Mammography Less Accurate After Breast Cancer

Zosia Chustecka

February 23, 2011

February 23, 2011 — Women who have had breast cancer are at increased risk of developing a second breast cancer, and it is recommended that they undergo an annual mammogram.

However, a new study shows that mammography screening is less accurate in women who have had breast cancer than in those who have not, and the authors suggest that "a more tailored strategy . . . may be warranted."

The finding appears in the February 23 issue of the Journal of the American Medical Association.

The researchers examined results from more than 100,000 mammograms and found that, "in general, screening did not perform as well in women with a personal history of breast cancer."

Screening sensitivity in such women was lower than in women without a personal history (65.4% vs 76.5%), largely because of a lower sensitivity for detection of invasive breast cancer (61.1% vs 75.7%), the study found.

The study was conducted by Nehmat Houssami MBBS, PhD, and Les Irwig, MBBCh, PhD, from the University of Sydney in New South Wales, Australia, working closely with American researchers from the Breast Cancer Screening Consortium and Group Health Research Institute in Seattle, Washington.

They believe that this is the "the first comprehensive study" of this issue, and that it "provides evidence to inform practice and guide recommendations on mammography screening in women with a personal history of breast cancer."

This is an important article, and many of these findings are new.

"Our findings support annual mammography screenings in women [with a personal history of breast cancer], but they also highlight issues needing further evaluation," they conclude.

"I think this is an important article, and many of these findings are new," said Michael Linver, MD, FACR, director of mammography at X-Ray Associates of New Mexico PC, and clinical professor in the Department of Radiology at the University of New Mexico School of Medicine, in Albuquerque, who was approached for comment.

Details of the Findings

For the study, Dr. Houssami and colleagues analyzed data on women who had been screened from 1996 to 2007 at 5 facilities in the United States that were associated with the Breast Cancer Surveillance Consortium, which is funded by the National Cancer Institute.

Dr. Houssami explained to Medscape Medical News that the Breast Cancer Surveillance Consortium was the only potential source that she could find for comprehensive high-quality data on women with a personal history of breast cancer, and that she and her colleague Dr. Irwig worked very closely with the American researchers by email and phone. She was prompted to investigate outcomes in this group of women because they represent an "increasing group in my practice," she reported.

The team analyzed results from 58,870 mammograms for 19,078 women with a history of early-stage breast cancer (in situ or stage 1 to 2 invasive), and matched them with 58,870 mammograms from 55,315 women without a history of breast cancer.

The mammograms were matched for breast density, age group, mammography year, and registry location. This is a strength of the study design, Dr. Houssami noted, and it means that, "from a global perspective, if you work in a screening context that has different screening outcomes from those in the United States for some performance measures, the relative screening outcomes of the women [with a personal history of breast cancer vs no history] are still informative and relevant."

Among the women with a history of breast cancer, there were 655 second cancers (499 invasive, 156 ductal carcinoma in situ [DCIS]) within 1 year of screening mammography. The cancer rate was 10.5 per 1000 screens.

Among the women without such a history, there were 342 cancers (285 invasive and 57 DCIS). The cancer rate was 5.8 per 1000 screens.

The cancer rate, the cancer detection rate, and the interval cancer rate (the number of cancers detected in the interval between screenings) were 1.3 to 2.6 times higher in women with a history of breast cancer.

Women with a history of breast cancer were more likely to have additional imagining (with mammography or ultrasound), and a recommendation for fine-needle aspiration, biopsy, or surgical consultation after assessment, the researchers reported.

Despite the higher underlying cancer rates and the higher rate of additional evaluation, screening has a lower accuracy in women with a history of breast cancer than in women without, they conclude.

Effects of Previous Treatments

The study also found that treatment for breast cancer had an effect on the accuracy of mammograms in women with a history of breast cancer.

For example, the specificity was higher and the abnormal interpretation rate was lower in women who had undergone a mastectomy than in those who had undergone breast-conserving therapy.

Radiation treatment was associated with a very small but significant reduction in specificity and an increase in the abnormal interpretation rate.

Plus, women who had received chemotherapy were significantly less likely to have their cancer detected by mammography than women who did not. This lower screening sensitivity in women who had chemotherapy remained evident even when the analysis was adjusted for age, cancer stage, and other relevant variables, Dr. Houssami said.

Our work is the first to demonstrate this.

"It is not entirely surprising that some of the therapeutic factors were associated with mammography accuracy, although our work is the first to demonstrate this to my knowledge," Dr. Houssami told Medscape Medical News.

"This is the first time that I have heard of this," agreed radiologist Wendie Berg, MD, PhD, FACR, from the American College of Radiology Imaging Network, in Lutherville, Maryland. She will be taking up a new academic position at the University of Pittsburgh, in Pennsylvania, shortly, and was approached for comment by Medscape Medical News.

"It is intriguing that chemotherapy reduces the accuracy of the mammogram, but unclear as to why this would happen," she said, speculating that chemotherapy might hold the cancer in check only for a while. However, there are not enough data here to make any new recommendations for this subset of patients with a history of breast cancer who received chemotherapy.

