Sebaceous Carcinoma: A Retrospective Multicenter Analysis of 213 Cases Authors: Sebastian A. Wohlfeil, Jochen S. Utikal, Christiane Bauer-Auch, Irina Surovtsova, Tilo Vogel, Anna-Lena Koy, Philipp Morakis Published: 2026 ABSTRACT Simple SummarySebaceous carcinoma is a rare cutaneous malignancy originating from the adnexal glands of the skin. This retrospective multicenter study analyzed a German cohort of patients with sebaceous carcinoma. The aim was to determine whether findings from patient cohorts in the USA and the UK also apply to our population. The analysis showed that sebaceous carcinoma predominantly occurred in elderly female patients (median age: 79 years) and did not reveal any gender-related differences in overall survival. Despite its reputation as an aggressive adnexal tumor, lymph node excisions or systemic therapy were only performed in a minority of cases. These findings are relevant for the routine management and follow-up of patients with sebaceous carcinoma in Germany and across Europe. 1. Introduction Sebaceous carcinoma (SC) is a rare malignant cutaneous malignancy [1]. Among adnexal tumors at the head and neck it is the most common. In addition, SC is one of the most frequent tumors of the eyelid. SC of the skin is classified into a periocular subtype, accounting for around 75% of the cases, and an extraocular subtype [2]. Clinically, SC often presents as a fast growing, yellow-colored, ulcerated nodule [3]. The population at risk are patients older than 60 years, but also a few pediatric cases with SC were reported [4]. Epidemiologic analyses show a strong increase in recent years, mostly in male patients with fair skin [5]. The incidence ranges from 0.1 per 100.000 person-years in England to 0.23 in the US [5,6]. There is still a controversy regarding overall gender-specific incidences [4,5]. Nevertheless, women tend to be diagnosed more often with periocular SC [5]. The exact pathophysiology remains elusive. Both UV irradiation and radiotherapy may be causative risk factors [7]. In addition, patients with immunosuppression, caused by hematological malignancies, infection with human immunodeficiency virus (HIV) or immunosuppressants, were prone to develop SC [7]. Furthermore, viral oncogenesis, such as human papilloma virus, Merkel cell polyoma virus and Epstein–Barr virus, were discussed as explanatory factors of SC [8]. Molecular profiling defines three subtypes: UV-mutated, pauci-mutated and microsatellite instable (MSI) SC [9]. Periocular SC are often pauci-mutated, while extraocular SC show UV signatures or MSI. Potential drivers of SC include mutations along the phosphoinositide 3-kinase (PI3K) pathway [10]. Periocular SC specifically often shows mutations of the TP53 and RB1 genes. In contrast, extraocular SC reveal hereditary mutations, MSI or mismatch repair (MMR) genes. Interestingly, SC also occurs as part of the Muir–Torre syndrome (MTS) which is a phenotypic variant of the Lynch syndrome [11]. It results from autosomal dominant inherited mutations in MMR genes, such as MSH2. Patients with this condition are prone to develop sebaceous neoplasms, including SC, or visceral malignancies like colorectal carcinoma. Typically, periocular SC is not associated with MTS. The Mayo MTS risk scoring may help to screen patients with extraocular SC for MTS [12]. It includes the age of the patient, number of sebaceous tumors and malignancies associated with Lynch syndrome in their own and family history. Genetic testing should be performed in patients with a score ≥ 2. The primary treatment of SC is surgery, with micrographic surgery as the gold standard [1,13]. Numerous studies showed that micrographic surgery is superior to local wide excision which was often performed in the past [1,14,15,16]. In addition, this is an advantage for aesthetic areas such as the eyelid [17]. For periocular SC stage T2c and higher, but not for extraocular SC, sonography of the lymph nodes (LN) and sentinel lymph node (SLN) biopsy are recommended [1,13,18]. If nodal involvement was detected, a LN dissection should be discussed with the patient. SC are classified according to the UICC classification for skin cancer and its respective localization. Radiotherapy is considered an option for patients who decline surgical treatment [19]. Data on adjuvant radiotherapy is limited but may be performed in SC with certain risk factors such as incomplete resection, recurrent lesions or perineural invasion [19,20]. For systemic treatment of SC, no randomized controlled clinical trials have been performed, as only few cases with metastasized SC have been reported. Systemic treatment comprised conventional chemotherapy or immune checkpoint inhibition [21,22]. Currently, enfortumab vedotin, an antibody-drug conjugate, is being discussed as a novel systemic therapy of SC due to its expression of Nectin-4 [23]. Neoadjuvant application of a platinum-based chemotherapy proved effective for periocular SC in a case series [24]. Recent analyses of the American National Cancer Database report an increased mortality in extraocular compared to periocular SC [5], while old analyses did not find any differences [4]. Therefore, our multicenter retrospective study on SC compared real-world data from German cancer centers to add evidence on the presentation and clinical course of SC in Europe. 