Incidence of NMSCs in the United States

A 1971 to 1972 National Cancer Institute (NCI) survey from 4 geographic regions in the United States estimated the national incidence of NMSCs to be around 300,000 new cases per year, representing about 50% of the incidence in total cancers in the United States. Because of concerns of ozone depletion and the association between UV exposure and skin cancer, the NCI conducted a second survey from 1977 to 1978 that expanded to 8 geographic regions to evaluate dose-response effects between UV exposure and skin cancer incidence. The incidence of NMSCs from this period was estimated to be 233 per 100,000 per year (480,000 persons).

A 1994 update used a data registry from Kaiser-Permanente, a health maintenance organization (HMO) in Portland, OR, USA, to project age-adjusted incidences of BCCs and SCCs using linear and log-linear regression analyses. These analyses showed that national estimates for the incidence of NMSCs ranged from 0.9 to 1.2 million, again nearly totaling the incidence of all other cancers combined.

The most recent national estimates of NMSC incidence were determined to be 2,152,500 treated persons and 3,507,693 total number of NMSCs in 2006 based on 2 Medicare databases (2006 Total Claims Data Set and the Medicare Limited Data Set Standard Analytic File 5% Sample Physician Supplier Data) as well as the National Ambulatory Medical Care Survey (NAMCS). These databases provided information on skin cancer procedures for Medicare patients and NMSC-related office visits.

Incidence by Skin Cancer Type

BCCs account for approximately 80% of NMSCs, although the ratio of BCCs to SCCs declines in regions closer to the equator, with increasing age, and with male gender. For BCCs, the male to female ratio is 1.4 to 1.6, whereas for SCCs, this ratio is 2.8. The overall lifetime risk for developing NMSC is 1 in 5. The lifetime risk for developing BCCs (28%–33%) is higher than for developing SCCs (7%–11%).

Approximately 80% of BCCs are found on the head and neck in the United States, with the trunk as the second most common site ; this suggests that chronic sun exposure and intermittent sun exposure from recreational activities increase the risk of developing BCCs. Variations in BCC anatomic distribution can differ by gender, ethnicity, and geographic location.

SCCs have a greater capacity for invasion and metastasis and account for most of NMSC-related mortality. About 70% to 80% of SCCs occur on the head and neck, followed by the upper extremities as the next most common location, indicating that SCC distribution is more strongly correlated with sun-exposed skin. In combination with the increased SCC incidence with age and decline in latitude, cumulative UV exposure is a strong risk factor for cancer development. However, SCCs occur with greater frequency in areas of non–sun-exposed skin in nonwhite populations, particularly African Americans.

Time Trends: NMSC Incidence is Increasing

Numerous reports indicate that NMSC incidence is rising. Two of the eight geographic regions surveyed in 1977 to 1978 were also previously surveyed in 1971 to 1972. During this 6-year period, there was a 15% to 20% increase in the age-adjusted rates of NMSCs. In the 2006 national estimates, the total number of NMSC-related procedures increased 76.9% from 1,158,298 in 1992 to 2,048,517 in 2006.

Other geographic regions have also reported increases in age-adjusted BCC and SCC incidence rates with varying effects by gender and anatomic distribution ( Table 1 ). Overall, BCC incidence increased from 0.95% to 5.30% per year, whereas SCC incidence increased from 4.0% to 23.3% per year. Increases in NMSC incidence, however, are not geographically universal and may be due to more aggressive treatment of precancerous actinic keratosis and increased sun-protective behavior in higher risk areas.

These trends have been attributed to greater UV exposure from ozone depletion, greater sun-seeking behavior and exposure, and longer lifespans of the general population. It is also recognized that greater public awareness of skin cancer as well as changes in medical practice, such as more routine biopsies of suspicious lesions, could have led to a perceived increase in incidence. Limitations in data collection may overestimate and underestimate NMSC cases, which can lead to inaccurate conclusions of changing patterns.

NMSC Incidence and Geographic Location

Given the correlation between skin cancer and degree of UV radiation, geographic locations with a greater proximity to the equator and a subsequent higher UV-B index have higher rates of NMSCs. Of the 4 locations surveyed from 1971 to 1972, the age-adjusted incidence of NMSCs was highest in Dallas-Fort Worth at 379 per 100,000 persons. Furthermore, patients in Dallas-Fort Worth had more SCCs relative to BCCs, faster rates of incidence at younger ages, and a significantly higher peak in those aged 75 to 84 years, again underscoring the relationship between UV exposure and NMSCs.

