Prognostic Factors for Orbital Exenteration, Local Recurrence, Metastasis, and Death from Disease in Conjunctival Squamous Cell Carcinoma

Abstract

Purpose:

Information regarding risk of metastasis and disease-related death (DD) from conjunctival squamous cell carcinoma (SCC) is relatively scarce. We explored prognostic factors for orbital exenteration, local recurrence, nodal metastasis, and DD in patients with conjunctival SCC.

Design:

Retrospective cross-sectional study.

Methods:

All consecutive patients with conjunctival SCC treated by the senior author at MD Anderson Cancer Center during1999–2018 were included. Survival curves were estimated using the Kaplan-Meier method, and survival differences were assessed using 2-sided log-rank tests.

Results:

The study included 44 patients (24 men;20 women); median age was 64 years (range, 40–90). T categories at presentation were as follows: Tis, 20 patients; T2, 8; T3, 9; and T4, 7. Eighteen patients (41%) had tumors exclusively in the bulbar conjunctiva; 26(59%) had non-bulbar conjunctival involvement. The median follow-up time was 29.2 months (95% CI: 21.8–44.3). Orbital exenteration was performed in 10 cases (23%) and was associated with T3 or more advanced disease at presentation (p<0.001). Seven patients developed local recurrence during follow-up. History of organ transplant correlated with local recurrence and orbital exenteration (p<0.01). Nodal metastasis was present in 1 patient at presentation and occurred in 3 patients during follow-up, for an overall nodal metastasis rate of 9% (4/44). By end of follow-up, 2 patients had died of disease, 4 had died of other causes, and 38 were alive with no evidence of disease. Our results suggest that both orbital exenteration and nodal metastasis are independent variables associated with DD.

Conclusions:

In patients with conjunctival SCC, orbital exenteration and nodal metastasis are associated with DD and organ transplantation is associated with orbital exenteration.

Précis:

In patients with conjunctival squamous carcinoma, overall risk of nodal metastasis is about 10%. Orbital exenteration and nodal metastasis are independent variables associated with disease-related death.

Introduction

Conjunctival squamous cell carcinoma (SCC) is the most common malignant epithelial tumor of the conjunctiva.¹ In one of the largest series to date of conjunctival neoplasms which included 5002 cases, 14% were SCC, and the authors reported that the incidence of SCC varied widely by geographic location and ethnicity.² In more developed Western countries, the most frequently encountered clinical presentation of conjunctival SCC is in situ disease, which is also referred to as ocular surface squamous neoplasia (OSSN)³; invasive conjunctival SCC is relatively rare and is more commonly seen in immunocompromised individuals.⁴

Most previous reports on OSSN or conjunctival SCC have been focused on local treatments, mostly for bulbar conjunctival disease; few authors have reported on the risk of orbital invasion, the rate of orbital exenteration, and the risks of local recurrence, nodal metastasis or death from disease. To our knowledge, only 1 previous report on conjunctival SCC attempted to validate the prognostic value of the latest (8th edition) American Joint Committee on Cancer (AJCC) TNM criteria for this disease.^{5, 6}

The purpose of the study reported here was to explore prognostic factors for orbital exenteration, local recurrence, nodal metastasis, and death from disease in a cohort of patients with conjunctival SCC. Furthermore, we aimed to validate the prognostic value of the AJCC 8th-edition TNM criteria for conjunctival SCC. (Supplemental Table 1)

Methods

This retrospective study was carried out following the tenets of the Declaration of Helsinki and is compliant with the regulations and conditions set forth in the Health Insurance Portability and Availability Act. Approval for the study was obtained from the Institutional Review Board at The University of Texas MD Anderson Cancer Center.

