Purpose

Sebaceous gland carcinoma (SGC) is an uncommon disease representing 1.0-5.5% of eyelid malignancies globally [1]. It arises from the meibomian and Zeis glands that produce oil in the eyelid, and can present as a painless and yellowish lesion often seen at the upper eyelid [2]. SGC is an aggressive neoplasm, with a high propensity for nodal spread and local recurrence despite treatment. Since it mimics common benign conditions, diagnosis may be challenging, leading to delayed treatment. In the Philippines, SGC is as common as basal cell carcinoma, accounting for 30% of cases, with a higher incidence among females [3]. The most effective definitive treatment remains radical surgical excision. However, not all patients are suitable candidates for invasive surgery [4]. Since complete excision of eyelid tumors is difficult to achieve without causing significant functional and cosmetic impairment, radiotherapy may be a treatment option for these patients. A recent long-term retrospective study by Laskar et al. demonstrated the efficacy of high-dose-rate (HDR) brachytherapy for patients with eyelid malignancies. The majority of cases were SGCs, showing a 100% loco-regional control rate and excellent cosmetic outcomes [5]. Whereas for patients who refuse or are not suitable candidates for interstitial brachytherapy, plesiotherapy is a non-invasive treatment option [6].

In the current case report, HDR plesiotherapy was used for treating an eyelid SGC in an elderly patient ineligible for surgical treatment. Fabrication of a customized applicator, treatment planning, delivery, and long-term outcomes are presented, highlighting HDR plesiotherapy as a suitable alternative to surgery for superficial eyelid or facial carcinomas.

Material and methods

Case description

A 84-year-old female (ECOG 0) presented in March 2020 with an 8 mm slow-growing nodule on the left upper eyelid, unrelieved by topical antibiotic (Figure 1). Excision biopsy revealed SGC with palpebral conjunctival involvement and close posterior and inferior margins (< 1 mm). PET scan was negative for nodal and distant metastases, confirming stage IB (T2aN0M0) disease. History revealed a previous phacoemulsification of the right eye, pre-existing cataract on the left eye, and the use of anticoagulants for pacemaker-dependent bradyarrhythmia. Induration in the lateral canthus and the lateral third of the lower eyelid persisted post-biopsy. Given the patient’s cardiac risk, multidisciplinary team recommended non-invasive plesiotherapy prescribed at 14 × 3.5 Gy BID, according to the GEC-ESTRO guidelines and institutional protocol.

Applicator design

With the patient in supine, neutral neck position, a thermoplastic mask was prepared as per facial contours. Boundaries of the lesion were traced with black ink on the outer surface of the mask, and radiopaque markers were placed corresponding to the tracings, which allowed reproducible placement of the mask per fraction. On the outer side of the mask, four plastic needles were arranged equidistant to each other, encompassing 5-10 mm beyond the lesion, and secured using dental putty (Figure 2).

CT simulation and volume delineation

A plain CT scan was taken with the patient in supine, neutral neck position, and the customized mask was locked into the baseplate. The scan was obtained from the vertex to C1 vertebra at a 1 mm slice thickness (Figure 3). Lens shield placement was precluded by the limited opening of the eyelids due to local induration and scarring from the excision biopsy. High-risk clinical target volume (HR-CTV) encompassed the area of induration in the left eyelid and lateral canthus, also corresponding to the traced boundaries. Intermediate-risk CTV (IR-CTV) involved 5 mm around the HR-CTV. The left globe and left lens were considered organs at risk. However, dose constraint for the left lens was relaxed due to the pre-existing cataract, and the dose was prescribed to cover at least 90% of the HR-CTV (HR-CTV D₉₀). Also, the dose to at least 95% of the IR-CTV (IR-CTV D₉₅) was monitored.

Results

Planning, dosimetry, and treatment delivery

Treatment was planned to deliver a total dose of 49 Gy, with each fraction of at least 3.5 Gy delivered to the HR-CTV D₉₀, administered twice daily, with a minimum 6-hour interfraction interval, for a total of 14 fractions. The final plan was generated using the Oncentra® Brachy system (v. 4.5.2.152). Three channels were used (labelled 1 to 3, with the topmost channel labelled as 1), a single unidirectional stepwise moving source, and high-dose-rate remote afterloading. The accepted plan delivered 3.74 Gy to the HR-CTV D₉₀ and 2.66 Gy to the IR-CTV D₉₅ per fraction, to total doses of 52.36 Gy and 37.24 Gy to the HR-CTV and IR-CTV, respectively. The treatment dosimetry parameters demonstrated an acceptable target coverage and conformity, with the HR-CTV and IR-CTV being 0.74 cc and 1.93 cc, respectively, and achieving a conformity index of 1.05, while sparing adjacent organs at risk (Table 1). Assuming no tumor repair within 6-hour intervals and an α/β ratio of 10, these doses corresponded to HR-CTV and IR-CTV EQD₂ of 73.9 Gy and 46.4 Gy, respectively. For normal tissues, assuming an α/β ratio of 3, D_0.1cc of the lens and cornea were 24.4 Gy and 56.3 Gy, respectively (Figure 4). For each fraction, correct applicator placement was verified on orthogonal radiographs using radiopaque markers and dummy wires in catheters as references. Overall, the treatment was administered over 11 days.

