Why Nomograms Provide Optimal Risk Assessment of Recurrent DCIS

Kimberly J. Van Zee, MS, MD, FACS

With advances in cancer detection, the resulting increased diagnosis of ductal carcinoma in situ (DCIS) — considered an early stage, non-invasive form of breast cancer — poses challenges to the clinician when it comes to future risk assessment post initial treatment, said Kimberly J. Van Zee, MS, MD, FACS.

DCIS now accounts for more than 20 percent of all newly diagnosed breast cancer, noted Van Zee. The problem with DCIS for women who choose breast-conserving surgery alone, explained Van Zee, is that the local recurrence (LR) rate is relatively high. Women under 40 who undergo BCS for DCIS also face a greater risk of developing invasive breast cancer. In this climate, it’s imperative for the clinician to be able to provide as accurate an LR risk assessment as possible to the patient, said Van Zee.

“A clinician’s assessment of risk is about the same as flipping a coin,” said Van Zee. In order to do better than chance, Van Zee argued for the use of nomogram over other LR risk assessment tools, including 12-gene assay panel, calling nomogram “the perfect tool for estimating DCIS risk.”

A surgical oncologist at the Memorial Sloan Kettering Cancer Center and a professor of surgery at the Weill Medical College of Cornell University, Van Zee presented her findings on the benefits of nomogram during a panel on “Breast Pathology for the Non-Pathologist” at the 18^thannual Lynn Sage Breast Cancer Symposium in Chicago.

Van Zee explained that the online assessment tool, which takes into account the 10 known factors that influence recurrence of DCIS, is very user friendly. More importantly, she said, nomogram can handle the complex relationships between risk factors, providing a graphical representation of risk based on a mathematical equation, integrating numerous low and high risk characteristics to provide an individualized risk estimate. “In a perfectly calibrated tool, we expect to see a slope of 1 percent between the predicted risk and the actual, and we see that with nomogram,” said Van Zee.

“Nomogram satisfies the two main measures of how predictive a tool is when it comes to risk—good calibration and discrimination,” said Van Zee. Four independent studies have externally validated nomogram’s 5-year and 10-year LR predicted probability. In addition, compared to other tools, said Van Zee, nomogram is incredibly cost effective because it’s free.

The other option, which Van Zee said she does not recommend because of its independently-verified weakness in calibration, is the 12-gene assay panel, a tool specifically developed to quantify LR risk in DCIS patients after BCS.

Van Zee explained the deficit: “In a perfect model, you expect it to be going up, but with the 12-gene assay panel, studies show it peaks with intermediate risk assessment and then you see a downward slope. The intermediate risk was higher than the high DCIS score. Keep in mind that a high DCIS score is predictive of less invasive risk of cancer.”

The panel may incorrectly weight risk factors creating a DCIS risk score that’s not predictive of future harm. Age is one of those risk factors that the 12-gene assay panel may incorrectly weight, what Van Zee jokingly called a “novel” predictor of DCIS risk because it perfectly follows a predictive slope of 1 percent. As a woman gets older, her risk for DCIS decreases, explained Van Zee.

“The oldest woman had an 80 percent reduction in recurrence over time than the youngest patients, those under 40. Younger women were also more likely to have an invasive recurrence versus a DCIS recurrence,” noted Van Zee.

The time factor—the year in which a woman had her BCS—is also important to a woman’s overall risk assessment, and another factor calculated well by nomogram, but not represented in the panel test, said Van Zee.

“Clinical attributes, which include age, seem to trump the DCIS risk score from the 12-panel assay test,” said Van Zee.

“In this era of individualized care, we want to give the best options for women, based on her priorities and values. Having an accurate risk estimate helps her align those to her situation,” Van Zee summarized. “Having the most accurate risk assessment tool is particularly important in an era where we’re starting to see a fall in breast-conserving surgery and an increase in unilateral and even bilateral mastectomies in DCIS. So it’s important we have the best ways for women to estimate their real risk.”