MAMMOGRAPHY BEFORE ULTRASOUND

THERE ARE MANY MORE BREAST CANCERS GROWINGIN THE POPULATION THAN ARE DIAGNOSED EACH YEAR

Many years ago we developed a very simplemodel of tumor growth that can explain a number of observations (1).  It explained how tests like Magnetic Resonance Imaging (MRI) are able to detect so many more breast cancers than mammography (2).  It also explains why breast cancers in young women tend to be faster growing and more aggressive.  If we assume that cancers double in cell number from 60 days for fast growing cancers to 120 days for an intermediate growth cancer and as long as 180 days or more for slow growing cancers, and since it takes approximately 1 million cells to make a 1 millimeter tumor, and if all cancers have an “in situ” period of growth, it may take anywhere from 10 to aslong as 30 years to go from a single cell to a 2 cm., clinically detected cancer.  An intermediate growth cancer might take only 20 years to reach the 2 cm. size while a fast-growing cancer may only take 10 years to become 2 cm..  If we assume that most cancers, prior to screening, were detected at around the 2 cm. size each year, this means that, in order to sustain the annual rate of clinically evident cancers, there must be a very large number of cancers that are smaller in size, and growing undetected in the population to sustain the annual rate of breast cancers diagnosed in the population.

The good news is that metastatic disease is likely a function of the number of cells in a cancer.  This is clearly a statistical phenomenon since there are women who present with metastatic disease with a cancer that is so small that it cannot be found by whole breast pathological sectioning, while at the other end of the spectrum are women with very large tumors where the cancer is totally confined to the breast. There is almost certainly a bell-shaped curve with the majority of cancers developing between these two extremes. The more cells in a cancer the greater the chance that it has become successfully metastatic (3).  In general, the smaller a cancer is at the time of diagnosis, the greater the chance that is has not yet become successfully metastatic and the greater the chance that it can be cured.

HOW DO WE KNOW THAT EARLY DETECTION SAVESLIVES

The trial that experts agree is the only way to prove a benefit from an intervention is the “randomized, controlledtrial” (RCT).  RCT’s with a mortality endpoint eliminate the biases that compromise observational analyses.  With regard to breast cancer screening, thousands of women are randomly divided into two groups. If the allocation it truly random, and the number of participants issufficiently large, the same number of women in each group will develop breast cancer and the same number of women will die from breast cancer if nothing is done differently in either group.  In a screening trial one group is tested with the intervention while the other group is provided with the “usual” care.  If the tested group has fewer women who, ultimately, die from breast cancer and if the differences in deaths is “statistically significant” then the ability of the test to reduce deaths has been “proven”.

The only randomized, controlled trials ofan imaging test for breast cancer screening have involved mammography as the technology.  The Health Insurance Plan of New York showed that compared to women who were not being clinically examined,Clinical Breast Examination (CBE) could actually save lives, but the subsequent RCT’s of mammography screening proved that even earlier detection by mammography could continue to reduce deaths from breast cancer.

In part, the reason that mammography screening is the gold standard for early detection is that it is the only test that has been directly proven in RCT’s to lower the death rate from breast cancer.  I am fairly certain that Ultrasound screening can also contribute to the decline in breast cancer deaths because it can detect some cancers earlier that are not evident on mammography.  Perhaps the best way to screen for breast cancer would be the use of Magnetic Resonance Imaging.  MRI detects breast cancers that are not evident on mammography or ultrasound. However, aside from its high cost and low accessibility and the need to inject gadolinium intravenously, MR has also never been tested in an RCT. 

DETECTION VS. DIAGNOSIS

Years ago Moskowitz wrote “Screening is notdiagnosis” (4).  This is a fundamentally important point.  A fishing net may be very good at catching herring, but it catches other fish as well. The “catch” must be sorted to remove other fish.  The net is a screening test.  Diagnostic tests do the sorting.  In fact, although mammography is an excellent screening test it is not a good “diagnostic” test.  Mammography is efficient and effective atfinding “things”, many of which prove to be cancers that can cured (screening),but it finds other “things” that are not cancer and is not very good at telling them apart (diagnosis).  In fact, the only reason to use mammography is for screening to detect breast cancer earlier at a time when cure is more likely. 

Mammography is the only breast cancer screening test that has been proven to save lives.  In fact, it could be argued that the only role for mammography is for screening. Mammography’s primary (and perhaps only) use is in detecting lesions inthe breast, many of which represent early breast cancers.  However, its role as a “diagnostic” test is very limited.  For a given lesion that raises concern, mammography is often not able to differentiate benign from malignant with sufficient accuracy to obviate the need for a tissue diagnosis.  It cannot, for example,differentiate a cyst from a solid mass. 

