The Spectrum of Eyelid Lesions: Histopathologic Classification of Benign, Malignant, and Non-Neoplastic Findings

Abstract

Introduction

Extensive literature exists on characterizing and differentiating eyelid neoplasms that may mimic benign lesions. However, there are limited epidemiologic data on the most common benign and malignant neoplasms of the eyelid, especially within the USA.

Methods

A retrospective cohort study was performed of all histopathologically confirmed eyelid biopsy specimens between January 2000 and December 2021 at a large tertiary care academic medical center. Age, sex, lesion site and laterality, and diagnosis were collected. Specimens were stratified into benign and malignant eyelid lesions.

Results

Five hundred sixty-one eyelid cases (male = 270, female = 291) were analyzed, 460 benign (82.0%) and 101 malignant (18.0%). Mean age for benign lesions was 60.6 ± 15.2 years. Mean age for malignant lesions was 67.6 ± 12.5 years. Most common benign lesions were papillomas (n = 146: 31.7%), hidrocystoma (n = 65: 14.1%), and epidermal inclusion cyst (n = 41: 8.9%). Most common malignant lesions were basal cell carcinoma (n = 72: 71.3%), squamous cell carcinoma (n = 17: 16.8%), and melanoma (n = 5: 5.0%).

Conclusion

The number of benign eyelid lesions outweighs those of malignant lesions. Benign lesions were commonly seen at an earlier age compared to malignant lesions. Epidemiologic data presented adds to knowledge about lesions and can help inform clinical differential diagnoses.

Introduction

Eyelid tumors compromise a wide spectrum of clinical presentations and behaviors. While extensive literature exists characterizing clinical features of eyelid tumors, there remains a notable lack of epidemiologic data quantifying the relative frequency of benign and malignant lesions, particularly within the USA. Accurate epidemiologic data are essential for informing differential diagnoses, guiding biopsy decisions, and aiding pathologists in accurate histopathologic interpretations.

Several international studies have provided valuable insights into the incidence of eyelid tumors and geographic variations. A 21-year study analyzing 1,166 eyelid cancer cases in Taiwan reported a predominance of basal cell carcinoma (BCC), while a study in Sri Lanka noted a predominance of sebaceous carcinoma [1, 2]. Studies from Turkey, Switzerland, Istanbul, and Iran further highlight regional variations in both diagnostic distribution [3–6]. These reports underscore the need for geographically specific data to inform clinical decision-making.

However, studies within the USA remain limited. One large-scale study based on national registry data provided epidemiologic on malignant eyelid lesions based on reported diagnosis but did not include histopathologic confirmation [7]. A smaller study in Minnesota focused solely on malignant lesions within a single county in the 1990s, while another study in Florida examined 206 malignant lesions but excluded basal and squamous cell carcinomas [8, 9]. Notably, none of these studies included benign lesions.

Our study aimed to fill this gap by analyzing both malignant and benign lesions with confirmed histopathology, providing a more complete and accurate representation of diagnoses within a US population. By quantifying and ranking these lesions, we aimed to assist clinicians in refining clinical differentials, potentially reducing unnecessary biopsies and their associated morbidity. Additionally, our findings may support pathologists in correlating clinical impressions with histopathologic diagnoses, ultimately improving diagnostic accuracy and patient outcomes.

Materials and Methods

This retrospective cohort study was performed of all histopathologically confirmed eyelid biopsy specimens between January 2000 and December 2021 at the Medical College of Wisconsin in Milwaukee, WI. Cutaneous pathology records of the eyelid were extracted from Epic, Cerner, Pathologix, and MCW databases using the keywords “eyelid(s),” “lateral canthus,” “medial canthus,” “eye,” and “eyebrow.” Age, sex, lesion site and laterality, and diagnosis were collected. All information was collected from the pathology reports including reported diagnosis. Lesions from the pediatric population, multiple re-excisions of margins following Mohs surgery, and all other diagnoses except those definitively on the eyelids and lateral/medial canthus were excluded. Chi-squared analyses were performed to assess differences in the distribution of eyelid lesions by laterality or site.