2. Materials and Methods 2.1. Data Collection and Study Population A retrospective multicenter analysis was performed on a data set obtained from the clinical Cancer Registry Database (Klinisches Landeskrebsregister, KLR GmbH) of the German Federal State Baden Württemberg (BW). Cases with diagnosis of SC between 2008 and 2024 were extracted from the registry. Data cutoff was on the 1st of December 2024. All cases with histologically confirmed SC (ICD-O-3 code 8410/3) were extracted. Tumor localization was defined according to ICD-10 codes and categorized into ocular and extraocular sites. Ocular tumors were restricted to the eyelid (C44.1). Extraocular tumors included lesions of the head and neck region (C44.0, C44.2, C44.3 and C44.4) as well as tumors of non-head and neck sites, including trunk, extremities, and genital regions (C44.5*, C44.6, C44.7, C44.8, C44.9, C51*, C63.2). Nine cases with a multiple synchronous SC were excluded from the analysis. 2.2. Statistical Analyses Baseline characteristics were compared using Fisher’s exact test or the chi-square test, as appropriate. Only patients residing in Baden-Württemberg were included in the survival analyses. Overall survival (OS) was defined as time from diagnosis to death from any cause. Survival outcomes were analyzed using Kaplan–Meier estimates with the log-rank test, while multivariable Cox proportional hazards models were applied to assess independent prognostic factors. Results are presented as median OS rates and hazard ratios (HRs) with 95% confidence intervals (CIs). Statistical significance was set at p < 0.05. All statistical analyses were performed with R (version 4.2.1) and the tidyverse, survival or survminer packages. 3. Results A total of 213 patients were included, comprising 172 with extraocular and 41 with ocular tumors (Figure 1, Table 1). The extraocular tumors were most commonly located in C44.3 (81; 38.0%), followed by C44.4 (35; 16.4%), C44.5 (21; 9.9%), C44.2 (13; 6.1%), and C44.6 (10; 4.7%). Less frequent sites included C44.9 (5) and C44.7 (3), while all other locations (C44.0, C44.59, C51.9, C63.2) were rare (n = 1 each) (for a detailed mapping of ICD-10 codes to anatomical locations please refer to Supplementary Table S1). The mean age of the overall cohort was 75.1 years (extraocular: 74.5; ocular: 77.5; p = 0.213), with a median age of 79 years (extraocular: 77.5; ocular: 83; p = 0.054). Age distribution did not differ significantly between groups (p = 0.087), although patients older than 80 years were more frequent in the ocular group (63.4% vs. 44.8%), (Figure 1A–C, Table 1). Male patients predominated overall (67.6%), with a significantly higher proportion in the extraocular group compared to the ocular group (71.5% vs. 51.2%; p = 0.021). Tumor grading, T-, N-, and M-stage, as well as UICC stage, were available for a subset of patients and were analyzed descriptively. Based on available UICC staging, the majority of patients presented with stage I disease (80.0%), with no significant differences between ocular and extraocular groups (p = 0.838), (Figure 1D,E). Only five cases of extraocular SC were in stage III (Supplementary Table S2), and two were in stage IV (Supplementary Table S3). Among cases with available data, tumor grading was comparable between groups (p = 0.638), with grade 2 being the most frequent (44.2%). Both extraocular and ocular SC showed no obvious differences regarding grading of SC in male and female patients (Figure 1F,G). The T stage tended to be more advanced in ocular tumors; this difference did not reach statistical significance (p = 0.055). Mean tumor thickness was 3.92 mm overall and did not differ significantly between extraocular and ocular tumors (4.04 vs. 2.98 mm; p = 0.554). Regarding nodal or distant metastasis only few cases were reported (Table 1). Nodal disease occurred more frequently in ocular SC than in extraocular SC (p = 0.03), while there was no difference regarding distant metastasis (Table 1). However, this must be interpreted with caution, given the small number of ocular SC cases. Treatment of SCs was mainly surgical (Figure 2A). In 81.4% (118 cases), a microscopically controlled excision was performed. When comparing margin-free (R0) resections, results were similar to wide local excisions (Figure 2B). However, the exact safety margins of the wide local excisions had not been documented in the registry data. Regarding extraocular SC, the percentage of margin-free resected SC (R0) was slightly higher in male (95.3%) as compared to female (77.4%) patients (Figure 2C). In ocular SC a lower percentage of complete tumor resections (R0) was found in male patients (72.7%) in relation to female patients (91.7%) (Figure 2D). Among 63 patients with available postoperative outcome data, those primarily treated with microscopically controlled surgery showed a lower rate of progressive disease (9%; five cases) compared to patients with wide local excision (18%; two cases) (Figure 2A). However, the number of patients treated with wide local excision was low (27 cases in total). In