Article Text

Clinical and pathological risk factors for worse stage and mortality of eyelid and periocular squamous cell carcinoma
  1. Yingxiu Luo1,2,
  2. Huifang Chen1,
  3. Shengfang Ge1,2,
  4. Yamin Rao3,
  5. Jie Yu1,2,
  6. Xianqun Fan1,2,
  7. Renbing Jia1,2,
  8. Shiqiong Xu1,2,
  9. Yefei Wang1,2
  1. 1 Department of Ophthalmology, Ninth People's Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China
  2. 2 Shanghai Key Laboratory of Orbital Diseases and Ocular Oncology, Shanghai, China
  3. 3 Department of pathology, Ninth People's Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China
  1. Correspondence to Dr Renbing Jia, Department of Ophthalmology, Ninth People's Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China; renbingjia{at}sjtu.edu.cn; Dr Yefei Wang; paper34{at}163.com; Dr Shiqiong Xu; 215769592{at}qq.com

Abstract

Background The clinical and pathological risk factors for worse T stage and prognosis in eyelid and periocular squamous cell carcinomas (SCCs) remain unclear. P63 was reported to predict a worse prognosis in other SCCs; however, this correlation was not validated in eyelid and periocular SCCs.

Methods We reported on a retrospective case series of 85 consecutive patients with eyelid and periocular SCCs from 1995 to 2019. Cox proportional hazards regression models and logistic regression models were applied for risk factor analysis.

Results Thirty-nine (45.8%) patients were diagnosed with T4 SCCs. Four (5.1%) patients developed nodal metastasis, and five (6.4%) patients developed distant metastasis during the follow-up. 2-year and 5-year disease-specific survival rates were 95.3% and 86.4%, respectively. Poorly or moderately differentiated eyelid and periocular SCCs were associated with worse T stage (p=0.001; p=0.008). Poor differentiation was associated with a higher risk of recurrence (p=0.024). Disease-specific death was more common in patients with T4 stage SCCs (p=0.038, HR=9.05). P63 expression was more common in patients with T3c or worse stage (p=0.008, OR=3.77). P63 expression alone was associated with worse differentiation (p=0.029), higher risk of perineural invasion (p=0.042, OR=4.61) and metastasis (p=0.009, HR=3.99). P63 expression (p=0.012, HR=7.80), coexpression of P63 and Ki67 (p=0.007, HR=9.21) and distant metastasis (p=0.001, HR=11.23) were associated with disease-specific death.

Conclusion Patients presented with more aggressive orbital invasion features and a higher rate of distant metastasis in this cohort. P63 and coexpression of Ki67 predicted a worse stage, differentiation and prognosis, including metastasis and death due to disease.

  • eye lids
  • neoplasia
  • pathology

Data availability statement

Data are available on reasonable request. The datasets used and analysed during the current study are available from the corresponding authors on reasonable request.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Background

Eyelid and periocular squamous cell carcinoma (SCC) is a rare malignancy with metastatic potential that may be fatal. It accounts for 3.4%–18.0% of all eyelid carcinomas among Asian patients.1–6 UV light exposure was widely identified as an extrinsic risk factor for SCC.7 It is a malignant tumour with a high recurrence rate ranging from 3.6% to 10.0%.6 8–10 Nodal metastasis has ranged from 0.8% to 24.3%,6 10 11 and distant metastasis has ranged from 0.8% to 6.3%.6 10 12 The mortality rate in different countries has varied, 8.3% in the USA,10 1.6% in Australia13 and 4.0% in India.6 Previous studies have also sought to identify risk factors for poor prognosis. For example, poor differentiation,14 previous recurrence, a worse T stage and chronic immunosuppression were reported to indicate a higher risk of local recurrence.8 T2c or worse T stage,10 poor differentiation15 and previous recurrence16 have suggested an increased risk of metastasis. T2c or worse T stage10 and overall metastasis15 increased the risk of a worse disease-specific survival (DSS).

P63 expression has been reported to be high and asscociated with poor prognosis in cutaneous SCC (cSCC)17 and head and neck SCC (HNSCC).18 A previous study revealed that ZNF750 is a critical target of P63 during SCC development.19 However, little information is available about p63 expression and its predictive value for the prognosis in eyelid and periocular SCC.

This study aimed to evaluate the clinical features and outcomes of eyelid and periocular SCC in Chinese, explore the risk factors for a poor prognosis.

Methods

Clinical data

This study adhered to the tenets of the Declaration of Helsinki, and consent from patients was achieved. Patients with eyelid or periocular SCC who visited Shanghai Ninth People’s Hospital, Shanghai JiaoTong University School of Medicine for treatment from January 1995 to June 2019 were included. Those with incomplete baseline information, tumours originated from other than eyelid or primary orbital SCC were excluded. For the prognosis study, cases with fewer than two follow-ups were excluded. The diagnosis of eyelid or periocular SCC and nodal metastasis was based on pathologic examination. For frozen section tissue processing technique, en face histological margin control was adopted to evaluate both the deep and continuous peripheral margins.