Acute and late toxicity

Weekly monitoring was performed until acute toxicity peaked and improved to grade 1 (Figure 5). At two weeks follow-up, the patient developed grade 2 eyelid function disorder and conjunctivitis (as per CTCAE v. 5.0), which were adequately managed with short-course oral steroids, antibiotic eye drops, and artificial tears. No late toxicities were observed.

Disease control and survival

Following initial excision biopsy, radiation therapy was delivered with curative intent. At 2-year follow-up, the patient developed nodal recurrence on the ipsilateral neck level III, confirmed by PET CT. Given the patient’s advanced age and pre-existing medical comorbidities, a palliative approach was pursued, with the nodal disease managed expectantly per patient preference. At 5 years, the patient remained asymptomatic without local recurrence.

Discussion

The definitive management of SGC is surgery, but may be associated with unacceptable functional and cosmetic outcomes [7]. For skin cancers, radiation therapy can be delivered with electrons, external beam radiotherapy (EBRT), or brachytherapy. EBRT is widely available, but is not ideal due to its broad dose fall-off, which may lead to unnecessary exposure of surrounding ocular tissues. Although electrons can deliver a high surface dose with rapid dose fall-off [8], achieving a homogeneous dose on the curved mobile eyelid is difficult. In addition, eyelid lesions require a highly conformal dose distribution that can be achieved with high-dose-rate (HDR) brachytherapy. Pasha et al. highlighted HDR brachytherapy as an effective approach to preserve ocular anatomy and function while maintaining excellent oncologic control [9]. Interstitial brachytherapy (ISBT) can be employed for deep and irregular tumors, demonstrating excellent local control and good functional preservation for deeper eyelid lesions [7, 10]. In a comprehensive review, Guinot et al. emphasized that modern HDR brachytherapy allows precise 3D dose optimization compared with LDR brachytherapy, achieving excellent local control with minimal toxicities [11]. Also, Krengli et al. reported 96% local control with good cosmesis and functional preservation in up to 80% of cases using HDR-ISBT [7, 10].

Adding to the novelty of our case, SGC of the eyelid is a rare malignancy as opposed to more common basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), which have well-established treatment guidelines. Deep periorbital lesions of these histologies can be effectively treated with more invasive ISBT under anesthesia, but may not be suitable for elderly patients or those with multiple comorbidities. Surface molds are a non-invasive alternative for these patients, but most published reports on outcomes focused on non-SGC cancers and usually rely on ocular shields. The eyelid is a particularly difficult site to treat due to its thin, mobile, and curved contour, as well as its direct proximity to critical ocular structures. Furthermore, SGC is an aggressive tumor with high regional recurrence rates; however, reports on alternative non-invasive treatment options and delivery are sparse.

Given our patient’s old age and pre-existing comorbidities, a non-invasive modality, particularly HDR plesiotherapy, was preferred. Plesiotherapy is a type of brachytherapy, where superficial lesions are treated using customized molds, establishing contact between HDR radioactive source and underlying lesion [12]. This technique was selected over EBRT and electron therapy because the mold fitted the exact shape of the lesion, thus ensuring precise target coverage while maximizing organ sparing.

Conejo et al. demonstrated low local recurrence rates with manageable toxicity among patients with skin cancer treated with plesiotherapy [13]. Similarly, Vavassori et al. showed the utility of plesiotherapy molds with ocular tungsten shielding, reporting mostly mild skin and conjunctival site effects, easily managed late toxicities, and no recurrences [6]. Literature on plesiotherapy for eyelid SGC remains scarce due to the rarity of this malignancy. Patro et al. elaborated on its safety profile by reporting only mild dry eye at nine weeks in a patient with eyelid SGC treated with 52 Gy in 13 fractions using lead eye shielding [14].

An important limitation of plesiotherapy is its steep dose fall-off, leading to dose heterogeneity between superficial and deep tumor regions. The inverse square law creates a steep dose gradient, limiting the treatment depth in plesiotherapy to avoid very high surface doses. The GEC-ESTRO guidelines for brachytherapy of skin cancers recommend using a bolus to shift the steep dose-gradient zone into the bolus, ensuring safer surface dosing while treating up to a maximum depth of 5 mm. Flat lesions can also benefit from customized molds, with taped plastic tubes at a distance of 3-5 mm from the skin to avoid overdosing. In our case, the thermoplastic mask acted as both a spacer and bolus. By maintaining a fixed distance from the source to the skin surface, a uniform therapeutic dose is delivered to the tumor using the sharp dose fall-off to protect underlying ocular tissues [15]. The American Brachytherapy Society recommended a total HDR dose equivalent to 60.0 Gy LDR prescribed at 5 mm depth, using individualized fractionation. For our patient, a total dose of 52.36 Gy was delivered over 14 fractions [16]. No literature has firmly established the α/β ratio for SGC. However, Takagawa et al. demonstrated that SGCs respond to long-course radiotherapy of 60.0 Gy in 30 fractions, suggesting a high α/β ratio, similar to epithelial cancers [17].

