Original Article
Risk of second primary malignancy after non-small cell lung cancer: a competing risk nomogram based on the SEER database
Abstract
Background: With the improvement of survival for non-small cell lung cancer (NSCLC), research focused on second primary malignancy (SPM) in NSCLC survivors is becoming urgent. This study aimed to estimate the risk of SPM in NSCLC patients.
Methods: We retrospectively analysed NSCLC patients diagnosed between 2004 and 2010 in SEER database. We firstly evaluated the crude and cumulative incidence of SPM. SPM incidence in NSCLC survivors compared to that in the reference population was calculated as standardized incidence ratio (SIR). A competing risk nomogram was also built, to predict the incidence of SPM.
Results: The crude and 10-year cumulative incidences of SPM were 4.04% and 5.05%, respectively, while the SIR was 1.62. The nomogram was well calibrated and had good discriminative ability, with c-index of 0.80. It showed a significantly wide interval of SPM cumulative incidence between the first and tenth-decile according to the risk model (1.04% vs. 16.70%, P<0.05). The decision curve analysis indicated that the clinical net benefit of risk model was larger than that in other scenarios (all-screening or no-screening) in a range of threshold probabilities (1% to 20%).
Conclusions: Our study firstly performed a systematic estimation of the incidence of SPM in NSCLC, which implied the necessity of a risk predicting model. We developed the first competing risk nomogram to predict the risk of SPM, which performed well in the evaluation and might be helpful for individualized SPM screening.
Methods: We retrospectively analysed NSCLC patients diagnosed between 2004 and 2010 in SEER database. We firstly evaluated the crude and cumulative incidence of SPM. SPM incidence in NSCLC survivors compared to that in the reference population was calculated as standardized incidence ratio (SIR). A competing risk nomogram was also built, to predict the incidence of SPM.
Results: The crude and 10-year cumulative incidences of SPM were 4.04% and 5.05%, respectively, while the SIR was 1.62. The nomogram was well calibrated and had good discriminative ability, with c-index of 0.80. It showed a significantly wide interval of SPM cumulative incidence between the first and tenth-decile according to the risk model (1.04% vs. 16.70%, P<0.05). The decision curve analysis indicated that the clinical net benefit of risk model was larger than that in other scenarios (all-screening or no-screening) in a range of threshold probabilities (1% to 20%).
Conclusions: Our study firstly performed a systematic estimation of the incidence of SPM in NSCLC, which implied the necessity of a risk predicting model. We developed the first competing risk nomogram to predict the risk of SPM, which performed well in the evaluation and might be helpful for individualized SPM screening.