After careful review of the charts and pathological reports of these patients, the data of gender, age, tumour location, tumour size, primary or recurrent disease, time from ocular sign to diagnosis and pathological features were documented. The date and sites of metastasis, recurrence and death (if occurred) were collected. The time to events (recurrence, metastasis and death) was defined as the period from the date of treatment to the occurred events. Overall metastasis included both distant and nodal metastasis at any time (at present and the follow-up). The eighth edition of the American Joint Committee on Cancer (AJCC) staging system for eyelid carcinoma was applied to stratify the patients according to their clinical presentations.

Immunohistochemistry

The pathology charts of these patients were reviewed, and information concerning pathological differentiation, Ki67 and P63 was documented. Tumours were classified as low, intermediate and highly proliferating according to the value of Ki67 labelling index of ≤10%, 10%–30% and >30%, respectively.20 21

Statistical analysis

Kaplan-Meier survival analysis was applied to depict the overall survival (OS), DSS, metastasis and recurrence. Logistic regression was applied to identify risk factors for a worse T stage. Cox proportional hazards regression model was used to identify risk factors for recurrence, metastasis and disease-specific death. The HR (HR) and OR (OR) were used to describe the impact of risk factors. All statistical tests used a significance level of 5.0%, that is, p<0.05 was significant.

Results

Baseline clinical features and management

Ninety-two patients were referred to our clinic from 1995 to 2019, and seven were excluded because of insufficient information for T stage. The clinical and tumour features of the 85 patients included in this study are summarised (table 1). The AJCC8 TNM designations at presentation are listed in online supplemental table 1 and 39 (45.8%) patients presented at T4 stage. The median follow-up time was 29.1 months (range 1.7–225.5 months). T3c or worse category was significantly associated with poor differentiation (p=0.001).

Table 1

Patient and tumour characteristics

High recurrence, metastasis and death rates were observed

Seven patients were lost to follow-up after initial treatment at our hospital. Prognosis analysis in the remaining 78 patients was performed. By the last contact, nine (11.5%) patients had recurrence and the median time from diagnosis to local recurrence was 10.5 months (range 1.6–23.0 months). Local recurrence was significantly associated with poor differentiation (p=0.024).

At presentation, 12 (14.1%) patients had nodal metastasis. One (1.2%) patient had both nodal and brain metastasis. This patient died of disease 18.1 months after the initial diagnosis. During the follow-up, four (5.1%) patients developed nodal metastasis with a median time to metastasis of 32.3 months (range 25.8–41.1 months), three of which had T4b stage and one had T4a stage at presentation. One died 28.3 months after nodal metastasis. Five (6.4%) patients developed distant metastasis. Two metastasised to the lung, two had liver involvement and one had brain metastasis. The median time to distant metastasis was 21.1 months (range 4.6–50.1 months). Three of them died of disease with time from metastasis to death of 62.0 months, 5.5 months and 2.5 months, respectively (table 2).

Table 2

Status of metastasis and presentation and during follow-up

In total, 11 (14.1%) patients died of other diseases by the end of follow-up. Eight (10.3%) patients died of disease at the end of follow-up, including seven with T4 SCC and one with T3a SCC. Kaplan-Meier analysis revealed 2-year and 5-year OS rates of 95.3% and 69.0%, respectively. Additionally, the 2-year and 5-year DSS rates were 95.3% and 86.4%, respectively (online supplemental figure 1). Overall metastasis (nodal and distant) was associated with disease-specific death (p=0.011, HR=15.15, 95% CI 1.86 to 123.53, figure 1). A significant correlation was found between distant metastasis and disease-specific death (p=0.001, HR=11.23, 95% CI 2.80 to 44.99, figure 1).

Figure 1

(A) Kaplan-Meier plot of disease-specific survival stratified by overall metastasis. (B) Kaplan-Meier plot of disease-specific survival stratified by distant metastasis.

P63 predicts a worse T stage and prognosis

P63 is expressed in one-third of eyelid and periocular SCC. In this cohort of 85 patients, 29 (34.1%) of patients were documented as P63 positive.

P63 positivity was associated with a higher risk of perineural invasion (PNI) (p=0.042, OR=4.61, 95% CI 1.06 to 20.04) and worse differentiation (p=0.029, Kendall’s τ-b value=0.27). Patients with P63 and poor differentiation also showed expression of the downstream P63 target ZNF750 (figure 2).

Figure 2

Immunohistochemistry of p63 and ZNF750. Well-differentiated case presenting with negative p63 and ZNF750 expression. Poorly differentiated cases presenting with positive p63 and ZNF750 expression.