The clinical records of all consecutive patients treated by 1 author (BE) between May 1999 and August 2018 for histologically proven conjunctival SCC were reviewed for age, sex, race/ethnicity, clinical information needed for disease classification at presentation (tumor location; clinical tumor size; involvement of the eyelid, orbit, and surrounding structures; presence and location of lymph node and/or distant metastasis), treatment information (surgical treatments, adjuvant topical chemotherapy or radiotherapy), and outcomes of interest (orbital exenteration, local recurrence, lymph node metastasis, distant metastasis, death from disease, and death from any cause). Histopathological records were also reviewed for tumor size based on the surgical specimen and perineural invasion. Each tumor was classified when the patient sought treatment at our center according to the AJCC 8th-edition criteria for conjunctival carcinoma.

Patient characteristics were summarized using descriptive statistics. Overall survival was defined as the time interval between the start date of treatment and the date of death and was censored at the last follow-up date for patients who were alive. Disease-specific survival was defined as the time interval between the date of treatment and the date of death for patients who died of disease and was censored at the last follow-up date or the date of death of unrelated disease for patients who were alive or died of other causes, respectively. Local-recurrence-free and nodal-metastasis-free survival were defined as the time interval between the date of treatment and the date that the event (local recurrence or nodal metastasis) occurred and was censored at the last follow-up date or date of death, whichever occurred first, for patients who did not present the event. Associations between patient characteristics and outcomes diagnosed during follow-up were studied using survival analysis. Survival curves were estimated using the Kaplan-Meier method, and differences in survival among groups were assessed using 2-sided log-rank tests. Univariate Cox proportional hazards regression models were used to estimate and evaluate effect sizes of risk factors on survival. Associations between patient characteristics and outcomes diagnosed at presentation to MD Anderson and by the end of the study were assessed using Fisher’s exact tests. For the analyses of risk factors for outcomes diagnosed by the end of the study, we used the last information available for each patient, regardless of their follow-up time. Statistical analyses were conducted in R version 3.4.2.

Results

A total of 44 patients were included in the study. The patients’ clinical characteristics are summarized in Table 1. There were 24 men and 20 women. Thirty-seven patients (84%) were White. The median age at presentation to our center was 64 years (range, 40–90) (Table 1). Thirty-four patients (77%) presented to our center with a primary lesion; the other 10 presented with recurrent lesions. Twenty-seven patients (61%) had conjunctival SCC of the left eye only. Twenty-four patients (55%) had invasive conjunctival SCC; the other 20 presented with in situ disease (OSSN). Among the patients with invasive conjunctival SCC and available information (N=11), the median tumor thickness was 2 mm (range, 0.47–18 mm). The median largest tumor diameter was 14 mm (range, 4–45 mm).

Table 1.

Demographics and tumor characteristics of 44 patients with conjunctival squamous cell carcinoma (SCC)^a

Characteristics	Value
Sex
Female	20 (45)
Male	24 (54)
Race/ethnicity
African American	3 (7)
Hispanic	4 (9)
White	37 (84)
Age at presentation to MD Anderson, median (min, max), years	64 (40, 90)
Presented to MD Anderson with primary lesion
No	10 (23)
Yes	34 (77)
Side of tumor
Both	1 (2)
Left	27 (61)
Right	16 (36)
TNM per AJCC 8th edition at presentation
TisN0M0	20 (45.45)
T2N0M0	8 (18.18)
T3N0M0	9 (20.45)
T4aN0M0	6 (13.64)
T4aN1M0	1 (2.27)
TNM per AJCC 8th edition by end of follow
TisN0M0	20 (45)
T2N0M0	8 (18)
T3N0M0	6 (14)
T3N1M0	2 (5)
T4aN0M0	6 (14)
T4aN1M0	2 (5)
Definitive treatment
Chemotherapy alone	4 (9)
Excision alone	6 (14)
Excision with adjuvant chemotherapy and radiation	2 (5)
Excision with adjuvant chemotherapy	14 (32)
Exenteration	10 (23)
Neoadjuvant chemotherapy and excision	3 (7)
Neoadjuvant chemotherapy, excision, and adjuvant chemotherapy	5 (11)
Exenteration
No	34 (77)
Yes	10 (23)
History of non-cutaneous malignancy
No	42 (95)
Yes	2 (5)
History of organ transplant
No	41 (93)
Yes	3 (7)
Active or past history of smoking
No	37 (84)
Yes	7 (16)
SCC in anophthalmic Socket
No	41 (93)
Yes	3 (7)
Perineural invasionb
No	10 (42)
Unknown	12 (50)
Yes	2 (8)

^aValues in table are number of patients (percentage) unless otherwise indicated.

