National Performance Indicators for the Breast Cancer Screening Program for the 2011–2012 Period
Download the tables of all national indicators for the combined years 2011 and 2012, broken down by rank (xls - pdf).
Download the tables of key indicators broken down by technology type for the combined years 2011 and 2012, by rank (xls - pdf).
The organized breast cancer screening program aims to detect the disease at an early stage, in order to enable appropriate treatments that will reduce mortality as well as other serious consequences of late detection of the disease. To maximize benefits and minimize risks—that is, to be effective—an organized screening program must be of high quality. It must therefore be evaluated. We present here the quality and early effectiveness indicators for the organized breast cancer screening program for the 2011–2012 period.
Data from all screenings performed as part of the organized screening program across all departments (a total of 99, with the two departments of Corsica evaluated together) were included to form the 2012 national database, comprising 2,466,910 screenings. At the same time, updated data were received regarding all positive screening results from 2008 to 2011, for all departments.
Since women are encouraged to undergo mammography screening for breast cancer every two years, the results are presented over a cumulative two-year period. During this 2011–2012 period, 4,899,223 screenings were performed. Note that the overall results summarized here mask significant geographic disparities, as can be seen by examining the results by region or by department.
Quality Indicators (Screening Results)
The data provide information on the number of screenings performed, their results, and diagnostic assessments for the 2011–2012 period. Indicators T02 through T05 and P07 through P09 present the corresponding rates and percentages.
For the 2011–2012 period, 4% of screenings were first organized screenings with no prior history of mammography (hereinafter referred to as “true initial screening”), 15% were first organized screenings with a history of mammography, and 81% were subsequent screenings.
The rate of positive mammograms (i.e., those judged abnormal or suspicious) on initial reading before further evaluation (L1) is 14 per 100 women undergoing a “true” initial screening and 6.8 per 100 women undergoing a follow-up screening. These rates are slightly higher than those for the 2010–2011 period. An immediate diagnostic evaluation (BDI) is performed for 97% of positive mammograms in L1.
The rate of positive mammograms on second reading (L2) is 2 per 100 women undergoing a "true" initial screening and 1.2 per 100 women undergoing a follow-up screening. These rates are similar to those for the 2010–2011 period. Among positive mammograms in L2, a deferred diagnostic evaluation (BDD) is performed for 77% of “true” initial screenings and 83.6% of follow-up screenings.
Ultimately, the rate of positive mammograms on first or second reading prior to evaluation (equivalent to the “European recall rate”) is 15.9 per 100 women undergoing a “true” initial screening and 7.9 per 100 subsequent screenings. The rate of positive screenings after evaluation (BDI or BDD) is 8% for "true" initial screenings and 3.1% for subsequent screenings. These rates are slightly higher than those for the 2010–2011 period.
Program Performance Indicators
The incidence data provide information on cancer incidence and characteristics for the 2011–2012 period. The TC01 and PC02 to PC10 indicators present the corresponding rates and percentages.
At the time of data extraction, 36,776 cancers (invasive cancers and in situ cancers, excluding lobular carcinoma in situ) had been recorded for the 2011–2012 period, representing a rate of 7.5 cancers per 1,000 women screened. This rate is 13.4 per 1,000 women undergoing an initial "true" screening and 7.1 per 1,000 subsequent screenings. These rates are slightly higher than those for the 2010–2011 period. Based on evaluations from previous years, it is estimated that approximately 5% of cancers will be added to these figures during updates of cancers detected following screening in 2012.
The percentages of cancers with a good prognosis among all detected cancers are indirect indicators of the program’s effectiveness.
Among women undergoing a "true" initial screening, 10.9% of cancers with known invasive/in situ status were in situ cancers, 23.8% of cancers with known size were 10 mm or smaller, and 63.5% of cancers with known lymph node status had no lymph node involvement.
Among women undergoing follow-up screening, 15.1% of cancers with known invasive/in situ status were in situ cancers, 39.2% of cancers with known size were 10 mm or smaller, and 77.4% of cancers with known lymph node status had no lymph node involvement.
Finally, among 100 cancers detected in 2011–2012, 3.5 were detected through the second reading among women undergoing a “true” initial screening and 6.9 among those undergoing follow-up screenings. These percentages are slightly lower compared to the 2011–2012 period.
Positive Predictive Value (PPV)
The indicators represent the positive predictive values before and after diagnostic evaluations. Calculating the positive predictive values of screening within the screening program requires knowing the outcome of the screening procedure. At the time of data transmission to the InVS, this information may be either definitive (cancer, normal, benign, woman definitively reported as lost to follow-up or deceased) or provisional (follow-up still ongoing or screening management facility awaiting information). At the time of analysis, the information was definitive for 98.9% of screened women, and 93.9% for women who had positive screening results prior to diagnostic evaluation.
The positive predictive value of positive screening results prior to diagnostic evaluation—PPV01—(a specification indicator), i.e., the percentage of cancers detected following a positive mammogram or clinical examination relative to the number of women with a positive mammogram prior to diagnostic evaluation (immediate or delayed), is 8.4%.
The PPV of the screening process completed, i.e., the percentage of cancers detected relative to all positive screenings following diagnostic evaluation—PPV02—(immediate or delayed) is 19.6%.
These two PPVs are stable compared to the 2010–2011 period. Based on evaluations from previous years, it is estimated that approximately 5% of cancers will be added to these figures during updates of cancers detected following screening in 2012. This will slightly increase these PPVs.
Digital Mammography
Since 2008, digital mammography has been authorized in the national program in addition to analog technology. Two main types of digital technologies are used: direct radiography (DR) and computed radiography (CR). Over the 2011–2012 period, 86% of screenings were performed using digital technology, broken down into 42% CR and 58% DR. The rate of technically correct images is approximately 98% for all three technologies.
The crude positivity rates on first reading prior to further evaluation, as well as the rates of detected cancers, are higher for digital than for analog. And among digital methods, these rates are higher for DR than for CR. Consequently, the contribution of the second reading appears to be lower with DR digital technology, since the rate of positive mammograms on second reading prior to evaluation and the percentage of cancers detected by the second reader among all detected cancers are lower. These results hold true regardless of the screening round
If we focus solely on subsequent screenings—that is, 81% of screenings from the 2011–2012 period—we observe that cancers with a good prognosis are more numerous with digital technology, particularly with digital DR. These better results for DR are also observed for “true” initial screenings, except for the rate of in situ cancer. The PPVs of positive screenings prior to further evaluation are higher.
It should be noted that the differences observed between the three types of mammograms do not correspond solely to differences in technological performance: confounding factors, notably the experience of radiologists, also play a role.
Interval cancers
A project was carried out in 2014 that enabled the cross-referencing of data from cancer registries with data from administrative systems, in order to comprehensively identify, according to a common definition, interval cancers within the area covered by a registry, namely 13 departments. The results of this cross-referencing are currently being processed and will be published at a later date.
