Effects of kinship and timing in familial cancer

Family cancer history is one of the most important risk factors for cancer. We investigated in depth the effect of family cancer history, in particular, the effects of kinship and timing on cancer incidence and prognosis using Swedish population-based registers.

In Study I, we investigated how the type of kinship and sex affect the familial risk of adult chronic lymphatic leukemia (CLL). We found the highest relative risks for sisters of affected women and sons had a much higher risk than daughters if the affected parent was the mother.

In Study II, we developed a unified model for familial risk by extending a Cox regression model to estimate the detailed effects of all specific kinships, using data from all family members simultaneously. This enabled a formal comparison of the risk to different relatives. We illustrated the method with applications to adult leukemia and non-Hodgkin’s lymphoma and found sisters of female patients had significantly higher risk for both cancers than other relatives.

We next investigated how the risk pattern of four main cancers in Sweden (colorectal cancer, breast cancer, prostate cancer and melanoma) changes with age and elapsed time from diagnosis of the same cancer in a sibling. Results were presented graphically in Study III. For all four cancers, siblings of cancer patients had higher cancer incidence at all ages compared to siblings of cancer-free individuals. Relative risks were especially high in siblings who were young when the first cancer was diagnosed in the family. The relative risks were relatively constant up to 20 years after the cancer diagnosis in siblings for all cancers except prostate cancer, where the hazard ratio decreased steeply during the first few years. We found evidence that this may be due to a screening effect for prostate cancer while there was no evidence of a screening effect in breast cancer.

In Study IV, we examined how family cancer history affects prognosis for patients with several major cancers in Sweden and we further investigated for cancers whose prognosis has been affected by family cancer history whether these effects are associated with histological type or subtype or tumor stage at cancer diagnosis. For breast and prostate cancer patients, family cancer history played a protective role in cancer survival, and this may be associated with medical surveillance of family members. However, daughters or sisters of ovarian cancer patients had poorer cancer survival, which is consistent with a higher proportion of diagnosed later stage tumors and of an aggressive histological type of ovarian cancers.

In conclusion, our investigation in this thesis used Swedish population-based registers to highlight the effect of family cancer history on several outcomes, including cancer risks for family members and cancer patient prognosis. Our findings provide evidence that could help tailor screening programs for relatives of cancer patients and inform the design of genetic biomarker studies.

List of scientific papers

I. Rebora P, Lee M, Czene K, Valsecchi MG, Reilly M. High risks of familial chronic lymphatic leukemia for specific relatives: signposts for genetic discovery? Leukemia 2012; 26: 2419-2421.
https://doi.org/10.1038/leu.2012.103

II. Lee M, Rebora P, Valsecchi MG, Czene K, Reilly M. A unified model for estimating and testing familial aggregation. Statist. Med. 2013; 32: 5353-5365.
https://doi.org/10.1002/sim.6025

III. Lee M, Czene K, Rebora P, Reilly M. Patterns of changing cancer risks with time since diagnosis of a sibling. Int. J. Cancer 2015; 136(8): 1948-56.
https://doi.org/10.1002/ijc.29239

IV. Lee M, Reilly M, Lindström LS, Czene K. Differences in survival for patients with familial and sporadic cancer. [Manuscript]