only 2.3% of all cases, specifically five cases with extraocular SC, a SLN biopsy was performed (Supplementary Table S4). A LN dissection was only documented in 1.9% of our cohort (four patients), with no difference between extraocular and ocular SC patients (Supplementary Table S4). Radiotherapy was performed in two cases as definitive therapy for extraocular stage III and IV SC, and four SC cases received an adjuvant radiotherapy (Supplementary Table S5). Systemic therapy with platinum-based chemotherapy or taxanes, often in combination with radiotherapy, was only performed in three patients, two patients with locoregional disease and one case with lung metastasis (Supplementary Figure S1A). Overall survival analysis was further conducted to explore factors associated with prognosis in patients with SC. This analysis was restricted to patients residing in Baden-Württemberg (n = 186), with a median observation time of 3.2 years. The median OS in the entire cohort was 61.4 months (95% CI 46.8–95.8). We first assessed whether tumor localization at initial presentation influenced the OS. Patients with extraocular SC had a median OS of 64.8 months compared to 53.7 months in those with ocular SC, with no significant difference observed between groups (p = 0.490) (Figure 3A). Multivariable Cox regression analysis confirmed the findings of the univariate analysis, showing no significant association between tumor localization and overall survival (extraocular vs. ocular: HR 1.4, 95% CI 0.85–2.4; Figure 3B). Age remained a strong independent prognostic factor: compared to patients <60 years, those aged 70–79 years had an HR of 4.4 (95% CI 1.01–19.2), and those aged ≥80 years an HR of 16.1 (95% CI 3.91–66.1). For extraocular and ocular SC there were also no gender-specific differences among the patients (Supplementary Figure S2A,B). Second, the histologic grading was investigated. No differences were observed among patients with SC from grade 1 to 3 (Supplementary Figure S3A). As SC might also occur as part of MTS [11], we were interested in secondary malignancies. The Mayo MTS score is a helpful tool to identify patients at risk [12]. Unfortunately, due to the retrospective analysis of our study and the nature of the registry data, this could not be applied to our cohort. Therefore, we selected patients who suffered from malignancies before diagnosis of SC. Patients with multiple SCs were excluded from the whole study (Supplementary Table S6). In total, 113 cases had a history of a malignancy before the diagnosis of SC (Figure 4). Commonly, this was non-melanoma skin cancer (NMSC) (37.2%; 42 cases). A total of 34 patients (30.0%) had at least one additional non-cutaneous malignancy (Supplementary Table S7) including 13 cases (38.2%) with cancers like bladder cancer or rectal cancer that might be associated with MTS (MTS-like) and six cases (17.6%) with hematological neoplasia, such as lymphoma or leukemia, were identified. In addition, 29 patients (25.7%) had a broader history of malignancies including skin cancers (C44) (Supplementary Table S8). Within this subgroup, seven (24.1%) had MTS-like cancers and 13 patients (44.8%) had hematological malignancies. We further assessed the impact of prior malignancies on overall survival. In univariate analysis, patients with MTS-like cancers showed a trend toward improved survival compared to those with other malignancies (p = 0.073; Supplementary Table S9 and Supplementary Figure S4A), although this did not reach statistical significance and was not confirmed in multivariable Cox regression (HR 1.2, 95% CI 0.53–2.7; Figure 5A). Hematological neoplasms were associated with significantly worse OS in univariate analysis compared with patients with other prior malignancies (p = 0.0056, Supplementary Table S10 and Supplementary Figure S4B), but this association was not retained in multivariable analysis (HR 1.7, 95% CI 0.91–3.3; Figure 5B). 4. Discussion Since SC is a rare adnexal tumor [1], data from large patient cohorts remain limited. Therefore, our multicenter retrospective study aimed at analyzing the epidemiology, treatment and outcome of 213 patients with either extraocular or ocular SC in the German federal state of Baden-Württemberg, which has a population of over 11 million (as of 2019). SC predominantly affected the elderly with a median age of 79 years—higher than that reported in previous reports [4,25,26]. In addition, our data revealed a male-to-female ratio of 2:1 among all SC patients. Published studies confirm this predominance in men [5,25,26]. The incidence in men seems to be rising in comparison to earlier data from the Surveillance, Epidemiology, and End Results (SEER) program [4]. However, this could be biased by enhanced histopathological diagnostics and disease awareness. With respect to the localization, extraocular SC was more common in elderly men, while ocular SC was overrepresented in female patients older than 65 years. Epidemiologic data from the USA, England, Norway or Taiwan are in line with our findings [6]. As recommended by guidelines [1,13], primary treatment of SC was surgical, with microscopically controlled surgery used in 81.4%. Follow-up data suggest a benefit