Closing the Mammography gap
Author and Curator: Dror Nir, PhD
There are 40 million women seeking mammography breast-screening every year in the USA, out of which 15 million are women with heterogeneously dense or extremely dense breasts. USA epidemiology statistics show that 6 out of 7 missed cancers at mammography occur in women with dense breasts. It is also known that the majority of women presenting with mammography-dense breasts are below 45 years old.
The Oct. 22 issue of the American Journal of Roentgenology ( AJR) publishes results of a study showing that ultrasound is superior to mammography in evaluating symptomatic women 30-39 years of age [1].
This study was conducted by researchers at the Seattle Cancer Alliance and University of Washington. Patients were recruited between January 2002 and August 2006. 954 women ranging from 30 to 39 years old who presented for diagnostic breast imaging evaluation were examined, and it was found that sensitivity (probability for cancer detection) of ultrasound was 95.7 percent compared to 60.9 percent for mammography. A very important result of this study is the calculated Negative Predictive Value (the probability to have negative pathology if the imaging-test is negative) which was similar for both modalities: 99.9% for ultrasound and 99.2% for mammography.
Show case in images (All images courtesy of the American Roentgen Ray Society.):
In regards to which imaging modality should be used when screening such a population, the conclusion of the investigators is very clear: “Ultrasound has high sensitivity (95.7%) and high NPV (99.9%) in this setting and should be the primary imaging modality of choice. The added value of adjunct mammography is low.”
When reading this article I noted a gap to overcome if we want to successfully replace mammography with ultrasound. The Positive Predictive Value (the probability of detecting a cancer) calculated for ultrasound in these study settings was lower than that calculated for mammography: 13.2% for ultrasound and 18.4% for mammography. This is because mammography detected one additional malignancy in an asymptomatic area in a 32-year-old woman who was subsequently found to have a BRCA2 gene mutation. Mammography could do that because it scans the whole breast, whereas the investigators in this study used ultrasound just for scanning the suspicious lumps. A solution is offered in using the recently introduced ultrasound modalities, which are able to perform automatic full breast ultrasound scans [2], preferably enhanced by real-time tissue characterisation capability – a technology I’m working to develop.
References:
- Accuracy and Value of Breast Ultrasound for Primary Imaging Evaluation of Symptomatic Women 30-39 Years of Age,Constance D. Lehman1,2, Christoph I. Lee1,2, Vilert A. Loving1,2, Michael S. Portillo1,2, Sue Peacock1,2 and Wendy B. DeMartini1,2, Oct. 22 issue of the American Journal of Roentgenology
2. Using Automated Breast Sonography as Part of a Multimodality Approach to Dense Breast Screening, Vincenzo Giuliano, MD, RDMS, RVT1, Concetta Giuliano, DO1, Journal of Diagnostic Medical SonographyJuly/August 2012 28: 159-165,
1Novasoutheastern University, Winter Springs, FL, USA Written by: Dror Nir, PhD.
Thanks Dror. An insightful and informative piece. I appreciate your sharing this with me.
Kind Regards,
Gina
Dr. Nir,
a few questions: 1) Does the 65% prediction rate also include the newer digital forms of mammography? 2) Is there a stratification of prediction rates using ultrasound for the different types breast cancer i.e. in-situ, triple negative?. 3) Also does this benefit of ultrasound over mammography only relate to dense breast tissue and younger women possibly at high risk? I really like the post and would like to get more information. In addition this would be a great subchapter for a cancer book (New Trends in Cancer diagnosis). Thanks
Hi,
1) Recruitment to the study took place between 2002 to 2006. At that time, low-dose mammography systems were not yet available. In the paper’s material & methods section it is mentioned that “Diagnostic mammograms were acquired using a Senographe DMR+ (2002-2004) or a Senographe Essential unit (2004-2006) (both, GE Healthcare). All examinations were interpreted by one of seven radiologists subspecialized in breast imaging.”
2) The paper includes information on detection rates according to BI-RADS assessment groups only. The mentioned sensitivity and specificity in my post are related to the whole cohort.
3) This study was conducted in women 30-39 years of age. The issue of dense breasts is not specifically addressed. Nevertheless, women at this age range usually are with mammography dense breasts. I think that women with all kind of breast tissue can benefit from screening by ultrasound. In regards to applying ultrasound to screening of patients with elevated risk for having breast cancer: Indeed, as I mentioned in my post, in this study ultrasound was applied to the suspicious lump only. Therefore, the conclusion was that it has limited benefit for such group. As said, I think that if ABUS was used, the conclusion might have been different.
I hope I answered all your questions in a satisfactory manner. Otherwise, please do not hesitate comment again.
