ABSTRACT
Background
Advances in genetic testing allow clinicians to identify patients with pathogenic variants that increase their risk of skin cancer. Whether this information results in more frequent dermatology screening visits as recommended by the National Comprehensive Cancer Network (NCCN) guidelines is unknown.
Objectives
Our objective was to evaluate compliance with NCCN dermatology screening guidelines among patients with skin cancer‐predisposing pathogenic variants and to determine whether visiting a Hereditary Cancer Clinic (HCC) would improve adherence to skin cancer screening.
Methods
We employed a retrospective observational cross‐sectional design in a consecutive sample of 230 patients with pathogenic variants in BRCA1, BRCA2, Lynch syndrome genes (MLH1, MSH2, MSH6, PMS2) and other variants (ATM, CDKN2A, FLCN, MITF, POT1, PTEN, TP53) at the HCC at the Medical University of South Carolina. The sample was predominantly female (80%), Caucasian (82%) and non‐Hispanic/Latino (94%) with a bimodal age distribution showing peaks around age 40 and 70. Compliance with NCCN dermatologic screening guidelines with a 6‐month buffer period was measured, both before and after an HCC visit. Statistical significance was determined using the chi‐squared proportion test.
Results
Overall compliance with dermatology visits improved from 11.1% before HCC visits to 37% after visiting the HCC (p < 0.001). Factors associated with increased compliance included a personal (p = 0.09) and family history (p < 0.01) of non‐melanoma skin cancer and age between 45 and 50 years old (p < 0.01). Among those receiving biopsies, 37.2% exhibited non‐benign lesions. Lynch syndrome variants accounted for 50% of non‐benign skin lesions.
Conclusions
HCC visits and personal and family history of skin cancer were associated with improved adherence to NCCN dermatology screening guidelines among patients with skin cancer‐predisposing variants. Thus, dedicated HCCs could improve early detection and intervention of premalignant and malignant skin lesions. Additional studies that include a larger number of minority patients are required to confirm these findings.
Keywords: compliance, genetic testing, hereditary cancer clinic, screening, skin cancer
Patients with genetic variants that increase skin cancer risk are recommended to undergo regular dermatology screenings per National Comprehensive Cancer Network guidelines, but compliance is uncertain. In a retrospective study of 230 patients with such variants, compliance with screening visits significantly improved after visiting a Hereditary Cancer Clinic, from 11.1% to 37% (p < 0.001). Factors linked to higher adherence included personal or family history of skin cancer. Lynch syndrome variants accounted for 50% of non‐benign skin lesions.
1. Introduction
Genetic testing advances and the expansion of population‐wide genomic screening allow identification of patients carrying pathogenic variants that increase the risk of skin cancer [1, 2, 3]. The National Comprehensive Cancer Network (NCCN) suggests skin cancer screening guidelines for individuals with BRCA and Lynch syndrome pathogenic variants: annual and biennial full body skin examination (FBSE), respectively [4, 5]. While genetic counseling and knowledge of genetic variants improve compliance with screening recommendations and treatment outcomes in the context of breast and ovarian cancers [6, 7], adherence to NCCN skin cancer screening guidelines is unknown. Further, though it is presumed that dedicated Hereditary Cancer Clinics (HCCs) could enhance patient compliance to screening [8], no data exist in this area. In this retrospective observational study, we examined compliance with NCCN skin cancer screening guidelines in patients seen in HCCs and assessed the results of biopsies among patients who attended dermatology screening visits.
2. Methods
2.1. Study Design and Population
This study employed a retrospective observational cross‐sectional design. The study population consists of patients carrying pathogenic variants that increase skin cancer risk seen in the HCC at the Medical University of South Carolina (MUSC). A total number of 252 patients were included.
2.2. Outcomes
To assess compliance with NCCN‐recommended skin cancer screening among patients carrying pathogenic variants increasing skin cancer risk, and whether a visit to the HCC led to greater adherence to dermatology screening.
2.3. Data Collection and Organization
Data were collected via chart review (IRB approved, Pro00137694) and included age, race, ethnicity, legal sex, census‐tract geocoded social vulnerability index (SVI), median income, distance to HCC (i.e., straight line distance from home address), date of genetic test results, genetic variant(s), date of HCC visit, personal history of cancer, dates of dermatology appointments and biopsy findings. For fifteen patients with > 1 listing for race and gene variants, each listing was treated individually in the analysis. Gene variants were classified into four categories: BRCA1, BRCA2, Lynch syndrome (MLH1, MSH2, MSH6 and PMS2) and Other (which contained ATM, CDKN2A, FLCN, MITF, POT1, PTEN and TP53). Skin biopsy results were classified into five categories: Benign (including melanocytic nevi, inflammatory lesions and other benign lesions), actinic keratoses, atypical nevi, melanoma and non‐melanoma skin cancer (which included basal cell carcinoma, squamous cell carcinoma and cutaneous T‐cell lymphoma).