The 1977 to 1978 NCI study surveyed 8 regions that ranged from 47.5° latitude to 30.0° latitude. Geographic trends showed that BCC incidence was 2 to 3 times greater in southern latitudes than at sites of northern latitude, while SCC incidence was about 5 times greater in southern regions. Similar to the 1971 to 1972 study, the ratio of BCCs to SCCs decreased with increased UV-B index. Incidence rates also increased more sharply at younger ages in areas of higher UV-B index.

Individual population studies also support the correlation among geographic region, UV exposure, and skin cancer. By the mid-1980s, multiple epidemiologic studies in Hawaiian populations indicated that published rates of NMSCs in Kauai, a latitude of 22° north, were the highest in the country (922 per 100,000 persons). Compared to a population from Rochester, MN, USA, men in Kauai had a BCC incidence rate 3.3-fold greater than men in Rochester, whereas women in Kauai had a rate 2.4-fold greater. Incidence of Bowen disease was 10-fold greater in Kauai compared with the Rochester population.

The relationship between UV dose response and NMSC incidence can be quantified and used to predict rate increases in different geographic locations. Exponential models correlating incidence with UV-B index indicate that a 1% increase in UV-B index could lead to a 2% increase in NMSC incidence. Rates of SCC are subject to even greater increases by a UV gradient, with incidence doubling for every 8° to 10° decline in latitude.

NMSC Incidence and Age

The average age of NMSC diagnosis ranges from 59 to 65 years for men and 60 to 66 years for women, depending on location. In the 1977 to 1978 NCI study, NMSC incidence was 4 to 8 times higher in those aged 55 to 75 years compared to those younger than 20 years. Incidence rates tended to increase significantly around the fifth decade and peak during the seventh and eighth decades. Rates of SCCs generally increase rapidly with age compared with BCCs. However, little is known about NMSCs in younger patients, and recent studies have presented conflicting data.

Estimated age-specific and gender-specific incidences of NMSCs in a Minnesota population younger than 40 years indicated that the incidence of both BCCs and SCCs had significantly increased during the study period between 1973 and 2003. Although greater public awareness and surveillance may lead to increases in incidence, the investigators note no association between tumor size and year of diagnosis or age at diagnosis. They hypothesize that increased sun exposure, the use of tanning, and tobacco use may be other contributors to higher rates of NMSC in younger populations.

In contrast, multivariate analysis using national data from the NAMCS during the same study period, 1973 to 2003, revealed no significant increases in NMSC-related outpatient office visits in individuals younger than 40 years, adjusted for gender, race, and geographic location. In this study, the mean age of a patient with an NMSC-related office visit was 64.7 years in 1979 and 69.0 years in 2003. Further epidemiologic studies in younger populations would no doubt be beneficial.

NMSC Incidence and Race/Ethnicity

One of the major limitations of prior estimates is that NMSC incidence generally refers to white populations because of significantly lower rates of incidence in nonwhite individuals. This limitation is illustrated in the 1978 NCI survey in which the annual age-adjusted incidence rate for African Americans was 3.4 per 100,000 persons compared to 232.6 per 100,000 persons for white populations. In Hispanics, NMSC rates were 6 to 8 times less than that of white populations in New Mexico and 11 to 14 times less in southeastern Arizona. Lower skin cancer rates are attributed to increased epidermal melanin production in darker-skinned individuals, which can confer a photoprotection factor of up to 13.4. However, the importance of epidemiologic studies in non-white populations is increasingly being recognized because morbidity and mortality are often greater.

In African Americans, SCCs are more commonly found than BCCs, particularly in non–sun-exposed areas such as the lower extremities, anogenital area, and feet. Specifically, SCCs can occur 8.5 times more frequently than BCCs, which suggests that UV exposure may not play a significant role in SCC development in African Americans. Rather, risk factors for SCC development in African American populations include chronic scars, burns, chronic leg ulcers, and chronic inflammation. The tendency of these cancers to arise in non–sun-exposed areas and in areas of chronic trauma or inflammation may be reasons for the greater morbidity and mortality in this population because cancers are more aggressive or are diagnosed at later stages.

