Decoding Health Care

Part 18: Cancer screening (part one)

When I was growing up, I never used a seat belt. My father thought they were a nuisance. He wasn’t a bad father. Nobody used seatbelts. In fact seatbelts weren’t standard equipment on cars until 1958, they weren’t required equipment until 1966, and the first seat belt mandates didn’t come until the 1980s. Given the ubiquitous use of seatbelts today it is impossible to imagine a world without them. And they reduce serious injury and save lives. Seatbelts reduce the likelihood of dying in a car crash by about 50% and the likelihood of severe injury also by about 50%. That equates to about 15,000 lives every year. That is nothing to sneeze at. More than 90% of Americans wear their seatbelts. But the truth is that the likelihood that you as an individual will benefit is really small since fatal car crashes are very rare. Don’t get me wrong, I am NOT arguing against seatbelts and I assure you I have no intention of stopping my own use of seatbelts. We have succeeded tremendously in getting people to use them but the absolute benefit to the individual is probably slightly exaggerated. Cancer screening has a lot in common with seat belts.

Like seatbelts, it feels as if cancer screening has been around forever, but it definitely has not. Mammography, the poster child of a successful screening program, became an accepted screening tool after the publication of the HIP study (Health Insurance Plan of New York). In this trial (launched in 1963 with final results published in 1971), 60,000 women between the ages of 40-64 were randomized to mammography or standard of care. As compared with women in the control group, women in the screening group more commonly had localized cancers and had better breast cancer specific survival. These results were greeted with acclaim by the American Cancer Society, and they made advancing breast cancer screening a top priority.

This was excellent timing from a political perspective. President Nixon had just launched the War on Cancer and there was a lot of support for projects that could make a big difference. With this in mind, the National Cancer Institute in collaboration with the American Cancer Society launched the Breast Cancer Detection Demonstration Project (BCDDP). This project enrolled 280,000 women across the US to a screening strategy of annual mammography for five years. It was NOT a research project. In truth, it was generally accepted that mammography’s benefit had been established. Despite the clear bias, results from the BCDDP were used to bolster support for mammography.

But there was not unanimity in the support of breast cancer screening (Controversies Surrounding Screening Mammography - PMC (nih.gov)). There was a lot of criticism of the HIP study for methodological flaws, including from within the NCI itself. Ultimately, seven randomized studies were completed in several countries over about 20 years. These studies have been exhaustively critiqued. The consensus is that mammography reduces breast cancer specific mortality in women between the ages of 50 and 70 by about 20%. The studies do not show a clear benefit in women younger than 50 or older than 70. But these studies are noteworthy in part because they point out the importance of false positives and overdiagnosis.

Much of the debate around mammography occurred in academic circles, and it really had no practical significance. Rates of mammography steadily increased over time. From 1987 to 2000, rates of mammography in eligible women increased from 29% to 70%; today it stands at about 75%. The Department of Health and Human Services established the United States Preventative Services Task Force (USPSTF) in 1984 in an effort to help primary care physicians focus on prevention strategies. If you look at the USPSTF web site (https://www.uspreventiveservicestaskforce.org/uspstf/), you will notice cancer screening is only one of the areas they address. They first tackled breast cancer screening in January, 1996 and they regularly update their recommendations (most recently within the last year). These recommendations frequently change, most commonly related to the frequency of testing and the age group appropriate for testing. But from the very beginning they have enthusiastically supported mammography. And the USPSTF has, to a considerable degree, quieted the mammography critics (at least until recently, as discussed below).

The USPSTF is composed of a group of primary care physicians, nurses, preventive medicine experts, epidemiologists and public health experts. The panel consists of 16 members who are volunteers and serve for a term of four years. They are appointed by the secretary of the Department of Health and Human Services (HHS). They operate completely independently of HHS. With the passage of the  Affordable Care Act (ACA), they gained significant influence due to a provision that mandated coverage of USPSTF endorsed measures by all insurance companies (including Medicare) without any cost to the insured. Prior to the ACA, unless Congress specifically passed a law mandating coverage of preventive services, Medicare didn’t cover it (a law mandating mammography coverage was passed in 1990).