Best Regards,
Dror Nir, PhD
Managing partner
BE: +32 (0) 473 981896
UK: +44 (0) 2032392424
web: http://www.radbee.com/
blogs: http://radbee.wordpress.com/ ; http://www.MedDevOnIce.com
If the negative predictive value is the same, a negative result in either case excludes CA. The sensitivity of 95.7% is the occurrence of a positive with the disease. The positive predictive value (not stated) is the presence of disease with a positive result, which must be somewhat less. A ROC curve would be helpful. In any case, the benefit of no exposure to radiation is something. So the US would be a good first choice. But we don’t know how it would stack up against the 3-D imaging, which limits exposure. You want the fewest false positives, and maximum negative predictive value.
Dear Larry,
At the end of my post I did mention the PPV and explained something very important in that respect..:)
There was no ROC analysis done, therefore I could not mention that.
In regards to comparing 3D Ultrasound with 3D MRI and 3D mammography – Tomographic imaging – I might post on that sometime soon.
Dr. Nir,
Thank you fou this very insightful post.
Ultra sound focused oh the lump area, only. How was the lump detected, via self exam or by mammogram?
Focusing on the lump area, is in sufficient, there are cases of several lamps, one bigger other smaller. At present ultrasound is accessible to patients only if a suspicious mammogram results are obtained via routine screening.
In light of the fascinating results, you presented for the middle age cohort, how a 35 year old can arrive to the Women Imaging Cenert for routine screening and request ultrasound instead of mammogram?
Routine screening is covered by the Insurance companies,annually and every 6 month for high risk cases.
Penetration of the technology you are working on will require a change in the Guidelines for Routine Breast screening, a national procedure, the legislation of which is in the hands of congress in the US.
Medical equipment obsolescence is another issue, to replace the entire intalled base of mammography equipment with new ultrasound machines for full breat screening would require capital investment on the part of many independent Radiology offices in the US.
In Boston, the Partners Hospitals have very updated equipment, Radiology departments in community hospital in Fitchburg, MA was very old equipment,
Decentralization of screening centers is characterized by high equipment obsolescence.
If tomorrow an ultrasound full breat machine is approved by FDA, the access to the new technology will be immediate by screening centers currently serviced by GE Health and Phillips, HP – winning the market with the new technology requires, first, change in legislation og the screening protocol, second, having the leading companies mentined above licensing the new technology for mass production fo the new equipment, replacement of an existing installed base.
Diffusion of innovations embedded in hardware has a life of its own. In mt first career on 25 years, I was the designer of economic-mathematical models for penetration of new models and upgrades for Amdahl’s highend mainframe computers which powered AT&T in mid 80s. Later on in mid 90s, as Head of Research for EDAC at MITRE, I was involved with the largest pricing model for addressing equipment obsolescence in the US Air Force.
I would like to see ubiquitous penetration of the full breast screening ultrasound equipment and full accessibility to this service to follow universal healthcare insurance and a change in the National guideline on Breast screening.
A patient who is coming to be screened for breast cancer as part of the routine (reimbursed) procedure in the USA cannot elect a different pathway. I guess that private paying patients will get what they are willing to pay for.
As you will learn from my coming post, an ultrasound system that is capable of performing automated full breast scan was recently approved for screening women presenting dens breast for breast cancer. It so happens, that I have been in contact with that company several times in the past trying to convince them to embed tissue characterisation in their system. Unfortunately, they refused…
Dror Nir, PhD
Managing partner
BE: +32 (0) 473 981896
UK: +44 (0) 2032392424
web: http://www.radbee.com/
blogs: http://radbee.wordpress.com/ ; http://www.MedDevOnIce.com
what is your comment on the above?
It is not only a matter of accuracy,but yes ultrasound is better as a screening method under 40 years old.The is another Question:serial mammograms before 40 Raise the risk of Breast Cancer!But after no!
Well, it all depends how long you live..:) and for how many years you are screened. Normally, after 70 people do not screen any more. Anyway, the rule is: The more you are exposed to X-rays radiation the higher is the probability to get cancer. BTW; the elevated risk is for lungs cancer more than any other.