Logistic regression showed that patients with a worse T stage were more likely to be P63 positive (p=0.012, OR=1.24, 95% CI 1.05 to 1.46). P63 positivity was more common in patients with a T3c or worse stage than in patients with a lower T category (p=0.008, OR=3.77, 95% CI 1.42 to 9.97). Moreover, 63.6% (7/11) of stage T4a cases and 46.4% (13/28) of stage T4b cases were P63 positive.

P63 predicts a worse prognosis of metastasis and death due to disease. Of the 19 patients with overall metastasis, nine tested positive for P63. Additionally, P63 was significantly associated with a higher risk of overall metastasis (p=0.009, HR=3.99, 95% CI 1.42 to 11.22). We further probed the correlation of P63 positivity and death due to disease. Seventy-five per cent (6/8) of patients who died of disease tested P63 positive, and P63 positivity was associated with a higher risk of death due to disease (p=0.012, HR=7.80, 95% CI 1.56 to 38.90) (figure 3).

Figure 3

(A) Kaplan-Meier plot of metastasis stratified by p63 expression. (B) Kaplan-Meier plot of metastasis stratified by the coexpression of p63 and Ki67. (C) Kaplan-Meier plot of disease-specific survival stratified by p63 expression. (D) Kaplan-Meier plot of disease-specific survival stratified by the coexpression of p63 and Ki67.

P63 and Ki67 are codependent and validate a poor prognosis

P63 was reported positive in one among 50 patients with low Ki67 expression (2%), five out of five patients with intermediate Ki67 expression (100%) and 23 among 30 patients with high Ki67 expression (76.7%). There is a significant correlation between higher Ki67 expression and P63 positivity (p<0.001, Kendall’s τ-b value=0.757).

Expression of Ki67 alone was not associated with death of disease (p=0.922, 22.2% in low Ki67 group, 40% in intermediate group and 25% in high group, respectively) or metastasis (p=0.305, 17.7% in low Ki67 group, 20.0% in intermediate group and 35.7% in high group, respectively). Multivariant logistic regression showed that patients with a worse T stage were associated with worse differentiation (p=0.002, OR=5.88, 95% CI 1.90 to 18.18), but not Ki67 (p=0.546) or PNI (p=0.999).

Cases that coexpressed Ki67 (intermediate and high) and P63 were more likely to be staged T3c or worse (p=0.013, OR=3.44, 95% CI 1.30 to 9.11). Coexpression of Ki67 and P63 was associated with a higher risk of overall metastasis (p=0.013, HR=3.86, 95% CI 1.33 to 11.25, figure 3). Furthermore, coexpression of Ki67 and P63 showed greater value in predicting death due to disease. Thus, patients with coexpression of Ki67 and P63 had a nine times higher risk of death due to disease, and P63 positivity only predicted a seven times higher risk of death due to disease (p=0.007, HR=9.21, 95% CI 1.82 to 38.90 vs p=0.012, HR=7.80, 95% CI 1.56 to 38.90 figure 3). Among eight patients who died of disease, six had coexpression of P63 and Ki67; in seventy patients who survived, only two had coexpression of P63 and Ki67.

Discussion

This study of 85 patients suggested that the expression of P63 and coexpression of P63 and Ki67 were both associated with a worse T stage and worse differentiation. Their potential to predict a worse prognosis including metastasis and death due to disease was also identified. Although half of this cohort presented with T4 stage, the time from diagnosis to nodal metastasis was longer than that reported previously, suggesting a more aggressive orbital invasion pattern.

P63 plays an essential role in the differentiation and homeostasis of squamous epithelium.22 23 Other studies have revealed higher and more diffused expression of P63 in both cSCC17 and HNSCC.18 To the best of our knowledge, no previous study revealed high P63 expression in the eyelid and periocular SCC. In our cohort, coexpression of P63 and Ki67 was detected in one-third of patients, particularly those with advanced malignancy, adding to the literature of P63 expression in eyelid and periocular SCC.

Ki67 was widely recognised as a marker of proliferation, yet the cut-off values of Ki67 varied in literature.21 Previous study reported correlation of Ki67 expression and proliferative clinicopathological features in eyelid cancers.24 However, the specific alarm value for worse prognosis was lacking. We tested different cut-off values of 10%, 15% and 30%, and discovered that coexpression of Ki67 (dichotomised at 10%) and P63 was more valuable for predicting worse prognosis. Accordantly, Ki67 value higher than 10% predicted higher risk of recurrence in other cancer.25

Regarding the prognosis, P63 alone and coexpression of P63 and Ki67 predicted higher metastasis and worse survival. In a study of HNSCC, patients with increased P63 had poorer survival rates than the group with reduced p63 expression (p<0.05), a finding that agrees with ours.18 In another study of cSCC, amplification of TP63 was identified in 7 (24.1%) patients with metastatic cSCC.26 However, we found that 47.4% (9/19) of metastatic patients tested positive for P63 in our cohort, a result that was one time higher than that reported in cSCC. This difference may be due to the location of the tumour and complex histology of the periocular component. Therefore, P63 may play a key role in the progression of eyelid and periocular SCC. Notably, one of its known downstream target, the epidermal differentiation factor ZNF750,19 also showed an increased trend in poorly differentiated samples in our cohort. Thus, P63, bound to ZNF750, represents a potential therapeutic pathway for this malignancy.