Recently, the USPSTF has come under fire (US Preventive Services Task Force Challenged in Federal Court | Guidelines | JAMA | JAMA Network). The legitimacy of the task force has been questioned, specifically the use of USPSTF endorsement to mandate insurance coverage. The argument revolves around whether HHS has the authority to empower USPSTF without congressional approval. This debate is NOT about cancer screening, but as you can see the current favorable coverage of cancer screening is jeopardized as a consequence. Let’s hope that this challenge just goes away since it could do a lot of damage.

Just as the USPSTF has weighed in on mammography, so has it weighed in on other areas of cancer screening. For the USPSTF, a grade of A or B means they endorse a preventive strategy (and as of now that means health plans must cover it without copay to its members). I refer you to the USPSTF web site for the most current recommendations; at present, the USPSTF recommends screening for breast, cervix, colorectal, and lung cancers. (https://www.uspreventiveservicestaskforce.org/uspstf/recommendation-topics/uspstf-a-and-b-recommendations). Of note, the USPSTF does NOT recommend prostate cancer screening.

From a very practical perspective, in order to make the USPSTF list two criteria need to be fulfilled: there needs to be a good test and early diagnosis must make a difference. These two criteria might seem to be self-evident, but I assure you that either one or the other has caused controversy when it comes to cancer screening.

Let’s take prostate cancer screening as an example. The PSA test was introduced in 1986 and approved by the FDA (in association with digital rectal exam) for the diagnosis of prostate cancer in 1994. We have been arguing about it ever since. Why? The PSA blood test just isn’t that great. You can certainly have prostate cancer with a normal PSA, and there are lots of causes of an elevated PSA that aren’t cancer. When you have an elevated PSA, presuming you are reasonably healthy, you head down a diagnostic odyssey that can save your life or make you wish you hadn’t bothered.

Let me explain. Prostate cancer is really common in American men. Somewhere between one third and one half have a cancer lurking in their prostate if they live to age 90. Most of these cancers are meaningless. But if you are <70 and have an elevated PSA, you will certainly be a candidate for a medical intervention, perhaps a course of antibiotics for prostatitis or more commonly a bunch of tests including some form of prostate imaging (like an MRI) or biopsy (which I am told by my patients really hurts). If either of those tests suggest a cancer, depending on how worrisome the findings are, you will receive a recommendation for “active observation” or surgery or radiation to eliminate the cancer.

Theoretically, the decision on how to proceed is made after an informed dialogue between the doctor and the patient. But this is a tough conversation. People are scared of cancer. And active observation doesn’t mean you just ignore the result; it means at the minimum serial blood tests and exams, or serial imaging or depending on the urologist, repeat biopsies. Most people don’t choose active observation. And most patients underestimate the risks of overtreatment. The reason this is a big deal is that a lot of those overtreated cancers really don’t put the man’s life at risk. Overtreatment means the patient runs the risk of impotence or urinary incontinence.

So PSA testing isn’t that great, and early diagnosis can but doesn’t always matter. What I just said is extremely controversial. Proponents of screening point out that prostate cancer specific mortality has often been shown to be lower in screened populations but these studies rarely consider the negative health effects of overtreatment. And there are a ton of studies. As an informed consumer, I have chosen not to personally undergo prostate cancer screening. The USPSTF grades prostate cancer screening as a D.

As we think about breast, colon, cervical and lung cancer screening, we ought to consider those criteria that disqualify (in my view) prostate cancer screening. How good is the test? Does early diagnosis mean more cures/longer life?

For breast cancer, it’s complicated. Mammography is an okay test. In some populations, like women between 50 and 70 it’s pretty good. In young women (age 40-50) it isn’t nearly as good, mostly because the breast density in this younger population makes interpretation tough. There is little doubt that, as with prostate cancer, overdiagnosis is a problem.

If a woman has an abnormal mammogram, a diagnostic odyssey is launched. Sometimes that means short interval repeat imaging. Sometimes it means an ultrasound or an MRI. Sometimes it means a biopsy. Most of these turn out to be negative. Even if follow-up testing is positive, there is good reason to believe that some of these cancers are indolent and will never pose a threat to the patient. It has been estimated that for every woman whose life is saved by screening, as many as ten have some sort of negative outcome (unnecessary biopsy, treatment of indolent tumor, etc.). This is an extremely sensitive topic to discuss; perhaps the most vocal voice in this overdiagnosis debate is Dr. Gilbert Welch (Breast-Cancer Tumor Size, Overdiagnosis, and Mammography Screening Effectiveness | New England Journal of Medicine (nejm.org)).

