Caruncular lesions in Denmark 1978–2002: a histopathological study with correlation to clinical referral diagnosis

Authors


Steffen Heegaard
Eye Pathology Institute
University of Copenhagen
Frederik V's Vej 11
DK-2100 Copenhagen
Denmark
Tel: + 45 35 32 60 70
Fax: + 45 35 32 60 80
Email: sh@eyepath.ku.dk

Abstract.

Purpose: To carry out a retrospective clinicopathological evaluation of caruncular lesions.

Methods: Data were collected from all surgically removed and histopathologically evaluated caruncular lesions registered by Danish pathology departments during the 25-year period 1978–2002.

Results: A total of 574 caruncular lesions were identified. The number of caruncular lesions increased significantly during the 25-year period. This was due to an increase in the number of benign lesions, whereas the number of premalignant and malignant lesions remained constant. A total of 550 (96%) of the lesions were benign. Naevus (n = 248, 43%) and papilloma (n = 131, 23%) were the most common neoplasms. Premalignant lesions (n = 10, 1.7%) were dominated by primary acquired melanosis (PAM) with atypia and epithelial dysplasia. Malignant lesions constituted a total of 14 neoplasms (2.4%), with basal cell carcinoma (n = 4, 0.7%) and lymphoma (n = 4, 0.7%) being the most frequent. The preoperative clinical diagnosis was correct in 286 (50%) of cases.

Conclusion: Caruncular lesions are predominately benign but the lesions are rare and diverse, making clinical diagnosis difficult. Referral of excised lesions for pathological examination is recommended.

Introduction

The caruncle is situated between the plica semilunaris and the medial angle of the eyelid aperture and represents a transitional zone between the skin and the conjunctiva. Caruncular lesions have been found to constitute about 1% of the total amount of excised conjunctival lesions (Shields et al. 1986; Santos & Gomez-Leal 1994).

The compound architecture of the caruncle gives rise to a high diversity of neoplasms (Luthra et al. 1978). An inconsistency between clinical and histopathological diagnoses in about half the cases has previously been reported (Shields et al. 1986; Santos & Gomez-Leal 1994). Misdiagnosis of malignant lesions may lead to severe morbidity due to orbital invasion. Furthermore, malignant melanoma involving the caruncle seems to metastasize early (Paridaens et al. 1994a, 1994b). Thus, in order to reduce morbidity, the correct diagnosis and handling of caruncular malignancy are important.

Differences in the distribution of lesions in time, geographic location and race have been identified. Norn (1984) found the prevalence of pigmentation of the plica semilunaris and lacrimal caruncle to be 2% among North Europeans, but 45% in Eskimos. Inflammatory and cystic lesions of the caruncle have been reported to have a particularly high frequency in North Africa (Wilson 1958).

Over the last three decades, marked increases in the incidence of both melanoma and non-melanoma skin cancers have been observed in Denmark (Sundhedsstyrelsen 2003). Due to the skin origin of the caruncle and its exposure to actinic rays, the frequency of premalignant and malignant caruncular lesions might also have increased. The aim of this study was to make a retrospective clinicopathological evaluation of all caruncular lesions examined histopathologically in Denmark during the 25-year period 1978–2002. In particular, we wanted to investigate changes in the distribution of benign, premalignant and malignant lesions. Furthermore, we wanted to compare the clinical referral diagnoses with the histological diagnoses.

Material and Methods

We reviewed the archives for 1978−2002 at the Eye Pathology Institute, University of Copenhagen. Data from the same period from all pathology departments in Denmark were identified by SNOMED (Systematized Nomenclature Of MEDicine) codes and reviewed. Thus, all surgically removed and pathologically verified lesions involving the caruncle were included in the study. The material was divided into benign, premalignant and malignant lesions (Table 1). A modified World Health Organization (WHO) nomenclature was used for histological classification (WHO 1998).