Results

A total of 561 eyelid cases (male = 270, female = 291) were analyzed, 460 benign (82.0%) and 101 malignant (18.0%). Mean and median age was 63.0 and 64 years, respectively. The right eye (n = 269: 48.0%) and left eye (n = 266: 47.4%) were involved at approximately equal rates. The lower lid was the most common lesion site (n = 210: 37.4%), followed by the upper lid (n = 174: 31.0%). Detailed age distribution, laterality, and site information for the total study population can be found in Table 1.

Table 1.

Study population

Parameter	Frequency		%
Total cases	n =	561	​
Gender	Male	270	48.1%
	Female	291	51.9%
Mean age, years	Mean	63.0	​
	SD	15	​
Median age, years	Median	64	​
	IQR	54–74	​
Age distribution	0–20 years	2	0.4%
	21–40 years	48	8.6%
	41–60 years	174	31.0%
	61–80 years	276	49.2%
	81+ years	60	10.7%
Laterality	Right	269	48.0%
	Left	266	47.4%
	Bilateral	11	2.0%
	Unspecified	15	2.7%
Site	Upper lid	174	31.0%
	Lower lid	210	37.4%
	Medial canthus	26	4.6%
	Lateral canthus	42	7.5%
	Multiple sites	2	0.4%
	Unspecified	107	19.1%
Type of lesion	Benign	460	82.0%
	Malignant	101	18.0%

Of benign lesions, the mean age of presentation for benign lesions was 61.9 ± 15 years (range: 16–91 years). There were 217 males (47.2%) and 243 females (52.8%). The age range of 61–80 years comprised 48.7% of benign lesions. Combining the 41–60 years and 61–80 years age ranges, 80.9% of benign lesions are included. There was largely no difference in laterality of benign lesion or in site frequency between upper eyelid (n = 159: 34.6%) and lower eyelid (n = 160, 34.8%). The top 5 most common benign lesions of the eyelid were papilloma (n = 146: 31.7%), hidrocystoma (n = 65: 14.1%), epidermal inclusion cyst (n = 41: 8.9%), benign nevi (n = 32: 7.0%), and inflammation (n = 25: 5.4%), comprising 67.1% of all benign lesions. Representative pathology images of papilloma, hidrocystoma, and epidermal inclusion cyst are shown in pathology (Images 1–3), respectively. Additional information regarding laterality, site, and diagnosis for benign eyelid lesions is shown in Table 2.

Fig. 1.

Papilloma (seborrheic keratosis) at 2× (a) and 10× (b) magnification stained with H&E.

Fig. 3.

Epidermal inclusion cyst at 2× (a) and 10× (b) magnification stained with H&E.

Table 2.

