A simple risk calculator may help predict low-risk patients with atypical ductal hyerplasia for whom observation is a safe alternative to surgery, according to new research.
“Some studies have begun to challenge the current paradigm in an attempt to avoid unnecessary surgery in low-risk patients; however, it’s quite difficult to predict which patients will have benign disease and which will upstage to cancer,” Daniel Lustig, MD, a surgical resident at the University of British Columbia, in Vancouver, said.
“We set out to identify risk factors associated with ADH [atypical ductal hyperplasia] upstaging to cancer on core needle biopsy [CNB], and then to develop a risk calculator capable of stratifying women into low- and high-risk categories.”
Using a prospectively designed database, Dr. Lustig and his colleagues looked at the association between 10 clinical, radiological and pathologic variables and the likelihood of ADH upstaging to cancer in 290 women between 2013 and 2017.
After univariate regression analysis, the researchers identified five variables associated with progression to cancer:
- suspicion for ductal carcinoma in situ (DCIS);
- more than one high-risk lesion;
- size on mammogram;
- size on ultrasound; and
- whether all calcifications had been removed after CNB.
“The variables most significantly associated with upstaging to cancer were suspicious for DCIS and one or more high-risk lesion, with odds ratios of 9.5 and 3.6, respectively,” Dr. Lustig said.
Using those five variables, Dr. Lustig and his colleagues developed a calculator, assigning a score of 1 for every metric present, with a scoring system ranging 0 to 5. Patients with a score of 0 are low risk, with only 2% likely to develop cancer; those with a score of 1 or higher are high risk, with 17% to 100% likely to upstage depending on the number of metrics present.
“As the risk score increases, so does the risk of finding a malignancy after the final surgical excision,” Dr. Lustig said.
To prospectively validate the calculator, they examined a cohort of patients who had undergone surgery for ADH. They found that the calculator’s predictions aligned closely with the actual patients’ progression rates.
“In our study, 50 women (17%) may have been able to avoid unnecessary surgery,” Dr. Lustig said.
“The current recommendation for BI-RADS [Breast Imaging-Reporting and Data System] category 3 lesions, which have a 2% or less risk of being cancer, is not a tissue diagnosis by surgery but short-term follow-up. We propose offering observation and six months to patients with a score of 0, and recommend operations for those with a risk score of 1 or more.”
Dr. Lustig and his team are working on refining the calculator by applying different score weightings to the various metrics. He presented the research at the American Society of Breast Surgeons’ Virtual Education Series, and received recognition for Best Scientific Presentations.
Carla S. Fisher, MD, an assistant professor of surgery at Perelman School of Medicine of the University of Pennsylvania, in Philadelphia, called the calculator “a brilliant tool” for a diagnosis that breast surgeons see commonly.
“While we don’t want to miss a cancer diagnosis, we know we are doing a lot of unnecessary surgery. Unfortunately, we’ve really struggled to find a best-prediction model. One of the nice things about this study is that they retrospectively identified risk factors and then prospectively validated their prediction tool, which makes it more robust.”
Dr. Fisher, who is the program director of the breast fellowship at the Hospital of the University of Pennsylvania, in Philadelphia, looks forward to further refinement of the calculator, but said it would be useful as is.
“I think most breast surgeons would agree that these variables are things we already use to discuss with our patients; for instance, in patients with ADH, I always review whether all calcifications have been removed after CNB. But this is a simple tool, easy to use, and you can share it with patients to support a decision for or against surgery.”
Please log in to post a comment