FFDM vs. CRm - Is There More to the Story?

PACSman: You have to love the consistency of inconsistency especially when it deals with clinical studies. They are typically used to either promote one vendor’s technology or downgrade another’s. But what if the exact same study comes up with different results? That is what seems to have been the outcome of a recent report on a study comparing FFDM and CR mammography.

On October 10, 2008 the headlines blared “Flat-panel FFDM produces better image quality than CR mammography, study says,” yet a February 23, 2004 article in the same journal indicated, “Austrian study finds FFDM, CR equivalent for breast cancer detection.” Huh? Instead of these two distinguished doctors reporting their results I think it was Oliver North pinch hitting instead. You remember Ollie don’t you – the guy who stood up in front of a hearing on the Iran Contra affair back in 1987 and answered, “I wasn’t lying, Senator. I was presenting a different version of the facts.” And so too is what seems to have happened here. Same study, different interpretation.

The older study found the differences between the two technologies to be statistically insignificant. Interestingly enough the newer report showed the same finding, saying, “flat panel FFDM and CR mammography were comparable in diagnostic efficacy with similar sensitivity specificity positive predictive value (PPV) and negative predictive value (NPV)…” Where they seemed to differ is when the report stated that FFDM detected more lesions with calcifications than CR mammography. But is that the end of the story? Nope.
The new report said that “flat panel FFDM was rated better than CR mammography for each comparative aspect” and listed seven categories:1) brightness; 2) contrast; 3) sharpness; 4) noise; 5) artifacts; 6) detection of anatomic structures; and 7) detection of calcifications. Each of these was rated on a five point scale with 1 being excellent and 5 unacceptable….Interestingly they identified 7 categories, but clump all the ratings together, and not breaking them out as they should have...OK, let’s recap. Seven categories with 5 points each they could be rated on. This should provide you with a total point value of 5,250 (7 x 5 x 150), yet when you add the reported scores up you get a total point value of 6,750 points. The only logical answer is 2 extra categories snuck in there some how allowing a total of 45 possible points to be scored, not 35). Sooooo…again, what’s the deal here?

The original publication of this study rated 10 categories: 1) quality; 2) contrast; 3) sharpness; 4) blackness; 5) noise; 6) artifacts; 7) skin; 8) retromammiliary space; 9) parenchymal structures; 10) detection of calicifications. Why the change? And why the different results?
The first (2003) report indicated that image quality was significantly better with FFDM, except blackness, BUT for BI-RADS categorization CR mammography did better. FFDM has a sensitivity of 97% classification with a confidence interval of 91-97%, while DSPM (CR mammography) had a sensitivity of 100% with a confidence interval from 95-100%. The positive predictive value (PPV) of FFDM was 54% (46-63% CI) and 56% for DSPM (CR mammography). This was found not to be statistically significant.
Interestingly enough, while CRM did worse in Category B: scattered fibroglandular densities, they scored better in Category C: heterogeneously dense. From my very limited knowledge of mammography there are a helluva lot more C’s out there than B’s....(or maybe I’m thinking cup size…never mind (laugh).

So, why was some of this left out or reported differently in the new “report”? Good question. Could it possibly be to put a negative spin on DSPM vs. FFDM with having to rebut the previous findings? Scully was right – trust no one….

The third chart in the new study seems to have told what seems to be the truth – “FFDM and CR mammography were comparable in diagnostic efficacy, with similar sensitivity, specificity, positive predictive value (PPC) and negative predictive value (NPV) figures.”

The only thing that differed here was who reported this and when. The original study was reported by Dr. Wofgang Matzek at RSNA 2003. The new report was published in the European Journal of Radiology in September 2008 and attributes the findings to a team led by Dr. Gerd Schueller. I have to assume Dr. Matzek was part of this team as well since it is the same study…So why the five year delay in re-reporting old stats? Good question.
150 people also really can’t be considered much of a study...1,500 maybe, 15,000 much better, not 150....The technology has also leapfrogged in the five years since this study was conducted. If our esteemed counterparts in Vienna were serious about this all, they would redo the study today and see what the outcome is…

So do we shoot the publications for reporting it? Heck no. There is no requirement that a journal do any in-depth analysis of a report before publishing it. They get the news, they print it. Even the New York Times used as its banner “All the News That’s Fit to Print” and while this is no Sarah Palin fiasco, it does question the credibility of the research team that conducted the study. After all, this is a five year old study that was already presented at RSNA 2003. We’re also not talking a blown call on an NFL game here, but rather bringing into question a technology that can help save lives…


The facts seem to support that FFDM and CRM are indeed equal – and even Ollie North can’t change that…


Ms. PACS: Hey, PACSman, have you ever heard the phrase “penny wise, pound foolish”? Well, believe it or not, that is a pretty good description of how healthcare is run in this country. Preventative medicine - what’s that? It’s all about preventative upfront costs. That does seem to be consistent.

So, if early detection of breast cancer were the only consideration when adopting a technology, then couldn't the majority of women just get breast MRIs and forego the mental anguish of so often receiving inconclusive results from a mammogram. Then we could avoid having to schedule a return visit for another breast ultrasound – for which, despite the fact that the physician required the follow-up breast ultrasound, the insurance company refuses to pay. And have you heard from the latest studies – researchers found breast ultrasound is overly sensitive, generating a large number of false positives and unnecessary biopsies. Ouch! That hurts your wallet and you – or just women I should say.

As good as it may be, breast MRI is still not a replacement for mammography or ultrasound, even though it is used to identify early breast cancer not detected through other means, especially in women with dense breast tissue and those at high risk for the disease. And even though, according to www.radiologyinfo.org, breast MR allows, “MRI of the breast offers valuable information about many breast conditions that cannot be obtained by other imaging modalities, such as mammography or ultrasound.” And even though in another study, "MRI scans of women who were diagnosed with cancer in one breast detected over 90 percent of cancers in the other breast that were missed by mammography and clinical breast exam at initial diagnosis. Given the established rates of mammography and clinical breast exams for detecting cancer in the opposite, or contralateral breast, adding an MRI scan to the diagnostic evaluation effectively doubled the number of cancers immediately found in these women."*

As confusing as it may seem, breast MRI is only a supplemental tool. It still has some drawbacks - it cannot always distinguish between cancerous and non-cancerous abnormalities, which can lead to unnecessary breast biopsies, and it has historically been unable to effectively image calcifications, while mammography can image calcifications, which are often associated with early-stage breast cancers such as ductal carcinoma in situ (DCIS).

The fact is all of these imaging exams – FFDM, ultrasound, CR, breast MR, etc. - have pros and cons or inconsistencies. Since clinical efficacy sometimes takes a back seat to cost-effective solutions, and when vendors are driven to get a new technology out there in the marketplace, expect to see CR mammography make headway in the clinical setting. Financially it is viable - you can change over a screen-film mammography site to digital for half the cost of its current competitors. Some estimates say a single-room mammography setup can go to digital for less than $200,000. CR is also a manageable approach to mobile mammography because a reader can be easily installed on a van. Plus, CR offers a hybrid approach where CR and flat-panel mammography systems co-exist, similar to CR and DR in general X-ray.

So, after more studies are completed, whether or not flat panel FFDM proves to be better than CR mammography, if you want to know what will be the technology of choice, just follow the money. Even if it ends up costing more in the end.

*The American College of Radiology Imaging Network (ACRIN) study, supported by the National Cancer Institute (NCI), part of the National Institutes of Health (NIH), appears in the March 29, 2007 issue of the New England Journal of Medicine.* (http://www.cancer.gov/newscenter/pressreleases/MRIContralateralRelease)