Clinical and histopathologic characteristics of metastatic and locally advanced cutaneous basal cell carcinomas

Abstract

Locally advanced or metastatic basal cell carcinomas (laBCC, mBCC) are rare, with few case series providing information on their epidemiology. We aimed to describe the clinical and histologic features of locally advanced and metastatic basal cell carcinomas. 40 cases of laBCC or mBCC were identified by searching Vanderbilt’s database from 1984 to January 2019. A retrospective chart review was performed. Pathology slides were available for 23 cases (13 mBCC, 10 laBCC). 21 of 23 cases were Clark Level IV or V, with a mean depth of invasion of >7mm for both types. The mean mitotic rate was 4.4 mitoses/mm² for laBCC and 3.3 mitoses/mm² for mBCC. Ulceration was identified in seven laBCC and eight mBCC cases. Perineural invasion was present in two laBCC and six mBCC cases, with three mBCC invading nerves >0.1mm. Of 13 mBCC cases, histologic subtypes included infiltrative (n=9), nodular (n=7), morpheaform (n=4), and superficial (n=2), with multiple patterns present in some specimens. 10 of 13 mBCC patients had local recurrence before metastasis. In summary, we identified several potential markers of high-risk BCC, including perineural invasion, deep invasion, elevated mitotic rate, and local recurrence of the primary tumor.

INTRODUCTION

Basal cell carcinoma (BCC) is the most common form of cancer in the United States.¹ Although more than 4 million cases of BCC are diagnosed in the U.S. annually, metastatic spread is extremely rare, with an estimated incidence of 0.0028% to 0.55%.² Historically, metastatic BCCs have conferred an extremely poor prognosis with median survival of less than one year from the diagnosis,^2,3 although a recent case series showed a more favorable median overall survival of 7.3 years.⁴

The rarity of metastatic or locally advanced BCC (mBCC, laBCC) has limited the ability to study associated histopathologic features.^5,6 Through mostly single case reports and small series, several clinical risk factors have been posited, including size of primary tumor, multiple recurrences, and certain BCC histologic subtypes.^3,7–10 Morgan et al. reported invasion beyond fat and tumor diameter ≥4 cm as predictors of metastasis and/or death in primary BCC ≥2 cm.¹¹ A study of the immunohistochemical markers in mBCC found that primary lesions of metastatic BCCs had decreased CD44 and increased Twist1 nuclear expression relative to non-metastatic lesions.¹² However, none have performed a comprehensive evaluation of the histologic parameters of mBCCs and laBCCs. We studied mBCC and laBCC cases from 1984 to 2019 to provide a detailed description of the clinical and histopathological characteristics observed.

MATERIALS AND METHODS

Following approval from the Institutional Review Board, we used the Research Derivative, a searchable mirror image of the electronic medical record of Vanderbilt University Medical Center (VUMC), to identify potential cases of locally advanced or metastatic BCC occurring between 1984 and January 2019, using the search terms “metastatic basal cell carcinoma”, “metastatic BCC”, “basal cell carcinoma metastatic”, and “BCC metastatic”.¹³ Presumptive cases were reviewed and were included if there was documentation of metastatic or locally advanced BCC and the presence of a VUMC pathology report of the primary tumor excision.

A retrospective chart review was performed to collect information regarding patients’ demographics, vital status, and clinical parameters such as primary tumor location, primary and secondary treatment modalities, recurrences, and date of metastasis. Accession numbers obtained from the pathology reports of the biopsy or resection of the primary tumor were used to retrieve the available pathology slides. Two board certified dermatopathologists (JPZ, ASB) reviewed the slides and documented the histologic subtype of BCC, tumor grade, depth of invasion, Clark level, mitotic rate, whether negative margins were achieved, and any additional histologic tumor characteristics. Clinical and histopathologic information were used to classify risk status at diagnosis (low risk, high risk, advanced disease) according to National Comprehensive Cancer Network (NCCN) guidelines.⁶