The most recent USPSTF breast cancer screening guidelines, published in 2024, recommend that women start getting mammograms at age 40. This change was not at all based on any new evidence. Rather, it was based on statistical modeling, as well as an attempt to improve screening in young women of color who have both lower screening rates as well as poorer breast cancer outcomes. This has created a lot of controversy (https://www.statnews.com/2024/05/02/breast-cancer-screening-mammogram-women-40s/) given the multiple negative screening studies in younger women. We will see what happens. There is no way this fixes the diagnostic challenges of using mammography in young women, and we have no idea whether it will impact racial disparities in the diagnosis and treatment of breast cancer.

At the end of the day, proponents of screening argue that screening is doing its job if it reduces death due to the cancer you are screening for. So does it? As I mentioned earlier, the party line is that mammography reduces breast cancer mortality by 15-20%.

The impact of screening for colon cancer is a little bit confusing because the screening test has changed over time. Commonly used tests include fecal testing (like fecal occult blood testing or FOBT), colonoscopy and more recently stool based tests for DNA (Cologuard). These are summarized in the USPSTF report published in 2021 (https://www.ncbi.nlm.nih.gov/books/NBK570913/). These studies show about a one third reduction in the risk of colon cancer death (especially for colonoscopy and Cologuard, a little less for FOBT). There is little question that early diagnosis of colon cancer yields a better outcome, and concerns of overdiagnosis are much less. This is largely because colonoscopy eliminates precancerous polyps and does so at the time of the diagnostic procedure.

Cervical cancer screening, using PAP smears and HPV testing, is amazingly effective. It is estimated that this approach might reduce cervical cancer mortality by up to 70%. This is very important from a global health perspective. Cervical cancer is becoming less and less of an issue in the US due to widely accepted screening plus HPV vaccination, but this is not the case globally. Cervical cancer screening recommendations are currently being updated by the USPSTF. Again, overdiagnosis is not much of an issue, and treatment of early stage cervical cancer (or pre-cancerous cervical lesions) is dramatically effective.

Lung cancer screening is the most interesting of all. Lung cancer is the most common cause of cancer death in the US. For years, this was largely the result of cigarette smoking. And although the number of smokers in the US has dropped dramatically, we still are seeing lung cancer cases due to smoking and there is a lot of lung cancer in non-smokers for a bunch of reasons. It is pretty clear that diagnosing lung cancer early makes a huge difference in outcomes. Half of all lung cancers are advanced (and largely incurable) at the time of diagnosis and although systemic treatments have resulted in these patients living longer and better lives, metastatic lung cancer is still ultimately fatal. And there is little or no overdiagnosis in lung cancer as far as we know.

Over the years there were many attempts to develop a lung cancer screening strategy. These focused on smokers. Chest X rays and sputum cytologies proved inadequate. Only when chest CT scans were used were we able to show benefit of lung cancer screening. For the first time in 2013, the USPSTF recommended annual low dose chest CT (LDCT, which refers to the radiation dose to the patient from the scan) for people age 50-80 with a 20 pack year smoking history who had smoked within the last 15 years.

CT screening of lung cancer reduces lung cancer mortality by 20%. What is more interesting is that in the largest lung cancer screening trial, CT screening reduced ALL CAUSE mortality by 7%. There are no other cancer screening tests that reduce all cause mortality. Why this is the case is not clear. Obviously you get a look at the coronary arteries, lung parenchyma, aorta and other vascular structures but the etiology of this mortality reduction is quite impressive. There is a significant downside to CT screening for lung cancer, and that is the finding of pulmonary nodules of uncertain significance. Sophisticated algorithms and follow-up strategies have been developed to manage these. Given the influence of AI on imaging I wouldn’t be surprised if AI increases the ability to discriminate worrisome from incidental.

This is where cancer screening stands today. We have done a lot better with seat belts. Next week, we will discuss how we can move cancer screening forward.