^bFor the 24 patients with T2-T4 disease (i.e, with invasive carcinoma).

At presentation to our center, 8 patients had T2 tumors, 9 had T3 tumors, 7 had T4 tumors (Figure 1), and 20 patients had Tis tumors. Eighteen patients had tumors exclusively involving the bulbar conjunctiva. Thirty patients underwent excision with or without neoadjuvant and/or adjuvant therapy, 4 underwent chemotherapy alone and 10 patients underwent orbital exenteration. Only 2 patients (5%) had a history of a non-cutaneous malignancy, and 3 (7%) had a history of organ transplant. Seven patients (16%) were smokers or had a past history of smoking.

Figure 1.

(A) 79 year old female presented with multiply recurrent squamous cell carcinoma of conjunctiva treated with repeated surgical excisions and over two years of topical interferon treatment without success. At the time of presentation to our hospital, the tumor was growing rapidly and invaded the orbit (B) and the patient also was found to have ipsilateral parotid nodal metastasis (C) during the staging workup at our hospital. After two cycles of immunotherapy using cemiplimab without any measurable response, an orbital exenteration, a free flap reconstruction, and parotidectomy and neck dissection were performed followed by adjuvant radiation therapy to the orbit and parotid and neck area.

One patient had nodal metastasis at presentation, and 3 additional patients developed nodal metastasis during the follow-up period. The median time to nodal metastasis among patients who developed nodal metastatic during follow-up was 36.5 months (range, 14.9–70.2 months). Two patients (8%) had perineural invasion detected on the surgical specimen.

No patient had documented distant metastasis at baseline or during follow-up; however, chart review showed that 2 patients died of disease, likely with distant metastasis. Seven patients developed local recurrence during the follow-up period. The median time to local recurrence among patients who developed local recurrence was 9.46 months (range, 2.79–41.46 months).

By the end of follow-up, 2 patients had died of disease (one at 50 months and the other at 97 months after presentation to our center), 4 had died of other causes, and 38 were alive with no evidence of disease.

Associations between patient characteristics

Table 2 shows the associations between T category at presentation and other patient characteristics, including orbital exenteration. There was no significant association between T category and history of non-cutaneous malignancy, history of organ transplant, or side of tumor. However, the proportion of patients who underwent orbital exenteration increased with increasing T category (p<0.001, Figure 2).

Table 2.

Association between T category at the end of follow-up (AJCC 8th edition) and selected other patient characteristics

Other characteristic	T2	T3	T4	Tis	p-value
History of noncutaenous malignancy
No	7 (88)	9 (100)	7 (100)	19 (95)	0.609
Yes	1 (12)	0 (0)	0 (0)	1 (5)
History of organ transplant
No	8 (100)	8 (89)	6 (86)	19 (95)	0.547
Yes	0 (0)	1 (11)	1 (14)	1 (5)
Orbital exenteration
Yes	1 (12)	2 (22)	6 (86)	1 (5)	<0.001
No	7 (88)	7 (78)	1 (14)	19 (95)
Side of tumor
Left	5 (62)	5 (62)	4 (57)	13 (65)	>0.99
Right	3 (38)	3 (38)	3 (43)	7 (35)

Figure 2.

Association between T category at the end of follow-up and orbital exenteration.

Orbital exenteration was also significantly associated with a history of organ transplant (p<0.01); 17% (7/41) of the patients without a history of organ transplant had orbital exenteration, compared to all (3/3) of the patients with a history of organ transplant.