2.4. Overall Compliance Analysis
Overall compliance with dermatology screening was determined by adherence to NCCN guidelines [4, 5] with a 6‐month buffer period: 30 months for Lynch syndrome variants (MLH1, MSH2, MSH6 and PMS2), 12 months for CDKN2A and 18 months for BRCA1, BRCA2 and all Other variants with increased skin cancer risk (ATM, FLCN, MITF and POT1) [9, 10]. Compliance was measured within two intervals: from the date of the genetic test result to the date of the HCC visit, and from the HCC visit to the study endpoint. Patients with < 6 months available for dermatology follow‐up were excluded from the analysis. When a shortened interval of 6–18 months (e.g., BRCA variants) or 30 months (e.g., Lynch syndrome variants) was available for dermatology follow‐up, the previous dermatology appointment was used as the starting point. If a dermatology visit with FBSE was observed within the interval, then that visit was the starting point for the next interval (i.e., variable intervals) for overall compliance analysis.
2.5. Analysis of Expected and Observed Visits
The same intervals listed above were used to calculate expected visits. However, these intervals were fixed (i.e., the starting point of the next interval is the same date as the endpoint of the previous interval). Expected visits were quantified until the study endpoint, and truncated intervals ( < 12 months) were excluded. Observed visits include the total number of dermatology visits with FBSE within the expected intervals.
2.6. Data Visualization and Statistical Analyses
Data was imported into RStudio for manipulation, visualization and analysis using ‘RODBC’ and ‘tidyverse’ packages [11]. The chi‐square test for proportions was used to determine differences between groups. One‐sided chi‐square tests were employed for directional hypotheses such as skin cancer history and compliance, whereas two‐sided chi‐square tests were used to detect overall differences. The Fisher's exact test was used for small samples (n < 5). Logistic regression was used to model the odds of compliance as a function of SVI, median income and distance to HCC.
3. Results
3.1. Sample Demographics
Of the 252 patients seen in the HCC, 22 patients were excluded due to inadequate time for screening follow‐up. Eighty percent were female (Figure 1A) with a bimodal age distribution showing peaks around age 40 and 70 (Figure 1B). Most younger patients were female, whereas older patients ( > 65 years old) showed a more even distribution among males and females (Figure 1C). The majority of our sample was Caucasian (Figure 1D) and non‐Hispanic/Latino (Figure 1E). Socioeconomic variables are depicted in Figure S1A‐C. Most genetic variants represented in our sample were BRCA1, BRCA2 and Lynch syndrome genes (i.e., MSH6, PMS2, MLH1, MSH2; Figure 1F). Most patients with a history of breast cancer exhibited BRCA1 and BRCA2 pathogenic variants, and those with a history of colon and rectal cancer showed variants associated with Lynch syndrome (Figure 1G, black text). However, patients with a personal history of skin cancer showed a near‐equal distribution of the variants studied (Figure 1G, red text).
Figure 1.
Sample demographics. (A) Bar graph showing number of patients included in the study separated by legal sex. (B) Histogram plot showing the distribution of age among included patients. (C) Histogram plot showing the distribution of age separated by legal sex. (D) Pie chart depicting race among included patients. (E) Pie chart depicting ethnicity among included patients. (F) Pie chart depicting the distribution of pathogenic variants among included patients. (G) Bar chart showing the past medical history of cancers with grouped distribution of pathogenic variants.
3.2. Compliance With Dermatology Visits
Overall compliance was 11.1% in the interval from discovery of a genetic variant to HCC visit (n = 72) and improved to 37% after HCC visit (n = 230; p < 0.001; Figure 2A). The ratio of observed to expected (O/E) visits to the dermatologist was significantly higher after HCC visit (0.61) compared to before HCC visit (0.30; p < 0.001; Figure 2B). Additional factors that were associated with improved compliance include a personal (p = 0.09) and family history (p < 0.01) of non‐melanoma skin cancer (Figure 2C) and age between 45 and 50 years old (p < 0.01; Figure 2D). Single‐parent households (p = 0.183) and lack of a personal vehicle (p = 0.184) showed weak relationships with noncompliance (Figure S1D–F).
Figure 2.