In contrast, BCCs in darker-skinned individuals tend to have an anatomic distribution similar to lighter-skinned populations. Of the BCC cases seen in African Americans in a Howard University Hospital study, 60% of patients were considered to be of fair complexion and 7% of olive complexion. This is compared to the control group in which only 10% of patients were of fair or olive complexion, highlighting the association between UV exposure and BCC cases. In other ethnic groups, comparisons of ethnic Japanese in Kauai standardized to Japanese populations show that rates of BCCs and SCCs were at least 45 times greater in Hawaii but 12-fold and 5-fold lower than Caucasian Kauaiians, respectively. Histologically, pigmented BCCs are more common in darker-skinned populations, which make it difficult to differentiate from other lesions. Thus, despite the lower incidences of NMSCs in non-white populations, atypical presentations and distributions of NMSCs in these individuals warrant further consideration during differential diagnosis.

Methods and Challenges to Estimate Incidence

Estimating NMSC incidence has evolved over the past 40 years. One of the major challenges of NMSC incidence is how to account for multiple lesions, that is, should incidence be reported by the number of persons with skin cancer, regardless of multiplicity, or should incidence reflect the total number of NMSC lesions. The incidence estimates from 1994 attempted to account for multiplicity by including cancers occurring greater than 1 year after the primary cancer as new primary neoplasms and excluding those occurring with less than 1 year as a recurrence. Patients, therefore, could be counted multiple times. The investigators recommended reporting NMSC in 2 ways: by person and by the number of tumors.

In addition to concerns over multiplicity, another epidemiologic challenge is accurately identifying NMSC cases. This issue was addressed in the most recent 2006 estimates by using fee-for-service Medicare claims and associated International Classification of Diseases, 9th Edition, Clinical Modification ( ICD-9-CM ) codes from 2 Medicare databases to determine the number of approved NMSC-related treatment procedures. Although this report has perhaps provided the most accurate estimates so far, it too carries limitations. The study assumed that 1 NMSC treatment equals 1 NMSC lesion and does not take into account lesions completely removed by biopsy, which could underestimate incidence. Conversely, counting retreatment of lesions with positive margins as a new case could overestimate incidence. Lesions treated nonsurgically also lead to underestimation. Furthermore, the Medicare population may not accurately reflect rates of NMSC in patients younger than 65 years or non-Medicare patients older than 65 years.

With the advent of electronic medical records, HMO databases, and other health care system databases, it may be easier to identify patients with NMSCs, particularly those younger than 65 years. In a study by Eide and colleagues, HMO and health care system databases from Michigan were used to identify patients with NMSCs in 3 ways. The first algorithm used ICD-9-CM to identify cases. The second algorithm used Current Procedural Terminology (CPT) codes to identify NMSC-related treatment procedures. The third algorithm used both ICD-9-CM and CPT codes to identify cases. Using a random sample of 965 cases from all-payers from 1988 to 2007, NMSCs were verified in 47% of patients identified using ICD-9-CM codes only, 73.4% of patients identified using CPT codes, and 94.95% of patients identified using ICD-9-CM and CPT codes. Among HMO members only, the positive predictive values of identified NMSCs were 96.5% of patients using ICD-9-CM codes, 98.3% of patients using CPT codes, and 98.7% of patients using both codes.

Misclassification of codes leading to false-positive results is an inherent limitation of using administrative databases to identify NMSC cases. This study highlights that there are no standardized or validated algorithms to date to identify NMSC lesions, unlike those in place for other cancers such as breast cancer. Nevertheless, using the combined ICD-9-CM /CPT code led to a positive predictive value in identifying NMSC cases at a rate consistent with a similar breast cancer ascertainment study.