Table 1.  Caruncular lesions in Denmark 1978–2002.
CategoryNumberPercentageRecurrences/
patients
Gender
M/F
Mean age
(years)
Range
(years)
Benign
Degeneration
 Elastoid degeneration20.31/11/17372–75
 Amyloid deposits10.200/173
Inflammation
 Chronic inflammation254.41/110/155014–94
 Pyogenic granuloma71.202/5283–55
 Chalazion40.71/12/26352–81
 Foreign body30.503/04942–56
 Abscess10.200/175
 Folliculitis10.201/041
Neoplasia
Cystic
 Inclusion cyst356.11/116/195015–82
 Ductal cyst91.62/25/45937–84
Epithelial
 Papilloma12722.121/2079/48365–84
 Sebaceous gland hyperplasia315.42/219/125627–81
 Oncocytoma162.807/97351–89
 Ectopic lacrimal gland71.204/35023–80
 Sebaceous gland adenoma30.502/16362–63
 Hidrocystoma10.200/169
Lymphoid
 Reactive lymphoid hyperplasia20.301/06562–68
Mesenchymal
 Leiomyoma10.201/044
Melanocytic
 Naevus24843.25/5107/141404–89
 Primary acquired melanosis50.93/31/47150–85
Vascular
 Haemangioma91.606/34714–81
 Lymphangioma40.703/14510–73
Miscellaneous
 Normal caruncular tissue50.902/36053–71
 Dermoid tumour10.201/073
 Fibrolipoma10.200/135
 Xanthoma10.200/158
Premalignant
Neoplasia
Epithelial
 Papilloma with dysplasia40.703/14718–80
 Dysplasia20.302/13721–54
Melanocytic
 Primary acquired melanosis with atypia40.71/13/16955–85
Malignant
Neoplasia
Epithelial
 Basal cell carcinoma (primary)10.21/10/161
 Basal cell carcinoma (secondary)30.502/16561–67
 Sebaceous gland carcinoma20.302/07162–81
Lymphoid
 Malignant lymphoma40.702/26035–75
Melanocytic
 Malignant melanoma20.301/17670–82
Vascular
 Kaposi's sarcoma20.301/14036–45
Total574100.039/38289/285453–94

The following data on each case were collected: age, gender, affected side, duration, indication for surgery (malignancy, discomfort, cosmetic or fear of cancer), year of surgery, location of surgery (hospital/private practice), clinical referral diagnosis, histopathological diagnosis and recurrence.

To identify tumours with pathologically confirmed secondary invasion of the orbit or systemic spread, the National Orbital Database and the SNOMED databases were used.

Approval for the obtaining and registering of material and relevant data were obtained from the Danish Data Protection Agency and the Danish Scientific Ethical Committee.

The Year 2000 US standard population was used for age standardization. Statistical analysis was performed using Poisson regression analysis. A level of probability of p < 0.05 was considered significant.

Results

A total of 574 excised and pathologically examined lesions involving the caruncle were identified for the 25-year period 1978–2002. Of the lesions 475 (83%) were collected from the database of the Eye Pathology Institute and 99 (17%) from the SNOMED database. A total of 58% of the lesions had been excised in an ophthalmological department and 42% by ophthalmologists in private practice.

The overall frequency of surgically removed and histopathologically evaluated caruncular lesions increased significantly during the 25-year period (p < 0.001) (Fig. 1). The mean frequency was 11.4 lesions per year over the first 5-year period, which increased to 34.2 lesions per year over the last 5-year period.

Figure 1.

The frequency of caruncular lesions in Denmark during 1978–2002, showing the total number of lesions, the number of lesions registered at the Eye Pathology Institute and the number of lesions from other pathology departments. The total increase is significant (p < 0.001).

Table 1 presents the distribution of lesions, number of recurrences, gender and average age according to the 35 pathological diagnoses identified. The largest category was neoplasia (n = 522, 91%), followed by inflammation (n = 41, 7%).

The overall benign : malignant ratio was 39 : 1. The numbers of premalignant and malignant lesions were not found to increase significantly during the study period. However, the increase in the number of benign lesions was highly significant (p < 0.001) (Fig. 2).

