Neoadjuvant Chemotherapy in Non-Metastatic Eyelid Sebaceous Gland Carcinoma: A Report in 10 Cases

Abstract

Introduction

Eyelid sebaceous gland carcinoma (SGC) is an aggressive malignancy. Surgical excision is the standard treatment for non-metastatic eyelid SGC. This study aimed to evaluate treatment outcomes with use of neoadjuvant chemotherapy (NACT) and any change in ease/difficulty of surgical treatment in such cases.

Methods

This was a prospective interventional study conducted over 24-month period. Histopathologically, confirmed cases without any systemic metastasis were included. Clinico-demographic details were collected for 30 patients. 10 patients received NACT using cisplatin and 5-FU. Tumour response was evaluated using RECIST criteria. An ease of surgery questionnaire was used to assess difficulty of surgical treatment before and after NACT.

Results

Of the 30 patients evaluated for recruitment, 37% had recurrent SGC and 72% had advanced tumour stage. Ten patients were recruited for NACT. There was partial response in 80% and complete response in 10% cases. Tumour T category downstaging was seen in 50% of cases. While tumour dimensions/volume reduced substantially, surgical ease improved in only 40% cases.

Conclusion

A significant proportion of SGC patients evaluated in our study presented with recurrent and/or advanced disease. NACT caused tumour regression in 90% of cases. However, surgical ease improvement was limited, pointing to a need for surgical modification in such cases. Corneal ulceration was noted in 2 cases with large tumours causing a complete mechanical ptosis. Overall, the study introduced an ease of surgery questionnaire and provided insights into benefits and challenges of using NACT for eyelid SGC management.

Introduction

Eyelid sebaceous gland carcinoma (SGC) is a highly malignant and potentially lethal tumour. While it accounts for 1–5.5% of all eyelid malignancies in the Western population [1, 2], it is the most common malignant eyelid lesion contributing to 37% of all eyelid malignancies in the subcontinent. Tumour recurrence after surgical excision is a major problem in eyelid SGC. Shields et al., in a previous study, reported 18% recurrence at 5 years. Eyelid reconstruction after a repeat surgical excision can be complicated. Studies show that patients have delayed presentation and advanced eyelid tumours in the low- and middle-income countries [3, 4]. This further makes surgical excision and eyelid reconstruction challenging in these regions.

Use of systemic chemotherapy for eyelid SGC is sparsely described in the literature. Metastatic SGC is routinely treated with cisplatin/carboplatin and 5-fluorouracil (5-FU)-based chemotherapy protocols [5–7]. Recently, role of systemic chemotherapy has been explored in a neoadjuvant setting for treatment of locally advanced large eyelid SGC [8–10]. Herein, we describe treatment outcome with use of NACT in 10 surgically challenging cases of non-metastatic eyelid SGC with an intent to achieve complete tumour excision and improve ease of surgery.

Methodology

This was a prospective interventional study conducted over 24-month period (June 2020 to June 2022) at a tertiary care eye centre. The Institute Ethics Committee approval was obtained for the study (IECPG-586/October 21, 2020). All patients presenting with eyelid SGC and confirmed on histopathology were evaluated for inclusion. All cases of SGC with tumour stage greater than/equal to T2b (AJCC 8th edition) or any tumour stage T with N1/Mo or recurrent SGC (previously excised) were included in the study. Those patients not consenting/not fit to receive NACT or with haematogenous metastasis (any M1) or those previously treated with chemotherapy were excluded.

Clinical and Demographic Details

Clinico-demographic data were collected including detailed clinical history, age, laterality, risk factors, and previous treatment details. Systemic and ocular examination parameters including visual acuity, IOP, eyelid and adnexal involvement, and anterior and posterior segment findings on slit lamp were noted. Tumour dimensions were noted along vertical and horizontal eyelid axis. Tumour was classified according to 8th AJCC TNM classification. Patients with tumour category greater than/equal to T2b were evaluated for suitability of NACT. Contrast-enhanced computed tomographic scan of orbit was performed to evaluate orbital extent of disease in cases of tumours reaching fornix/canthus. USG neck, chest X-ray, L.F.T & USG abdomen was done to rule out metastasis. A baseline ease of surgery questionnaire was filled by the evaluating surgeon, and surgical difficulty category was noted (Table 1).