RESULTS

The initial database query yielded 74 individuals with any of the search keywords. Manual chart review of each of these identified 40 patients with confirmed metastatic (n = 23) or locally advanced BCC (n = 17). Patient characteristics are shown in Table 1. Most cases occurred in males (M:F ratio of 4:1), arising most commonly on the head and neck (73%), the upper extremities (10%) (Figure 1), and the trunk (7.5%). The median age at diagnosis of the primary tumor was 51 years (range, 26 – 79 years) for patients with mBCC and 66 years (range, 52 – 76 years) in patients with laBCC. Upon retrospective risk classification of the primary tumors using NCCN guidelines,⁶ among 23 mBCC cases, there were 1 (4%) case of low-risk disease, 15 (65%) cases of high-risk, and 5 (22%) cases of advanced disease at diagnosis. Among 17 laBCC cases, there were 2 (12%) cases of low-risk disease, 13 (76%) cases of high-risk disease, and 1 (6%) case of advanced disease at diagnosis. The median time interval between the onset of primary tumor and metastasis was 5 years.

Table 1.

Characteristics of 23 patients with metastatic basal cell carcinoma (mBCC) and 17 patients with locally advanced basal cell carcinoma (laBCC).

Characteristics	mBCC (n = 23)	la BCC (n = 17)	Total (N = 40)
Male	20/23 (87)	12/17 (71)	32 (80)
Age of onset of primary tumor, median (range), y	51(26 – 79)	66 (52 – 76)	60 (26 – 79)
Location of primary tumor
Head and neck	17 (74)	12 (71)	29 (73)
Upper limb and shoulder	3	1	4 (10)
Lower limb and hip	1	1	2 (5)
Trunk	1	2	3 (7.5)
Other/unknown	1	1	2 (5)
Risk status at diagnosis
Low	1 (4)	2 (12)	3 (8)
High	15 (65)	13 (76)	28 (70)
Advanced disease	5 (22)	1 (8)	6 (15)
Unknown	2 (9)	1 (8)	3 (8)
Interval before metastasis, median (range), y	5 (<1 – 21)	NA
Age at metastasis, median (range), y	64 (46–92)	NA

Abbreviations: mBCC, metastatic basal cell carcinoma; laBCC, locally advanced basal cell carcinoma; y, years.

Data are presented as No. (%) unless otherwise indicated.

Figure 1.

Large, ulcerated nodular basal cell carcinoma on the left arm. The patient ultimately died of a lung metastasis from this lesion

Pathology slides were available for 13 mBCC and 10 laBCC cases. Slides from the excisional specimens were examined when available (11 mBCCs, 9 laBCCs); for 2 mBCCs and 1 laBCC, the pathology slides from biopsy were examined (Figures 2 and 3). Histopathological features are summarized in Table 2. Of 13 mBCC cases, the distribution of histologic subtypes included infiltrative (n= 9), nodular (n=7), morpheaform (n=4), and superficial (n=2), accounting for multiple patterns present in specimens (Figure 2). There were 2 mBCC cases that contained exclusively low-risk histologic subtypes (superficial and nodular). Perineural invasion was noted in two laBCC and six mBCC cases, with three mBCC cases invading nerves > 0.1mm. Ulceration was present in seven laBCC and nine mBCC cases. Tumor infiltrating lymphocytes were noted in two laBCC and one mBCC cases. 21 of the 23 cases were Clark Level IV or V (Figure 2), with a mean depth of invasion of >7mm for both types. The mean mitotic rate was 4.4 mitoses/mm² for laBCC and 3.3 mitoses/mm² for mBCC. Primary tumors recurred at least once in 6 of 10 (60%) laBCC cases and 10 of 13 (77%) mBCC cases before documentation of metastasis.

Figure 2.

Primary tumor histology of metastatic and locally advanced basal cell carcinomas. A, B) Large, broadly invasive BCC; C) perineural invasion; D) Invasion of BCC into muscle; E) Invasion around fat; F) Invasion around parotid gland.

Figure 3.

Histology of basal cell carcinoma in the lymph nodes

Table 2.

Histologic features of 13 metastatic and 10 locally advanced basal cell carcinoma cases.

Characteristics	Metastatic BCC (n=13)	Locally Advanced BCC (n=10)	Total (N=23)
Location of primary tumor
Head and neck	11 (85%)	6 (60%)	17 (74%)
Trunk	0	2 (20%)	2 (9%)
Lower limb	1 (8%)	1 (10%)	2 (9%)
Upper limb	0	1 (10%)	1 (4%)
Unknown	1 (8%)	0	1 (4%)
Risk Status at diagnosis
Low	1 (8%)	2 (20%)	3 (13%)
High	10 (77%)	7 (70%)	17 (74%)
Advanced disease	2 (15%)	1 (10%)	3 (13%)
Histologic subtype
Infiltrative	9	6	17
Nodular	7	8	15
Superficial	2	2	4
Morpheaform	4	4	8
Tumor grade
Well differentiated	1 (8%)	1 (11%)	2 (9%)
Moderately differentiated	10 (77%)	8 (89%)	18 (82%)
Poorly differentiated	1 (8%)	0	1 (5%)
Undifferentiated	1 (8%)	0	1 (5%)
Tumor Characteristics
Ulcerated	8	7	15
Acantholytic	1	0	1
Perineural invasion	6	2	8
Perineural invasion >0.1mm	3	0	3
Tumor infiltrating lymphocytes	1	2	3
Present at base	8	7	15
Present at edges	9	8	17
Satellite lesions	2	1	3
Pigmented	0	1	1
Tumor Clark Level
III	1	1	2
IV	6	4	10
V	5	5	10
Depth of invasion mm, mean (median, range)	7.8 (7.2, 0.7 – 17)	7.9 (7.4, 1.9 – 15)	7.8 (7.2, 0.7 – 17)
Mitotic rate, mean (median, range)	3.3 (2, 0 – 10)	4.4 (3, 1 – 11)	3.8 (2, 0 – 11)
Thickness mm, mean (median, range)	6.4 (4.7, 0.7 – 17)	8.8 (10, 3.2 – 15)	7.5 (5, 0.7 – 17)
Width mm, mean (median, range)	12.6 (12, 2 – 24)	16.7 (16.5, 11 – 22)	14.6 (16, 2 – 24)

DISCUSSION

In this study spanning nearly 35 years, we identified and characterized 40 cases of metastatic or locally advanced BCCs. We found that over half of the mBCC cases exhibited an infiltrative subtype, demonstrated perineural invasion, or recurred locally before metastatic spread.

Historically, metastatic BCC occurs more frequently in males, most commonly arising on the head and neck region (64–85%),^2–3,11 which is consistent with the results of our study, where 87% of mBCCs occurred in men and 74% originated from tumors on the head and neck. The overall frequency of BCC histologic subtypes is estimated to be 50–79% nodular, 15% superficial, and 5–10% morpheaform, sclerosing, or infiltrative followed by the rarer micronodular and metatypical subtypes.¹⁴ In our cohort, infiltrative subtype was overrepresented in both mBCC (69%) and laBCC (60%). Mochel et al.¹² and Bransen et al.¹⁵ similarly identified a high proportion of infiltrative subtype in metastatic BCC cases at 89% and 50%, respectively. Morpheaform BCCs have a reported metastatic rate of 0.28% in a review of 5270 cases and was found in 11 of 12 mBCC cases in another case series,^3,16 compared with 2 of the 13 mBCC specimens in this series. Although a previous literature review suggested that metatypical morphology was present 14% of mBCCs and associated with increased risk of metastasis,² there were no cases of metatypical histology among the 13mBCC cases in this series. Of note, our chart review found 2 laBCC cases with metatypical features. However, the pathology slides for these 2 cases were not available for review.

According to NCCN guidelines, aggressive growth pattern is defined as having infiltrative, micronodular, morpheaform, basosquamous, sclerosing, or carcinosarcomatous features in any part of the tumor.⁶ In this study, recurrence occurred in 2 of 3 BCCs that were classified as having low risk for recurrence by NCCN guidelines. Furthermore, 2 mBCC cases contained exclusively low-risk histologic subtypes (nodular and superficial). Of these two cases, one was 2.2 cm in width and ulcerated. Examination of the second case was limited to the pathology slides from the initial biopsy. We were thus unable to identify additional risk factors for metastasis, including potential deeper aggressive patterns. An additional 2 purely nodular mBCCs were among the cases for which pathology slides were unavailable. The one with more complete data was a large, ulcerated nodular BCC that had been neglected for years in the setting of no insurance and untreated hepatitis C infection (Figure 1). These findings encourage further study into the prognostic value of BCC risk classifiers.

Metastatic BCCs have also been associated with tumor diameter.^3,11 Morgan et al. found that 14 of 16 (88%) metastatic BCC cases originated from primary tumors ≥ 4cm and that no metastatic cases emerged in the 248 BCCs <2 cm.¹¹ In contrast, our study found that nearly all of the primary tumors were < 2.5 cm, with a median tumor size of 1.2 cm, indicating that smaller tumors also have metastatic potential.

