What’s the best way to screen dense breasts?

Screening smackdown: 3D mammogram vs. abbreviated MRI

The study, conducted at 47 institutions in the U.S. and one in Germany, included nearly 1,500 women, ages 40 to 75, all of whom had dense breast tissue. Participants were randomized first to receive either DBT or AB-MRI. Then they were screened, first with one modality, then the other, at study start (both within a 24-hour period) and again after a year (researchers will do a three-year follow-up).

Results were compelling.

All told, 17 out of 1,444 women were diagnosed with invasive breast cancer (either ductal or lobular) and six women were found to have DCIS, or ductal carcinoma in situ, sometimes referred to as stage 0 breast cancer.

AB-MRI detected 22 out of these 23 breast cancers. It found invasive breast cancers in all 17 cases where it was present and found DCIS in 5 of the 6 women. DBT or 3D mammography, on the other hand, found 9 out of 23 possible breast cancers, accurately detecting invasive breast cancers in 7 out of 17 women and DCIS in 2 of 6 women.

Overall, the detection rate for finding invasive cancers (the study’s primary endpoint) for AB-MRI was 11.8 per 1,000 women versus DBT’s detection rate of 4.8 per 1,000 women.

“Abbreviated MRI caught over two times more cancers. That’s very, very promising,” said Rahbar, clinical director of breast imaging at the Hutch’s clinical-care partner, Seattle Cancer Care Alliance. “It confirms that when we add contrast and use a higher quality technique like MRI, we catch more cancers. But the abbreviated MRI also had a higher biopsy rate and more false positives than DBT.”

Weighing the harms and benefits

You might call that sticking point No. 1. The AB-MRIs are more sensitive so they found more suspicious spots that could potentially be cancer.

In the study, AB-MRI called for 107 biopsies, a procedure where breast tissue is excised and analyzed for cancer cells, which was nearly four times the rate of DBT (it called for 29). Biopsies can leave the breast bruised and sore for a few days, and as always, complications can happen.

Rahbar said with abbreviated MRIs “only about 20% of the biopsies were actually cancer,” referring to the positive predictive value of each test, or how often it was right in identifying a suspicious lesion as a cancer. “In 3D mammography, it was about 35%,” he said.

In other words, the majority of women who have a potential breast cancer identified by AB-MRI will undergo an unnecessary biopsy.

“There’s a higher number of false positives with abbreviated MRI,” Rahbar said. “While the high sensitivity is very exciting — they’re catching more cancers — we’d also be recommending more biopsies. And they will be more expensive.”

It’s also unclear how frequently women would need AB-MRIs. The trial was a two-year study, but the paper reported only the first round of screening.

“It’s possible fewer cancers will be identified in subsequent rounds of MRI, so it’s not clear if this test should be done yearly,” Rahbar said.

How much will it cost? It’s complicated

That brings us to sticking point No. 2: cost.

Currently, the cost for an MRI can run anywhere from $400 to $3,500 depending on where in the country it’s done and what part of the body is being imaged. The business data site Statista lists the average cost in 2017 at $1,430.

“Breast MRI is expensive,” said Dr. Gary Lyman, a health economics researcher and policy expert with the Hutchinson Institute for Cancer Outcomes Research, or HICOR. “That’s why the guidelines limit it to high-risk settings. Currently, breast MRI is only approved for patients with a family history or a risk of 20% or greater for developing breast cancer.”

In use since the late 1970s, MRIs can take anywhere from 30 to 90 minutes and involve an IV-injected contrast agent or dye and multiple images taken over a period of time. The abbreviated MRIs performed in the study also used a contrast agent but took fewer images. They were also much shorter — each took less than 10 minutes.

But as it now stands, “there’s no standard for an abbreviated MRI,” Rahbar said. “This is the first study to examine it across many sites in a clinical trial. We don’t even have a realistic price for it because there’s no insurance code.”  

That code, known as a CPT or “current procedural terminology” code, is what’s used for medical billing. No medical billing code? No insurance coverage.  

“This situation is not dissimilar to other new technologies when they first get developed, like tumor sequencing or biosimilars,” Lyman said. “When there’s no code to begin with, it will take a year or two for CMS [the Centers for Medicare and Medicaid Services] to see what’s happening in practice and create a code so it can then be billed.”

Generally, CMS — also known as government-sponsored public health insurance — leads the charge with regard to reimbursement for physician services, with private insurers soon following suit.

Until that happens, Lyman said, it’s pretty much the Wild West. Health centers may have to “fly on their own” in terms of billing and establishing best practices, he said. And patients could potentially be charged full price for an MRI, even though they’re getting a truncated version.

“My guess is CMS will very quickly get a code and be put into the regulations so clinicians can bill for it,” Lyman said. “But this is unlikely to happen in less than a year and could take longer. The results of the next [study] screening will further clarify the performance of this approach and its value to patients, health systems and society. These are the types of studies that we need.”

Dr. Christopher Comstock of Memorial Sloan Kettering Cancer Center, lead author of the paper, said only a few cancer centers currently offer AB-MRI.

“Our hope,” he said, “is that the study will provide the impetus for more centers to provide this new test and that insurers will see its benefit to improve early detection.”