Benign lesions

Parameter	Frequency		%/B	%/total
Total cases	n =	460	​	82.0%
Gender	Male	217	47.2%	38.7%
	Female	243	52.5%	43.3%
Mean age, years	Mean	61.9	​	​
	SD	15	​	​
Median age, years	Median	64	​	​
	IQR	53–72	​	​
Age distribution	0–20 years	2	0.4%	0.4%
	21–40 years	44	9.6%	7.8%
	41–60 years	148	32.2%	26.4%
	61–80 years	224	48.7%	39.9%
	81+ years	41	8.9%	7.3%
Laterality	Right	226	49.1%	40.3%
	Left	210	45.7%	37.4%
	Bilateral	11	2.4%	2.0%
	Unspecified	13	2.8%	2.3%
Site	Upper lid	159	34.6%	28.3%
	Lower lid	160	34.8%	28.5%
	Medial canthus	13	2.8%	2.3%
	Lateral canthus	31	6.7%	5.5%
	Multiple sites	2	0.4%	0.4%
	Unspecified	95	20.7%	16.9%
Diagnosis	Papilloma	146	31.7%	26.0%
	Seborrheic/verrucous keratosisa	79	17.2%	14.1%
	Acrochordona	27	5.9%	4.8%
	Verruca vulgaris	20	4.3%	3.6%
	Squamous papilloma	17	3.7%	3.0%
	Filiform wart	3	0.7%	0.5%
	Hidrocystomaa	65	14.1%	11.6%
	Epidermal inclusion cysta	41	8.9%	7.3%
	Nevib	32	7.0%	5.7%
	Inflammationa	25	5.4%	4.5%
	Benign skin and/or adipose	23	5.0%	4.1%
	Xanthelasma	13	2.8%	2.3%
	Actinic keratosisa	12	2.6%	2.4%
	Cystc	11	2.4%	2.0%
	Chalazion	9	2.0%	1.6%
	Milia	8	1.7%	1.4%
	Dermal edema	7	1.5%	1.3%
	Cystadenoma	6	1.3%	1.1%
	Foreign body reaction	6	1.3%	1.1%
	Syringoma	6	1.3%	1.1%
	Solar lentigo	5	1.1%	0.9%
	Angiofibroma/fibrous papule	4	0.9%	0.7%
	Epidermal hyperplasia	4	0.9%	0.7%
	Molluscum contagiosum	3	0.7%	0.5%
	Sebaceous hyperplasia	3	0.7%	0.5%
	Granulomatous dermatitis	2	0.4%	0.4%
	Hemangioma	2	0.4%	0.4%
	Lipoma	2	0.4%	0.4%
	Nodular amyloidosis	2	0.4%	0.4%
	Pyogenic granuloma	2	0.4%	0.4%
	Otherd	21	4.6%	3.7%

%/B: percent of benign lesions; %/total: percent of total study lesions.

^aContains few double-counted lesions given multiple lesions within a specimen.

^bNevi include common blue, compound, and intradermal.

^cCyst includes steatocystoma, follicular, trichilemmal, vellus hair, and epidermal.

^dContains multiple single lesions with n = 1, such as keratoacanthoma, trichilemmoma, angioleiomyoma, spiradenoma, spongiotic dermatitis, among others.

Fig. 2.

Hidrocystoma at 2× (a) and 10× (b) magnification stained with H&E.

The mean age of presentation for malignant lesions was 68.1 ± 14 years (range: 42–90 years), notably greater than that of benign lesions. There were 53 males (52.5%) and 48 females (47.5%). The age range of 61–80 years comprised 51.5% of malignant lesions. There was no significant difference in laterality of malignant lesions, with 43 (42.6%) right eye lesions and 56 (55.4%) left eye lesions. However, there was a significant difference in upper lid (n = 15: 14.9%) vs. lower lid (n = 50: 49.5%) malignant lesions when performing a Chi-squared analysis (χ² [1, n = 65] = 18.84, p < 0.001). The top 3 most common malignant lesions of the eyelid were BCC (n = 72: 71.3%), squamous cell carcinoma (n = 17: 16.8%), and melanoma (n = 5: 5.0%), comprising 93.1% of all malignant lesions. Representative pathology images of BCC, squamous cell carcinoma, and melanoma are shown in pathology (Images 4–6), respectively. Additional information regarding laterality, site, and diagnosis for malignant eyelid lesions is shown in Table 3.

Fig. 4.

BCC at 4× magnification stained with H&E.

Fig. 6.

Melanoma in situ at 2× (a) and 10× (b) magnification stained with H&E.

Table 3.

