Aggressive Squamous Cell Carcinomas in Persons Infected With the Human Immunodeficiency Virus

Objectives  To illustrate the potential for aggressive growth of cutaneous squamous cell carcinomas (SCCs) in patients infected with the human immunodeficiency virus (HIV) and to determine the factors associated with increased morbidity and mortality from aggressive SCCs in HIV-infected patients.

Design  Retrospective nonrandomized case series.

Setting  University-based dermatologic referral center.

Patients  A consecutive sample of 10 patients infected with HIV who had "aggressive" SCC based on the following criteria: diameter larger than 1.5 cm, rapid growth rate, local recurrence, and/or evidence of metastasis.

Main Outcome Measures  Morbidity and mortality.

Results  Five patients died of metastatic SCC within 7 years of their initial diagnosis despite treatment. Human immunodeficiency virus stage and the degree of immunosuppression were not associated with increased morbidity and mortality. Patients initially undergoing combination surgery and radiation therapy or radical neck dissection had the best outcomes.

Conclusions  Patients infected with HIV can develop rapidly growing cutaneous SCCs at a young age, with a high risk of local recurrence and metastasis. High-risk SCCs should be managed aggressively and not palliatively in patients infected with HIV.

ACTINICALLY INDUCED cutaneous squamous cell carcinoma (SCC) in individuals with fair skin and long-term sun damage has been typically slow growing, easily treated, and rarely metastatic (0.3%-3.7% for low-risk cutaneous tumors).¹ More aggressive behavior has been correlated with multiple factors, including anatomic site, size at presentation, growth rate, histologic features, and recurrence after treatment. Higher rates of aggressive growth and metastasis occur in immunocompromised patients, especially organ transplant recipients and patients with lymphoproliferative disorders.^1-6 To our knowledge, similar increased rates of SCCs or more aggressive behavior have not been reported in human immunodeficiency virus (HIV)–infected patients.

We reviewed the medical records of 10 patients who had "aggressive" SCC based on the following clinical criteria: rapid growth rate noted clinically (doubling in size over several months), a diameter larger than 1.5 cm, a history of recurrence, and/or evidence of metastasis.³ The 10 patients identified were seen at the Department of Dermatology, University of California, San Francisco, between January 1, 1996, and December 31, 1998.

In the 10 HIV-infected patients identified with aggressive SCCs, information regarding sex, age, race, and HIV status (CD4⁺ cell count, history of opportunistic infections, history of Kaposi sarcoma, and antiretroviral medications) was recorded. Data on other cancers, including previous basal cell carcinomas (BCCs) and "nonaggressive" SCCs, their location, and their management were documented. Data on the aggressive SCCs, including location, depth, diameter, growth rate, and previous treatment, were recorded. Length was defined as the longest diameter of the tumor or final micrographic surgical defect. Width was defined as the longest measurement perpendicular to the length. Thickness was measured microscopically from the surface of the lesion to the deepest part. Growth rate was determined by subjective observation by either the patients or their physicians.

One pathologist (M.E.M.) performed the histologic examination of the primary lesion and regional metastasis. Squamous cell carcinomas were classified according to their differentiation: well, moderately, or poorly differentiated. Well-differentiated tumors had large polygonal cells with a refractile eosinophilic cytoplasm. Moderately differentiated tumors had more infiltration, less well-demarcated cells or spindle cells, and a less refractile cytoplasm. Poorly differentiated cells were highly infiltrative, poorly identifiable cells, often requiring immunostaining techniques for identification. The invasion of adjacent structures was determined by histologic and/or clinical observation. The following 2 patients were representative of patients who had different outcomes from their aggressive SCCs.

This patient was a 44-year-old HIV-infected white man (CD4⁺ cell count, 391/µL) who died of metastatic SCC. His medical history was notable for psoriasis, hypertension, and hepatitis A, B, and C; he had no history of opportunistic infections or other conditions defining the acquired immunodeficiency syndrome. He was not taking antiretroviral medications.

The patient first noted a lesion in his upper lip in 1993. This was diagnosed clinically as an actinic keratosis and was treated with liquid nitrogen and tretinoin (Retin-A). In June 1993, the lesion recurred. A skin biopsy specimen showed a well-differentiated SCC with perineural involvement. A reexcision demonstrated invasion of the underlying skeletal muscle, with tumor extending to the surgical margins. In January 1994, the area was again surgically excised and the results of a microscopic examination showed an SCC with positive peripheral margins and perineural spread. The patient underwent radiation therapy of the left nasal vestibule ala and upper lip until April 1994. Because the tumor continued to enlarge, a magnetic resonance imaging scan was obtained and confirmed massive recurrence, with tumor in the left maxillary area and metastasis in the left submaxillary lymph node. A soft tissue mass was invading the anterior wall of the left maxillary sinus and projected into the sinus cavity. Extension was also noted across the left upper lip passing the midline, with a 1-cm lesion crossing over to the right upper lip. The patient underwent palliative radiation therapy to the left maxillary field (3000 rad [30 Gy]) and to the left neck field (3000 rad [30 Gy]), with minimal response.

