NYT: In Medicine, It's Man vs. Machine

Last week, a new study showed that radiologists don’t need no stinking machines. Radiologists read mammograms, at $120 a pop, to look for breast cancer. One out of every 200 films shows a suspicious pattern the radiologist recognizes from years of experience. It’s tedious work in a dark room with a magnifying glass. And mammograms are often read twice — another $120. A new system called computer-aided detection, or CAD, can identify the same patterns a doctor can, by referring to a database of known cancer films. And now a third of all second mammogram readings are done by CAD, for $20 each. But the study from the University of California, Davis, showed that a radiologist with CAD was statistically no better than a radiologist alone in finding cancer. Round 1: Doctors.

Yet what’s amazing about this is not that the dog can’t speak English so well, but that the dog can speak at all. A machine can read mammograms! The value of CAD is not in assisting radiologists but in getting rid of them. A better study would use the CAD system to read all mammograms and, in the few cases with suspicious patterns, call in the radiologist for a second read. But what radiologist is going to volunteer for that study?

Ever cheaper technology is making its way into medicine in fits and starts. In the many industries it’s transformed, technology works by getting rid of the people in its way — operators, tellers, travel agents, librarians, stock traders and on and on.

In medicine, doctors have always had all the expertise. The knowledge gathered during seven years of medical school is embedded in their brains. They stick a tongue depressor down your throat, take your blood pressure and still rip out the rubber hammer to bang your knee. Physicals haven’t changed much since “Dr. Howard, Dr. Fine, Dr. Howard.” 

But today we have the technology to look inside our bodies more closely than any doctor can. This is still expensive, but as cheaper chips and better software are developed, more screening will be done by machines than by doctors. And that could keep us all healthy longer.

What would happen to healthcare-related stocks if we all stayed healthy? Drug stocks alone are worth $1.1 trillion, biotech another $350 billion, and medical devices and equipment some $400 billion. Who would need that stuff if we had fewer heart attacks, strokes and diagnoses of cancer? Good question.

Today, doctors check your risk of heart attack by measuring your blood pressure and cholesterol. If the numbers are too high, you’re put on an expensive regime of statins like Lipitor (itself a $25-billion business). But CT scanners that can look for blockages in your arteries directly are becoming more and more sophisticated. Before 1998, they could image only one “slice” of the body per rotation. The technology then advanced so that it was possible to scan four slices per rotation, and then (in 2002) 16. For a few years now the state of the art has been the 64-slice scanner.

CT scanners, in other words, are keeping up the same pace of innovation as silicon chips have. A 256-slice CT scanner, sheduled to be out later this year, will scan a heart in 4/10ths of second and create a high-definition, three-dimensional image that a cardiologist can visually fly through.

If such a diagnostic tool were used on, say, everyone over the age of 40, it would cut the market for cholesterol-lowering drugs to only the population that truly needs them. (My guess is that the market would shrink by 75 percent.) What do you think that would do to Pfizer and Merck, pharma companies that are already getting slapped around by the stock market?

The scans now cost about $1000 each, so you can see why insurance companies don’t pay for them. It’s cheaper to let nature do the screening. But what if the price of scans were to drop to $300, or even $100. Considering that it costs about $30,000 to patch up a patient who suffers a heart attack, insurance companies may decide it’s a good idea to make scans part of every physical. Round 2: Machines.

For cancer, the technology that may change the game is likely to be biomarker chips. Cancer drugs are expensive. Herceptin, Genentech’s breast cancer drug, costs $48,000 per year. So it is important to find new ways to detect cancer before it spreads. Mammograms aren’t enough; even 256-slice CT scans can’t find the tiniest tumors. And the expensive biomarker tests we use today — the P.S.A. test for prostate cancer, for example, and the CA-125 for ovarian cancer — are at best 80-percent accurate.

But on the horizon are silicon chips with arrays of antibodies that can scan a patient’s blood for cancer cells. The antibodies lock onto the proteins the cancer cells create, and when a protein and an antibody connect, it produces a tiny electrical charge that the chip easily detects. The first of these $10 chips will have only one antibody, but the number will multiply to 4, 16, 64 — you know the pattern. And the cost per cancer test is expected to drop 30 percent annually until we can do thousands of antibodies for the same $10. Round 3: Silicon.

When cancer cells are found in a patient’s blood, techniques still in the works like molecular imaging will be able to pinpoint the microscopic tumors five years earlier than doctors once could — so they can be zapped, frozen or cut out before growing to a size where they can kill the patient or need to be fought with $48,000 worth of drugs. Think of Genentech’s stock in this brave new world of healthcare.

New technologies are always expensive at first, but engineers figure out how to bring the price down. Once medical imaging technology is cheap enough — I’m not talking overnight, this will be a 20-year transition – spending on early detection for heart disease, stroke and cancer will overtake spending on chronic care. Money is often made in the first five years of these transitions. It’s early, but keep your eyes open.

http://kessler.blogs.nytimes.com/2007/04/11/in-medicine-its-man-vs-machine/

Comments

[Genomik]

I agree that we are on the cusp on transformative medical technologies. But it is not necessarily game over when these technologies come around. Other battles will need to be fought. For example, antibiotic resistant bacteria are making a comeback and will demand new antibiotics, which in turn the bacteria will then become resistant. Perhaps by then doctors will need to engineer new therapies for multi drug resistant bacteria such as RNAinterference or even more advanced technologies such as bacteria eating bacteria. Round 4 Doctors, Machines, Silicon, and bacteria = God

[prozac zoloft uk]

Prozac (FLUOXETINE) is a Serotonin Reuptake Inhibitor (SSRI) that helps patients with depression by increasing the availability of serotonin in the brain. Scientists believe serotonin affects many types of activity in the brain, including the regulation of mood.
http://www.nordmed.com/generic-prozac-medication.php

Zoloft (SERTRALINE) is a selective Serotonin Reuptake Inhibitor (SSRI) used to treat depression, panic disorder, obsessive - compulsive disorders (OCD), post-traumatic stress disorder (PTSD, and a severe form of premenstrual syndrome (Premenstrual Dysphoric Disorder - PMDD).
http://www.nordmed.com/generic-zoloft-medication.php