A common thread in my scribblings are the inconsistencies, bias and logical fallacies that fly in the face of so-called evidence-based medicine – in other words, those situations where the p-value is great, but the conclusion foolish.
And one of those inconsistencies is the attitude – no, fervor – about contralateral mastectomy (yes or no) vs. preventive mastectomy, the former patient having one cancer, the latter having no cancer. I’m speaking about the party-line directive that minimizes risk of contralateral disease in one patient while maximizing risk of the first primary cancer in another.
When the “epidemic” of contralateral preventive surgery began, everyone rushed to blame somebody or something – e.g., patient ignorance, poor counseling by physicians who overestimate risk, overly enthusiastic plastic surgeons, pre-op MRI, etc. It was inconceivable that women would opt to have the other side removed electively when there was no survival advantage. (Of course, if survival is the only endpoint, we’re all in trouble).
These stunned critics of contralateral prevention were likely too young to have lived through a true surgical epidemic, that of subcutaneous mastectomies in the 1970s. Shortly after implants were introduced, many thousands of women had both breasts removed for pain and cysts alone (the much-dreaded fibrocystic disease) with only secondary consideration for cancer prevention (which failed because too much breast tissue was left behind in order to achieve optimal cosmesis).
As an aside, more than other organ systems, it is my contention that physicians – both female and male – project their own reverence for breast tissue onto patients and their desires, thus clouding the objective assessment of why patients do what they do.
“If women only knew what their true contralateral risk was, they wouldn’t opt for contralateral mastectomy so often.” One still hears this today, as if this decision is based on a single issue (risk level), in spite of all evidence that it is a complex multi-factorial decision, based on different reasoning by different people.
In a 1990 academic Morbidity and Mortality conference at the University of Oklahoma, shortly after I joined the faculty, one of the general surgeons noted, “What’s going on around here? It seems like all our patients are having either lumpectomy or bilateral mastectomy? What happened to the standard unilateral mastectomy?”
What was happening was an exhaustive informed consent in very knowledgeable breast cancer patients, younger than average, coupled to a cutting-edge reconstruction team that introduced pedicled TRAMs early on, then free TRAMS later. The majority of our mastectomy patients were reconstructed with flaps rather than implants. And since the flap approach was a one-time operation, we skewed the experience toward bilateral where both sides could be reconstructed symmetrically, although some still opted for unilateral surgery. Many other reasons prompted the choice for contralateral prevention, including an accurate risk assessment, but all of them legitimate in my view, always patient driven, and none of them related to a survival advantage.
But I digress from my main point:
My point has to do with the incredible inconsistency in practice today whereby we try to minimize the contralateral risk in a newly diagnosed cancer patient, while we maximize the risk for women with no history of cancer. Why?
I don’t know “why,” but consider this inconsistency. Breast MRI is widely accepted for high-risk screening in women with no prior cancer, but not routinely accepted (condemned by some) for the newly diagnosed breast cancer patient to check the opposite side. Yet, the probability of finding an unsuspected contralateral cancer with MRI is greater in the cancer patient (3-4%) than it is in the high-risk population (2-3%) or in the modest risk population where a great deal of MRI screening takes place (1-2%). Dismissive comments about pre-op MRI work best on the ipsilateral findings, but when contralateral cancer is found, it makes “leaving an untreated cancer behind” a reality, magnified in the DCIS patient who has an occult invasive cancer on the opposite side.
There are no “calls to action” to stop high-risk MRI screening — quite the opposite. We use mathematical models that often inflate risk beyond reality in order to maximize the probability that MRI screening will be covered by insurers — “anything” to reach that magical 20% lifetime, even though these models are not very accurate at the individual level (good calibration for cohorts, but only 0.7 c-stat for discrimination). Yet, in the process of overestimating risk (and through multi-modality imaging and its callbacks), we drive some women to preventive mastectomy. Where is the same outrage that we see with cancer patients opting for contralateral prevention?
In sharp contrast to the high-risk patient, we don’t overinflate contralateral risk in the cancer patient. In fact, we use numbers for counseling that are sometimes lower than reality, not too much different than baseline risk in a non-cancer patient. For instance, tables and tools tell us to use 0.5%/year in the typical ER-positive patient, but this number assumes total compliance with endocrine therapy when, in reality, only about half of women actually complete 5 years of endocrine therapy. Without this protection, the contralateral risk is at least 0.75%/year, maybe 1%/year in the ER+ patient (that is, without other supporting family history or gene-positivity). For a 40 y/o cancer patient, non-compliant with her endocrine therapy, that’s 30-40% lifetime risk of contralateral cancer, without considering the mortality rate of her first cancer…
(I always found it hard to say, “Your contralateral risk would be 30%, but…since you have a 50% mortality rate from your current cancer over that same course of time, the true risk that you’ll ever live to see a contralateral cancer is 15%.” So, I ignore the models that include mortality, and simply say: “Assuming you beat this first cancer, the odds that you will face a second….”)
We can ratchet up the risk level in high-risk (gene-negative) patients to numbers comparable to BRCA-positivity when tissue risks are coupled with Level D density, which are then coupled further to SNP scores. Since no one has actually observed, let’s say, 80% of women getting cancer based on atypia, Level D and SNPs, it begs a cautionary note: “Are we counting 3 different manifestations of the same thing?” Our mathematical models for the first primary are suspect without prospective validation. In contrast, our risk models for contralateral cancer are drawn from direct observation. Regardless, the push is to get that number as high as possible in the no-cancer patient for MRI coverage, and ignore the backfire that might send the patient toward preventive mastectomies.
So, we end up with this:
To the 40 year-old cancer patient: “Don’t worry about your opposite breast. The risk of getting cancer on that side is very low, only 0.5% per year or 20% lifetime, and it won’t improve survival if you remove it now.”
And to the 40-year-old no-cancer patient: “You are at very high risk for the development of breast cancer – 20% lifetime – and you should undergo annual MRI screening in addition to mammography. (Then, 5 years later, after high anxiety due to frequent callbacks and benign biopsies, she opts for bilateral preventive mastectomies.)
After 40 years of practice, I’ve had plenty of patients return with contralateral cancer to say, “I told you I wanted that other side removed.” I’ve never had a single patient say, “I wish I’d never done the opposite side.”