Childhood cancer survivors face markedly worse overall survival after diagnosis with breast cancer, melanoma, or colorectal cancer

Abstract

Background: Childhood cancer survivors (CCS) are at elevated risk of secondary malignancies (SM). Enhanced screening for SM is recommended, but compliance is poor. We hypothesized that CCS with adult-onset SM (colorectal cancer [CRC], melanoma, or breast cancer [BC]) would present with more advanced disease and have decreased overall survival (OS).

Methods: The Surveillance, Epidemiology, and End Results Program was queried for patients diagnosed with cancer at age less than or equal to 18 also diagnosed with adult-onset CRC, melanoma, or BC. A cohort without a history of prior malignancy was likewise identified. Tumor features and clinical outcomes were compared.

Results: CCS with a SM (n = 224) were compared with patients without a childhood cancer history (n = 1,392,670). CCS were diagnosed younger (BC = 37.6 vs. 61.3, p < 0.01, CRC = 35.0 vs. 67.1, p < 0.01, melanoma = 29.6 vs. 61.3 years old, p < 0.01). CCS with BC were more likely to have Stage III or IV disease (25.2% vs. 16.5%, p = 0.01). Hormone-receptor expression also differed; CCS were less likely to develop Luminal A-type tumors (48.6% vs. 66.9%, p = 0.01). After age-adjustment, CCS had worse OS (Hazard ratio: CRC = 2.449, p < 0.01, melanoma = 6.503, p < 0.01, BC = 3.383, p < 0.01).

Conclusion: CCS were younger when diagnosed with a SM. After age-adjustment, OS was diminished. Heightened surveillance may be necessary for CCS diagnosed with SM.

Keywords: breast cancer; colorectal cancer; melanoma; secondary malignancy; survivorship.

Figure 1:. Overall Survival after CRC Diagnosis.

A) Kaplan-Meier survival curve comparing OS of CCS with colorectal cancer patients without a history of childhood cancer. Shading indicated 95% CI. Unadjusted median OS 92 vs. 72 months, respectively. Log-rank p-value = 0.272. B) Cox-Multivariate Regression modeling of OS with Age as a Co-variate. Survival probability of CCS vs. non-CCS patients 160 months after CRC diagnosis: 10.2% vs. 33.4%, p-value <0.01, median OS 19 vs. 74 months. Hazard ratio (95% CI) of CCS = 2.449 (1.543-3.888).

Figure 2:. Overall Survival after Melanoma Diagnosis.

A) Kaplan-Meier survival curve comparing OS of CCS with melanoma patients without a history of childhood cancer. Shading indicated 95% CI. Unadjusted median OS not reached at 160 months. Log-rank p-value = 0.532. B) Cox-Multivariate Regression modeling of OS with Age as a Co-variate. Survival probability of CCS vs. non-CCS patients 160 months after CRC diagnosis: 16.2% vs. 60.1%, p-value <0.01, median OS 32 months vs. not reached. Hazard ratio (95% CI) of CCS = 6.503 (3.251-13.011).

Figure 3:. Overall Survival after Breast Cancer Diagnosis.

A) Kaplan-Meier survival curve comparing OS of CCS with breast cancer patients without a history of childhood cancer. Shading indicated 95% CI. Unadjusted median OS not reached at 160 months. Log-rank p-value = 0.7662. B) Cox-Multivariate Regression modeling of OS with Age as a Co-variate. Survival probability of CCS vs. non-CSS patients 160 months after BC diagnosis: 24.6% vs. 59.8%, p-value <0.01, median OS 67 months vs. not reached. Hazard ratio (95% CI) of CCS = 3.383 (2.210-5.179).

Figure 4:. Cox-Regression Model of Overall Survival by Stage after Breast Cancer Diagnosis.

A) Stage I, age-adjusted OS probability at 160 months: 17.8 vs. 70.6 months, p-value<0.01. B) Stage II, age-adjusted OS probability at 160 months: 34.9 vs. 62.8 months, p-value=0.07. C) Stage III, age-adjusted OS probability at 160 months: 15.1 vs. 43.3 months, p-value<0.01. D) Stage IV, age-adjusted OS probability at 160 months: 4.3 vs. 8.2 months, p-value=0.66.