What kind of cancer does radiation cause


If cancer occurs after previous exposure to radiation, the question arises as to whether the cancer that has occurred was caused by radiation. Since there is currently no way to differentiate naturally occurring cancer from radiation-induced cancer, the relationship between the cancer case and radiation exposure can only be expressed in terms of probabilities. The methods and the software program ProZES are developed to calculate this correlation probability Z. Z is the probability that the observed case of cancer was caused by previous radiation exposure and is defined as:

Z = radiation-induced incidence rate / total incidence rate.

ProZES can calculate Z according to low-LET radiation (including gamma and X-rays) for all solid cancers, as well as for leukemia and lymphoma. Risk models for lung cancer after exposure to radon are also implemented. Particular emphasis is placed on determining the uncertainties. ProZES can, for example, support experts in the case of compensation procedures after occupational radiation exposure.

The cancer risk models are a central part of ProZES. Most of these models were newly developed for ProZES or re-evaluated on cohort data. For low-LET radiation, the models are mainly based on the incidence data of atomic bomb survivors from Hiroshima and Nagasaki, which is the world's largest study of radiation-induced cancer. For radon, separate models for lung cancer by exposure have been implemented in mining and indoor areas. ProZES contains specialized risk models of the most common radiation-induced cancers, for cancers of the lung, female breast, colon, stomach and thyroid. The risk models for the other cancers were developed for groups of functionally related cancers.

The calculation of the probability of association involves a number of methodological challenges. Risks from the Japanese population have to be transferred to a current German population. The method of multi-model inference is used to take different dose dependencies into account equally and to reduce selection bias. Various sources of uncertainty are estimated using Monte Carlo methods. ProZES is being developed as a Windows program with a graphical user interface in German and English.

Current status

The software program was handed over to the Federal Office for Radiation Protection (BfS), which is the owner of ProZES. ProZES is to be further developed in the long term. The program has been available since January 2017. ProZES will run as a trial version for one year. After completion of the test phase, a new version is to be created, taking into account the experiences from the test phase.

Documentation and download of ProZES are available at:


ProZES screenshot.