Morbidity and Mortality

Although cure rates of NMSCs reach 96% to 99%, they can be associated with significant morbidity, such as high rates of recurrence. For instance, BCCs have been reported to recur in 20% of patients within 1 year of treatment. The average 5-year BCC recurrence rate is 8.7%, with varying rates based on treatment modality. Five-year recurrence rates were determined to be 1% after Mohs micrographic surgery, 10.1% after surgical excision, 7.7% after electrodessication and curettage (EDC), 8.7% after radiation therapy, and 7.5% after cryosurgery. BCC metastasis is rare, and reported rates have ranged from 0.0028% to 0.55%. Metastatic risk seems to correlate with tumor size. Tumors that are 3 cm in size are associated with a 2% risk. Tumors larger than 5 cm have a 25% metastasis risk, and tumors larger than 10 cm have a 50% risk.

SCCs have an even greater capacity to metastasize, with rates varying from 0.1% to 13.7% depending on study design, practice, location, and length of follow-up. SCCs arising in scars have higher metastatic rates ranging from 26.2% to 37.9%. Risk factors associated with metastasis include location, tumor size, tumor depth, histologic grade, perineural invasion, immunosuppressed state, and recurrence. An examination of SCC prognostic factors revealed that long-term (>5 years) recurrence rates were lower for skin, ear, and lip lesions treated by Mohs micrographic surgery compared with non-Mohs modalities (3.1% vs 7.9%, 5.3% vs 18.7%, 2.3% vs 10.5%, respectively). Furthermore, Mohs surgery also achieved greater cure rates, regardless of tumor size or grade. Overall long-term survival after metastasis was noted to be 26.8%, with combined surgery and radiation therapy achieving greater long-term survival.

In a more recent analysis, 3-year cumulative recurrence risk after an index SCC case was 18%, whereas 3-year overall survival and disease-specific survival have been reported to be 70% and 85%, respectively. Many of the risk factors for SCC metastasis are also associated with disease-related mortality and include recurrence, subcutaneous tissue invasion, perineural invasion, size greater than 4 cm, and depth of invasion. In a prospective trial, patients with at least 1 risk factor had a disease-specific survival of 70%, whereas patients without any risk factors had a 100% disease-specific survival.

Increases in incidence, particularly of SCCs, have led to concerns over mortality trends. However, data collected from death certificates indicate an overall decline in mortality rates by 45% in men and 76% in women between 1950 and 1986. This apparent disparity and the need to better characterize NMSC-related deaths have since prompted population-based studies to examine this issue.

A 1991 Rhode Island study examined death certificates for NMSC-related deaths between 1979 and 1987 and found the age-adjusted mortality rate of nongenital SCCs to be 0.26 per 100,000 and of BCCs to be 0.1 per 100,000. SCC age-adjusted mortality for men was 3.9-fold greater than that for women, whereas BCC age-adjusted mortality was 2.3-fold greater for men. At time of death, 80% of SCC-related deaths had become metastatic compared with only 1 metastatic BCC case. Nearly 50% of SCC-related deaths resulted from a primary lesion of the ear; head and neck primary lesions accounted for 60% of SCC deaths, suggesting that UV exposure is an important factor in SCC-related mortality.

A follow-up study of Rhode Island residents examined NMSC-related deaths between 1988 and 2000 and found a decrease in nongenital SCC mortality to 0.21 per 100,000 and a decrease in BCC mortality to 0.05 per 100,000 compared with their previous study. More than 25% of nongenital SCC deaths originated with a primary lesion on the ear. However, unlike the previous study, which excluded genital skin cancers, this follow-up study found that almost 50% of NMSC mortality was caused by vulvar and penile carcinomas. Of note, mortality rates for women were 3.1 times greater for vulvar carcinomas than nongenital SCCs, whereas this trend was reversed for men in whom mortality rates were 2 times greater for nongenital SCCs than for penile cancers.

However, determining accurate mortality rates from death certificates is challenging. An evaluation of 130 NMSC-related deaths in California showed that only 30% of these deaths were confirmed to be caused by NMSCs. The investigators attributed these misclassifications to ambiguous terminology and arbitrary code classification. Weinstock and colleagues corroborated these high rates of inaccuracies. Of the 110 deaths coded as NMSC by ICD-9 – CM , 54% were misclassified. Of these misclassified deaths, 83% were caused by head and neck mucosal SCCs, highlighting the need for clearer and more consistent coding guidelines. Misclassification continued to be a significant factor in death inaccuracies in the Rhode Island follow-up study, accounting for 44% of reported NMSC-related deaths.