Figure 2.

The frequency of benign and malignant lesions of the caruncle in Denmark during 1978–2002. A significant increase (p < 0.001) in the number of benign lesions is observed, whereas the numbers of malignant and premalignant lesions remain constant.

A total of 550 (96%) lesions were classified as benign.

Naevus (n = 248, 43%) (Fig. 3A, B) and papilloma (n = 127, 22%) were the most common neoplasms, followed by cysts (n = 44, 8%) and sebaceous gland hyperplasia (n = 31, 5%) (Fig. 3C, D). The inflammatory lesions were most commonly chronic (n = 25, 4%) or appeared as pyogenic granuloma (n = 7, 1%).

Figure 3.

(A) A young woman with a pigmented nodule of the caruncle. (B) Histological survey showing subepithelial nests of pigmented naevoid cells (asterisks) consistent with a naevus of intrastromal type. (Haematoxylin-eosin; original magnification × 30.) (C) An elderly man with a yellow, lobulated, solid tumour of the caruncle. (D) The tumour is composed of lobules → of closely packed cells with light cytoplasm and central, round nuclei consistent with a sebaceous gland adenoma. (Haematoxylin-eosin; original magnification × 30.) (E) A middle-aged woman with an indolent, slow growing, salmon-coloured tumour. (F) The lesion consists of closely packed lymphoid cells consistent with a mucosa-associated lymphoid tissue (MALT) B-lymphoma. (Haematoxylin-eosin; original magnification × 30.)

Premalignant lesions involving the caruncle (n = 10, 1.7%) were dominated by primary acquired melanosis (PAM) with atypia and epithelial dysplasia. Malignant neoplasms constituted a total of 14 (2.4%), with basal cell carcinoma (n = 4, 0.7%) and lymphoma (n = 4, 0.7%) (Fig. 3E, F) being the most frequent lesions in this group.

The indications for surgery/biopsy in order of decreasing frequencies were: to verify possible malignancy (in cases of a change in pigmentation, an enlarging lesion, feeding vessels, necrosis, ulceration) in 199 cases; discomfort (in cases of epiphora, impaired vision, mechanical irritation) in 50 cases; cosmetic in 20 cases, and fear of cancer in three cases. In the remaining cases no specific indication for surgery was stated.

The preoperative clinical diagnosis was found to be correct by histological examination in 286 (50%) cases. In 207 (36%) cases no clinical diagnosis was stated and 81 (14%) lesions were misdiagnosed. Table 2 shows the correlation between the histopathological and clinical diagnoses of caruncular lesions. The histopathological diagnoses of various clinically misdiagnosed lesions are shown in Table 3.

Table 2.  Correlation between histopathological and clinical diagnoses of caruncular lesions.
Pathological diagnosisClinical diagnosisNumber
 CorrectIncorrectNot specified 
Benign and premalignant
 Naevus121(49%)36(14%)91(37%)248
 Papilloma91(69%)5(5%)35(27%)131
 Cysts26(59%)7(16%)11(25%)44
 Sebaceous gland hyperplasia10(32%)5(16%)16(52%)31
 Oncocytoma0(0%)2(13%)14(87%)16
 Other28(31%)23(26%)39(43%)90
Malignant
 Malignant lymphoma4(100%)0(0%)0(0%)4
 Basal cell carcinoma3(75%)1(25%)0(0%)4
 Malignant melanoma2(100%)0(0%)0(0%)2
 Kaposi's sarcoma1(50%)1(50%)0 (0%)2
 Sebaceous gland carcinoma0(0%)1(50%)1(50%)2
Total286(50%)81(14%)207(36%)574
Table 3.  Histopathological diagnoses of clinically misdiagnosed caruncular lesions.
Clinical
diagnosis
Pathological
diagnosis
Number
  • *

    Various other diagnoses.