Compliance with dermatology visits. (A) Schematic with pie charts depicting overall compliance of dermatology visits before and after the HCC visit. (B) Lollipop plot showing the number of expected visits (circles) compared to the number of observed visits (triangles) among patients before the HCC visit (top panel) and after the HCC visit (bottom panel). Each line emanating from the x‐axis represents an individual patient. Noncompliant patients are shown in orange, and compliant patients in blue. (C) Bar chart showing overall compliance (dark blue bar) and compliance associated with patient factors (light blue bars). Numbers indicate sample size for each bar, **p < 0.01. (D) Bar chart showing percent compliance across the distribution of age. Dashed line indicates overall compliance, **p < 0.01.
3.3. Biopsy Results From Dermatology Visits
Among the 230 patients studied, 43 patients received a biopsy (average number of biopsies = 0.75, range 0–24). Nearly all patients receiving biopsies were non‐Hispanic white, and 16 of these 43 patients received non‐benign biopsy results (Figure 3A). Further, half of all patients with non‐benign biopsy results carried a pathogenic variant associated with Lynch syndrome (Figure 3B).
Figure 3.
Biopsy results from dermatology visits. (A) Stacked bar chart showing the number of biopsies for each patient (one bar indicates one patient), coloured by diagnosis of biopsy from pathology reports. Coloured circle indicates race/ethnicity of patient. (B) Stacked bar chart shows the distribution of benign and non‐benign (atypical naevus, actinic keratosis, NMSC, melanoma) skin lesions among different genetic pathogenic variants. Numbers indicate sample size for each bar.
4. Discussion
Our study demonstrates improved compliance with NCCN skin cancer screening guidelines after a visit to an HCC compared to patients who did not visit an HCC (37% vs. 11.1%). These findings suggest that dedicated HCCs improve adherence to NCCN cancer screening guidelines [4, 5]. Although patients with Lynch syndrome accounted for 37% of our sample, they accounted for 50% of non‐benign skin lesions. Melanoma and squamous cell carcinoma in patients with Lynch syndrome may arise through microsatellite instability pathways seen in colorectal cancers in these patients [12, 13, 14], yet further studies are needed to clarify skin cancer risk and whether more frequent dermatology screening visits (e.g., annual instead of biennial) would be warranted in this population. More than one‐third of patients attending screening visits were found to have non‐benign skin lesions, and over 95% of biopsies were performed on non‐Hispanic white patients. Eighty percent and 82% of our sample were female and Caucasian, respectively. Though lighter skin phototypes are associated with a higher incidence of skin cancer [15], further studies with greater representation of males, as well as Hispanic, African American and other minority groups are warranted.
Author Contributions
S. J. Gadomski, G. De Jesús and K. S. Hughes organized the database, analysed data, prepared figures and wrote the manuscript. J.S. Obeid assisted with obtaining socioeconomic data. C. Rowley provided guidance for IRB approval.
Ethics Statement
This study was exempt from IRB review (Category 4), and all patient data was deidentified and stored on a secure, encrypted network.
Conflicts of Interest
K.S.H. receives honoraria from Hologic, Targeted Medical Education (TME), MedNeon and Myriad Genetics. K.S.H. has a financial interest in CRA Health and is the co‐creator of Ask2Me.Org. Other authors declare no conflicts of interest.
Supporting information
Supplemental Figure 1: Socioeconomic Variables. (A) Histogram plots showing distribution of social vulnerability index (SVI) as defined by the CDC. (B) Histogram plot showing distribution of median income. (C) Histogram plot showing distribution of distance to the HCC. 2 outliers were excluded. (D) Boxplots depicting odds ratio and 95% confidence intervals of SVI among non‐compliant (red) and compliant (blue) patients. (E) Boxplots depicting odds ratio and 95% confidence intervals of median income among non‐compliant (red) and compliant (blue) patients. (F) Boxplots depicting odds ratio and 95% confidence intervals of distance to HCC among non‐compliant (red) and compliant (blue) patients.
Acknowledgements
Guiu Puigcercos i Vilar from MUSC Department of Public Health Sciences assisted with socioeconomic data collection.
Contributor Information
Kevin S. Hughes, Email: hughkevi@musc.edu.
Graciela De Jesús, Email: dejesusg@musc.edu.
Data Availability Statement
All original code has been deposited at Github and is publicly available at https://github.com/gadomskisj/DermatologyCompliance/blob/main/OverallCompliance. Any additional information required to reanalyse the data reported in this paper is available from the lead contact upon request.
References
- 1. Allen C. G., Judge D. P., Levin E., et al., “A Pragmatic Implementation Research Study for in Our DNA SC: A Protocol to Identify Multi‐Level Factors That Support the Implementation of a Population‐Wide Genomic Screening Initiative in Diverse Populations,” Implementation Science Communications 3 (2022): 48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Stark Z., Dolman L., Manolio T. A., et al., “Integrating Genomics Into Healthcare: A Global Responsibility,” American Journal of Human Genetics 104 (2019): 13–20, preprint at 10.1016/j.ajhg.2018.11.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Cust A. E., Drummond M., Kanetsky P. A., et al., “Assessing the Incremental Contribution of Common Genomic Variants to Melanoma Risk Prediction in Two Population‐Based Studies,” Journal of Investigative Dermatology 138 (2018): 2617–2624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Gupta S., Wise J. M., Axell L., et al. NCCN Guidelines Version 1.2024 Genetic/Familial High‐Risk Assessment: Colorectal, Endometrial, and Gastric Continue, 2024, https://www.nccn.org/home/.
- 5. Daly M., Pal T., AlHilli Z., et al. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines ®) Genetic/Familial High‐Risk Assessment: Breast, Ovarian, and Pancreatic, 2024, https://www.nccn.org/home/member-.
- 6. Buchanan A. H., Voils C. I., Schildkraut J. M., et al., “Adherence to Recommended Risk Management Among Unaffected Women With a BRCA Mutation,” Journal of Genetic Counseling 26 (2017): 79–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Bobbili P., Olufade T., DerSarkissian M., et al., “Adherence to National Comprehensive Cancer Network Guidelines for BRCA Testing Among High Risk Breast Cancer Patients: A Retrospective Chart Review Study,” Hereditary Cancer in Clinical Practice 18 (2020): 13, preprint at 10.1186/s13053-020-00144-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Allen C. G., Donahue C., Coen E., et al., “Implementation Mapping for Managing Patients at High Risk for Hereditary Cancer,” American Journal of Preventive Medicine 66 (2024): 503–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Zocchi L., Lontano A., Merli M., et al., “Familial Melanoma and Susceptibility Genes: A Review of the Most Common Clinical and Dermoscopic Phenotypic Aspect, Associated Malignancies and Practical Tips for Management,” Journal of Clinical Medicine 10 (2021): 3760, preprint at 10.3390/jcm10163760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Leachman S. A., Lucero O. M., Sampson J. E., et al., “Identification, Genetic Testing, and Management of Hereditary Melanoma,” Cancer and Metastasis Reviews 36 (2017): 77–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Ripley B. and Lapsley M., “RODBC: An ODBC Database Interface,” CRAN: Package RODBC (2023): 0, 10.32614/CRAN.package.RODBC. [DOI] [Google Scholar]
- 12. Ponti G., Losi L., Pellacani G., et al., “Malignant Melanoma in Patients With Hereditary Nonpolyposis Colorectal Cancer,” British Journal of Dermatology 159 (2008): 162–168. [DOI] [PubMed] [Google Scholar]
- 13. Ykema B. L. M., Adan F., Crijns M. B., et al., “Cutaneous Squamous Cell Carcinoma Is Associated With Lynch Syndrome: Widening the Spectrum of Lynch Syndrome‐Associated Tumours,” British Journal of Dermatology 185 (2021): 462–463, preprint at 10.1111/bjd.20139. [DOI] [PubMed] [Google Scholar]
- 14. Zhong C. S., Horiguchi M., Uno H., et al., “Clinical Factors Associated With Skin Neoplasms in Individuals With Lynch Syndrome in a Longitudinal Observational Cohort,” Journal of the American Academy of Dermatology 88 (2023): 1282–1290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Garbe C., Forsea A. M., Amaral T., et al., “Skin Cancers Are the Most Frequent Cancers in Fair‐Skinned Populations, but We Can Prevent Them,” European Journal of Cancer 204 (2024): 114074. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 1: Socioeconomic Variables. (A) Histogram plots showing distribution of social vulnerability index (SVI) as defined by the CDC. (B) Histogram plot showing distribution of median income. (C) Histogram plot showing distribution of distance to the HCC. 2 outliers were excluded. (D) Boxplots depicting odds ratio and 95% confidence intervals of SVI among non‐compliant (red) and compliant (blue) patients. (E) Boxplots depicting odds ratio and 95% confidence intervals of median income among non‐compliant (red) and compliant (blue) patients. (F) Boxplots depicting odds ratio and 95% confidence intervals of distance to HCC among non‐compliant (red) and compliant (blue) patients.
Data Availability Statement
All original code has been deposited at Github and is publicly available at https://github.com/gadomskisj/DermatologyCompliance/blob/main/OverallCompliance. Any additional information required to reanalyse the data reported in this paper is available from the lead contact upon request.