With the increasing numbers of patients with multiple primary cancers, survival from subsequent cancers is of growing interest. The majority of the analyses on the subject so far have, however, suffered from methodological difficulties.
A new model is now proposed for estimating relative survival of patients with subsequent primary cancer. The model is an extension to that proposed earlier by Estève et al. for estimating relative survival using individual patient data. The model is illustrated with real data on patients with one or two primary breast cancers and used in comparing the excess hazards between first and subsequent breast cancer.
For patients with multiple cancers, the traditional analysis of relative survival can be made cancer-specific. The excess hazards are different between the first and subsequent breast cancer: The excess hazard of the subsequent breast cancer tends to decrease with increasing age when compared to the corresponding hazard of the first breast cancer.
Relative survival of patients with subsequent cancer can be modelled facilitating studies on different hypothesis on the excess hazards of a first and subsequent cancer.