Incorrect
CarcinomaSebaceous gland hyperplasia2
CarcinomaChronic inflammation2
CystBasal cell carcinoma1
Cyst with haemorrhageKaposi's sarcoma1
Hypertrofic caruncleNaevus3
Hypertrofic carunclePapilloma2
MelanomaNaevus3
MelanosisNaevus11
PapillomaNaevus15
PapillomaChronic inflammation4
Sebaceous gland hyperplasiaSebaceous gland carcinoma1
XanthelasmaSebaceous gland hyperplasia2
Other*34
Total 81

In the malignant group 10 out of 14 cases (71%) were given a correct clinical diagnosis. Two sebaceous gland carcinomas, a Kaposi's sarcoma and a primary basal cell carcinoma were misdiagnosed (Table 3). The basal cell carcinoma showed orbital invasion.

Two cases of mucosa-associated lymphoid tissue (MALT) lymphoma, one case of mantle cell and one of small cell B-lymphoma were identified. Systemic involvement was later confirmed in one of the MALT lymphomas and in the case of the small cell B-lymphoma.

Discussion

The 574 caruncular lesions were sampled for a study period of 25 years and included all data available. The frequency of reported caruncular lesions increased significantly during the period under study, due to an increase in the frequency of benign lesions. Of the lesions, 83% were sampled from the Eye Pathology Institute and this proportion alone showed a significant increase. A minor growth in the population from 5.1 to 5.4 million was observed over the period (Danmarks Statistik 2004). Other causes contributing to the increase might be higher levels of awareness of neoplasms among patients and ophthalmologists leading to an increased surgical treatment and a higher degree of referral of surgical tissue for pathological examination.

During 1978−2002 a two-fold increase in the incidence rates of both melanoma and non-melanoma skin cancers was observed in Denmark (Sundhedsstyrelsen 2003). As the caruncle is skin-derived and is exposed to actinic rays, the frequency of premalignant and malignant caruncular lesions might have increased similarly. However, although the present material included five times more lesions than the largest series of caruncular lesions reported (Table 4), no increase in malignant caruncular lesions was observed.

Table 4.  Distribution of caruncular lesions in the present and previous studies.
Caruncular lesionsPresent studyLuthra et al. (1978)Shields et al. (1986)
 n = 574n = 112n = 57
  • *

    Primary acquired melanosis with and without atypia.

  • Two cases of acquired melanosis, which both later gave rise to malignant melanoma.

Benign and premalignant
 Naevus248(43.2%)48(42.9%)14(24.6%)
 Papilloma127(22.1%)15(13.4%)18(31.6%)
 Inclusion cyst35(6.1%)4(3.6%)4(7.0%)
 Sebaceous gland hyperplasia31(5.4%)9(8.0%)1(1.8%)
 Chronic inflammation25(4.4%)5(4.5%)4(7.0%)
 Oncocytoma16(2.8%)4(3.6%)2(3.5%)
 Ductal cyst9(1.6%)0(0.0%)0(0.0%)
 Haemangioma9(1.6%)3(2.7%)0(0.0%)
 PAM*9(1.6%)0(0.0%)0(0.0%)
 Pyogenic granuloma7(1.2%)3(2.7%)5(8.8%)
 Other44(7.7%)18(16.1%)6(10.5%)
Malignant
 Malignant lymphoma4(0.7%)0(0.0%)1(1.8%)
 Basal cell carcinoma4(0.7%)0(0.0%)1(1.8%)
 Sebaceous gland carcinoma2(0.3%)1(0.9%)0(0.0%)
 Kaposi's sarcoma2(0.3%)0(0.0%)0(0.0%)
 Malignant melanoma2(0.3%)2(1.8%)0(0.0%)
 Squamous cell carcinoma0(0.0%)0(0.0%)1(1.8%)
Total574(100.0%)112(100.2%)57(100.2%)

The indication for surgery was most often to rule out malignancy. A total of 71% of the lesions were from people younger than 60 years of age, with a maximum in the age group 20–29 years. Conversely, all the malignant lesions, with the exception of two cases of Kaposi's sarcoma and two of malignant lymphoma, were diagnosed in people over the age of 60 years. Fear of cancer was rarely stated as the reason for the surgical procedure.