Findings Are Not Practice Changing

Overall, there are not enough data in the paper to make any new recommendations, Dr. Berg noted. The study provides, for the first time, fundamental information about what has seemed intuitive to those in the field (i.e., that mammograms are less accurate in women with a personal history of breast cancer), she explained. It also provides "a good starting point with baseline data to show how we are doing in this patient population," she added.

However, there is not enough information here to make recommendations about additional screening, Dr. Berg said. The interval cancer rate (34%) was high in these women with a history of breast cancer, but the cancer was mostly caught early when it was still treatable. It was not clear from the information given in the paper to what extent the interval cancers were discovered by supplemental screening (ultrasound or magnetic resonance imaging [MRI]), or whether they were discovered clinically, Dr. Berg reported. It appears that the researchers themselves were unable to ascertain this information, and this is a limitation of the study. Without that information, the benefit of, or need for, supplemental screening is unclear, she said.

Dr. Houssami noted that it is possible that some of the interval cancers were found on screening (e.g., with MRI, although this is not recommended for routine screening in women with a history of breast cancer) being done between mammograms. "However, this is unlikely to explain all or even most of it, as the timeframe of our study was before the popularity of screening MRI," she added.

"There are not really any data here that would cause anyone to change the practice of screening women with [a personal history of breast cancer] with mammography," agreed Dr. Linver, who was also approached by Medscape Medical News for an outside comment on the paper.

"Mammography should still be performed at least yearly in this group," he said. However, he suggested that "adding yearly MRI or ultrasound to the screening regimen should be considered as well, in my opinion. This is one of the groups at highest risk for breast cancer, and deserves to be screened as carefully as possible to find their second (or more) cancers early enough to allow for additional treatment to be optimally effective in curing the disease."

Dr. Houssami and colleagues are not recommending changes in practice, but they say that the new findings should prompt evaluation (in well-designed studies) of the potential role of tailored screening in this patient population.

"I would not argue for more frequent mammography," Dr. Houssami told Medscape Medical News. "I am thinking of adjunct screening in some groups of women with [a personal history of breast cancer] . . . which may involve ultrasound or MRI."

"But it is very important to understand that I am referring to evaluation and not a change in current practice, at least not until we have evidence on whether adjunct screening in these groups of women [with a personal history of breast cancer] translates into a reduction in interval cancers and the resulting false-positive trade-off to achieve this," Dr. Houssami emphasized. "It is possible that adjunct screening in these women may lead to a lot of false positives, which would not be acceptable in screening."

Role of Scar Tissue

Dr. Linver also commented on the impact that scar tissue from the surgical treatment of a previous breast cancer might have on later screening.

"It occurs to me that almost all these patients have prominent postoperative scars at the site of the previous cancer, adding greatly to the level of difficulty in finding a new or recurrent cancer. Since postoperative scars themselves have the appearance of cancer, the task of finding a new cancer in or near the scar is much more challenging, compared with images of breasts where there is no large scar present. The situation is analogous to the comparison between images of dense breast tissue and fatty breast tissue. We know the ability to find a breast cancer in a fatty background is infinitely easier than in a dense tissue background," he explained.

However, Dr. Houssami disagreed with this explanation. "I don't think this is the reason for the markedly lower sensitivity in women [with a personal history of breast cancer], although it might explain some of the reduced specificity. When you look at our data, sensitivity was similar for the previously affected (conserved) breast and the contralateral breast. Therefore, the suggested explanation of scar tissue does not fit."

"This is an interesting paper," said Daniel Kopans, MD, senior radiologist at Massachusetts General Hospital, and professor of radiology at Harvard Medical School in Boston, who was approached for comment.

"That said, it has been known for decades and it is not surprising that the sensitivity of mammography to recurrences and new cancers on the side treated for cancer is diminished, since the treated breast is distorted. It is more difficult to position in the mammography device, and surgery and radiation changes can be confusing and mask recurrences and new cancers in the treated breast," Dr. Kopans told Medscape Medical News.

"The lower sensitivity on the untreated side should be investigated more completely," he added.

"The bottom line," Dr. Kopans said, "is that there are not many options" for these women with a history of breast cancer. There is clinical examination (not very good), breast self-examination (women still find a large percentage of their cancers), and mammography."

"MRI will likely be more sensitive in this situation, but it is also far more expensive and will likely lead to many more biopsies for what prove to be benign findings (unnecessary)," he added, echoing the comments made by Dr. Houssami when discussing further evaluation in this patient population.

Dr. Houssami and coauthors have disclosed no relevant financial relationships. Dr. Linver reports serving on the advisory board of the women's health division of Philips, and as a consultant to Hologic. Dr. Berg reports receiving compensation for reading, data analysis, and manuscript preparation for Supersonic Imagine, and serving on the medical advisory board of Philips.

JAMA. 2011;305:790-799. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: