Abstract
Introduction: The aim of the study was to report our experience in the use of Gamma Knife Stereotactic Radiosurgery (GKSTRS) for conjunctival squamous cell carcinoma (SCC) invading the orbit, as an alternative to exenteration surgery. Patients and Methods: Patients who had GKSTRS for conjunctival SCC invading the orbit but sparing the bone (AJCC eighth ed. T4a) with a minimum of 1-year follow-up were included. Treatment failure was defined as no decrease in tumour size 3 months post-treatment, or further growth during the follow-up period. Patients were followed up 3-monthly for 2 years and 6-monthly afterwards with a minimum of yearly MR Imaging of orbit. Results: Six patients met the inclusion criteria. MR imaging was used to identify the extent of orbital involvement of SCC. Stereotactic radiosurgery utilizing the Leksell Gamma Knife® Perfexion was delivered in a single session in which patients received 18–20 Gy to the 45–50% isodose. The median follow-up was 29 months. Four patients responded to the treatment and had no evidence of recurrence at their most recent follow-up. The treatment failed in 2 patients, in 1 of whom the tumour was larger and extended deeper into the orbit. The other patient developed a recurrence away from the treated area at 9 months, suggesting a multifocal disease. Both patients had poorly differentiated SCC. Conclusion: GKSTRS is a potential alternative to orbital exenteration surgery for conjunctival SCC extending into the anterior orbit in a selected group of patients refusing orbital exenteration or has a second blind eye. Nevertheless, the disease is more likely to recur than with exenteration surgery, hence lifelong monitoring and low threshold for surgical intervention or retreatment is prudent.
Introduction
Ocular surface squamous neoplasia is a spectrum of premalignant and malignant squamous epithelial lesions of the conjunctiva and the cornea. Invasive squamous cell carcinoma (SCC) lies at the severe end of the spectrum and is believed to have an incidence that ranges from 0.03 to 1.9 per 100,000 person/year in the USA and Australia [1-4]. Conjunctival SCC has a fair to good prognosis with a little tendency to metastasize (1–8%) [5-7] and a very low mortality rate but high risk of local recurrence (30–40%) [5, 8-10]. However, it can lead to orbital invasion and eventually brain invasion, through perineural spread, and subsequently death if left untreated [11, 12].
Treatment of conjunctival SCC includes a combination of surgical resection, cryotherapy, and topical therapies such as mitomycin C, 5-fluorouracil, and interferon alpha 2b. Once it invades the orbit, exenteration surgery is often the only option for effective local disease control. The purpose of this study was to examine the outcomes of stereotactic radiosurgery as an alternative to exenteration surgery in patients with conjunctival SCC extending into adjacent orbital structures.
Patients and Methods
We retrospectively reviewed the medical records of patients who had undergone Gamma Knife Stereotactic Radiosurgery (GKSTRS) for histopathologically proven conjunctival SCC invading adjacent orbital and periorbital structures (AJCC stage T4a N0 M0) at Sheffield Teaching Hospitals [13]. Patients with bone invasion, intracranial invasion, and lymph node and distant metastasis were excluded. Patients who had a history of SCC elsewhere were excluded when it was not possible to conclude whether the conjunctival disease was primary or metastatic. A minimum follow-up of 1 year was required. Patients who had prior debulking of the lesion, cryotherapy, topical mitomycin C, and intralesional interferon alpha 2b treatment to the conjunctival SCC were included in this study provided subsequent biopsy showed residual disease or recurrence.
The study adhered to the tenets of the declaration of Helsinki. The original ocular oncology and oculoplastics multidisciplinary team meeting recommendation was orbital exenteration surgery. Stereotactic radiosurgery was considered after each patient expressed their reluctance to undergo orbital exenteration. Prior to offering the treatment to the patient, each case was discussed again in both the ocular oncology and the stereotactic radiosurgery multidisciplinary team meetings and a management plan was agreed between the treating ophthalmologist and neurosurgeon. All the patients had a lengthy counselling session and were aware of the lack of evidence for the use stereotactic radiosurgery for conjunctival SCC. The risks and benefits of both stereotactic radiosurgery and conventional surgery were discussed with them. The patients were also warned that conventional surgery (i.e., orbital exenteration) might still be needed in the event of a treatment failure.
The patients were followed on a 3-monthly basis during the first 2 years and on a 6-monthly basis thereafter. They had a baseline post-treatment MR imaging of the orbit at 3 months followed by orbit and neck MR imaging at least once per year to look for disease recurrence and metastatic spread to lymph nodes. Treatment failure was defined as an increase in tumour size on orbital MR imaging or progression into a worse AJCC stage.
Stereotactic Radiosurgery Technique
The eye is immobilized with a retrobulbar injection of anaesthetic comprised of lidocaine 2% with adrenaline (1:200,000), bupivacaine 0.5% with adrenaline (1:80,000), and hyaluronidase. The stereotactic Leksell G Frame (Elekta) is attached to the outer table of the skull after subcutaneous local anaesthetic injections and a contrast enhanced stereotactic MRI scan is obtained. The Leksell Gamma Knife® Perfexion contains 192 Cobalt-60 sources distributed over a hemispherical source core, so that each beam of gamma radiation emitted is directed towards a common geometric focus in the centre of the hemisphere. Finely collimated beams of radiation are used to achieve a very high spatial accuracy; they are selectively deactivated to conform to the irregular 3-dimensional proportions of the tumour.
The treatment was planned so that at least 95% coverage of the tumour volume was in the prescription isodose line of 45–50%. All cases were treated with a prescription dose of 18–20 Gy. At the end of the procedure, the frame was removed, and the patient was discharged.
Results
Seven patients who received stereotactic radiosurgery for SCC of the conjunctiva between 2014 and 2018 were identified. One patient of the 7 had a history of lung SCC and because it was not possible to conclude whether the conjunctival disease is a primary or a metastatic tumour he was excluded. The age at the time of presentation for the remaining 6 patients ranged from 46 to 78 years (median 70.5). In addition to the bulbar conjunctiva, the forniceal conjunctiva was involved in all 6 patients whereas the caruncle was involved in 2. Five patients underwent prior treatment (Table 1), including debulking surgery in patients 1 and 2. Invasive SCC was confirmed on histopathology in all 6 patients. The SCC was well-differentiated in 2, moderately differentiated in 2, and poorly differentiated in 2. None of the patients had evidence of head or neck lymphadenopathy, brain invasion, or distant metastasis at presentation.
Orbital MR Imaging confirmed the extension of the SCC into the orbit in all patients. In patient 1, the SCC extended into the superolateral aspect of the orbit and encroached upon the lacrimal gland and the insertion of the lateral rectus muscle (Fig. 1a). Patient 2 had a 12-millimetre residual mass inferomedially within the orbit contiguous with the insertions of the medial rectus and the inferior rectus muscles following debulking surgery (Fig. 1b). Similarly, patients 3 and 4 had a mass lesion in the inferomedial aspect of the orbit. The mass had an ill-defined edge laterally as it appeared to involve the insertion of the inferior rectus muscle in patient 3 (Fig. 1c), while in patient 4 it involved both the inferior and the medial recti and extended further to lie adjacent to the lacrimal sac (Fig. 1d). Patient 5 had the least aggressive disease clinically but the most extensive on MR Imaging, which revealed a 2-cm-thick mass that extended beyond the equator of the globe. It appeared to be inseparable from the sclera and was contiguous with lateral rectus muscle (Fig. 1e). MR imaging of patient number 6 revealed soft tissue mass in the region of the medial canthus extending anteromedially around the globe but did not extend beyond the orbital septum nor into the nasolacrimal duct (Fig. 1f).
A variety of factors contributed to the decision to treat with stereotactic radiosurgery. Patient 2 was offered stereotactic radiosurgery, because she had conjunctival SCC in the other eye as well, which involved the fornix and put her at risk to develop orbital invasion. Patient 3 had severe amblyopia of the other eye, which meant exenteration surgery would have rendered him blind. The other 4 patients were adamant that they do not want to lose their eye from exenteration surgery.
The details of treatment are shown in Table 2. The age at treatment ranged from 46 to 78 years (median 72). All patients received a prescription dose of 20 Gy to the 45–50% isodose (Fig. 2) except for patient 2, who received 18 Gy. The median follow-up time was 29.5 months (range 15–51). The treatment was successful in 4 out of the 6 patients (Fig. 3a–d). In patient 5, the treatment failed to control the disease as MR Imaging 3 months following treatment showed further enlargement of the tumour and level 2 neck lymphadenopathy (Fig. 3e). This patient underwent exenteration surgery and radical neck dissection soon after but 3 months later he developed brain invasion. He received adjuvant chemotherapy (cisplatin) and external beam radiotherapy (EBRT) to his brain and neck. The skin lining his socket broke down several times over the following months. He eventually passed away 1 year after being diagnosed with brain invasion. MR imaging of patient 6 3 months following treatment showed initial response, but a repeat scan 9 months following treatment showed a second focus of orbital invasion (Fig. 3f) for which she underwent intensity modulated radiotherapy (IMRT).
Apart from patient 5, none of the patients developed lymph node or distant metastasis. However, they all developed local complications (Table 3). Patient 1 and 3 developed cataracts for which they underwent cataract surgery. Patients 2, 3 and 4, all of whom had medially located disease, developed punctal stenosis that was asymptomatic and did not require any surgical intervention. In addition, patient 2 developed trichiasis that required electrolysis. Patient 4 had the most severe local complications as she developed limbal stem failure with resultant recurrent corneal erosions and eventually severe corneal scarring and a visual acuity of hand movements. Similarly, patient number 6 developed recurrent corneal erosions (Table 2), for which she received Botox® injection to her levator superioris muscle. Patients 1, 2, and 3 maintained normal or near normal visual acuity and had no diplopia.
Discussion
Our study showed GKSTRS to be an effective sight and/or globe sparing alternative to orbital exenteration surgery in treating conjunctival SCC invading the anterior orbit in 4 of our 6 patients in whom local resection was unlikely to yield cancer free margins. Complications of radiosurgery occurred but were often manageable. Nevertheless, it was not successful in treating poorly differentiated conjunctival SCC, extensive posteriorly located orbital invasion, and multifocal ill-defined lesions.
SCC of the orbit usually results from extension of SCC of adjacent structures. In 1 study, the paranasal sinuses were the most common source of invasion in 61 of 102 patients followed by eyelid and ocular adnexa in 23 [14]. In another smaller study, the conjunctiva was the most frequent source of invasion in 28 of 30 patients [12]. The same study also found that a delay in seeking medical attention and incomplete initial surgical excision were the 2 most important factors leading to orbital invasion by conjunctival SCC. Orbital invasion resulting from fornix disease appears to be less amenable to local resection as evidenced by all patients in our study having tumours that were too posterior to allow clear margins with local resection. This could be explained by the more posterior location of the forniceal conjunctival in comparison to the bulbar and tarsal conjunctiva and hence its intimate relation with important orbital structures such as the extraocular muscles and the lacrimal apparatus.
SCC of the conjunctiva results in regional and distant metastasis in 1–8% of cases [5-7], but once it invades the orbit it may extend into the cranial cavity. Thus, aggressive treatment is necessary to avoid deadly complications. Miller and colleagues reported the use of local excision supplemented with interferon alpha 2b when margins were positive to control the tumour. They had a recurrence rate of 52%, which they treated by further local excision [15]. However, local excision of the tumour is often difficult due to the infiltrative nature of the tumour and the difficulty in achieving free surgical margins, which renders orbital exenteration inevitable. Orbital exenteration is successful in preventing recurrences in 80% of cases and is generally considered the gold standard treatment when free margins with local resection cannot be achieved [12, 16, 17], but it entails significant morbidity including the loss of vision and disfigurement, both of which may have psychosocial implications. This becomes more problematic in patients with bilateral disease or 1 good eye such as patients 2 and 3 in our study, respectively. Furthermore, orbital exenteration typically requires general anaesthesia, for which patients with multiple comorbidities might not be fit.
SCC limited to the conjunctiva is well known to be responsive to treatment with ionizing radiation in the form of brachytherapy and proton beam radiotherapy [18-22]. The radiation dose needed to treat orbital SCC is on average 50 Gy [23, 24], which exceeds the tolerance of the eye and periocular structures and is likely to cause radiation related complications if delivered using EBRT. We opted to use GKSTRS due to its superior dosimetry and the steep dose gradient achievable, and its ability to treat tumours precisely with minimal damage to surrounding structures. The tissues just outside the tumour received a dose of less than 20 Gy, which is well tolerable by the eye. All the patients still developed local complications related to radiotherapy but, with the exception of patient 4, those were either easily manageable or did not require any intervention. The occurrence of some radiotherapy related complications was unavoidable due to certain ocular and periocular structures being diseased and hence falling within the treatment zone. In patient 4, limbal stem cell failure and conjunctivalization of the cornea might have resulted from a combination of the disease itself, previous treatment and ionizing radiation rather than from ionizing radiation alone as the treatment plan in Figure 2d shows that the majority of the limbal area did not receive more than 10 Gy. Even though patient 4 developed a blinding complication, the outcome is still better than orbital exenteration given she kept her eye and remained pain free.
