The role of other tumors in breast cancer aggressiveness

While mammographic screening programs have been successful in reducing mortality from aggressive breast cancer, not every woman benefits from regular examinations. According to our results, those women are also more likely diagnosed with other tumors, either before or after breast cancer diagnosis.

Go to the profile of Felix Grassmann
Nov 01, 2019

Primarily, we wanted to identify novel (genetic, environmental and life-style) factors that are involved in breast cancer aggressiveness in order to identify women at risk and to shed further light on the underlying processes. We were particularly interested in risk factors for interval breast cancer, that is a breast tumor that was identified between two regularly scheduled mammographic screenings. Those tumors typically grow quite fast within the screening window to palpable size and are thus considered more aggressive with worse prognosis and survival. In contrast, those tumors that are detected at a regularly schedules mammographic screening (termed screen detected) are usually smaller and of less aggressive nature with excellent prognosis (see Figure 1).


Figure 1. Prognosis of breast cancer depending on the mode of detection. Interval breast cancer (IC) patients diagnosed between two regularly scheduled mammographic screenings have a lower survival rate than patients diagnosed with breast cancer at the mammographic examination (SDC). 

Our initial idea was to look at women that had survived a prior tumor since we thought they would be more likely to have beneficial genetics resulting in less aggressive and therefore slower growing tumors. However, we found that cancer survivors were more likely to be diagnosed with interval compared to screen detected breast cancer. Similarly, women which survived aggressive interval breast cancer were more likely to develop another primary tumor in a different tissue. In contrast to our initial hypothesis, we now realized that the underlying factor that resulted in multiple cancer diagnoses in an individual’s lifetime was also linked to breast cancer aggressiveness. 

Prior studies showed that family history of breast cancer in close relatives was more commonly reported in interval compared to screen detected patients. The underlying genetic variations that cause an increased occurrence of breast cancer in families can be attributed to either rare (and penetrant) or common (and weak) genetic changes. According to our results, though, common genetic changes are unlikely to explain the observed association. When we compared to combined genetic risk for other tumors from common variants between interval and screen detected patients, we found that interval breast cancer patients were actually at reduced genetic risk for those tumors. Therefore, rare cancer mutations are likely the cause for the occurrence of multiple tumor diagnoses in those women as well as the increased risk for interval breast cancer. 

Since we did not have comprehensive sequencing data available for all our study participants, we instead investigated history of other tumors (excluding breast cancer) in close relatives. Overall, a family history of other tumors was not found to be a risk factor for interval cancer when we looked at the whole cohort.  However, in women with a “second hit”, that is another tumor diagnosis either before or after breast cancer diagnosis, family history of other tumors was indeed reported more often in interval compared to screen detected patients.

To conduct those studies, we had access to two large Swedish cohorts of breast cancer patients with a rich collection of risk factors, tumor characteristics, information on drugs and other diseases as well as accurate mammographic screening dates. We previously found that the mammographic density (i.e. the density of the breast in mammographic images) can mask small tumors and thus result in false-positive interval cancer status. Those masked tumors are not true interval cancers that grow within the screening interval but instead were missed at the prior screen and more closely resemble screen detected cancer. Therefore, we restricted our analyses to interval cancer patients which hat low dense breasts at the prior screen.

To replicate our findings, we wanted to include an independent cohort of women with interval and screen detected breast cancer patients. We applied for access to the Women’s Health Initiative (WHI) via the DataBase of Gentotypes and Phenotypes (dbGAP) and tried to compute interval cancer status from the mammographic screening dates deposited in the database. However, we quickly realized that those dates were not comprehensive enough. Instead, we used the questionnaire data to infer the breast cancer mode of the detection. Women in the WHI were asked annually whether they have attended a mammographic screening in the prior year. This information allowed us to infer likely mammographic screening dates which proved crucial and sufficient to determine whether a woman was diagnosed with breast cancer at a mammographic screening or not. Indeed, the women with interval breast cancer in the WHI as determined by questionnaire had comparable tumor characteristics and survival as interval cancer patients in the Swedish cohorts ascertained from actual screening dates. Importantly, we were able to confirm our findings in breast cancer patients from the WHI resulting in a more robust and significant association.

Taken together, our results showed a role of other tumors in breast cancer aggressiveness and prognosis, likely due to rare genetic changes that increased the chance for multiple tumor diagnoses in a woman’s lifetime as well as her risk for interval breast cancer.

Go to the profile of Felix Grassmann

Felix Grassmann

Postdoctoral Fellow, Karolinska Institutet

No comments yet.