There was no statistically significant association between invasive SCC and side of tumor (p>0.99).

Associations between patient characteristics and outcomes by the end of follow-up

Table 3 shows local-recurrence-free survival, nodal-metastasis-free survival, overall survival, and disease-specific survival rates.

Table 3.

Local-recurrence-free survival, nodal-metastasis-free survival, overall survival, and disease-specific survival rates with 95% CI

Outcome	year 1	year 2	year 3	year 4	year 5
Local-recurrence-free survival	0.86 (0.76;0.98)	0.86 (0.76;0.98)	0.81 (0.68;0.97)	0.71 (0.52;0.97)	0.71 (0.52;0.97)
Nodal-metastasis-free survival*	1 (1;1)	0.97 (0.9;1)	0.97 (0.9;1)	0.91 (0.78;1)	0.91 (0.78;1)
Overall survival	0.98 (0.93;1)	0.94 (0.87;1)	0.94 (0.87;1)	0.94 (0.87;1)	0.67 (0.45;1)
Disease-specific survival	1 (1;1)	1 (1;1)	1 (1;1)	1 (1;1)	0.91 (0.75;1)

^*One patient with lymph node metastasis at presentation was excluded from the time-to-event analysis.

Local recurrence

There was a statistically significant association between local-recurrence-free survival and a history of organ transplant (hazard ratio=5.98 [95% CI: 1.12–32.02], log-rank test p=0.02; Figure 3). However, local-recurrence-free survival was not significantly associated with history of non-cutaneous malignancy (p=0.5), T category at presentation, (p=0.21), exclusive non-bulbar location of tumor (p=0.95), tumor with a non-bulbar component (p=0.79), or tumor size treated as a continuous variable (p=0.843) or a categorical variable (0–10, 11–20, >20, p=0.9). For patients who had orbital exenteration, the median time to orbital exenteration since initial presentation to MDACC was 1.87 months (range: 0.95–41.46).

Figure 3.

Kaplan-Meier curve for local recurrence by history of organ transplant. E, event; N, sample size.

Nodal metastasis

Because only 3 patients had nodal metastasis during follow-up, survival analysis to study the association between patient characteristics and nodal metastasis was not feasible. Therefore, we studied the association between patient characteristics and nodal metastasis by the end of follow-up (i.e., at presentation or during follow-up). Results are summarized in Table 4. No statistically significant associations were detected between any of the characteristics examined and nodal metastasis by the end of follow-up.

Table 4.

Association between N category by the end of follow-up^a (AJCC 8th edition) and selected other patient characteristics^b

Other characteristic	N0 (N=40)d	N1 (N=4)	p value
T category at presentation
T2	8 (40)	0 (0)	0.273
T3	6 (30)	3 (75)
T4	6 (30)	1 (25)
T category at presentation
T2	8 (40)	0 (0)	0.262
T3+T4	12 (60)	4 (100)
T category at presentation
T2+T3	14 (70)	3 (75)	>0.99
T4	6 (30)	1 (25)
T categoryc during follow up
T2	8 (40)	0 (0)	0.494
T3	6 (30)	2 (50)
T4	6 (30)	2 (50)
T categoryc during follow up
T2	8 (40)	0 (0)	0.262
T3+T4	12 (60)	4 (100)
T categoryc during follow up
T2+T3	14 (70)	2 (50)	0.578
T4	6 (30)	2 (50)
History of organ transplant
No	38 (95)	3 (75)	0.254
Yes	2 (5)	1 (25)
History of noncutaneous malignancy
No	38 (95)	4 (100)	>0.99
Yes	2 (5)	0 (0)
Exclusively nonbulbar location
No	27 (68)	3 (75)	>0.99
Yes	13 (32)	1 (25)
Nonbulbar component
No	18 (45)	0 (0)	0.133
Yes	22 (55)	4 (100)

^aBy the end of follow-up means at presentation or during follow-up.

^bValues in table are number of patients (percentage) unless otherwise indicated.