In our cohort, patients presented with more aggressive orbital invasion pattern. Nearly half of the patients were staged T4, and 40% of this cohort showed orbital invasion, none of whom had pathological evidence of PNI though. The median time from the presentation of the ocular sign to diagnosis in T4 cases was 9.5 months (range 0.6–121.0 months). Among these patients, four developed nodal metastasis within a median period of 32.3 months. Moreover, in five patients who developed metastasis, three were stage T4 and the median time from diagnosis to metastasis was 25.48 months. Kaliki et al from India reported 29% of T4 eyelid SCC, but they did not report the metastasis duration.6 The worst T stage from an earlier study by Sun et al including 254 patients was only 2.4% (6 patients in total) with T3b as the worst category (AJCC seventh edition) and one patient died of disease.13 A recent study from a large cancer centre by Esmaeli et al reported that 34.8% of patients were diagnosed with T4 SCC and 6.4% of patients developed nodal metastasis during follow-up with a median time to nodal metastasis of 5.2 months.10 Compared with the results of this large cancer centre, we have a higher proportion of patients staged T4 at presentation (45.9%), whereas the median time to nodal metastasis was much longer (32.3 months). In summary, the current series revealed more locally aggressive behaviour but a relatively lower risk of regional metastasis of eyelid and periocular SCC in Chinese patients.

According to our findings, patients with a T4 or worse category had a ninefold higher risk of disease-specific death than patients with a better T category. In contrast to the previous findings of Esmaeli et al,10 no significant associations were found between DSS and T2 or T3 stage or better T category. The explanations could be due to the relatively smaller proportion of T2 and T3 involvement in our cohort or that eyelid SCC of Chinese patients tends to behave more aggressively locally than systematically.

Our findings provided a similar nodal metastasis rate (16.8%) as reported previously in the USA (10.1%)10 and Australia (1.0%–21.0%).2 Our overall distant metastasis rate of 7.1% was higher than that reported in America (2.8%),10 India (4.0%)6 and Australia (0.8%).2 As mentioned previously, among six patients with distant metastasis, four were diagnosed with T4 stage. We believe that the higher rate of distant metastasis is likely due to the higher percentage of the T4 category in our study and our extensive follow-up. In our study, the 2-year and 5-year DSS rates were 95.3% and 86.4%, respectively. Similarly, the 2-year and 5-year DSS rates were 92.6% and 87.7% in the USA in 2019,10 and the 5-year DSS rates were 86% for the same centre in 2008,27 96% in India6 in 2019 and 98.4% in Australia in 2015.13

The limitations of our study included its retrospective nature and the relatively small numbers of cases in a single centre and that 8.2% of patients were lost to follow-up.

Conclusion

This retrospective study on eyelid and periocular SCCs comprised a fairly large cohort with a medium follow-up of 29.1 months. The patients exhibited a high percentage of T4 involvement and a high distant metastasis rate. The major findings were that patients with coexpression of Ki67 and P63 had a higher risk of a worse stage and prognosis. Our study suggests that more strict surveillance should be considered in patients with the coexpression of P63 and Ki67.

Data availability statement

Data are available on reasonable request. The datasets used and analysed during the current study are available from the corresponding authors on reasonable request.

Ethics statements

Patient consent for publication

Acknowledgments

We would like to thank all participants in this study for their valuable contribution. We would like to thank the pathological department for pathological diagnosis of the patients.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • YL and HC contributed equally.

  • RJ, SX and YW contributed equally.

  • Contributors YW, SX and XF designed this study. RJ and YW provided the data of patients. SX and JY did literature search. SG and YL designed the figures, did data collection and drafted the manuscript. SX and YL did data analysis and interpretation. YR provided pathological reports. RJ did proofread of the manuscript. YW, SX and RJ are the cocorresponding authors. All authors approved this manuscript.

  • Funding This work was supported by National Natural Science Foundation of China (grant 81802702), Shanghai Municipal Science and Technology Major Project(17JC1420100), Science and Technology Commission of Shanghai (grant 17DZ2260100), Innovative research team of high-level local universities in Shanghai(SSMU-ZDCX20180401).

  • Disclaimer The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Linked Articles

  • Commentary
    Bita Esmaeli
  • Highlights from this issue
    Frank Larkin