It's pretty much accepted fact that women who are genetic carriers or have a strong family history of breast cancer need to be monitored more closely than women who aren't at risk. And although MRI has shown considerable strength in detecting cancers not seen on mammography, one issue that still requires sorting out is whether high-risk women would benefit from multimodality screening.
Several presentations at the 2004 RSNA meeting in Chicago tackled this question. For the most part, they found that MRI came out on top, but the jury is still out on whether multimodality screening is erring on the side of caution or simply overkill.
DCIS in high-risk women
In the first study, Dr. Christiane Kuhl and colleagues from the University of Bonn in Germany used MRI, ultrasound, and mammography to screen for ductal carcinoma in situ (DCIS) in women who were proven or suspected BRCA carriers. Specifically, they looked at the features of DCIS on imaging.
The study population consisted of "a total of 530 women, out of 400 families, (who) have been followed up for an average of five years," Kuhl explained. "We do have a total of 1,700 complete screening rounds available where mammography, ultrasound, and MRI imaging were gathered in the same time point and read independently of each other to evaluate the possible contribution of these different imaging modalities."
Intraductal cancer was identified in 19 women (mean age 41). Nine of the 19 had pure DCIS, and all were clinically asymptomatic. The median size of the pure DCIS lesions was 13 mm. Six were high-grade DCIS, and three were low-grade. DCIS plus additional invasive cancer was found in 10 women.
While the three-modality screening protocol did find pure DCIS in eight of nine women and all 10 cases of DCIS with invasive cancer, individual modalities turned in varying results. In three patients with pure DCIS, mammography spotted suspicious mammographic calcifications. In eight patients with pure DCIS and all 10 with DCIS and invasion, the lesions were mammographically occult.
On MRI, pure DCIS appeared as segmental or linear enhancement that was unrelated to a mass in all but one woman. The MR imaging features were the same in the 10 cases of DCIS with invasive cancer.
Finally, none of the cancers were detected with ultrasound. All imaging missed one low-grade, pure DCIS, which was found incidentally during preventive mastectomy. The final sensitivity of each modality in this patient population was 33% for mammography, zero for ultrasound, and 89% for MRI. The group declared that MRI was the most helpful for identifying DCIS in this high-risk population, high-grade lesions in particular.
Because DCIS is not often associated with calcifications in this population, mammography's sensitivity is low, Kuhl added. RSNA session moderator Dr. Michael Racenstein asked Kuhl if MRI proved to be as sensitive in those rare cases in which DCIS was associated with calcifications. Kuhl said that they had cases of microinvasion that were deemed probably benign on mammography but spotted on MRI.
Surveillance with MR
Dr. Christopher Riedl and colleagues from the University of Vienna in Austria also found MRI to be superior to mammography and ultrasound for screening 234 high-risk women. The standardized protocol for this ongoing study is annual multimodality imaging with clinical examination. All of the 234 asymptomatic women (85 of whom were BRCA-positive) underwent imaging studies at least once.
Of this group, there were 17 breast carcinomas (confirmed with biopsy). All 17 were detected with MRI, 10 were found on mammography, and 12 on ultrasound, Riedl stated. Two of the 17 cancers were palpable.
The group reiterated that MRI should be considered the first-line imaging test for screening these high-risk women, based on the fact that, in their study, MRI turned in a sensitivity of 100%, a specificity of 93%, and an accuracy of 93%. In comparison, the sensitivity of mammography was 40%, the specificity was 96%, and the accuracy was 91%. Finally, ultrasound turned in a sensitivity of 29%, a specificity of 96%, and an accuracy of 90%.
Guiding multimodality screening
In a third study, researchers from the University of Pennsylvania in Philadelphia also found that multimodality imaging did have a significant cancer detection rate, but they discovered some pitfalls as well.
In this ongoing study, Dr. Mark Rosen, Ph.D., and colleagues used full-field digital mammography (FFDM), ultrasound, and MRI in combination or as single modalities for screening high-risk patients.
At the time of the RSNA meeting, 151 patients had been enrolled. The entry criteria were as follows:
- A history of prior contralateral breast cancer not treated with chemotherapy
- A Claus or Gail lifetime risk of > 25%
- A prior biopsy-proven atypical ductal hyperplasia, atypical lobular hyperplasia, or lobular carcinoma in situ
High-risk patients with heterogeneously dense breasts underwent imaging with all modalities, while patients with less-dense breasts were randomized to receive a single study or a combination study. All studies were interpreted independently, then in a consensus format.
Based on the results, four cancers were detected: three in high-risk patients, all with heterogeneously dense breasts, and one in the contralateral breast of a patient with scattered density. Three of the four cancers were caught only with MRI, while the fourth was seen on FFDM and MRI.
The cancer detection rate was 2.6%, while the biopsy rate was 21% with 41 biopsies recommended in 32 patients. The authors advised that the majority of biopsies (27) be performed under MRI guidance. Short-term follow-up was suggested in 22.5% of the women.
Rosen's groups cited these high biopsy and short-term follow-up rates as one of the drawbacks of multimodality images, despite its significant cancer detection rate. They stressed that further data was needed before any meaningful guidelines could be created for screening high-risk women with more than one test.
Italian experience
Finally, Dr. Francesco Sardanelli from the Instituto Policlinico San Donato in Milan shared the results of his group's study screening genetic carriers with more than one imaging modality. Unlike Rosen's group, these researchers expressed confidence in the benefits of a multimodality approach.
Eligible participants were women 25 years or older who were either BRCA carriers, had a first-degree relative who was a carrier, or had a strong family or personal history of breast cancer. Up to March 2004, 235 women had been enrolled at 17 centers. Eighty-eight of these women had a previous personal history of breast cancer. A total of 156 enrollees carried BRCA1 or BRCA2 mutations. Seventy-nine women had a strong family history of the disease.
Two annual rounds of screening with MR, ultrasound, and mammography were performed with 86 women undergoing both rounds. Contrast-enhanced MR exams were done on a 1.5-tesla scanner using a dynamic technique including 3D, T1-weighted GRE. Only lesions found on MR were worked up with additional MRI, ultrasound-guided biopsy, or MR-guided biopsy.
First-round multimodality screening found 11 patients with cancer, while second-round screening found another seven for a total of 18 cancers. For the women with a personal breast cancer history, the cancer incidence rate was 10.2%. In women with a strong family history, the cancer incidence rate was 8.9%. The overall cancer incidence rate was 7.7%.
Of the cancers that were found with multiple modalities, 40% were caught with MRI alone and 47% were detected with all three exams. Mammography and ultrasound had three false-positives each, which was two less than MRI.
Women with a genetic breast cancer risk would benefit from yearly multimodality screening, Sardanelli's group concluded. In particular, they found multimodality imaging to be adept at spotting infiltrating ductal carcinoma and infiltrating lobular carcinoma.
By Shalmali Pal
AuntMinnie.com staff writer
February 11, 2005
Related Reading
MRI finds breast cancers missed by mammography in high-risk women, December 2, 2004
Most mammographic features cannot predict high-risk DCIS, November 28, 2004
MR more likely than mammo to expose multiple malignancies in dense breasts, October 13, 2004
Women with genetic history of breast cancer benefit from MR screening, November 30, 2003
German study finds MRI superior for breast cancer detection in high-risk women, June 3, 2003
Copyright © 2005 AuntMinnie.com
