A Tool to Determine Individual Prostate Cancer Overdiagnosis

A personalized tool to predict the chances of a prostate cancer overdiagnosis has been developed by Roman Gulati, a statistician at the Fred Hutchinson Cancer Research Center, and colleagues. The tool could enable better decision-making and tailored patient treatment. The results of the study describing the new tool are published in the Journal of the National Cancer Institute.

Whether prostate cancer detected through screening is overdiagnosed can vary depending on patient characteristics, such as age and on the status of the tumor itself. An overdiagnosed cancer is one that would not become symptomatic and would not endanger the life of the patient. Because of the prevalence of prostate-specific antigen (PSA) screening and the general slow growth of many prostate tumors, overdiagnosis and overtreatment are a concern. The US Preventive Services Task Force recommendations now highlight these potential harms and state that careful diagnosis and discussion with the patient are crucial to help prevent overdiagnosis.

The model predicts overdiagnosis among patients with PSA-detected prostate cancer. The chances of overdiagnosis in the model range from 2.9% to 88.1%.

“Men with screen-detected prostate cancer are making decisions about treatment based on limited information about the chances that their cancer has been overdiagnosed,” said Gulati in a statement. “We think this is a useful tool for patients and their providers because it helps to tailor knowledge of the risks and benefits of different treatment choices to their individual situations.”

One 2009 study published in the Journal of the National Cancer Institute estimated that in the United States, anywhere from 23% to 42% of men between the ages of 50 and 84 who are screened have been overdiagnosed. But, the authors noted, the chance that any patient is overdiagnosed can vary widely depending on the patient’s age and tumor characteristics.

The new publication describes the development of a nomogram-a graphical calculating device-for overdiagnosis of prostate cancer, which the study authors said is different from nomograms that predict the presence of indolent cancer. An indolent tumor is characterized by its biology, but whether a tumor is overdiagnosed also depends greatly on the age and life expectancy of the patient. For very elderly patients, even a relatively aggressive tumor may be overdiagnosed and overtreated, as death from other causes are more likely to occur.

The new nomogram integrates the patient’s age, PSA level, and Gleason score to determine the chance that prostate cancer detected through screening has been overdiagnosed.

The authors used a virtual patient population of US men between the ages of 50 and 84. Data on PSA levels, biopsy, and cancer diagnosis patterns from the Surveillance, Epidemiology, and End Results Program were used to understand the trends in cancer progression among patients who have undergone screening and those who chose not to undergo screening. Data on PSA dynamics in non-cancer patients from the Prostate Cancer Prevention Trial were used. The information was then combined with screening and biopsy patterns to determine the timing of diagnosis either with screening or without screening. Whether each patient would have likely died of other causes was also modeled.

The calibrated model has 33% of men being diagnosed with prostate cancer in their lifetimes, and 38% of these diagnoses would have been made regardless of screening.

The major factor in overdiagnosis using this model is age.

“The results of this study extend our understanding of the range of risks of overdiagnosis in US men detected by PSA screening, and how they depend on patient and tumor characteristics,” stated the authors in their discussion. The tool can serve clinicians and patients in helping to weigh both the potential harms and benefits of different care strategies and can help make important decisions about whether to undergo treatment or to delay treatment of early prostate cancer.

“This information, coupled with the relatively high frequencies of overdiagnosis projected by the model for these patients, should provide a compelling reason to carefully consider the appropriateness of active surveillance for low-risk disease, particularly in older men,” concluded the authors.