Malignant lesions

Parameter	Frequency		%/M	%/total
Total cases	n =	101	​	18.0%
Gender	Male	53	52.5%	9.4%
	Female	48	47.5%	8.6%
Mean age, years	Mean	68.1	​	​
	SD	14	​	​
Median age, years	Median	70	​	​
	IQR	59–77	​	​
Age distribution	0–20 years	0	0.0%	0.0%
	21–40 years	4	4.0%	0.7%
	41–60 years	26	25.7%	4.6%
	61–80 years	52	51.5%	9.3%
	81+ years	19	18.8%	3.4%
Laterality	Right	43	42.6%	7.7%
	Left	56	55.4%	10.0%
	Bilateral	0	0.0%	0.0%
	Unspecified	2	2.0%	0.4%
Site	Upper lid	15	14.9%	2.7%
	Lower lid	50	49.5%	8.9%
	Medial canthus	13	12.9%	2.3%
	Lateral canthus	11	10.9%	2.0%
	Multiple sites	0	0.0%	0.0%
	Unspecified	12	11.9%	2.1%
Diagnosis	BCC	72	71.3%	12.8%
	Squamous cell carcinoma	17	16.8%	3.0%
	Melanoma	5	5.0%	0.9%
	Basosquamous cell carcinoma	1	1.0%	0.2%
	Sebaceous carcinoma	1	1.0%	0.2%
	T-cell lymphoma	1	1.0%	0.2%
	Adenocarcinoma	1	1.0%	0.2%
	Primary cutaneous mucinous carcinoma	1	1.0%	0.2%
	Poorly differentiated carcinoma	1	1.0%	0.2%
	Merkel cell carcinoma	1	1.0%	0.2%

%/M: percent of malignant lesions; %/total: percent of total study lesions.

Fig. 5.

Squamous cell carcinoma at 2× (a) and 10× (b) magnification stained with H&E.

Discussion

Accurate diagnosis of eyelid lesions is important for patient care and clinical decision-making but can be difficult due to the multiple tissue layers involved in the eyelid each with distinct possible neoplasms. Diagnosing eyelid lesions histologically can be challenging for a multitude of reasons including some benign lesions can closely mimic malignancy with overlapping microscopic features, or biopsy specimens may be small and altered by inflammation, prior procedures, or tissue processing. Common benign lesions such as papillomas, the most common within this dataset, may mimic more concerning pathology such as SCC. Squamous papillomas typically demonstrate orderly exophytic epithelial proliferation, while SCC may demonstrate a disordered epithelial architecture with cytologic atypia as well as invasion of the basement membrane. However, in cases where atypia is minimal, both lesions can exhibit exophytic or endophytic growth patterns, acanthosis, and squamous eddies making it a difficult diagnosis [10]. Among malignant eyelid tumors, BCC, being the most common in this study, histologically shows nests of basaloid cells with peripheral palisading and stromal retraction artifacts. Early, superficial, or fragmented biopsies may fail to capture these diagnostic features, making differentiation from benign basaloid proliferations difficult, especially on routine hematoxylin-eosin sections. This sampling error often becomes an even greater concern in a highly clinically sensitive area such as the eyelid. These factors highlight the importance of adequate sampling, careful histopathologic evaluation, and clinicopathologic correlation when diagnosing eyelid lesions.

Consistent with other studies, we report 82.0% of lesions were benign. However, the most common benign lesion varies largely by study and location. The most common benign lesion in this study was seborrheic keratosis, whereas in India, Israel, Turkey, and Taiwan, the most commonly identified included chalazions, fibromas, squamous cell papilloma, and intradermal nevus, respectively (Table 4) [3, 9, 11, 12]. This variation in frequently diagnosed benign lesions across countries may be related to varying genetic predisposition, access to care, and the likelihood of individuals seeking treatment.

Table 4.

