Methods

Charts of all patients with periocular BCC referred to the Orbital Unit of the University of Naples “Federico II” between January 1984 and December 2003 were reviewed. The diagnosis of BCC was confirmed by pathologic analysis in all patients. Patient demographics, previous treatments, tumor site, clinical presentation, duration of symptoms, and histologic subtype were analyzed and compared between all patients with BCC and those requiring exenteration to determine their significance as risk factors for orbital exenteration. Surgical procedures for local excision were performed by different surgeons (D.S., A.I., G.U., G.B.) using the following standard procedure. Clinically clear margins were marked around the tumor by noting the transition in surface contour, vascularization, skin color, and surface texture. The extent of tumor depth was judged by the macroscopic features of the excision plane, and margins of 3 mm were marked with the skin under tension. Fast-track paraffin section margin control was used in all patients treated with local excision. Reconstruction was performed immediately using direct closure when possible and was carried out with flaps or grafts after the results on margins were known. Exenteration was performed in patients with clinical and radiologic evidence of orbital invasion. The outcome measures were recurrence rates, tumor-related deaths, orbital infiltration, and rate of exenteration. Institutional review board approval was obtained, and each patient gave informed consent. Associations between categorical variables were analyzed using the chi-square test. All analyses were performed using MedCalc software version 11.2 (MedCalc Software, Mariakerke, Belgium).

Results

Data, including follow-up, were available for all 506 patients (52% male) identified during the defined study period. The mean patient age was 67.2 years (range, 47 to 84 years). All patients were white. The mean follow-up period was 3.2 years (range, 0.3 to 15 years). A total of 28 patients (5.5%) underwent orbital exenteration. Of these 28 patients, 15 were males (53.6%) and 13 were females (46.4%), with a mean age of 68.4 years (range, 46 to 84 years).

Before exenteration, 20 patients (71.4%) underwent 1 or more surgical excisions for BCC. For 16 patients (80%), initial therapy was provided elsewhere, whereas 4 patients (20%) received initial treatment at the study institution. The mean time from the first excision to the diagnosis of orbital invasion was 92 months (range, 6 months to 30 years). The average number of recurrences requiring surgical excision was 3.6 (range, 3 to 10). The remaining 8 patients (28.5%) had not undergone any surgery before exenteration. Six patients had extensive disease involving the orbit at presentation. In 1 patient, the previous treating physician had underestimated the extent of the tumor, and the remaining patient had been treated with external radiotherapy without excisional biopsy.

All of the patients requiring orbital exenteration had a visible or palpable mass. The most common signs of orbital invasion were bone fixation of the mass (40.6%), followed by limited ocular motility (31.3%), and globe displacement (18.7%). For 10 patients (31.3%), only a visible or palpable mass was noted, with no evident orbital involvement. Computed tomography imaging data were available for 26 patients, and orbital extension was apparent in 23 patients (82%). Signs of bone involvement were evident on computed tomography scans in 8 patients (28.5%).

Among all 506 patients, the most common periocular tumor site was the lower eyelid (58.1%), followed by the medial canthus (25.1%). The right and left side of the face were affected almost equally (49.6% and 50.4%, respectively). In the 28 patients who underwent exenteration, the most common initial tumor site was the medial canthus (53.6%), followed by the lower eyelid (35.7%), the upper eyelid (7.1%), and the lateral canthus (3.6%; P = .0007).

The distribution of histologic subtypes among the 28 patients undergoing exenteration was as follows: infiltrative (78.6%), basosquamous (14.2%), morpheaform or sclerosing (3.6%), and nodular (3.6%). This distribution of histologic subtypes in these 28 patients was significantly different compared with the histologic distribution for remaining 478 patients with periocular BCC ( P = .0001). Initial margin involvement data were available for 26 patients; in 19 patients (73%), the margins were not clear ( Table 1 ).

Twenty patients (71.4%) underwent exenteration after 1 or several conservative surgical excisions. In 8 patients (28.6%), exenteration was performed as the first surgical treatment because of the evident orbital invasion at the time of initial presentation; 2 of these patients (7.2%) underwent postoperative radiotherapy. In 14 patients (50%), orbital exenteration was carried out after 1 or more surgical excisions. Six patients (21.4%) underwent surgical excision and radiotherapy before exenteration ( Table 2 ).

In the 20 patients who underwent exenteration after other surgical excisions, 16 (80%) showed no margin control. In contrast, of the 478 nonexenterated patients who underwent 1 or more surgical excisions, 80 (16.7%) had uncontrolled margins ( P < .0001, between-group comparisons).

Postoperative orbital margin data were available for 24 of the 28 patients who underwent exenteration. In 12 patients (50%) the margins were clear of tumor. However, in the remaining 12 patients, the margins were not clear, and 7 of these patients were treated with adjuvant radiotherapy to control the disease. Four patients declined further surgical procedures.

Postexenteration recurrence was diagnosed in 8 patients (28.5%); 7 (87.5%) had positive margins at exenteration, and 4 had undergone radiotherapy. In these patients, further surgical excision was necessary. Of the 28 exenterated patients, 2 (7.1%) died as a result of BCC, both as a result of intracranial invasion. Both patients had a history of recurrent tumor, had a medial canthal infiltrative BCC, and had undergone radiotherapy after orbital exenteration.