for microscopically controlled surgery (9% recurrence) as compared to wide local excision (18.6% recurrence). Yadlapati et al. confirmed a superiority of Mohs micrographic surgery to wide local excision regarding local and distant recurrence in a systematic review [15]. In ocular SC, nodal disease was significantly more frequent than in extraocular SC which is in line with the literature [27]. Extended procedures such as SLN biopsy were only performed in 2.3% and LN dissection in 1.9% of all cases. Furthermore, only one case with lung metastatic SC was documented in our cohort and two cases with extended locoregional metastasis which were treated by systemic therapy with platinum-based chemotherapy or taxanes. One patient with stage IV SC (pT4b) received radiotherapy. It appears that SC is indeed a potentially aggressive adnexal tumor [13], but the metastasis rate would be low if it were adequately detected and treated. This notion is also supported by Chinese and Dutch cohorts that did not even detect any metastases in SC smaller than 10 mm [26,28]. The most important prognostic factor remains age. Especially, SC patients older than 70 years showed significantly decreased OS [29]. Notably, a recent study demonstrates that the increased mortality among SC patients is especially driven by cardiovascular, cerebrovascular, metabolic, kidney and lung diseases [30]. This is conceivable given the advanced age of the patients of SC study cohorts. Furthermore, solid organ transplantation is associated with a 25-fold increased risk to develop SC [7]. It is hypothesized that viral oncogenesis, immunosuppressive agents, medication inducing photosensitization or underlying medical conditions increase the risk of patients with solid organ transplantation to develop SC [31]. In this context, our findings suggest that immune dysregulation may also influence outcomes beyond the transplant setting. Patients with a history of hematological malignancies showed worse overall survival compared to those with other prior cancers, despite similar age distributions, possibly reflecting impaired tumor surveillance due to persistent immune alterations. However, this association was observed only in univariate analysis and was not confirmed in the multivariable model, indicating potential confounding or limited statistical power. These findings should therefore be interpreted with caution. A comparison of the localization of SC in our study showed that the OS was similar between ocular and extraocular SC patients, in contrast to the study by Tripathi and colleagues who detected a 1.4-fold increased mortality of extraocular SC patients in relation to ocular SC patients [5]. In addition, no gender-specific differences in overall survival were detected, which is also seen in cohorts from England [25] and in a SEER data set with ocular SC from the USA [29]. Last, no influence of the tumor grading on the OS was observed, as in a cohort of ocular SC [29]. Lately, it was debated whether MTS affects the OS of patients with SC [32,33]. Therefore, we analyzed the OS of SC patients with history of malignancies associated with MTS, such as gastrointestinal or urological cancers. This analysis was limited as neither information on family history nor genetic analyses were available. In our study we observed a tendency toward an improved OS in patients with SC and history of MTS-like cancers as compared to patients with SC and other malignancies. However, this likely reflects confounding by age, as these patients were significantly younger than the comparison group. Importantly, no association was observed in the multivariable model, indicating that MTS-like malignancies were not independently associated with survival in our cohort. Our data is in accordance with a study from Maloney and colleagues [33]. Our multicenter retrospective analysis represents a relatively large cohort of patients with sebaceous carcinoma; however, several limitations should be acknowledged. As is inherent to retrospective registry studies, documentation bias is likely, with initial diagnostic information generally well captured but subsequent clinical details sometimes incomplete. In addition, several relevant variables—including margin status, molecular features, adverse events, and comorbidities—were not systematically recorded. Furthermore, heterogeneity in treatment approaches across participating centers, including aspects of supportive care, may have influenced outcomes and limits direct comparability between groups. Despite these limitations, the large sample size and population-based design enhance the robustness of our findings and provide valuable insight into real-world clinical outcomes. External validation in independent cohorts is desirable and remains beyond the scope of the present study. 5. Conclusions In this relatively large multicenter cohort, sebaceous carcinoma mainly affected elderly patients and showed no survival difference between ocular and extraocular disease, with age being the only independent prognostic factor. Prior malignancies were common, including MTS-like and hematological tumors, but none showed an independent impact on survival despite divergent univariate trends.