When is it enough Data ?

One of the many ways in which I am fortunate is that my wife is constantly on the prowl for articles that she knows will be of interest to me. Admittedly, because of her Barnard training, she often corrects the grammar in the articles she passes on. But overall, it’s a win for me.

The following article deals with a decades old issue and unfortunately demonstrates that there are still basic questions that we have failed to answer related to critical and life-threatening diseases. This article from the New York Times discusses the value of breast biopsies, when there is a concern for cancer.

The complete evaluation of a woman in order to rule out early breast cancer is very extensive. There is the manual exam, mammography, ultrasound, possibly MRI, breast biopsy and interpretation, occasionally further excision for more tissue, and then if there is evidence of disease, some combination of radiation, chemotherapy, hormonal therapy and other treatments. Women go through this process, after the age of 50, on a yearly basis. At a certain point, women can stop screening. But there is a solid 2 to 3 decades when women are legitimately very aware of the health of their breasts.

The following quote summarizes the lack of certainty in a significant number of biopsies.

“The finding is of concern, because D.C.I.S. sometimes becomes invasive cancer, and it is often treated like an early-stage cancer, with surgery and radiation. Missing the diagnosis can leave a woman at increased risk for cancer — but calling something D.C.I.S. when it is not can result in needless tests and treatments.”

This paragraph sums up the basic problem with most screening programs. Firstly, no screening system is perfect and as such, despite a woman’s best efforts, a cancer can still be missed. Contrarily, there are a significant number of people identified as “problematic” by a screening system, that in fact are perfectly healthy. But the determination of the person’s health is only possible via, sometimes, invasive, painful and expensive tests. There is of course the psychological trauma of thinking that you have cancer only to find out later that the doctors were overly cautious and unnecessarily did further tests and treatments.

I don’t want to elaborate at this point on the whole issue of whether breast screening is at all helpful. It is hard to believe but there are major studies that question the value of present day breast screening programs. All I will say is that such research further complicates the issue and makes it all the harder for a physician to recommend a clear path for a patient.

The obvious question is, how do we overcome these issues? What can medical research do to make the assessment for breast cancer far more accurate? Ideally, checking for breast cancer should involve a non-invasive test that outputs a definitive “yes” or “no” answer to the question of whether breast cancer exists.

Technologies that are already available and that are coming down the line, will unquestionably help. The first thing that we could use is data. The more data we have about each and every person, the more capable we will be of determining the state of health of that patient. Breast cancer researchers dream of having databases with millions and even hundreds of millions of patient records, that have every detail about each patient. When I say “every detail”, I include whether that person drinks caffeinated or decaffeinated coffee. A priori, we don’t know all of the factors that increase or decrease the likelihood of breast cancer, or at the very least the likelihood of identifying it on a test.

Imagine  that for 20 million cases of women with breast cancer, we had hundreds of thousands of data points that detailed everything in their personal lives, professional environment, interpersonal relationships, and then of course every imaginable detail about their actual health. This database was of course also have thousands of quantified parameters related to the tests that were done. At present, a human pathologist looks at a slide and decides whether the visualized cells are consistent with a particular disease. Imagine instead that an automatic measuring device recorded thousands of descriptive points of information about the person’s cells.

What do you do with all of this data? You hand it over to a huge scale computer system that is designed to look for patterns that increase the likelihood of the disease in question. For all we know, it could be the combination of type of birth control, work environment and three distinct features of the patient’s breast cells that accurately predict the future development of full-blown breast cancer. Without using a computer, the number of possible combinations of factors that may accurately predict cancer is beyond a human’s ability to track. But with sufficient computer memory, CPU power and appropriate analysis algorithms, it is possible to sift through the billions of billions of data points to find the ones that matter.

Of late, new technologies have made it possible to identify one of a relatively small number of “smoking guns” that are responsible for a given cell transforming into a cancerous cell. This transformation is not specific to a given organ or location in the body. Something in the basic DNA of a cell can malfunction and manifest as a metastatic disease. Identifying the malfunction and blocking it or deleting it can stop the cancer, without even knowing what the cancer’s specific origin was.

This concept in Oncology was unthinkable until relatively recently, and herein lies a critical element of hope for anyone with any disease. It might very well be that in five years from now, we will be treating a whole group of diseases in a way that no one has even yet thought about. The field of medical research, backed by more and more data and ever more powerful computers, is changing so rapidly as to easily cause a physician to become dizzy. It is most likely that we will soon need a computer just to tell us what research is being done, how effective the research is [even at its earliest stages of testing] and whether a particular patient is suitable as a subject for the research. Human beings cannot manage all of this information and effectively hook up all of the institutions and all of the people necessary to help the patient. Computers will soon manage almost every step in the evaluation, screening, further testing, specific follow-up and treatment for a given patient. And yes, we will all be grateful for this fact.

Physicians of the world must reset their brains and recognize that their clinical training prepared them for a world that will soon exist only in a museum, next to the woolly mammoths. I imagine that there will be retraining courses offered to physicians to get them up to speed on all of the elements that will be fundamental to the care of all patients for all diseases. Some physicians will walk away and find something else to occupy their time. But most physicians will embrace this new world, and their patients will unquestionably benefit.

In the world of technology, the answer to any question is always “yes”, with the critical qualification that the answer will be dependent on “time and money”. We will cure cancer. We will cure heart disease. We will overcome aging. These things I know. I don’t know when these things will happen and how much it will cost. Maybe what medicine really needs is super smart accountants.

Thanks for listening

My website is at http://mtc.expert