In January 1995, the patient had a 6 × 8-cm crateriform lesion extending from the left side of his nose onto his left cheek and left upper lip. The patient's tongue and teeth were visible through the cavity. There was erythema and induration extending laterally over the entire left cheek to the border of the left eye. There was a 4 × 4-cm firm left submandibular mass. The patient died shortly thereafter of metastatic disease.

A 42-year-old white man with acquired immunodeficiency disease syndrome (with a CD4⁺ cell count of 174/µL and a history of Pneumocystis carinii pneumonia) presented with a nonhealing ulcer in the left preauricular area. For the past 5 months, he had been treated for diffuse facial molluscum with cryotherapy and curettage. When the treated area failed to heal, a biopsy was performed and a moderately differentiated SCC was found.

On examination, a 1.5-cm round granulating ulcer was present in the left preauricular area. There was no lymphadenopathy. Mohs surgery was performed; this cleared the lateral margins but revealed invasion of the SCC into the parotid fascia. Perineural inflammation was noted, but there was no definite evidence of perineural invasion. Two weeks after the surgery, the patient underwent radiation therapy (1500 rad [15 Gy]) directly to the wound bed.

During the next 2 months, the 3-cm surgical defect over the parotid gland healed completely by secondary intention. There was no evidence of recurrence or metastasis after 3 years. Since his surgery, the patient has started undergoing highly active antiretroviral therapy. He remains free of disease.

Demographic data are shown in Table 1. This study includes 10 HIV-infected white patients (9 men and 1 woman). The mean age of disease onset was 44 years. Eight patients met the Centers for Disease Control and Prevention criteria for acquired immunodeficiency syndrome. Four patients were asymptomatic; the remainder had symptomatic HIV disease. Four patients had a history of opportunistic infections, including 1 who had Kaposi sarcoma. The mean CD4⁺ cell count was 239.7/µL (median, 244.5/µL). Five patients had a CD4⁺ cell count greater than 200/µL. Seven patients were undergoing antiretroviral therapy, with 3 taking protease inhibitors.

Human immunodeficiency virus status and the degree of immunosuppression did not differ significantly among patients who died of their cancers and those who survived. They also did not differ among patients who developed metastatic or recurrent disease and those who did not. These tumors occurred in patients with all stages of HIV infection and a wide range of CD4⁺ cell counts.

The 10 patients had a total of 41 SCCs. Five patients had multiple SCCs; the largest number of SCCs in a single patient was 14. The SCCs occurred more commonly in the head and neck (31 [76%]) than on the trunk (7 [17%]) or the extremities (1 [2%]). Of the 41 SCCs among the 10 patients, 10 were considered aggressive; each patient had only 1 aggressive SCC. Three patients had a history of BCCs, 1 with 8 BCCs.

Figure 1 and Figure 2 show the appearance of 2 aggressive SCCs at initial clinical presentation.

Tumor characteristics are listed in Table 2. All of the aggressive tumors were located in the head and neck, 9 on the face and 1 on the scalp. Half of the tumors were located in an area of skin over the temporal bone extending to the zygomatic arch anteriorly and inferiorly and the external auditory canal posteriorly; one was located on the external ear. Five patients had a period of extremely rapid tumor growth. The average tumor diameter was 4.1 cm. The average tumor thickness on initial presentation was 7.2 mm (range, 2.0-16.5 mm). Five tumors were well differentiated, 4 had moderate differentiation, and 1 had poor differentiation. Half of the tumors had perineural invasion. All tumors extended to at least the level of the reticular dermis.

Seven patients had at least 1 local recurrence, and 3 had multiple recurrences. Of these 7 patients, 6 had proved nodal (n = 2) or distant (n = 4) metastasis; 1 additional patient had distant metastasis at initial presentation. Patients with recurrent or metastatic disease had larger thicker tumors and evidence of perineural involvement. The 5 patients who died of their metastatic cancers died within 7 years of their initial cancer diagnosis and treatment (range, 6 months to 7 years after the diagnosis). Two patients treated with surgery and radical neck dissection had no evidence of disease after follow-up. The mean follow-up for 9 of the 10 patients was 12.1 months.