Table 1.

Ease of surgery questionnaire

Q1. Subjective assessment by operating surgeon
Easy	1
Difficult	2
Very difficult	3
Not possible	4
Q2. History of previous surgery
Incisional biopsy	1
Excisional biopsy with direct closure	2
Excisional biopsy with Tenzel’s procedure	3
Excisional biopsy with lid-sharing procedure	4
Q3. Plan for reconstruction
Only advancement	1
Flaps (rotational)	2
Lid-sharing procedure	3
Healing by secondary intention where lid reconstruction is not possible/exenteration	4
Q4. Amount of lid tissue likely to be sacrificed
<1/3	1
1/3–1/2	2
>1/2–2/3	3
>2/3	4
Q5. Reliability of frozen section*
Reliable on all three sides	1
Only 2 sides	2
Only 1 side	3
Not reliable	4
Overall ease of surgery assessment
Easy	<8
Difficult	8–12
Very difficult	≥12

*A reliable frozen section is possible only when one is able to harvest full thickness eyelid tissue from any side. Reliability decreases once tumour resection margin extends beyond the tarsal plate.

Chemotherapy Protocol

Three cycles of NACT were administered using cisplatin (80 mg/m² on day 1) and 5-FU (1,000 mg/m² daily infusion over 6 h) (days 1–4) at 3-week intervals.

Response Evaluation

Tumour response was assessed in terms of percentage reduction of tumour size, which was measured as sum of horizontal and vertical dimensions measured with a scale. In cases where pre-NACT and post-NACT imaging were available, percentage reduction in tumour volume was calculated instead. Tumour response was classified according to RECIST criteria [11]. Tumour category (T) and ease of surgery score were reassessed post 3 cycles of NACT, and any change was noted.

Surgery and Follow-Up

For patients showing a partial response (PR), standard surgical excision was done with frozen section control of margins with eyelid reconstruction. Histopathology evaluation of excised tumour was undertaken. EBRT was advised in patients that had minimal residual orbital disease as suggested by positive resection margins. All patients were followed up 3 monthly for recurrences. The toxicity profile of NACT was studied using Common Terminology Criteria for Adverse Events (Version 4.0) given by the National Cancer Institute.

Statistical Analysis

The data were analysed using SPSS v.24 software. Descriptive data were noted as mean ± standard error. Correlation of tumour response to NACT with was done with different variables using Paired t test for non-categorical and Fisher exact test for categorical data.

Results

Thirty patients of eyelid SGC were treated at our centre during the study period (24 months). The mean age of the cohort was 56.13 ± 11.74 (range: 35–76 years). The mean duration of symptoms was 12.43 ± 11.66 months. Eleven of the 30 patients (37%) presented with recurrent tumour. Of the 30 cases, 8 (27%) had T1, 8 (27%) had T2, 8 (27%) had T3, and 6 (20%) had T4 tumour category according to the 8th AJCC TNM classification. Lymph node (LN) metastasis was noted in 3/30 (10%) cases. No systemic metastasis was seen in any case. Out of 22 cases eligible for recruitment, only 10 (45%) patients consented for NACT and were recruited for the study.

Clinico-Demographic Data (10 Patients)

The mean age of patients recruited for the study was 52.7 years. Of 10 patients, 6 (60%) presented for the first time, while 4/10 (40%) had recurrent SGC. The mean duration of complaint was 22.3 months. Upper eyelid was involved in 6/10 (60%), medial canthus in 3/10 (30%), bulbar conjunctiva in 1/10 (10%), and orbital extension was noted in 5/10 (50%) cases. Histopathological examination of incision biopsy showed features of moderately differentiated tumour in 50%, poorly differentiated tumour in 30%, and well-differentiated tumour in 20% cases. One case showed a collagenous tumour stroma with infiltrative growth pattern, and another one showed comedo pattern Table 2.

Table 2.

Clinical details, ease of surgery score, and response to neoadjuvant systemic chemotherapy in 10 patients of SGC

S. No.	Age/sex	Lat./eyelid	Primary/recurrent	Previous surgery	AJCC		% reduction/RECIST	Surgery planned		EOS score/category		TRT outcome and follow-up, months
								before after NACT		before after NACT
					before after NACT
1	40/F	RU	Recurrent	Tenzel’s flap	T2CN0M0	T1c N0M0	75/PR	Lid-sharing procedure	Direct closure	10/7	2/1	Alive (77)
2	44/M	RU	Primary	–	T3b N0M0	T2b N0M0	54.16/PR	Lid-sharing procedure	Hugh’s procedure	13/9	3/2	Alive (33)
3	42/M	LU	Recurrent	Excision with direct closure	T3c N0M0	T3c N0M0	88.7*/PR	Lid-sharing procedure	Cutler beard^	16/12	3/3	Alive (35)
4	70/M	LL	Primary	–	T4b N1M0	T4b N0M0	11.76/SD	Exenteration	Exenteration	17/16	3/3	##Died of disease (27)
5	65/M	RU	Primary	–	T4a N0M0	T3c N0M0	53.12/PR	Exenteration	Cutler beard	17/12	3/3	Lost to follow-up (26)
6	49/M	RL	Recurrent	Tenzel’s flap	T3c N0M0	T1c N0M0	CR	Lid-sharing procedure	Hugh’s procedure	17/12	3/3	Alive (32)
7	55/F	RL	Recurrent	Tenzel’s flap	T4a N0M0	T4a N0M0	>60* PR	Exenteration	Exenteration	20/20	3/3	Alive (31)
8	45/F	RU	Primary	–	T4a N0M0	T4a N0M0	>90%*/PR	Exenteration	Reverse cutler beard	17/11	3/2	Alive (23)
9	60/F	L	Primary	–	T3c N0M0	T2c N0M0	36.84/PR@	Lid-sharing procedure	Direct closure	12/9	3/2	Died of unrelated cause (22)
10	57/F	RU	Recurrent	Orbitotomy with tumour debulking	T4a N0M0	T4a N0M0	>90%*, PR#	Exenteration	Exenteration	17/17	3/3	Alive (23)

*On imaging.

**All lid tissue that appeared to have post-tumour regression changes was removed.

^$Orbital tumour disappeared completely, eyelid thickening reduced but persisted.

^#Thickening of the upper eyelid persisted.

^@Lid lengthening due to weight of the tumour allowed direct closure despite partial regression,

^#Exenteration done in view of pathological confirmation of residual orbital tumour.

^{^}More than 75% of length involved despite PR.

^##Denied radicle LN dissection and EBRT.

Baseline Tumour Details

Five out of 10 cases (50%) were T4 tumours (4-T4a, 1-T4b with extension into paranasal sinuses), 3/10 (30%) were T3c, and 1 (10%) each were T3b and T2c category. Clinically, pagetoid spread was noted in 1 case (case No. 10). Mean tumour height (vertical dimension) was 27.70 ± 22.11 mm (range: 5–78 mm) and mean tumour length (horizontal dimension) was 26.80 ± 11.18 mm (range: 10–45 mm). Mean ease of surgery score for the cohort was 15.6 (range: 10–20). The ease of surgery was graded as difficult in 1 (10%) and very difficult in 9 (90%). One patient had biopsy-proven LN metastasis to ipsilateral pre-auricular LN (case No. 4).

Tumour Response Evaluation and Ease of Surgery Assessment

All cases showed a decrease in tumour dimensions. 1 patient with LN involvement did not show any significant response in the LN metastasis. After 3 cycles of NACT, the mean tumour height reduced by 8.5 mm (30%) to 19.2 ± 12.24 mm and the mean tumour length reduced by 11.9 mm (44.4%) to 14.90 ± 10.50 mm. Case-wise percentage reduction in tumour dimensions/tumour volume are presented in Table 2.

Eight out of 10 (80%) patients showed P, 1 patient showed stable disease, and 1 showed complete response (CR) according to the RECIST criteria. There was a downstaging of tumour category in 5/10 cases. The mean percentage reduction in tumour dimensions for tumour not involving orbit was 62.67% (range 36.84–100%). The mean percentage volume reduction for cases evaluated on CT scan was 66.10% (range 11.76–91%). The mean ease of surgery score improved by 3–12.5, but improvement in ease of surgery category was noted in only 4/10 cases.

Treatment Outcome and Follow-Up

Eight (80%) patients received three, while 2 (20%) patients received four cycles of NACT as surgery was deferred owing to ongoing COVID-19 pandemic. patient with pagetoid spread was treated with topical chemotherapy with 0.04% mitomycin c in addition to NACT. There was complete resolution of pagetoid spread following 3 cycles of topical treatment. All 10 cases underwent surgical excision. 2/10 were reconstructed with direct closure, 5/10 required eyelid sharing procedures, while 3 were advised exenteration (1 refused surgery, 2 underwent exenteration). Radicle LN dissection was advised in patient with LN involvement. In 2 out of 5 (40%) cases that were planned for exenteration pre-NACT, globe salvage was possible after NACT. Two patients were advised adjuvant EBRT of which one received EBRT (Case No. 10), 1 was lost to follow-up (case No. 4). The mean follow-up was 32.9 months (range 22–77 months). Two patients died at 22 and 27 months of follow-up. 1 patient was lost to follow-up after completing treatment. At last follow-up, 8 patients were alive and free of disease.

Histopathological examination of excised tumour showed residual tumour in all cases. However, tumour islands significantly decreased in size and became well defined. All cases showed an increase in inflammatory cells and macrophages. Tumour giant cells were also noted in one case (Fig. 1c, d).

Fig. 1.

Contrast-enhanced computed tomographic (CECT) scan of orbit (sagittal cut, case 8) showing large eyelid tumour with orbital extension (a), post 3 cycles NAC showing >90% resolution with some residual tumour in eyelids and orbit (b); however, on HPE residual, tumour was found only in eyelid and not in orbit. c Microphotograph showing (H and E stain. ×200) small well-defined residual tumour island (arrow) after NAC. d (H and E stain. ×400) Inflammatory cell infiltration with lymphocytes and macrophages (arrows).

Side Effects of NACT

The most common adverse events noted were gastrointestinal disturbances like diarrhoea (54.5% patients), dyspepsia (45.5% patients), gastritis, and vomiting (27.3% each). Alopecia was complained of by 36.4% of patients. One patient (9.1%) developed internal jugular vein thrombosis. The patient was started on subcutaneous injection dalteparin (40,000 IU per week) followed by oral anticoagulants. The thrombus resolved at 5 months of treatment.

Sterile corneal melt developed in 2/10 patients after receiving second-cycle NACT. One patient was managed with cyanoacrylate glue and bandage contact lens, while other required evisceration.

Correlation of Response to NACT with Variables

Response to NACT was correlated with baseline tumour category and differentiation on histopathology, age, sex, and location of eyelid tumour. Tumour categories T2 and T3 had significantly better response to NACT as compared to T4 (p value = 0.0092). Also, younger patients (less than 50 years) had significantly better response to NACT than older patients (p value = 0.027).

Discussion

There are few case reports and case series in the literature that report good debulking of eyelid SGC with NACT. Koyama et al. [5] reported use of NACT (doxorubicin and cisplatin) in a case with eyelid SGC and reported PR and progression-free survival at a follow-up of 7 months. In another case report, Joshi et al. [6] reported CR and progression-free survival of 6 months after NACT with carboplatin and paclitaxel. Similarly, Hwa Jung et al. [7] and Orcurto et al. [12] administered cisplatin and 5-FU to achieve CR in 2 cases with eyelid SGC. In a retrospective study of 10 patients with advanced eyelid SGC with systemic metastasis in 2 cases and LN metastasis in 6, the authors reported a remarkable tumour shrinkage allowing globe salvage in 5/9 cases planned for exenteration [8]. In another retrospective study including 8 patients of eyelid SGC, authors reported a mean reduction of 71% in basal tumour dimensions with NACT. They reported CR in 2/8 and PR in 6/8 patients, thus allowing conservative surgery in 1 and eyelid sparing exenteration in 3/4 patients with T4 disease. NACT achieved CR of regional LN metastasis, thus sparing the patients from radical LN dissection in the same study [9]. In the current study, there was 30% decrease in height, 44.4% in horizontal dimension, and 66% reduction in tumour volume with NACT.

To the best of our knowledge, this is the first study to evaluate change in ease of surgery after NACT in eyelid SGC. Although the RECIST criteria gives a good measure of the response to chemotherapy, an improvement in ease of surgery measure can provide a more objective measure of tumour response for surgeons as it takes into account the likely surgery required for reconstruction.

We could not find any published questionnaires to evaluate ease of surgery. A questionnaire was therefore devised specifically for the study based on the subjective assessment of difficulty level by the surgeon and objective measures including the type of previous surgery, amount of eyelid tissue likely to be excised, surgical reconstruction technique being planned, and reliability of the frozen section. Since the same surgeon fills the questionnaire before and after NACT, the surgeon specific variables like experience and skill cancel out. We know that reliability of frozen section diminishes for larger tumours where excision margins extend beyond the tarsal plate. Here, eyelid lamellae become separate making full thickness eyelid tissue sampling difficult; also, resection margins become considerably large for complete evaluation. Therefore, surgeons often take representative samples from these margins.

Among 30 cases of SGC that presented at our centre during study period, 37% presented with recurrent tumour, and (22/30) 73% were stage T2b-T4. Thus, a significant percentage of patients presented with surgically challenging tumours that could potentially benefit from use of NACT. However, only 10 (45%) patients consented for NACT. Of these 8/10 (80%) patients showed PR, 1 patient showed SD, and 1 showed CR (Fig. 2a, b) according to the RECIST criteria.

Fig. 2.

Clinical picture of case 6 showing recurrent SGC in lower eyelid previously treated with excision and Tenzel’s flap (arrow: showing the area of Tenzel’s flap) (a), same case post 3 cycles NAC showing CR (b), case 7 with recurrent orbital SGC (arrow showing area of previous eyelid tumour excision and reconstruction) (c), same case showing complete clinical regression post 3 cycles of NAC (d); however, on palpation, there was residual tumour in upper eyelid, case 10 showing diffuse eyelid thickening and proptosis, with palpable mass in lateral upper eyelid (e), same case post 3 cycles of NAC showing complete resolution of proptosis and orbital mass (f); however, on palpation, there is residual eyelid thickening, case 8 showing large eyelid mass causing complete mechanical ptosis (g), same case post 3 cycles NAC showing almost complete resolution of eyelid mass (h); however, some eyelid thickening and complete ptosis persisted.

Three patients that appeared to have complete clinical resolution on inspection had induration of eyelid tissue on palpation (Fig. 2c, e). Imaging showed contrast-enhancing soft tissue lesion and histopathological examination confirmed residual tumour in all 3 cases. We therefore emphasise a careful examination of patients with a visibly complete tumour regression.