Sebaceous gland carcinoma carries a high-risk of regional lymphatic spread, with nodal involvement in up to 30% of cases. Pfeiffer et al. recommended baseline nodal staging with sentinel lymph node biopsy (SLNB), as detection of SLN-positive cancer may have important prognostic and therapeutic implications [18]. In the presented case, PET/CT scan was done instead of SLNB, according to the patient’s preference to minimize invasive procedures. While SLNB remains the gold standard for detection of micro-metastasis in clinically node-negative disease, modern PET/CT scan offers high diagnostic accuracy with excellent sensitivity, specificity, and negative predictive value in regional nodal staging. However, PET/CT may still have limitations in detecting very small tumors [19]. Although long-term local control was excellent, an asymptomatic level III neck nodal recurrence occurred at 2 years, which early intervention may have prevented.

An important limitation in this case is the relaxed dose constraints to the lens due to the pre-existing cataract, allowing delivery of the full target dose. Post-surgical scarring of the upper eyelid and canthal areas precluded placement of an ocular shield (e.g., lead or tungsten shields), exposing the cornea to radiation-induced damage, such as keratitis [20]. Ocular shielding reduces the scatter dose to the lens or cornea to prevent or delay the development of cataracts and keratitis, though low atomic number material coating (e.g., wax, acrylic) may be needed to avoid backscattering to the adjacent eyelid [21]. In the study by Hawari et al., customized ocular lead shielding was shown to reduce the cumulative dose to the eye during HDR brachytherapy while protecting adnexal structures [22]. Dosimetric verification was assessed by Zwierzchowski et al., and confirmed that shielding under topical anesthesia effectively reduced the dose to critical organs, including the lens, cornea, and retina, without compromising the dose to CTV [23]. However, literature regarding the use of shielding for plesiotherapy of an eyelid SGC is limited, hence there is the need to individualize dosimetry.

In terms of potential scleral toxicities, Kaliki et al. reported scleral necrosis in only < 1% among patients receiving plaque radiotherapy for uveal melanoma [24]. This low prevalence rate may be due to the avascular and hypocellular nature of the sclera conferring radioresistance [25].

A valuable strength of this report is the availability of a properly documented clinical workflow with rigorous monitoring of acute and late toxicities, as well as long-term outcomes. This case highlights plesiotherapy as a viable and non-invasive alternative for superficial eyelid malignancies, offering good local control with minimal toxicity despite no ocular shielding.

Conclusions

This case demonstrates the technical feasibility of HDR plesiotherapy for localized eyelid carcinomas in patients who are unable to undergo invasive treatment. While successful local control was achieved at 5 years, regional disease control was not maintained. Furthermore, findings from this study only represent an individual case. Therefore, the favorable toxicity profile, safety, and efficacy of treatment, despite the absence of ocular shielding, cannot be generalized to other patients. Larger-scale clinical studies are necessary to determine treatment guidelines, with emphasis on careful patient selection and dosimetric evaluation.

Privacy and confidentiality

Patient data and images were anonymized, de-identified, and kept confidential in compliance with the Data Privacy Act of 2012 and its Implementing Rules and Regulations of 2016. The access to original data and images was restricted to the investigators only, and soft copies of data were stored in a password-protected drive. All data will be safely stored for five years.

Informed consent process

The patient was invited to participate in this study by allowing the investigators to use her data in this study. An informed consent form was provided to the patient and family, with ample time to read, understand, and clarify the details presented in the form. The primary investigator supervised the informed consent process and answered necessary questions posed by the patient and her family. They were given a copy of the consent form. The patient affixed her signature to the certificate of consent, along with the primary investigator and an impartial witness. The principle of autonomy and respect for persons was maintained throughout the study.

Funding

This research received no external funding.

Disclosures

This study was reviewed by the Research Ethics Committee of the authors’ institution (approval No. REC-2025-11-211-TR-AP). It was conducted in accordance with the Declaration of Helsinki, WHO Operational Guidelines in Biomedical Studies 2011, Council for International Organizations for Medical Sciences 2016, Good Research Practice (GRP), National Ethical Guidelines for Research Involving Human Participants (NEGRIHP) 2022, and the Philippine Data Privacy Act of 2012 and its Implementing Rules and Regulations (IRR) of 2016.

The authors report no conflict of interest.