Risk Stratification Models for Cutaneous Malignant Neoplasms After Blood or Marrow Transplant.

Abstract

Importance: Blood or marrow transplant (BMT) recipients are at increased risk of developing subsequent cutaneous malignant neoplasms. There is a need to develop risk stratification tools to identify those at highest risk to inform dermatologic surveillance and evaluate risk mitigation strategies.

Objective: To develop and validate a risk stratification tool for subsequent cutaneous malignant neoplasms among BMT recipients that incorporates readily available patient and treatment information.

Design, setting, and participants: This prognostic study included participants in the multi-institutional, longitudinal BMT Survivor Study (BMTSS), which included individuals who underwent autologous or allogeneic BMT from January 1, 1974, to December 31, 2014, at 1 of 3 participating sites in California, Minnesota, and Alabama and survived 2 or more years after BMT. The present cohort was restricted to individuals who underwent a single transplant and completed at least 1 BMTSS survey. Data were analyzed from December 3, 2024, to June 3, 2026.

Exposures: Risk stratification models included the following candidate variables: age at BMT, sex, race, ethnicity, prior cutaneous malignant neoplasm, pretransplant exposure to monoclonal antibodies, total body irradiation, chronic graft vs host disease, and posttransplant immunosuppression.

Main outcomes and measures: The main outcomes were incident basal cell carcinoma (BCC), squamous cell carcinoma (SCC), or melanoma after BMT, ascertained based on participant report with medical record validation when available. Models for risk stratification were developed and validated using time-dependent area under the receiver operating characteristic (AUC) curve with repeated random partitioning into training and testing cohorts. Results were averaged to make a final model. Individuals were classified into low or high risk for subsequent cutaneous malignant neoplasm based on clinically meaningful risk stratification score thresholds.

Results: Of the 3448 included BMTSS participants (1917 [55.6%] male; mean [SD] age at BMT, 41.9 [19.2] years) who were alive without a cutaneous malignant neoplasm event 2 years after BMT, 282 (8.2%) were diagnosed with BCC, 183 (5.3%) with SCC, and 73 (2.1%) with melanoma. At 15 years after BMT, the test model AUCs were 0.81 (95% CI, 0.76-0.85) for BCC, 0.90 (95% CI, 0.86-0.93) for SCC, and 0.83 (95% CI, 0.74-0.90) for melanoma. The 15-year cumulative incidence of BCC, SCC, and melanoma in the low-risk group was 2.4% (95% CI, 1.5%-3.9%), 3.2% (95% CI, 2.4%-4.2%), and 0.8% (95% CI, 0.4%-1.5%), respectively. In the high-risk group, the 15-year cumulative incidence of BCC, SCC, and melanoma was 13.3% (95% CI, 11.6%-15.4%), 13.6% (95% CI, 11.2%-16.5%), and 4.4% (95% CI, 3.2%-6.1%), respectively.

Conclusions and relevance: In this prognostic study, risk stratification models demonstrated excellent performance, highlighting potential utility for identifying BMT recipients who are at greatest risk of subsequent cutaneous malignant neoplasms.