^cValues refer to the highest T category encountered for each patient during the study period (at presentation or during follow up)

^dFor the T categories, only the 24 patients with T2-T4 disease were considered (i.e, only patient with invasive SCC since the risk of nodal metastasis for Tis should be nil).

Death of disease

Because of the small number of disease-related deaths, a survival analysis was not feasible. However, Table 5 summarizes the associations between several patient characteristics and death of disease occurrence. There were significant associations between orbital exenteration (p=0.048) and N category at the end of follow-up (p<0.01) and death of disease occurrence. However, this result should be interpreted with caution since only 2 patients died of disease. No statistically significant associations were detected between any of the other variables tested and disease-related death.

Table 5.

Association between occurrence of death of disease and selected patient characteristics^a

Characteristicb	No (N=42)	Yes (N=2)	p value
Exenteration
No	34 (81)	0 (0)	0.048
Yes	8 (19)	2 (100)
History of noncutaneous malignancy
No	40 (95)	2 (100)	>0.99
Yes	2 (5)	0 (0)
History of organ transplant
No	40 (95)	1 (50)	0.133
Yes	2 (5)	1 (50)
T category at presentation
T2	8 (36)	0 (0)	0.308
T3	7 (32)	2 (100)
T4	7 (32)	0 (0)
T category at presentation
T2	8 (36)	0 (0)	0.536
T3+T4	14 (64)	2 (100)
T category at presentation
T2+T3	15 (68)	2 (100)	>0.99
T4	7 (32)	0 (0)
T category during follow up
T2	8 (36)	0 (0)	>0.99
T3	7 (32)	1 (50)
T4	7 (32)	1 (50)
T category during follow up
T2	8 (36)	0 (0)	0.536
T3+T4	14 (64)	2 (100)
T category during follow up
T2+T3	15 (68)	1 (50)	>0.99
T4	7 (32)	1 (50)
N category at -presentation
N0	41 (98)	2 (100)	>0.99
N1	1 (2)	0 (0)
N category during follow up
N0	40 (95)	0 (0)	0.006
N1	2 (5)	2 (100)
Exclusively nonbulbar location
No	28 (67)	2 (100)	>0.99
Yes	14 (33)	0 (0)
Nonbulbar component
No	18 (43)	0 (0)	0.505
Yes	24 (57)	2 (100)

^aValues in table are number of patients (percentage) unless otherwise indicated.

^bAll T and N categories are per the 8th edition of the AJCC Cancer Staging Manual.

Death of any cause

There was a statistically significant association between overall survival and exclusive non-bulbar location of tumor (p=0.02; Figure 4). The following variables were tested and found not to be significantly associated with overall survival: history of organ transplant (p=0.24), history of non-cutaneous malignancy (p=0.72), T category at presentation (p=0.94), T category at presentation ≥3 vs. <3 (p=0.83), N category at presentation (p=0.81), and tumor with a non-bulbar component (p=0.12).

Figure 4.

Kaplan-Meier curve for death of any cause by location of tumor with respect to bulbar conjunctiva. E, event; N, sample size.

The median follow-up time was 29.2 months (95% CI: 21.8–44.3). For patients alive at the time of analysis, the follow-up time ranged from 2.53 to 81.81 months. The median time to death was 94.6 months (95% CI: 57.8-NA).

Discussion

The major findings of our report include the significant association found between orbital exenteration and nodal metastasis and a higher risk of disease-related death. It is important to note that disease-related death occurred in only 2 patients (5% of cohort). Another novel finding of our study is that organ transplant was a significant risk factor for local recurrence and orbital exenteration. We also found that patients with higher T category at presentation were more likely to have an orbital exenteration.