Table 3 details the treatment history of the patients, and Table 4 shows the outcome by treatment method. A total of 18 procedures were performed on these tumors. Three patients treated initially with combination therapy were cured: 1 was treated by excision with radical neck dissection and parotidectomy, and 2 underwent surgery (including Mohs surgery) and radiation therapy. The other 2 patients who were initially treated with excision and radiation therapy died of their disease: one patient's tumor was poorly differentiated, and the other patient's tumor was located on the ear, a site of known increased metastasis and recurrence. Mohs surgery, local excision without radical neck dissection, and other single-treatment modalities used as initial treatment resulted in recurrence in all patients and metastatic disease in 3 of 4 patients. Of the 2 patients who underwent Mohs surgery, one had clinical evidence (ie, pain) of probable perineural invasion before therapy that was not identified by fine-needle aspiration or magnetic resonance imaging. The patient refused to undergo a biopsy. Later, bony invasion of the sinuses and orbit occurred. The other patient had a recurrence after Mohs surgery that was treated with radiation therapy. After radiation therapy, the patient was lost to follow-up. He died 4 years later. According to the patient's death record, he died of peritonitis and sepsis; a metastatic SCC was not part of his final diagnosis. Two patients who eventually developed metastatic disease were either initially misdiagnosed or treated nonsurgically. Of the 7 patients with recurrence after initial treatment, 4 underwent single-modality treatments. These patients experienced continued recurrence, developed metastatic disease, and eventually died. After an initial recurrence with nonsurgical therapy, one patient who then underwent surgery and radical neck dissection continued to do well 4 months after follow-up. One patient died after the initial recurrence, and one was lost to follow-up.

This study illustrates the potential for rapid growth of SCCs in the setting of HIV infection. The 10 patients in this series were relatively young and healthy, yet half of them died of their cancers. Patients who had high rates of recurrence and metastatic disease and eventually died of their cancers were not initially treated aggressively with either combination surgery and radiation therapy or surgery and radical neck dissection.

Studies^3,7-10 in patients who have undergone transplantation have shown an increased incidence of SCCs, an increased SCC/BCC ratio, and an increased number of SCCs per patient compared with the general population. In some studies,¹¹ the mortality from SCCs in patients who have undergone transplantation has approached 5% compared with less than 1% in immunocompetent patients. Although studies^12,13 have not shown an increased SCC/BCC ratio in the HIV-infected population, these SCCs have occurred at a much younger age (mean, 44 years) compared with those of the general population (mean age, 70 years) seen at our institution. Patients in our series also experienced significant morbidity and mortality from their SCCs during a short period (50% mortality within 41 months [range, 6-84 months] after the diagnosis).

In organ transplant recipients, the incidence of SCCs increased with the level of immunosuppression and the time after transplantation.^3,7-10 In this series, tumors occurred in patients with all stages of HIV disease. Local recurrence, metastasis, and survival did not correlate with the number of opportunistic infections and the CD4⁺ cell count. Thus, SCCs in HIV-infected patients should be treated aggressively regardless of the patients' degree of immunosuppression and HIV prognosis.

In this study, morbidity and mortality were more dependent on the initial control of local and metastatic disease than on the level of immunosuppression. In the general population, Mohs surgery has been the treatment of choice for high-risk tumors. The recurrence rate for those undergoing Mohs surgery was 2% to 3% compared with 13% for those undergoing non–Mohs surgery modalities. For recurrent tumors, the recurrence rate was 10% with Mohs surgery but 23.3% for all non–Mohs surgery modalities. The presence of perineural invasion has also prompted the addition of adjuvant therapy.^3,5,14 Similarly, high-risk SCCs in the HIV-infected population should be treated initially by ablative therapy with histologic control and, if necessary, adjuvant therapy. Ablative therapy without histologic control should be discouraged.

Although this is an uncontrolled retrospective case series, the strikingly aggressive nature of these tumors in a relatively young and healthy population warrants report. Patients in this series demonstrated a higher morbidity and mortality from their SCCs than from infection with HIV. Patients infected with HIV now live longer and healthier lives but can die of a metastatic SCC, a potentially curable disease. Thus, the treatment of cutaneous SCCs in HIV-seropositive patients should be at least as aggressive as the treatment in HIV-seronegative patients. Adjunctive modalities, such as local and regional radiation therapy and sentinel lymph node biopsy, should be considered for high-risk tumors, analogous to the management of SCCs in organ transplant recipients. Primary prevention with sunscreen and sun avoidance in addition to aggressive management of precancerous lesions should be recommended for all HIV-seropositive patients.

Accepted for publication October 3, 2001.

We thank Toby Maurer, MD, for her assistance in editing the manuscript.

Corresponding author: Kirsten Vin-Christian, MD, Department of Dermatology, University of California, San Francisco, 1701 Divisadero St, Third Floor, San Francisco, CA 94143 (e-mail: vinchristian@orca.ucsf.edu).