Because of such high rates of misclassification under the ICD-9-CM rubric 173.4 (NMSC of scalp and neck), ICD-9-CM 173.4 coded deaths were excluded to represent more accurate NMSC mortality rates. By doing so, temporal trends have become more apparent. Using data from the United States National Center for Health Statistics, national nongenital NMSC age-adjusted mortality rates between 1969 and 2000 have been 0.69 per 100,000 per year, with mortality rates for men being 2.2 times that of women and for white men being twice that of black men. Over time, nongenital NMSC mortality rates have declined by 27% in white men and by 30% in white women. In African American populations, declines in nongenital NMSC mortality are even greater at 43% in men and 51% in women. For genital NMSCs, age-adjusted mortality rates were higher in women at 0.54 than in men at 0.30. Furthermore, mortality rates of genital NMSC were at least 2 times greater in black men than white men, although recent mortality rates for these groups have converged. There were no differences in rates between white and black women, with mortality rates declining over time.

Although mortality from NMSC is low, a history of NMSC is associated with an almost 2-fold increased risk of subsequent noncutaneous cancers as well as a 20% to 30% increase in overall cancer mortality, suggesting that common exposures or susceptibilities, lifestyles, and even treatment of NMSC may be reasons for this association. This underscores the importance of understanding the impact of NMSC on the population.

QOL

Because NMSCs have such high cure rates, morbidity and mortality alone may not accurately reflect the health impact that these tumors have on the population. Instead, QOL assessments represent a significant health outcome that provides a better understanding of disease burden. For instance, treatments of NMSCs in the head and neck region have a substantial negative impact on function and cosmesis. Resulting facial disfigurement is associated with higher risks of depression and psychosocial dysfunction. Development of QOL instruments to measure and assess the psychological, social, and emotional impact of NMSCs have since been developed within the past 20 years.

The Facial Skin Cancer Index (FSCI) was the first NMSC-specific QOL instrument for patients with head and neck NMSCs. Development of the FSCI, now Skin Cancer Index (SCI), revealed that issues regarding appearance, scarring, and self-image were unique and specific concerns to NMSC-treated patients not covered by other health-related QOL instruments. Furthermore, emotional states such as anxiety and frustration were greater concerns than physical handicaps. The SCI comprises of 15 items encompassing the 3 domains of emotion, social factors, and appearance.

Using the SCI and other dermatologic QOL instruments, QOL outcomes for NMSC populations have begun to be explored. In a prospective study, patients received the SCI and a general dermatology QOL instrument, the Dermatology Life Quality Index (DLQI), before and 4 months after treatment with Mohs surgery. The SCI was able to capture improvements in total score as well as in each of the 3 subscales. This improvement was in contrast to the DLQI, which did not demonstrate any significant changes, suggesting that the SCI is a sensitive instrument for measuring clinical response and QOL in patients with NMSC. The study noted that predictors of lower QOL included female gender, populations younger than 50 years, and cancers of the lip. Greater improvements in SCI score occurred in younger patients, patients with lower incomes, patients with no prior history of NMSCs, and patients who underwent less reconstructive surgery.

Other studies have reported other predictors of QOL after NMSC treatment. Chren and colleagues used the Skindex, a validated measure of skin-related QOL that reports subscores on symptoms, emotional effects, and effects on functioning, in a prospective study to examine QOL before and after surgical treatment. Better pretreatment QOL scores were related to marital status, postgraduate education, incomes greater than $30,000 a year, fewer comorbidities, and better physical and mental health. Furthermore, patients with BCCs, tumors not on the head and neck, and smaller tumors also reported better pretreatment QOL. However, the strongest predictor of posttreatment QOL was pretreatment QOL. Other posttreatment QOL predictors included better mental health status, fewer comorbidities, and race.

The Skindex was also used to determine if type of surgical treatment (Mohs surgery, excision, and EDC) could predict posttreatment QOL. Patients treated by Mohs surgery or excision, but not EDC, experienced statistically significant improvements in QOL. There were no differences in QOL improvements between Mohs surgery and excision, which may have future implications when considering relative costs. In contrast, a recent report has suggested that type of treatment is not associated with QOL differences, although the investigators note that their cohort was underpowered to answer this question. Regardless, it is clear that QOL, possibly a more meaningful health outcome for patients with NMSC than morbidity and mortality, merits greater examination.