Lymph node metastasis is a well-established prognostic factor in many head and neck malignancies; however, the ophthalmic literature includes little information about lymph node metastasis associated with conjunctival SCC. The possibility of metastatic disease in patients with conjunctival SCC was first alluded to in 1942 without any documentation of the extent of metastatic disease.⁷ In 1986, in a series of 98 patients, by Erie et al. only 2 were found to have metastatic disease.⁸ In 1988, Tabbra et al. reported 10 cases of metastatic conjunctival SCC: 9 to the regional lymph nodes and 1 to bone and lungs.⁹ In 2000. Cervantes et al. reported only 2 cases of nodal metastasis in a series of 286 patients.¹⁰ We found the overall risk of nodal metastasis to be 9%. To our knowledge, no previous report established nodal metastasis as a prognostic factor for disease-related death in patients with conjunctival SCC. Our finding highlights the importance of a thorough search for regional nodal metastasis using both palpation and imaging (ultrasonography and/or computed tomography) both at presentation and during the surveillance follow-up period in patients with conjunctival SCC. We did not find any significant association between tumor location at presentation and nodal metastasis, in contrast to Chauhan et al, who previously reported a significant association between tumor location and metastasis; of note, however, those authors did not specify whether “metastasis” referred to regional lymph node or distant metastasis.¹¹ In addition, our findings corroborate the report by Cruzado-Sanchez et al. who described a metastasis rate of 8% in a series of 176 Peruvian patients: 13 patients had regional nodal metastasis and 1 patient had distant metastasis. While the latter study did not report on the association between metastasis and death, it described an absence of statistically significant correlation between the location of the tumor and the risk of recurrence or metastasis.¹² This finding is consistent with our report.

Presumed death related to conjunctival SCC was first reported in the USA, in 1942, by Ash et al. At the time, there were no evidence of disease-related death and it was speculated that that one patient in a series of 93 limbal epithelial tumors might have died of disease. The authors emphasize the non-lethal course of conjunctival squamous carcinoma and conclude that surgical excision without radiation represents an adequate therapy.⁷ Disease-related death occurred in only 2 patients (5%) in our cohort. Chauhan et al,⁷ who reported on patients treated in India, reported a higher risk of death: 12%: In their series of 64 patients with conjunctival SCC, 3 patients with T3 or T4 disease and 1 patient with T1 or T2 disease died of disease. A study from the United States published by Iliff et al in 1975 showed close to a 4% risk of death in a series of 27 patients with invasive conjunctival SCC.¹³ Of the 27 patients, only 1 died of generalized metastasis, 18 months after exenteration and radical neck dissection. In a more recent report from the United States, Tunc et al described 60 patients with conjunctival OSSN, 22 with intraepithelial tumors and 38 with invasive SCCs, and reported that none had died at the time of the analysis, with patient follow-up periods ranging from 18 to 226 months, with a mean of 56 months.¹⁴ In other major publications from the United States on prognosis of patients with conjunctival SCC and OSSN, no patients were reported to have experienced nodal metastasis or death.^6,8,10

To our knowledge, the association between occurrence of disease-related death and orbital exenteration has not been explored previously. Chauhan et al reported that patients with higher T category according to the 7^th edition of the AJCC Cancer Staging Manual were at increased risk of death, but these investigators did not report on an association between orbital exenteration and disease-related death.¹¹ In general, orbital exenteration is used only as a treatment of last resort in patients with recurrent conjunctival SCC in whom all previous forms of treatment such as eye sparing surgery and topical chemotherapy have failed; patients treated with orbital exenteration often have orbital extension of (T4 tumors). Ten patients (22.7%) in our series underwent an orbital exenteration; all had T3 or T4 tumors at presentation. This is a relatively high proportion of patients receiving an orbital exenteration and likely is explained by the fact that patients were treated at a major cancer referral center by an ophthalmologist specializing in orbital and oculoplastic surgery. This interpretation is supported by our finding of a significant association between a higher T category at presentation and orbital exenteration. It is possible that the smaller proportion of patients undergoing orbital exenteration reported in previously published large series of conjunctival SCC may be due at least in part to the fact that most previous reports have been from cornea/ocular surface or ocular oncology subspecialty practices rather than from an orbit and oculoplastic surgery subspecialty practice like ours.^{3, 6, 15, 16} However, Chauhan et al did report a similar rate of orbital exenteration, 28% (18 of 64 patients), with 94% (17 of 18) of the patients who underwent exenteration having T3 or T4 disease.¹¹.