Although there was a remarkable tumour shrinkage with NACT, downstaging of the tumour T category was noted in only 5/10 (50%) cases. The mean ease of surgery score improved with NACT, but ease of surgery category improved in only 4/10 patients. This can be explained by the observation that though tumour volume reduced drastically in most cases but the reduction in the length of eyelid involvement, which primarily determines the choice of eyelid reconstruction technique, did not change significantly. Also, we noted that after NACT, the eyelid may still show residual changes in the initial areas of tumour involvement that may be free on histopathology. Hence, a change in technique of surgical excision for residual tumour after NACT may be useful. The excision margins may be made flush with residual tumour edge, thus allowing salvage of valuable eyelid tissue.

In tumours with T4 category, residual orbital tumour on post-NACT imaging needs to be confirmed on histopathological examination. One patient (case No. 8) did not show any orbital extension on histopathology despite a radiology suggestive of same (Fig. 1).

We analysed the factors that may affect the tumour response to NACT. A younger age at presentation (<50 years) and a lesser T category had statistically significant better tumour response (p = 0.027, p = 0.009, respectively).

None of the patients in the current study developed systemic metastases till the last follow-up (mean 32.9 months, range 22–77 months). Previous studies recommend sentinel LN biopsy in SGC of tumour category more than T2b [13]; however, it is not available at most centres in low- and middle-income countries. It is likely that NACT may have treated any micro-metastasis in the study patients. Previous literature suggests a good response of regional LN metastasis to NACT [14].

We noted corneal ulceration in 2 patients after NACT. Although keratitis is known to occur with 5-FU-based chemotherapy, no previous report of such severe corneal affection is available in the literature. Both the patients developed corneal melt after NACT and before any surgery. Therefore, suture and lagophthalmos-related causes for corneal ulceration were ruled out. Investigations were also performed to rule out infectious causes or systemic inflammatory disease. Both cases had large upper eyelid masses causing severe mechanical ptosis. It is possible that in addition to effect of 5-FU, release of inflammatory mediators from the necrosing eyelid tumour and stagnation of same on the ocular surface due to absence of eyelid blink may have led to sterile corneal melts in these cases. Use of steroids and copious lubricants may therefore be considered in such cases.

To conclude, although there was a good response of non-metastatic eyelid SCGs to NACT in 90% cases, the ease of surgery improved only in 40% cases. This may point to a need to change the surgical technique in post-NACT cases. Corneal ulceration may be an important side effect of NACT, especially in large upper eyelid SGC causing severe ptosis. Lastly, this study introduces an “ease of surgery” questionnaire, which may be evaluated and modified in future studies.

Statement of Ethics

This study protocol was reviewed and approved by the Institute Ethics Committee for Postgraduate Research, All India Institute of Medical Sciences, New Delhi, approval number IECPG-586/21.10.202. Written informed consent was obtained from all participants to participate in the study and for publishing article and images.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

The authors have no funding source to declare.

Author Contributions

(ICMJE criteria for authorship) Conception and design: Rachna Meel, Sameer Bakhshi, Nipun Kumar, and Deepam Pushpam; acquisition of data: Rachna Meel, Nipun Kumar, Seema Kashyap, Seema Sen, and Sanjay Sharma; analysis and interpretation of data: Rachna Meel, Nipun Kumar, Seema Kashyap, and Seema Sen; drafting of the manuscript: Rachna Meel and Nipun Kumar; critical revision of the manuscript for important intellectual content: Rachna Meel; statistical expertise: Rachna Meel and Nipun Kumar; obtaining funding: N/A; administrative, technical, or material support: Rachna Meel, Sameer Bakhshi, Neelam Pushker, Mandeep Singh Bajaj, Seema Kashyap, Seema Sen, Sushmita Pathy, Sanjay Sharma, and Deepam Pushpam; supervision: Neelam Pushkar, Mandeep Singh Bajaj, and Sushmita Pathy; and final approval of the version to be published: Rachna Meel.

Funding Statement

The authors have no funding source to declare.

Data Availability Statement

All data generated or analysed during this study are included in this article. Further inquiries can be directed to the corresponding author.