Dror Nir, PhD
Managing partner
BE: +32 (0) 473 981896
UK: +44 (0) 2032392424
web: http://www.radbee.com/
blogs: http://radbee.wordpress.com/ ; http://www.MedDevOnIce.com
[…] discussed the potential role of ABUS in future breast screening in my recent posts: Closing the Mammography gap; Introducing smart-imaging into radiologists’ daily practice. As noted, in recent years, […]
[…] discussed the potential role of ABUS in future breast screening in my recent posts: Closing the Mammography gap; Introducing smart-imaging into radiologists’ daily practice. As noted, in recent years, […]
PUT IT IN CONTEXT OF CANCER CELL MOVEMENT
The contraction of skeletal muscle is triggered by nerve impulses, which stimulate the release of Ca2+ from the sarcoplasmic reticuluma specialized network of internal membranes, similar to the endoplasmic reticulum, that stores high concentrations of Ca2+ ions. The release of Ca2+ from the sarcoplasmic reticulum increases the concentration of Ca2+ in the cytosol from approximately 10-7 to 10-5 M. The increased Ca2+ concentration signals muscle contraction via the action of two accessory proteins bound to the actin filaments: tropomyosin and troponin (Figure 11.25). Tropomyosin is a fibrous protein that binds lengthwise along the groove of actin filaments. In striated muscle, each tropomyosin molecule is bound to troponin, which is a complex of three polypeptides: troponin C (Ca2+-binding), troponin I (inhibitory), and troponin T (tropomyosin-binding). When the concentration of Ca2+ is low, the complex of the troponins with tropomyosin blocks the interaction of actin and myosin, so the muscle does not contract. At high concentrations, Ca2+ binding to troponin C shifts the position of the complex, relieving this inhibition and allowing contraction to proceed.
Figure 11.25
Association of tropomyosin and troponins with actin filaments. (A) Tropomyosin binds lengthwise along actin filaments and, in striated muscle, is associated with a complex of three troponins: troponin I (TnI), troponin C (TnC), and troponin T (TnT). In (more ) Contractile Assemblies of Actin and Myosin in Nonmuscle Cells
Contractile assemblies of actin and myosin, resembling small-scale versions of muscle fibers, are present also in nonmuscle cells. As in muscle, the actin filaments in these contractile assemblies are interdigitated with bipolar filaments of myosin II, consisting of 15 to 20 myosin II molecules, which produce contraction by sliding the actin filaments relative to one another (Figure 11.26). The actin filaments in contractile bundles in nonmuscle cells are also associated with tropomyosin, which facilitates their interaction with myosin II, probably by competing with filamin for binding sites on actin.
Figure 11.26
Contractile assemblies in nonmuscle cells. Bipolar filaments of myosin II produce contraction by sliding actin filaments in opposite directions. Two examples of contractile assemblies in nonmuscle cells, stress fibers and adhesion belts, were discussed earlier with respect to attachment of the actin cytoskeleton to regions of cell-substrate and cell-cell contacts (see Figures 11.13 and 11.14). The contraction of stress fibers produces tension across the cell, allowing the cell to pull on a substrate (e.g., the extracellular matrix) to which it is anchored. The contraction of adhesion belts alters the shape of epithelial cell sheets: a process that is particularly important during embryonic development, when sheets of epithelial cells fold into structures such as tubes.
The most dramatic example of actin-myosin contraction in nonmuscle cells, however, is provided by cytokinesisthe division of a cell into two following mitosis (Figure 11.27). Toward the end of mitosis in animal cells, a contractile ring consisting of actin filaments and myosin II assembles just underneath the plasma membrane. Its contraction pulls the plasma membrane progressively inward, constricting the center of the cell and pinching it in two. Interestingly, the thickness of the contractile ring remains constant as it contracts, implying that actin filaments disassemble as contraction proceeds. The ring then disperses completely following cell division.
Figure 11.27
Cytokinesis. Following completion of mitosis (nuclear division), a contractile ring consisting of actin filaments and myosin II divides the cell in two.
http://www.ncbi.nlm.nih.gov/books/NBK9961/
This is good. I don’t recall seeing it in the original comment. I am very aware of the actin myosin troponin connection in heart and in skeletal muscle, and I did know about the nonmuscle work. I won’t deal with it now, and I have been working with Aviral now online for 2 hours.
I have had a considerable background from way back in atomic orbital theory, physical chemistry, organic chemistry, and the equilibrium necessary for cations and anions. Despite the calcium role in contraction, I would not discount hypomagnesemia in having a disease role because of the intracellular-extracellular connection. The description you pasted reminds me also of a lecture given a few years ago by the Nobel Laureate that year on the mechanism of cell division.
I actually consider this amazing blog , âSAME SCIENTIFIC IMPACT: Scientific Publishing –
Open Journals vs. Subscription-based « Pharmaceutical Intelligenceâ, very compelling plus the blog post ended up being a good read.
Many thanks,Annette
I actually consider this amazing blog , âSAME SCIENTIFIC IMPACT: Scientific Publishing –
Open Journals vs. Subscription-based « Pharmaceutical Intelligenceâ, very compelling plus the blog post ended up being a good read.
Many thanks,Annette
I actually consider this amazing blog , âSAME SCIENTIFIC IMPACT: Scientific Publishing –
Open Journals vs. Subscription-based « Pharmaceutical Intelligenceâ, very compelling plus the blog post ended up being a good read.
Many thanks,Annette