Jacky Luiten

Trends in incidence and tumor grade in screen‐detected DCIS and invasive breast cancer | 27 2 intermediate grade invasive carcinomas increased considerably, whereas the rate of high grade disease remained stable. Comparing initial and subsequent screens a significant part of the breast abnormalities was found at initial screening, which can be seen as the incidence of breast abnormalities in the normal population of asymptomatic woman under 50 years of age. At every subsequent screening round new abnormalities are detected. This finding suggests that new and clinically relevant abnormalities keep on developing and repeated screening therefore remains necessary. A decreasing, positive predictive value of recall was found during our 10‐year inclusion period. This decreased, positive predictive value may partly be explained by an increasing recall rate through the years. A Dutch optimization study stimulated screening radiologists to lower their threshold for recall, in order to increase CDR. 19 Also, the replacement of independent double reading by blinded double reading at the moment of the introduction of FFDM screening resulted in higher recall rate. 20,21 Finally, other possible explanations for the increased recall rates may be the routine use of two‐view screening mammography after the introduction of FFDM screening 22 , as well as a better and consistent image quality of digital mammography compared to screen‐film mammography. 17,18 Along with the increased recall rate, the overall CDR per 1000 screens increased gradually over the years. Our study has several strengths and limitations. A unique aspect of our study is that, to our knowledge, we are the first to investigate the correlation between screen‐detected DCIS and screen‐detected invasive breast cancer with complete information on tumor grade. Furthermore, our population‐based study design and our large sample size makes extrapolation of our results to other screening, comparable program possible. On the other hand, we only focused on women who attended the mammographic screening program, therefore we disregarded the women who did not attend and the abnormalities found in between two screening rounds (interval cancers). Moreover, extrapolation to the UK and US program may be limited, because in contrast to our biennial screening, women in the UK are screened every 3 years and in the US annual screening is offered. 23,24 In conclusion, at screening mammography mostly high grade DCIS and low or intermediate grade invasive carcinomas are detected. Except for the sharp increase of low grade DCIS and low grade invasive carcinomas following the introduction of FFDM screening, the grade distribution of DCIS and invasive carcinomas as well as the proportion of patients detected in the years after the introduction of FFDM remained rather stable compared to the era of SFM screening. By detecting high