Re: Changes in Breast Cancer Detection and Mammography Recall Rates After the Introduction of a Computer-Aided Detection System
Open accessThe DOI matches its Crossref record.
Abstract
In their recent article, Gur et al. ( 1 ) concluded that “The introduction of computer-aided detection…was not associated with statistically significant changes in recall and breast cancer detection rates for the entire group of radiologists as well as for the subset of seven radiologists who interpreted high volumes of mammograms.” Three important points must be made regarding this conclusion. First, the 95% confidence intervals (–11% to 19%) reported by Gur et al. allow for a wide range of detection rate changes. Even a 19% increase in detection rate would be consistent with their findings. The source of this large variation is not clear but may be a result of variability among readers. Second, we show that the use of computer-aided detection was associated with a substantial, but not statistically significant, increase in cancer detection rates for the subgroup of 17 low-volume radiologists in the Gur et al. study (Table 1 ). Among the 24 radiologists in their study, the seven high-volume radiologists each read more than 8000 mammograms during this 3-year study and the 17 low-volume radiologists interpreted an average of 1967 mammograms during the same period [table 1 in Gur et al. ( 1 ) ]. The authors analyzed the combined performance of all 24 radiologists and separately analyzed the performance of the seven high-volume radiologists [table 2 in Gur et al. ( 1 ) ]. However, they did not similarly analyze the performance of the 17 low-volume radiologists. As shown in Table 1 , we used the data reported by Gur et al. ( 1 ) to address this issue. We subtracted their data for high-volume radiologists from that for all radiologists and found that use of computer-aided detection by the low-volume radiologists was associated with a 19.7% increase in their detection rate (from 3.05 cancers per 1000 mammograms without computer-aided detection to 3.65 cancers per 1000 mammograms with computer-aided detection). Thus, use of computer-aided detection allowed the low-volume radiologists to increase their collective detection rate to a rate equal to that of the high-volume radiologists. However, because of the small sample size, this improvement was not statistically significant ( P = .37, chi-squared test). A sample size approximately five times larger than the one in the Gur et al. study would be needed to demonstrate that the nearly 20% increase in detection rate was statistically significant. Nevertheless, the increase we calculated using data reported by Gur et al. is similar to the 19.5% increase in cancer detection reported by Freer and Ulissey ( 2 ) in 2001 in their prospective study of computer-aided detection. Our third point relates to the authors' acknowledgement that during the study period, the percentage of women who were screened for the first time decreased from approximately 40% to 30%. This observation is clinically significant because a population undergoing periodic mammographic screening typically has lower cancer detection rates on second and third (i.e., incidence) screens than on the first (i.e., prevalence) screen ( 3 ). Gur et al. did not adjust for this confounding factor, which could obscure any evidence of benefit from computer-aided detection. Results of previous studies suggest that computer-aided detection should improve breast cancer detection rates and that the value of computer-aided detection may vary among different readers ( 4 , 5 ). The article by Gur et al. and the editorial by Elmore and Carney ( 6 ) should have mentioned that the breast cancer detection rates of low-volume radiologists benefited from computer-aided detection. The article by Gur et al. indicates that additional studies will be helpful to provide more insight into the benefits of computer-aided detection in different clinical settings. Mammography recall rates and breast cancer detection rates for high- and low-volume radiologists interpreting screening mammograms without and with computer-aided detection (CAD) Mammography recall rates and breast cancer detection rates for high- and low-volume radiologists interpreting screening mammograms without and with computer-aided detection (CAD)
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cited by
Diagnostic Accuracy of Digital Screening Mammography With and Without Computer-Aided Detection
Lehman, C. D., Wellman, R. D., Buist, D. S. M., Kerlikowske, K., Tosteson, A. N. A., & Miglioretti, D. L. (2015) · JAMA Internal Medicine · doi:10.1001/jamainternmed.2015.5231 · Cited by 678
The Potential of Radiomic-Based Phenotyping in Precision Medicine: A Review
Aerts, H. J. W. L. (2016) · JAMA Oncology · doi:10.1001/jamaoncol.2016.2631 · Cited by 669
Giger, M. L., Chan, H., & Boone, J. (2008) · Medical Physics · doi:10.1118/1.3013555 · Cited by 311
Computer aided detection (CAD): an overview
Castellino, R. A. (2005) · Cancer Imaging · doi:10.1102/1470-7330.2005.0018 · Cited by 284
Metadata
- Authors
- Feig, S. A., Sickles, E. A., Evans, W. P., & Linver, M. N.
- Venue
- JNCI Journal of the National Cancer Institute
- Publisher
- Oxford University Press (OUP)
- Published
- 2004-08-17
- Volume
- 96
- Issue
- 16
- Pages
- 1260-1261
- ISSN
- 0027-8874 (ISSN-L)
- 1460-2105
- OpenAlex
- W2083080230
- Cited by
- 56
- References
- 0
- Type
- journal article
- Language
- en
- Open access
- bronze
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Re: Changes in Breast Cancer Detection and Mammography Recall Rates After the Introduction of a Computer-Aided Detection System. https://app.cochat.ai/papers/10.1093/jnci/djh257
- The paper
- Feig, S. A., Sickles, E. A., Evans, W. P., & Linver, M. N. (2004). Re: Changes in Breast Cancer Detection and Mammography Recall Rates After the Introduction of a Computer-Aided Detection System. JNCI Journal of the National Cancer Institute. https://doi.org/10.1093/jnci/djh257
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.