One drawback to stereotactic radiosurgery is the high exit dose in comparison to other techniques such as proton beam radiotherapy, which theoretically could predispose to secondary tumours. Nevertheless, there was no increased risk of secondary brain tumours in patients who received stereotactic radiosurgery in a large study conducted by Rowe and colleagues [25]. Similarly, none of the patients in our study had developed a secondary tumour by their most recent follow-up visit. Proton beam radiotherapy is theoretically a viable alternative due to its Bragg peak, which allows precise localization of the tumour and minimizes irradiation of adjacent structures. In clinical practice; however, certain techniques are used to spread out the Bragg peak to encompass the tumour, which results in an increase in the entry dose and hence more irradiation to the ocular surface and ocular adnexa [26]. Another option is brachytherapy, which has been shown to be effective in treating a variety of orbital tumours and has the advantage of sparing the ocular surface of radiation unlike the aforementioned techniques [27], but it involves 2 surgical procedures to implant the radioactive material and remove it, requires a few days of hospitalization, and lacks the steep dose gradient achieved by stereotactic radiosurgery. Another treatment modality is intensity modulated radiotherapy (IMRT), which has been used to treat non-resectable SCC of the paranasal sinuses with a potential advantage over conventional EBRT [28]. Similar to stereotactic radiosurgery, it has a steep dose gradient. However, it requires multiple (fractionated) treatments unlike SRS in which the treatment is completed at 1 sitting.
A potential treatment approach could be surgical debulking followed by adjuvant stereotactic radiosurgery as evidenced by patients 1 and 2 in our study, both of whom did not develop a recurrence despite having the longest follow-up time of approximately 5 years. This approach further limits radiation related complications due to the smaller field of irradiation as evidenced by both patients developing very few complications none of which compromised their vision. This approach would still save the patient a general anaesthetic since the debulking could be performed under local anaesthesia, and the inconvenience and higher risk of complications of fractionated radiotherapy. Concurrent chemotherapy, particularly cisplatin, can be considered in order to achieve better results in patients with more extensive orbital involvement as it has been shown to improve cure rate and functional outcome in SCC of the head and neck region [29]. However, it can result in pancytopenia, ototoxicity, and peripheral neuropathy; hence, it is better avoided unless deemed extremely necessary.
There are a few limitations to our study. First, the number of patients was small, and they were different in terms of the severity of the tumour, its location within the orbit, and its differentiation. Hence, it is difficult to conclude at what extent the disease becomes too severe to respond to the treatment and whether or not different levels of differentiation respond differently. Also, the median follow-up is relatively short. Even though previous studies have shown that most recurrences following surgical intervention occur within the first 2 years following treatment, it is important to remember that late recurrences beyond the first 2 years are more likely with radiotherapy than with surgical excision [15]. Our study is also limited to conjunctival SCC and its results mat not necessarily apply on cutaneous SCC, for which a new less invasive treatment has emerged in the form of Cemiplimab, a recombinant human IgG monoclonal antibody known as a programed cell death 1 (PD-1) immune checkpoint inhibitor [30].
In conclusion, GKSTRS can be an alternative to orbital exenteration surgery in a selected group of patients with SCC of the conjunctiva invading the anterior orbit, who refuse orbital exenteration or has a second blind eye. It could also play a role in patients with multiple comorbidities who are not fit for general anaesthesia or for fractionated radiotherapy. Nevertheless, unlike with orbital exenteration the disease can recur and hence lifelong monitoring and low threshold for surgical intervention or retreatment is prudent. For better outcomes, surgical debulking prior to stereotactic radiosurgery should be considered. However, larger and comparative studies are needed to better understand the type of patients in whom it is more likely to be curative before it can replace orbital exenteration surgery, which remains the gold standard treatment when local resection with margins cannot be achieved.
Statement of Ethics
The study adhered to the tenets of the declaration of Helsinki. Prior to offering the treatment to the patient, each case was discussed in both the ocular oncology and the stereotactic radiosurgery multidisciplinary team meetings and a management plan was agreed between the treating ophthalmologist and neurosurgeon. All the patients had a lengthy counselling session and were aware of the lack of evidence for the use stereotactic radiosurgery for conjunctival SCC. The risks and benefits of both stereotactic radiosurgery and conventional surgery were discussed with them. The patients were also warned that conventional surgery (i.e., orbital exenteration) might still be needed in the event of a treatment failure. The patients provided a written consent to publish their cases. Ethic approval number: CEU 9456.
Conflict of Interest Statement
The authors have no conflict of interest to declare.
Funding Sources
This study was not funded by any financial body.
Author Contributions
Bashar M Bata designed the study, collected the data, wrote the manuscript and prepared the figures. Matthias WR Radatz wrote the section on the technique and parts of the discussion, prepared the treatment planning figures and revised the manuscript. Sachin Salvi contributed to the idea of the procedure and the paper, and revised the manuscript.