Evidence-Based Oncology

Overdiagnosis and Overtreatment: The Cancer Paradigm

Published Online: August 21, 2014
Based on presentations by Barbara L. Smith, MD, PhD; Rina S. Punglia, MD, MPH; Laura Esserman, MD, MBA,
Some experts argue that overdiagnosis (OD) and overtreatment (OT) of cancer is common and increasingly costly. Others argue that current cure rates are high because of the screening processes that are in place. Both viewpoints were debated during the session “Overdiagnosis and Overtreatment in Cancer: Point/Counterpoint,” conducted on June 1, 2014, the third day of the annual meeting of the American Society of Clinical Oncology (ASCO) held in Chicago.

“Cancer overdiagnosis and overtreatment, or appropriate attention to early disease?” was the question posed by Barbara L. Smith, MD, PhD, director of the breast program at Massachusetts General Hospital, codirector of the Gillette Center for Women’s Cancers, and associate professor of surgery at Harvard Medical School. So far, the goals of cancer care have been minimizing mortality, prolonging life, reducing morbidity of treatment, preventing primary cancers, and providing access to quality care for all. Of late, the discussion has widened to include cost effectiveness and avoiding excess treatment.

“For several years now, death rates from cancer have seen a dip and cancer incidence rates have plateaued—an indicator that the strategy is working. The cornerstones of success of modern care include earlier detection, improved treatments (multimodality care op-tions), a combination of early detection and better treatment, and also

improvement in overall population health. Healthier patients can work through the toughest treatments,” said Smith.

Smith questioned whether diminishing returns could be resulting in overkill. “Are cancer patients being administered more treatment than needed?” More importantly,

can comparable success be achieved with lighter treatment?

She provided an example of the increasing use of bilateral mastectomy in early-stage breast cancer. The numbers were at 3% back in 1998 but are up to 18% in 2007. The question to pose: Is it necessary? Patients want the procedure for peace of mind—to reduce their chances of recurrence. But what about the oncologist?

This is a challenge being discussed across a spectrum of disease states such as hypertension and cholesterol. Some of the questions that oncologists need to ask include:

• What is the optimal way to screen (sensitivity/specificity goals, cost, and limitations of diagnostic technologies)?

• What should the threshold for treatment be?

• What are the morbidity and mortality goals?

An ideal example is that of the recent debate about using the Papanicolaou (Pap) test cytology versus human papillomavirus (HPV) DNA testing. Pap has proved successful in reducing cervical mortality, but the role of HPV in disease has been recognized, and recently the cobas test was approved as a primary HPV screen.

The sensitivity of the screening is ideal—the problem lies in the fact that 10% of patients who are Pap-positive have precancerous lesions, but some cancers are HPV-negative. Additionally, there may be cancer-independent transient HPV infections that could result in falsepositive results. Only 10% of women with high-risk HPV develop persistent infection. The result: unnecessary colposcopies and biopsies following HPV testing.

However, a recent study showed that patients are the ones who want screening and would even overrule their physician in wanting to be screened.

Can improved treatment reduce the need for screening? Not yet, said Smith. We still cannot reliably distinguish between patients with indolent and aggressive cancer. “Cost is the elephant in the room—an unsustainable and sensitive topic that comes up with death panels, rationing, etc,” she said.

The other elephant in the room is liability for malpractice. Delays in diagnosis are common claims in malpractice lawsuits, and variations in guidelines (eg, annual vs biennial screening for breast cancer) can increase a physician’s exposure to such litigations.

What will it take to change practice? A change in physician perspective and a change in patient expectations may be the answer. Additionally, cost constraints need to be modified, guidelines need to be standardized, and medicolegal issues need to be addressed.

Smith concluded by saying that implementing small changes can lead to the bigger goals of:

• stretching screening intervals

• tweaking thresholds for intervention

• reducing cost of treatment

Rinaa S. Punglia, MD, MPH, associate professor, Department of Radiation Oncology, Harvard Medical School, presented “The Case Against Overdiagnosis and Overtreatment of Cancer.”

OD/OT are inherent in breast cancer screening, and OD is central to the mammography debate in breast cancer. “OD usually leads to OT. A treatment does not necessarily alter outcomes,” said Punglia.

The solution to this problem, she believes, lies in informed decision making (IDM), which is made possible by opening a dialogue. Decision to screen can lead to false positives, but also to early diagnosis and improved outcomes. IDM can be used to reduce OT.

For example, ductal carcinoma in situ (DCIS) has historically been treated with mastectomy, but that approach may be too drastic. Today, patients undergo breast preservation surgery, which can leave patients at risk for a second breast event in the preserved breast, but can be followed with radiation therapy (RT). Meaningful outcomes for a patient after RT include recurrence, treatment of recurrence, and breast preservation. However, on the downside, RT can be expensive.

So how does one decide if a patient receives RT? DCIS overestimates a patient’s risk of recurdrence, and making a decision is difficult for patients. Therefore, individual decision making is key—lay the options down and help patients make the right choice. Punglia shared knowledge on a decision-making tool being actively used at Dana-

Farber (www.onlineDeCISion.org), which helps patients and providers alike.

The final viewpoint in this debate, “The Case for Overdiagnosis and Overtreatment of Cancer,” was presented by Laura Esserman, MD, MBA, director, Carol Franc Buck

Breast Care Center; and coleader, breast oncology program, UCSF Helen Diller Family Comprehensive Cancer Center.

In March 2012, the National Cancer Institute convened a workshop to generate ideas for a research agenda around OD. According to Esserman, a critical shift in philosophy is needed to explain that the previous approach of OD saves lives.

The old paradigm of cancer was of inexorable progression—or of every patient having the same rate of disease progression. That has changed, and we know today that it is far from true. Oncologists recognize that there is variable progression in patients and that there are IDLE (indolent lesions of epithelial origin) lesions that may not affect mortality. Therefore, understanding the biology of the cancer is extremely important.

Although the spectrum of disease has been realized and understood, the definition of cancer has not evolved. If a patient thinks “cancer,” they think they will die if they are not treated.

“A very important question is: what do you know, when do you know, and how do you act on what you know?” said Esserman.

The need of the time is a new approach to cancer treatment rather than arguments on 30-year-old guidelines, Esserman emphasized. However, it is extremely important to understand outcomes before changing guidelines. “Increased detection saves lives” is way too simple: the concept is much more complex, she said.

Recognizing and gaining an improved understanding of IDLE disease is important said Esserman—for example, does a tumor with little potential for metastasis and with little chance of progression or death require aggressive treatment? “IDLE tumor gene signatures could help define ultra-low–risk tumors. This could be validated in various cohorts and used to stratify patients with high- and lowrisk tumors,” according to Esserman.

Esserman closed her argument by recommending that a better-tailored definition replace the word “cancer,” companion diagnostics be coupled with treatment, and observational registries be created to identify IDLE disease. EBO