I am surprised that some physicians screenwith ultrasound and then turn to mammography for further evaluation.  It makes far more sense to screen with mammography and turn to ultrasound to evaluate “things” that are detected by the mammogram.  Mammography is excellentfor screening, but poor at diagnosis.  It finds “things” in the breast many of which prove to be early breast cancer, but it is almost useless as a diagnostic test to evaluate a lesion in the breast to determine whether it is benign or malignant. 

On the other hand, ultrasound is anexcellent “diagnostic” test in that it can easily differentiate simple (benign)cysts from solid lesions in the breast. There are some breast cancers that are not evident by mammography that are evident by ultrasound, but the “specificity” of ultrasound is even lower than for mammography, and major concerns have been raised that the recall rate(women who need additional evaluation) is even higher for ultrasound than mammography. 

WHY MAMMOGRAPHY IS THE MAIN SCREENING TEST

The primary reason that mammography is the main screening test around the World is that screening using mammography has been directly proven through RCT’s to save lives for women ages 40-74 (the ages of the women who participated in the trials). Mammography screening has been tested repeatedly.  We understand its advantages and limitations.  It has been developed to permit screening very large numbers of women efficiently, and effectively.  With digital images, a study can be reviewedby unlimited numbers of experts anywhere in the World.   It has been completely standardized and automated where as tests like ultrasound have a high degree of operator dependence.     

Mammography has the highest spatial resolution of all of the technologies used for breast evaluation.  It is the best way to detect“microcalcifications” which are the main way that the earliest form of breast cancer, ductal carcinoma in situ (DCIS), is detected.  The importance of DCIS is still a major topic of discussion, but the incidence of invasive breast cancer and the death rate from breast cancer have declined in the U.S. since the institution of mammography screening.  It is likely that the rate of invasive cancers has declined, and the death rate has continued to fall, in part, because of the removal of DCIS due to mammography screening (5). Many invasive cancers have a component of DCIS.  Almost 20% of invasive breast cancers are detected because of the microcalcifications in their associated DCIS.  This is another reason why mammography is the preferred screening test.

Mammography has continued to improve as ascreening test with the development of Digital Breast Tomosynthesis (DBT) ([6],[7]).  DBT detects additional cancers that are hiddenby normal breast tissue on 2D mammography while at the same time reducing there call rate.  It has the unusual advantage of increasing both sensitivity and specificity - they usually go inopposite directions.

BOTTOM LINE

Mammography screening is the only test that has been proven to save lives.  It has been highly developed and standardized to facilitate the efficient and effective screening of large numbers of women. This is necessary because we are not yet able to accurately predict who will develop breast cancer and who will not develop breast cancer.  The vast majority of women who evelop breast cancer have none of the known factors that elevate risk.

If resources were unlimited and technology readily available (and if gadolinium accumulation is of no consequence) it is likely that Magnetic Resonance Imaging with intravenous contrast is the best way to detect most breast cancers at a time when cure is possible.  It remains to be seen if dual energy mammography with intravenous iodinated contrast (so called “spectral imaging”)can improve our ability to detect breast cancers.  Ultrasound screening could be greatly facilitated if it could be performed in the DBT unit so that the simultaneous collection of both studies could be facilitated, and, by having both image sperfectly registered, inter pretation of both studies could be greatly facilitated.

The best way to screen for breast cancer at this time is using Digital Breast Tomosynthesis.

REFERENCES

1. Kopans DB,Rafferty E, Georgian-Smith D, Yeh E, D’Alessandro, H  Moore R, Hughes K, Halpern E. A Simple ModelOf Breast Cancer Growth May Provide Explanations For Observations Of Apparently Complex Phenomena.  Cancer 2003;97:2951-2959.

2. Kuhl CK, StrobelK, Bieling H, Leutner C, Schild HH, Schrading S. Supplemental

3. Breast MR Imaging Screening of Women with Average Risk of Breast Cancer.Radiology. 2017 May;283(2):361-370.

4. Michaelson JS, Halpern E, Kopans DB. Breast cancer: computer simulation method for estimating optimal intervals for screening. Radiology. 1999 Aug;212(2):551-60.

5. Moskowitz M. Screening is Not Diagnosis. Radiology 1979;133:265-267.

6. Kopans DB. Arguments Against Mammography Screening Continue to be Basedon Faulty Science. The Oncologist 2014;19:107–112.

7. Wu T, StewartA, Stanton M, McCauley T, Phillips W, Kopans DB, Moore RH, Eberhard JW,Opsahl-Ong B, Niklason L, Williams MB. Tomographic mammography using a limited number of low-dose cone-beamprojection images.  Med Phys. 2003;30:365-80.

8. Kopans DB.Digital breast tomosynthesis from concept to clinical care. AJR Am J Roentgenol. 2014 Feb;202(2):299-308.