PROLARA: prognosis-based lifetime attributable risk approximation for cancer from diagnostic radiation exposure

To evaluate the impact of the reduced life expectancy of patients (compared to a nonpatient group with the same age distribution) on their nominal risk of developing cancer from the diagnostic use of radiation.

We define a "prognosis-based lifetime attributable risk modifier" (PROLARM) as the ratio of the risks for nonpatients and patients, a dimensionless quantity which indicates how strongly the life-time attributable risk (LAR) is reduced due to a patient's prognosis. An approximation to this ratio can be given (named PROLARA) which depends only on the patient's age at exposure and his/her life expectancy, but is independent of the exact choice of values for the baseline risk of cancer incidence and the excess relative risk (ERR) from radiation exposure. PROLARM and PROLARA were computed for a cohort of 4,285 female patients with metastatic breast cancer, for whom all necessary input data were available.

LAR of solid cancer was significantly decreased in these patients: PROLARM >20 for age at exposure < or = 65 years. For any reasonable choice of function for ERR, the approximation PROLARA gave a lower estimate of the reduction in risk. The risk for a patient from the above cohort, exposed at age 50 years, is decreased by a factor of 29 (PROLARM) and 27 (PROLARA). In other words, 50 mSv in a patient with metastatic breast cancer corresponds risk-wise to only 2 mSv in a nonpatient of the same age.

A major proportion of the total dose from diagnostic medical exposures does not constitute an additional cancer risk due to the poorer prognosis of patients compared to nonpatients of same gender and age. Our new PROLARA concept allows an estimation of the reduction in risk for any pathology when the associated survival is known.