Diane Chang BS, Presenter: Nothing to Disclose

Robert Mark Nishikawa PhD, Abstract Co-Author: Shareholder, Hologic, Inc, Bedford, MA Royalties, Hologic, Inc, Bedford, MA Research funded, Hologic, Inc, Bedford, MA Consultant, FUJIFILM Holdings Corporation, Stamford, CT Scientific Advisory Board, Dexela Limited, United Kingdom Consultant, Carestream Health, Inc, Rochester, NY Research grant, Carestream Health, Inc, Rochester, NY

Lorenzo Pesce PhD, Abstract Co-Author: Nothing to Disclose

To perform a meta-analysis to compare the performance of screening mammography when using computer-aided detection (CADe) as opposed to using independent double reading.

We included peer-reviewed studies that were performed prospectively and measured the number of cancers detected and the recall rate with and without CADe. We included independent double reading studies performed in the USA, as all CADe studies were done in the USA.

Seven matched CADe studies qualified for this analysis and included a total of 55937 women screened and four double reading studies that included 162980 women. The relative increase in cancers detected when CADe was used compared to when CADe was not used (odd ratio, OR) was 1.10 with 95% CI of (1.06-1.15). The OR for recall rate was 1.13 with 95% CI of (1.12-1.14). There was no statistically significant difference in OR for cancer detection and recall rate. For the four independent double reading studies, the OR was 1.08 (1.06-1.11) for cancers detected and 1.17 (1.16-1.17) for the recall rates, indicating that the increase in recall rate was higher than the increase in cancer detection. Between the two modalities, the 95% CI for difference in detection rates was (-0.028,+0.061) and (-0.050,-0.024) for recall rates.

Both CADe and double reading increase the number of cancers detected and the recall rate. For both modalities the increase in cancer detection is comparable. However, the increase in recall rate for independent double reading is higher than for CADe.

Using CADe in screening mammography gives comparable or better results than independent double reading.

Chang, D, Nishikawa, R, Pesce, L, Meta-Analysis of Computer-aided Detection in Screening Mammography and Independent Double Reading.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6013543.html