1755–1761.
23. Rudkin AK, Dodd T, Muecke JS. The differential diagnosis of localised amelanotic limbal lesions: a review of 162 consecutive excisions. Br J Ophthalmol. 2011;95:350–354.
24. Tabin G, Levin S, Snibson G, et al.. Late recurrences and the necessity for long-term follow-up in corneal and conjunctival intraepithelial neoplasia. Ophthalmology. 1997;104:485–492.
25. Kao AA, Galor A, Karp CL, et al.. Clinicopathologic correlation of ocular surface squamous neoplasms at Bascom Palmer Eye Institute: 2001 to 2010. Ophthalmology. 2012;119:1773–1776.
26. Semenova EA, Milman T, Finger PT, et al.. The diagnostic value of exfoliative cytology vs histopathology for ocular surface squamous neoplasia
. Am J Ophthalmol. 2009;148:772–778.
27. Nanji AA, Moon CS, Galor A, et al.. Surgical versus medical treatment of ocular surface squamous neoplasia
: a comparison of recurrences and complications. Ophthalmology. 2014;121:994–1000.
28. Zaki AA, Farid SF. Management of intraepithelial and invasive neoplasia of the cornea and conjunctiva: a long-term follow up. Cornea. 2009;28:986–988.
29. Gunduz K, Ucakhan OO, Kanpolat A, et al.. Nonpreserved human amniotic membrane transplantation for conjunctival reconstruction after excision of extensive ocular surface neoplasia. Eye. 2006;20:351–357.
30. Peksayar G, Altan-Yaycioglu R, Onal S. Excision and cryosurgery in the treatment of conjunctival malignant epithelial tumours. Eye. 2003;17:228–232.
31. Sudesh S, Rapuano CJ, Cohen EJ, et al.. Surgical management of ocular surface squamous neoplasms: the experience from a cornea center. Cornea. 2000;19:278–283.
32. Karp CL, Galor A, Lee Y, et al.. Pegylated interferon alpha 2b for treatment of ocular surface squamous neoplasia
: a pilot study. Ocul Immunol Inflamm. 2010;18:254–260.
33. Karp CL, Moore JK, Rosa RH. Treatment of conjunctival and corneal intraepithelial neoplasia with topical interferon alpha-2b
. Ophthalmology. 2001;108:1093–1098.
34. Kheirkhah A, Mahbod M, Farzbod F, et al.. Repeated applications of impression cytology to increase sensitivity for diagnosis of conjunctival intraepithelial neoplasia. Br J Ophthalmol. 2012;96:229–233.
35. Parrozzani R, Lazzarini D, Dario A, et al.. In vivo confocal microscopy of ocular surface squamous neoplasia
. Eye. 2011;25:455–460.
36. Tole DM, McKelvie PA, Daniell M. Reliability of impression cytology for the diagnosis of ocular surface squamous neoplasia
employing the Biopore membrane. Br J Ophthalmol. 2001;85:154–158.
37. Vann RR, Karp CL. Perilesional and topical interferon alfa-2b for conjunctival and corneal neoplasia. Ophthalmology. 1999;106:91–97.
38. Besley J, Pappalardo J, Lee GA, et al.. Risk factors for ocular surface squamous neoplasia
recurrence after treatment with topical mitomycin C and interferon alpha-2b
. Am J Ophthalmol. 2014;157:287–293.
39. Chen HC, Chang SW, Huang SF. Adjunctive treatment with interferon alpha-2b
may decrease the risk of papilloma-associated conjunctival intraepithelial neoplasm recurrence. Cornea. 2004;23:726–729.
40. Galor A, Karp CL, Chhabra S, et al.. Topical interferon alpha 2b eye-drops for treatment of ocular surface squamous neoplasia
: a dose comparison study. Br J Ophthalmol. 2010;94:551–554.
41. Karp CL, Galor A, Chhabra S, et al.. Subconjunctival/perilesional recombinant interferon 2b for ocular surface squamous neoplasia
: a 10-year review. Ophthalmology. 2010;117:2241–2246.
42. Kim HJ, Shields CL, Shah SU, et al.. Giant ocular surface squamous neoplasia
managed with interferon alpha-2b
as immunotherapy or immunoreduction. Ophthalmology. 2012;119:938–944.
43. Munoz de Escalona Rojas JE, Garcia Serrano JL, Cantero Hinojosa J, et al.. Application of interferon alpha 2b in conjunctival intraepithelial neoplasia: predictors and prognostic factors. J Ocul Pharmacol Ther. 2014;30:489–494.
44. Ramasubramanian A, Shields CL, Sinha N, et al.. Ocular surface squamous neoplasia
after corneal graft. Am J Ophthalmol. 2010;149:62–65.
45. Schechter BA, Koreishi AF, Karp CL, et al.. Long-term follow-up of conjunctival and corneal intraepithelial neoplasia treated with topical interferon alfa-2b. Ophthalmology. 2008;115:1291–1296.
46. Shah SU, Kaliki S, Kim HJ, et al.. Topical interferon alfa-2b for management of ocular surface squamous neoplasia
in 23 cases: outcomes based on American Joint Committee on Cancer classification. Arch Ophthalmol. 2012;130:159–164.
47. Shields CL, Kaliki S, Kim HJ, et al.. Interferon for ocular surface squamous neoplasia
in 81 cases: outcomes based on the american joint committee on cancer classification. Cornea. 2013;32:248–256.
48. Shields CL, Kancherla S, Bianciotto CG, et al.. Ocular surface squamous neoplasia
(squamous cell carcinoma) of the socket: management of extensive tumors with interferon. Ophthal Plast Reconstr Surg. 2011;27:247–250.
49. Galor A, Karp CL, Oellers P, et al.. Predictors of ocular surface squamous neoplasia
recurrence after excisional surgery. Ophthalmology. 2012;119:1974–1981.
50. Adler E, Turner JR, Stone DU. Ocular surface squamous neoplasia
: a survey of changes in the standard of care from 2003 to 2012. Cornea. 2013;32:1558–1561.