The incidence of mixed histology in basal cell carcinomas have been reported to be 30% and 38.5%.^17,18 Sloane observed a 50% recurrence rate in tumors with mixed patterns (3 of 6 BCCs).¹⁹ In our study, mixed histology was seen in 2 of 10 laBCCs (20%) and 9 of 13 mBCCs (69%). The presence of mixed histology, particularly of aggressive growth patterns, likely plays a role in metastasis and could explain our observation of metastasis in smaller tumors. Of note, Cohen et al.¹⁷ found that of the basal cell carcinomas that were found to have mixed histology on excision, only 10% of the original biopsy pathology reports had descriptions of multiple patterns of growth. In our study, 3 cases (1 laBCC, 2 mBCC) had only biopsy slides available for review. Morpheaform features were observed in the biopsy slides of 1 mBCC; the biopsy slides for the other 2 cases contained a mix of superficial and nodular features. It is possible that more aggressive histology was present in the deeper sections of the tumor that were not captured by biopsy.

Locally advanced tumors and metastatic ones had similar mean depth of invasion (7.9mm and 7.8mm, respectively). One study of more than 4,500 BCCs found a mean depth of invasion of aggressive subtypes to be only 1.51mm.²⁰ Depth of invasion could potentially be a useful predictor of aggressive behavior, although further study is needed to confirm a potential association.

Perineural invasion (PNI) is a well-known and described high-risk feature in cutaneous squamous cell carcinoma (SCC).²¹ Similarly, PNI in BCC is presumed to be found in higher risk tumors. A histologic review from one institution of all BCCs with aggressive histologic subtypes showed PNI in 3% of 507 cases.²² Both Massey et al.²³ and Morgan et al.¹¹ found that PNI did not independently predict BCC outcomes, though the latter study reported a 30% risk of metastasis/death in BCC lesions with PNI (9 of 30 cases). Our data showed PNI was a common finding in metastatic BCC (9 of 13 cases,69%).

Evaluation of mitotic rate is rarely reported in BCCs and even more poorly described in aggressive tumors. Curiously, metastatic tumors had a lower mitotic rate (3.3 mitoses/mm²) than locally advanced ones (4.4 mitoses/mm²). Both values fall within the range that has been previously reported for non-metastatic BCC (1.7 – 5.4),^24,25 limiting the utility of mitotic rate in predicting metastasis. Both Mooney et al. and Al-Sader et al. had sought to elucidate the reason behind BCCs’ limited metastatic potential.^24,25 Mooney et al. found that compared to melanomas, BCCs have higher apoptotic index, which was hypothesized to account for the overall more indolent nature of BCCs.²⁴ Al-Sader et al. found that compared to SCCs, BCCs have similar mitotic index and apoptotic index, but lower proliferation rate, suggesting that the difference in cell proliferation contribute to the lower risk of metastasis in BCCs.²⁵ Neither publication, however, detailed additional tumor characteristics such as subtype or clinically aggressive behavior.

Immunosuppression and prior radiotherapy at the site are considered risk factors for BCC development.^26–30 None of the patients included were immunosuppressed at the time of development of the mBCC nor had prior radiation to the area that the tumor arose from. Furthermore, none of the patients had a documented history of any genetic syndrome that would predispose them to developing numerous BCCs.

One limitation of this study was our specific use of the search term ‘metastatic’ and omission of the term ‘locally advanced’. Thus, it is possible that additional laBCC cases were not identified through database review. Our primary goal was to describe the histologic features of mBCC, not treatment modality or long-term outcome. While we collected treatment information where available, the timing of treatment such as chemotherapy or radiation was not compared to date of pathology specimen collection. Any potential effect of treatment on tumor pathology was not assessed. Many of the lesions occurred >10 years ago, and thus neither pathology slides nor detailed data regarding clinical characteristics or treatment could be obtained.

In summary, we provide a comprehensive report of the histopathologic features observed in metastatic BCCs in one of the largest series of metastatic and locally advanced BCC to date. We describe several potential markers of high-risk BCC, including perineural invasion, deep invasion, elevated mitotic rate, and local recurrence of the primary tumor.