Topical 5-fluorouracil 1% as primary treatment for ocular surface squamous neoplasia

Abstract

Introduction

Ocular surface squamous neoplasia (OSSN) is a term that encompasses a spectrum of epithelial squamous malignancies, ranging from dysplasia to invasive carcinoma.¹ It represents the most common non-pigmented tumor of the ocular surface.² Traditional treatment for OSSN involves excision using a no-touch technique, with adjuvant alcohol and cryotherapy.^3-6 Unfortunately, recurrences do occur after surgery, presumably due to microscopic disease beyond the clinically-identified lesion.⁷ Even with clear margins on pathology, recurrences of up to 33% have been reported.⁷ Furthermore, extensive surgical intervention carries the risk of subsequent limbal stem cell deficiency and symblephara formation.

In order to potentially avoid these risks and treat the entire ocular surface, medical treatment as a primary therapy for OSSN has increased in popularity.^{8, 9} Chemotherapeutic agents used for treatment of OSSN include interferon-alpha-2b (IFN), mitomycin-C (MMC), and 5-fluorouracil (5FU), and all have been shown to be effective.^{4, 6, 9-16} IFN is an endogenous protein with anti-viral and anti-neoplastic properties.^{17, 18} It is often the first choice of therapy in OSSN given its favorable side effect profile. Adverse reactions to topical IFN drops are uncommon and if present, usually mild.¹⁹ A literature search of primary treatment of OSSN with IFN (as opposed to adjuvant therapy after surgical removal), identified six studies with a total of 137 patients of which 92.5% of patients had successful resolution of their OSSN.^20-26

MMC is a potent antimetabolite that was first used as a surgical adjunct before being adapted as a primary treatment for OSSN.^{6, 27} It has an unpleasant side effect profile, which includes pain in the majority of patients and rare but serious sequelae such as corneal and scleral melts.¹⁹ Of 10 studies with 212 patients treated with MMC as primary therapy, 90% of patients had successful resolution of their OSSN. ^{11, 20, 28-35}

5FU is a structural analogue of thymine, which inhibits DNA formation by blocking the enzyme thymidylate synthetase. Rapidly multiplying cells, such as tumor cells, require more DNA and RNA than normal cells and therefore take up larger amounts of 5FU, allowing selective targeting of cancerous lesions.³⁶ Solutions of 5FU are stable for several weeks at 25 °C, and therefore it does not need refrigeration. ³⁷

While 5FU has been used as an adjunct to surgery³⁸, literature reporting primary topical 5FU as a treatment for OSSN is more limited than that evaluating MMC and IFN. For primary treatment, only 37 cases in literature report successful resolution of OSSN ranging from 85% to 100%, mean 95.6% in these reports.^{4, 15, 39-41}

5FU has several theoretical advantages over IFN and MMC including a lower cost of treatment, no need for refrigeration³⁷, and less frequent dosing regimen. Given a limited amount of data on the use of 5FU as primary treatment in OSSN, we performed this study to evaluate the efficacy and side effect profile of 5FU as a primary treatment in OSSN.

Methods

Results

Demographics of the study population are presented in Table 1. As in previous studies, the majority of patients were white males in their 7^th decade of life (Table 1).

Resolution of OSSN

Of the 44 eyes, 36 lesions (82%) completely resolved with 5FU therapy (Figure 1 and ​and2).2). Of these, the median number of cycles was 4 (range 2 to 9, mean 3.8).

Open in a separate window

Figure 1

a: 71 year old male with extensive ocular surface squamous neoplasia of the right eye. b: High resolution optical coherence tomography revealed thickened hyperreflective epithelium (caliper) c: Dramatic resolution of tumor after 4 weekly cycles of 5 fluorouracil. d: High Resolution OCT shows smooth thin normal epithelium (caliper).

Open in a separate window

Figure 2

a: 70 year old male with recurrent OSSN of the right eye. A conjunctival lesion (dashed arrows) is present as well as corneal component (solid arrows) b: Ultra high resolution OCT shows thickened hyperreflective epithelium (arrow) c: Resolved tumor after 4 weekly cycles of 5 fluorouracil 1% d: High resolution OCT shows thin, normal epithelium.

OSSN recurrences after treatment

A total of 4 recurrences were seen in the 36 patients who initially responded to therapy; one of which was in a patient with squamous cell carcinoma. One recurrence was at 3.7 years, another at 1.4 years, and two at 3 months. Using Kaplan-Meier survival analysis, the 1 and 2 year recurrence rates were 6% and 15%, respectively (Figure 3). The presence of nodularity increased the risk of recurrence approximately 11 fold (95% confidence interval (CI) 0.97 to 122, Cox proportional hazards p=0.05). None of the other demographic or clinical factors associated with recurrence.

Open in a separate window

Figure 3

Kaplan Meyer survival curve shows disease free survival at 1 and 2 years to be 94% and 85% respectively.