NMSC Expenditures

Although NMSCs are the most common cancer, related costs represent less than 1% of total cancer costs, largely because of effective management in the outpatient setting. As NMSC incidence increases, these costs are expected to correspondingly increase. In addition, as a disease more common in the elderly, Medicare costs due to NMSCs are significant. Indeed, NMSCs were determined to be the fifth most costly cancer to Medicare after lung, prostate, colon, and breast, representing about 4.5% of all Medicare expenditure. Furthermore, between 1992 and 1995, NMSC-related costs to Medicare increased by 41%, while payments for treatment of actinic keratosis increased by 91%. Total cost of NMSCs to Medicare have ranged from $285 to $426 million per year. National costs were estimated to be $650 million per year.

NMSC expenditure is largely based on treatment setting and treatment modality. 76% of NMSC-related costs were because of outpatient office visits between 1992 and 1995. The average cost for outpatient procedures was $492. For ambulatory surgery visits and inpatient procedures, the average costs were $1043 and $5537, respectively. Expenditure by specialty revealed that the average cost per episode was $957 for surgeons, $877 for multiple specialists, $296 for primary care physicians, and $521 for other physicians. Dermatologists managed about 50% of identified episodes, with the average cost per episode being $604. In another study, dermatologists managed 82% of NMSC episodes in which performed treatments and procedures included biopsies, excisions, destruction, Mohs micrographic surgery, and reconstructive repairs. Based on such cost differentials, the investigators conclude that dermatologists provide convenient, effective, and economical management for NMSCs.

The type of treatment is another major factor in NMSC expenditure. Cook and Zitelli showed that costs for Mohs micrographic surgery ($1243) were comparable to surgical excision with permanent section margin control ($1167) and less expensive than surgical excision with frozen section margin control ($1400) or ambulatory surgical excision ($1973). Although most of their cases were NMSCs, their study also encompassed non-NMSCs, including malignant melanoma. A subsequent study by Bialy and colleagues, had similar conclusions after evaluating cost comparisons between Mohs surgery and surgical excision of facial and auricular NMSCs only. Mohs surgery costs were comparable to surgical excisions with permanent sections ($937 vs $1029) and significantly less expensive than excisions with frozen sections ($937 vs $1399). A study using a computer-simulated probabilistic decision model determined that Mohs surgery was $292 less expensive than traditional surgical excision and offered a 0.056 quality-adjusted life year advantage, which translated to about 3 weeks of optimal QOL.

These results are in contrast to Medicare treatment costs in which Mohs micrographic surgery was significantly more expensive at $899 per patient than either excision at $239 per patient or local destruction at $221 per patient. One reason for this discrepancy is that these Medicare costs do not include associated pathology or subsequent repair. Indeed, Cook and Zitteli noted that procedure and pathology fees for excisions are unbundled, whereas reimbursements for Mohs surgery include both procedure and pathology. Furthermore, the investigators suggest that greater tissue defects by surgical excision leads to more complex reconstructive repair, contributing to overall increased costs.

A more recent study estimated treatment costs based on assigned 2008 relative value unit values and compared costs for a BCC cheek lesion and SCC arm lesion of different sizes by the following treatment modalities: EDC, imiquimod immunotherapy, Mohs micrographic surgery, traditional surgical excision with permanent section, surgical excision with frozen section, and radiation therapy. For all treatment modalities, a BCC cheek lesion was more expensive to treat than an SCC arm lesion. In addition, treatment costs increased with tumor size. EDC was the cheapest procedure at an average cost of $471 for a BCC cheek lesion and $392 for an SCC cheek lesion. Mohs micrographic surgery costs were $1263 for cheek BCC and $1131 for arm SCC, which were approximately 25% more expensive than surgical excisions with permanent section and immediate repair. Like prior studies, facility-based treatments and hospital-based treatments were significantly more expensive, ranging from $2200 for an arm SCC treated in an ambulatory surgery setting to $3085 for a cheek BCC treated in a hospital setting. Radiation therapy was the most expensive modality at $3460 for a cheek BCC and $3431 for an arm SCC.