The compound architecture of the caruncle results in a high level of diversity among lesions, which makes clinical diagnosis difficult. In the present study the overall rate of precision in clinical diagnosis was 50%, corresponding to equivalent rates in other studies (Shields et al. 1986; Santos & Gomez-Leal 1994).

An important exception was found in the group of malignant lesions. In this group 10 out of 14 (71%) lesions were diagnosed correctly. In comparison four out of six malignant lesions were clinically misdiagnosed in the study by Santos & Gomez-Leal 1994).

Naevus was the most frequent lesion found in the caruncle (Table 1). The number of correct diagnoses was surprisingly low but consistent with a previous study on conjunctival naevi (Gerner et al. 1996). The clinicians had difficulty differentiating benign naevi from premalignant and malignant pigmented lesions (Table 3). During the study period the incidence of this group of lesions increased significantly (p = 0.003). An increase was also demonstrated by Gerner et al. (1996) for the period 1960–80.

The proportion of papillomas differs from 13% to 32% between studies (Luthra et al. 1978; Shields et al. 1986). A Danish study previously found a strong association between conjunctival papilloma and certain types of human papilloma virus (HPV) (Sjö et al. 2001). Viral aetiology might explain the variation between different series and the high recurrence rate found in the present study (Table 1). The frequency of papilloma increased significantly (p < 0.0001) during the observation period.

A total of 14 oncocytomas (eosinophilic cystadenoma) were identified as representing the largest series of caruncular oncocytoma to be published. In agreement with the existing literature they were clinically characterized as reddish−bluish, slow-growing nodules found in patients over 50 years of age (range 51–89 years). The present material featured an insignificant preponderance of women, while an earlier study found that 11 out of 18 cases of oncocytoma of the ocular adnexa occurred in women (Biggs & Font 1977). In 12 out of 14 cases no clinical diagnosis was stated and two cases were mistaken for a naevus and a papilloma, respectively. None of these lesions recurred, as also found in the study by Shields et al. (1986).

Of the malignant lesions, both the malignant melanomas were diagnosed correctly (Table 2). One developed from PAM as a caruncular nodule in a 70-year-old woman. The other was found in an 82-year-old man with dysplastic naevus syndrome. Malignant melanoma involving the caruncle has a poor prognosis compared to epibulbar conjunctival lesions (Paridaens et al. 1994b).

Basal cell carcinoma accounted for four out of 14 cases (29%). The proportion was high compared to the zero found in two other series (Luthra et al. 1978; Santos & Gomez-Leal 1994) and might be a consequence of the overall increase in basal cell carcinoma observed during the study period (Sundhedsstyrelsen 2003). In the present material one basal cell carcinoma was primary (Østergaard et al. 2005), whereas three invaded from the pericaruncular skin.

Two sebaceous gland carcinomas were misdiagnosed. One was diagnosed as a benign sebaceous gland hyperplasia, the other characterized as a lobulated tumour with slow growth. Periocular sebaceous gland carcinoma is well known to masquerade as benign lesions (Doxanas & Green 1984). In a series of 88 cases of sebaceous gland carcinoma, eight were located in the caruncle (Boniuk & Zimmerman 1968).

In conclusion, caruncular lesions are rare and diverse, which makes clinical diagnosis difficult. The preoperative diagnosis was found to have been correct in 50% of cases, and 29% of malignant lesions were clinically misdiagnosed as benign. Referral of excised lesions for pathological examination should therefore be mandatory.

Acknowledgements

The authors thank Henrik Jørgensen PhD and Jens Lindegaard PhD-student for valuable advice regarding statistical analysis. Thanks are also due to Inge Gram for generous assistance in retrieving data from Patobank. This study was supported by Synoptik Fonden.

Ancillary