Discussion

5FU was first reported as a treatment for OSSN in 1986.³⁹ Since then, a few case series and small reports have been published with an overall response frequency ranging from 57-100% for 5FU treatment which seems rather low.^{15, 39-41} However when examining only cases treated with 5FU as an exclusive, primary therapy (as opposed to adjuvant therapy after surgical removal), the response frequency improves and ranges from 85% to 100%.^{4, 15, 39-41} In our study, representing the largest group of patients treated with 5FU as primary therapy, we found that 82% of patients experienced complete resolution of OSSN. Our response frequency of 82% for 5FU is similar to what has been reported in larger series of IFN and MMC treated patients.^{11, 20-26, 28-35}

One issue in comparing treatment response between 5FU studies is that varying regimens have been used to treat OSSN. Nutt et al. reported that only 10 of 24 (41.6 %) tumors completed resolved with 5FU as primary therapy but in this series, no patient received more than 4 cycles.⁴³ Some authors favor continuous administration of 1% 5FU 3 or 4 times a day, for 4 weeks. ^{4, 39-41, 15} Yeatts et al administered 1% 5FU 4 times a day for 2-4 days and then repeated it after an interval of 30-45 days with success.¹⁵ A series by Parrozzani et al that included 41 patients of which 22 patients were treated primarily with 5FU 4 times a day daily, for a month.⁴¹ In these patients, all tumors completely resolved after mean of 1.9 monthly cycles (range 1-5).), however 3 patients (7.3%) developed early recurrence. In our series there were 4 recurrences and 2 were noted within 3 months following cessation of therapy. These early recurrences likely represent persistent subclinical disease (incomplete treatment) and not true recurrences. In more recent years, we use high resolution optical coherence tomography (HR-OCT) imaging to guide length of treatment, a practice that may decrease the frequency of premature termination of medical therapy.⁴⁴

Our side effect profile is also similar to what has been reported in the literature. A large review comparing data from all three topical treatments reported side effects for 5FU from a total of 26 patients, of whom 11 (42%) had both hyperemia and keratitis.¹⁹ Encouragingly, a confocal microscopy study showed no long-term differences between the treated and control (fellow) eyes after treatment with 5FU.⁴¹

Our response frequency of 82% for 5FU is similar to what has been reported in larger series of IFN and MMC treated patients. The side effects from 5FU are worse than with IFN but better than those of MMC. For IFN, success for primary treatment ranges 82% to 100%^{20, 22-26}, has minimal reported side-effects such as follicular conjunctivitis and mild irritation.⁴⁵ MMC, on the other hand, has a similar response frequency, ranging from 76% to 100%^{11, 20, 28-35}, with a much higher frequency of adverse reactions, in approximately 76% of patients.¹⁹ These range from acute effects such as conjunctival hyperemia, ocular allergy, superficial punctate keratitis, ocular pain, and epiphora, to long term effects such as recurrent corneal erosion, limbal stem cell deficiency and punctual stenosis. ⁴⁵ Punctal plugs are used routinely with MMC but not required in 5FU. Punctual stenosis has not been reported with topical 5 FU use.

As with all studies, our findings must be considered in light of its limitations, which include its retrospective nature, limited patient numbers, small number of SCC lesions, and evolving management of OSSN. For example, we now give patients 4 cycles of 5FU prior to declaring nonresponse to treatment. However, when first using 5FU, some patients switched to an alternative therapy prior to completing 4 cycles. In addition, ultra-high resolution anterior segment OCT (HR-OCT) has altered the way we diagnose and manage tumors. We suspect that the two recurrences we had at 3 months had sub-clinical disease and premature termination of therapy. We are now better able to follow the complete resolution of the tumor with the HR-OCT to ensure normalization before cessation of therapy. The effect of this imaging modality on recurrence rates, however, is not yet known. Finally, the ideal dosing regimen is still unknown, and other protocols might have different outcomes.

Despite these limitations, this study adds important knowledge regarding the role of topical 5FU as a primary treatment of OSSN. In our institution, topical IFN remains our first choice of medical therapy given its high tolerability. There are a few considerations, however, when 5FU is our preferred choice. These include (1) an inadequate or incomplete response to IFN; (2) economic considerations as 5FU ($40/month) is much less costly than IFN ($240/month) and MMC ($190/month); (3) refrigeration issues as both IFN and MMC need to kept refrigerated at all times but not 5FU; and (4) when patients prefer a cycling regiment as the ease of one week per month is easier than the daily dosing of IFN. This study demonstrated that topical 5FU is a good choice in these circumstances as it has a similar efficacy as both IFN and MMC and a side effect profile in the middle of the two agents. It is interesting, that although the response frequencies appear similar between the 3 topical agents, there are some patients who respond to one agent and not the other.^{46, 47} As these agents have different mechanisms of action, more research is needed to predict which patient is most likely to respond to which treatment modality.

Acknowledgments

Footnotes

References