Our report supports the previously reported aggressive natural history of conjunctival SCC in immunosuppressed patients. While in our series, all immunocompromised patients were organ transplant recipients, previous reports described a similar course regardless of the cause of immunosuppression: organ transplant, HIV infection, or other chronic disease.^{13, 17} It is interesting to note that while there was no association between higher T stage and organ transplant, there was an association between higher T stages and orbital exenteration, and there was also an association between organ transplant and exenteration. This may be explained by the fact that each of our 3 organ transplant patients had a different T stages yet, all 3 underwent an exenteration. Immunocompromised patients require close follow-up, and when an immunocompromised patient presents with signs and symptoms that indicate potential conjunctival SCC, a high level of suspicion and a low threshold for intervention to save the patient’s eye and life are appropriate.

In contrast with our finding of no association between higher T category and risk of local recurrence, some previous studies using the 7^th edition of AJCC classification have validated the T category as the main prognostic factor in patients with conjunctival SCC, particularly as a risk factor for local recurrence.^{3, 16, 18, 19} An important factor to consider is that many of these previous reports classified conjunctival tumors only as T1, T2, T3, or T4 without any Tis category, even though the authors reported that at least some of the patients described had in situ disease or even mild to moderate dysplasia.^{3, 11, 15} Another issue with some of these studies is lack of availability of histologic information for classification of tumors. In other words, the diagnosis and staging of conjunctival carcinoma were based on clinical examination only without histologic confirmation. In our opinion, this can be an issue: it is impossible to stage in situ versus invasive disease without histological confirmation. Furthermore, the lack of any patients with nodal or distant metastasis in large cohorts of patients previously published in the ophthalmic literature ^(6,8,10) raises the question of whether there was adequate surveillance for nodal or distant metastasis or possibly due to the fact that most of these series were generated from corneal practices with few invasive carcinomas.

To our knowledge, only 1 previous study has specifically evaluated the prognostic value of the AJCC 8th-edition criteria for conjunctival SCC.⁶ In our current report, we did not establish a significant correlation between T category at presentation, tumor size, or tumor location and risk of local recurrence. The lack of correlation between T category and local recurrence may be partly due to the definition of “T3” for conjunctival carcinoma in the 8th edition of AJCC Cancer Staging Manual, which includes any invasive carcinoma that “involves the cornea, tarsus, fornix or caruncle”.⁵ This is a potentially problematic definition in that it equates a conjunctival SCC with even slight involvement of the cornea at the limbus with a conjunctival SCC that may involve large parts of the fornix, tarsal and palpebral conjunctiva, and caruncle. In future modifications of the AJCC staging system, investigators should consider categorizing the T3 tumors differently and at least distinguishing minimal corneal involvement from involvement of the palpebral and tarsal conjunctiva and caruncle. Despite the potential shortcoming of the current AJCC criteria for conjunctival SCC, we advocate routine use of the AJCC classification to facilitate consistent comparisons of various patient cohorts and treatment outcomes. In particular, the N category (nodal metastasis or not) is an important prognostic factor for disease-related death, and nodal disease should be evaluated and documented at presentation and during the follow-up surveillance period, especially given the recent availability and approval of immune checkpoint inhibitors such as cemiplimab for treatment of locally advanced and metastatic SCC.²⁰

The limitations of our study include its retrospective nature, the relatively small number of patients, and perhaps the disproportionate number of patients with locally advanced conjunctival SCC with orbital and adnexal invasion. However, this latter unique feature of our cohort also served as an opportunity for us to gain insights into the relatively infrequent but important cases of recurrent or locally advanced conjunctival SCC that extend beyond the bulbar conjunctiva.