Seeking Solutions to the Dilemmas of Overdiagnosis and Overtreatment

Get Permission

As clinicians, we need to recognize that overdiagnosis occurs, which can lead to overtreatment if unrecognized…. [W]e should try doing less and learn from that restrained clinical behavior.

— Laura Esserman, MD, MBA

Widespread use of screening technologies has markedly increased early detection rates of cancer, saving countless lives. However, while screening technologies have remarkable sensitivity, their  inability to identify which tumors will progress and which will not has created the phenomenon of overdiagnosing cancers that might otherwise not go on to cause symptoms or deaths, resulting in potentially unnecessary treatments and burgeoning costs.

Over the past several years, due in part to increased scrutiny over rising health-care expenditures; the value of certain cancer screening methods is increasingly being questioned. Laura Esserman, MD, MBA, examined the overdiagnosis dilemma at the ASCO Annual Meeting and offered a prescription for change.1 “Overdiagnosis by itself isn’t the problem; the problem is that we often overtreat cancer, and by doing so, we cause harm and psychological distress over risk of recurrence that is extremely unlikely to our patients,” said Dr. Esserman.

Dr. Esserman noted that part of the problem is rooted in the old paradigm. “Patients assume that cancer left untreated will kill them, and so do most doctors who are still operating under the premise that stage determines outcome in a disease that inexorably progresses in an orderly manner through a discrete set of stages,” said Dr. Esserman. This system dictates that early detection should result in reducing mortality, but this is not always going to be true, she added. “Some of the lesions we detect would not necessarily progress, so early detection would lead to overtreatment, and some would spread early despite being small. So early detection is not necessarily the road to cure.”

Dr. Esserman explained that cancer is heterogeneous and varies by subtype, which in turn determines fate, rate of growth, and response to therapy. “Biology, as it turns out, trumps the size of the tumor. Clearly there are very small cancers that are lethal because of their capacity to metastasize. But we now have longitudinal studies that challenge the notion that all cancers will grow and progress. Moreover, screening can reveal indolent disease, in which intervention is a net negative. We know that the more we look, the more we find. The question becomes: What is the value of the finding?” said Dr. Esserman.

Is New Terminology Needed?

Dr. Esserman pointed out that some tumors do well in spite of multiple positive nodes, using thyroid cancer as one example. “This is an opportunity to use some of our evolving molecular markers to put tumor biology into a better clinical context, and in the process adjust cancer taxonomy and change the terminology. For example, precancerous, indolent lesions need a renaming convention without the word “cancer” in the label. Maybe we need to look for a critical threshold before a disease is labeled cancer, ” said Dr. Esserman.

She continued, “I don’t think that a disease that if excised is associated with a 98% chance of survival should be labeled cancer. Lesions with a low risk of progression, that are suitable for active surveillance, and have a 90%-plus chance of never progressing should perhaps be relabeled indolent lesions. Again, the impact of early intervention will be minimal if the disease that develops is indolent.”

Risk vs Reward in Screening

Dr. Esserman stressed that it was important to fully understand the clinical benefits of screening and treatments. “Early detection from screening is likely to have the most benefit in slower-growing disease that is potentially lethal. However, for disease that is lethal and rapidly growing, the chance of hitting it at the right time is low. In those cases, early detection with screening is not likely to have much value. It is equally important for women not to ignore a new mass if they had a normal mammogram 6 months ago; the more rapidly growing cancers often show up between normal scans, as they can grow quickly,” said Dr. Esserman, commenting that in indolent disease, especially when not clinically apparent, screening is of no value and often leads to unnecessary treatment.

Using breast cancer screening—a $10 billion per year expenditure—as a model, Dr. Esserman focused on ductal carcinoma in situ (DCIS), a disease state she feels offers the best example for change. “Everyone understands the prostate cancer screening story, but breast cancer screening is part of a similar narrative. In in situ disease, the detection rate has increased about 500-fold since the advent of mammographic screening. So, we have to ask, have we identified the right precursor lesions? Because it is unlikely that the majority of DCIS detected on screening will develop into significant cancers,” she said. “The detection rate has increased 500-fold since the advent of mammography screening. And despite taking out more than 50,000 in situ cases per year for well over 10 years, we have not seen a concomitant drop in the incidence of invasive cancers. In fact, the rate of new cancers has risen over time.”

Dr. Esserman posed the question: If the goal of breast cancer screening is to reduce morbidity and mortality, how did DCIS become the focus of that effort? “I submit that we need to back off the model of assuming that all cases of DCIS are cancers and should be treated as we do stage I cancer. First, low- and intermediate-grade DCIS should not be a target of screening. And lesions with a 20% to 35% risk of progression to invasive cancer should be considered high-risk lesions, not cancer,” said Dr. Esserman.

Prescription for Change

“The data indicate that all cases of DCIS are not the same, so I am suggesting that we rename low- to intermediate-grade DCIS to better reflect their clinical path. We should also participate in trials that change our treatment approaches in low-grade lesions. Time can be used as a discriminator to determine which women do indeed need surgical intervention. And in high-grade lesions, there’s no reason to rush into surgery,” commented Dr. Esserman (see sidebar on page 13). “We can use a period of time (1 to 2 months) prior to surgical excision as an opportunity to test new approaches to prevention for high-grade cancers,” she said.

The central message moving forward is that DCIS is neither an emergency, nor a life-threatening cancer, and we must enable an approach that allows us to learn and better inform our patients. “As clinicians, we need to recognize that overdiagnosis occurs, which can lead to overtreatment if unrecognized. We shouldn’t look so hard for DCIS, especially in older women. In short, we should try doing less and learn from that restrained clinical behavior,” concluded Dr. Esserman.■

Disclosure: Dr. Esserman reported no potential conflicts of interest.


1. Esserman L: Avoiding overdiagnosis and overtreatment of common cancers: Ductal carcinoma in situ. 2012 ASCO Annual Meeting. Education Session. Presented June 2, 2012.

Related Articles

SIDEBAR: Mammography Study Stokes Overdiagnosis Debate

Overdiagnosis and the harms associated with unnecessary procedures is becoming a vibrant subject in today’s health-care dialogue, with serious implications for providers and patients alike. A new study from the Norwegian Screening Program concluded that 15% to 25% of breast cancers identified on...