Comparison of present study findings with existing literature

Study	# Lesions reviewed	Geographic location	Mean age	Benign lesions	Most common benign, %	Most common malignant, %	Reference number
Cao DY et al. (2025)	581	Wisconsin, USA	Benign: 61.9, malignant: 68.1	82.00%	Papilloma (31.7%)	BCC (71.3%)	x
Nanayakkara DPS et al. (2023)	654	Sri Lanka	All: 46, malignant: 64	49.20%	Seborrheic keratosis (30.4%)	Sebaceous carcinoma (32.4%)	[2]
Eren MA et al. (2020)	911	Ankara, Turkey	Benign: 46.5, malignant: 62.3	73.10%	Squamous cell papilloma (15.2%)	BCC (21.0%)	[3]
Deprez et al. (2009)	5,504	Lausanne, Switzerland	N/A	84.00%	Squamous cell papilloma (26%)	BCC (86.0%)	[4]
Bagheri A et al. (2013)	182	Tehran, Iran	Benign: 46.4, malignant: 63.9	45.10%	Melanocytic nevus (35%)	BCC (83%)	[6]
Levinkron O et al.	1,423	Israel	Benign: 59, malignant: 76	85.00%	Soft fibroma (20.1%)	BCC (9.2%)	[12]
Huang YY et al.	4,521	Taiwan	Benign: 55.4, malignant: 72.5	95.00%	Intradermal nevus (21.1%)	BCC (57.8%)	[13]
Yu SS et al. (2018)	2,228	Tianjin, China	Benign: 49.7, malignant; 63.5	85.70%	Melanocytic nevus (33.8%)	BCC (56.5%)	[14]
Wang L et al.	5,146	Zhejiang, China	Benign: 49.6, malignant: 71.2	85.10%	Nevus (33.07%)	BCC (48.7%)	[15]

Malignant lesions made up 18.0% of neoplasms, which was lower than a center in Iran which demonstrated 54.9% malignant lesions [6]. This could be suggestive of differences in access to care or clinical biopsy practices. The mean age at time of diagnosis was higher for malignant lesions than benign consistent with other studies regardless of most common identified malignant neoplasm [2, 7]. The most common malignant lesion was BCC (71.3%), similar to other studies [3, 9, 13, 14]. Of the 101 malignant lesions biopsied in this study, there were 7 diagnoses with single confirmed incidences and even broader variation than within the benign neoplasms. These findings highlight the importance of maintaining a broad differential when correlating clinical and histopathologic impressions.

In this study, malignant lesions were more commonly identified on the lower lid, whereas benign lesions showed no preference for upper versus lower lid. There was an increased incidence of malignant lesions on the left eye over the right. Interestingly, a study in India noted an increased incidence of right-sided malignant lesions with the hypothesis that it was in relation to driving on the right side [11]. Consistent with American driving patterns, chronic UV exposure on the left side could potentially impact the increased incidence of left-sided malignant lesions. However, given the limited current evidence, further investigation is needed before recommending changes in clinical practice.

Limitations

This study is limited by the fact that not all benign eyelid lesions are routinely biopsied, which may lead to underreporting and an artificially lower incidence of benign neoplasms in the dataset. Additionally, in the pathology reports, race and ethnicity are not reported. These additional findings may give further insight into the genetic predisposition to certain eyelid neoplasms.

Conclusions

This study provides epidemiologic data on the most common benign and malignant eyelid neoplasms from a US-based medical center. The findings are consistent with international studies but highlight important geographic variation in both the frequency and types of eyelid tumors, likely influenced by regional genetic predisposition, UV exposure, and patterns of healthcare access. Given the higher prevalence of malignancy in older patients and in lesions located on the lower, left eyelid, clinicians should maintain heightened suspicion in such presentations. These results emphasize the importance of integrating geographic context into diagnostic and management strategies for eyelid neoplasms to improve early detection and treatment outcomes.

Acknowledgments

The authors would like to thank Nathan Duncan for assisting with aggregating data from multiple databases.

Statement of Ethics

This study protocol was reviewed and the need for approval was waived by the Medical College of Wisconsin Institutional Review Board. Written informed consent from participants was not required in accordance with local/national guidelines.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This article has no funding sources.

Author Contributions

David Y. Cao: conceptualization, data acquisition and analysis, and writing – original draft and editing. Alyssa M. Jobe: writing – original draft and editing. Karolyn A. Wanat: conceptualization and writing – reviewing and editing.

Funding Statement

This article has no funding sources.

Data Availability Statement

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