CHICAGO, III.
Model just as effective in detecting prostate cancer as current "gold standard" screening
Using a predictive computer model could reduce unnecessary prostate biopsies by almost 38 percent, according to a study conducted by Oregon Health & Science University researchers. The study was presented at the American Society of Clinical Oncology's annual meeting in Chicago, Ill.
"While current prostate cancer screening practices are good at helping us find patients with cancer, they unfortunately also identify many patients who don't have cancer. In fact, three out of four men who undergo a prostate biopsy do not have cancer at all," said Mark Garzotto, M.D., lead study investigator and member of the OHSU Cancer Institute. "Until now most patients with abnormal screening results were counseled to have prostate biopsies because physicians were unable to discriminate between those with cancer and those without cancer."
Current screening practice standards include a digital rectal exam (DRE) and a prostate specific antigen (PSA) blood test. Up to 25 percent of men who receive prostate screenings have an abnormal DRE, an elevated PSA or both. The majority of these men undergo prostate biopsies to determine if the abnormalities are cancerous.
"Seventy-five percent of men who have biopsies don't have prostate cancer," Garzotto said. "Oncologists currently don't have tools to limit the number of these procedures."
Prostate biopsy is an outpatient procedure that requires local anesthesia. It can cause patient discomfort, bleeding and infection, and can burden the health care system with extra costs.
In this study, researchers used Classification and Regression Tree (CART) analyses to predict whether a biopsy would find cancer. CART is a computer program that can build a decision tree or predictive model.
"While CART itself is not new, before this study it had not been examined as a tool for prostate cancer diagnosis," Garzotto said. "CART told us that there are certain patients who don't need biopsies."
CART analyses indicated that men with a PSA level below 1.55 ng/ml did not necessarily need to be biopsied. Using this cutoff, nearly 97 percent of tumors were detected, regardless of DRE results. CART analyses determined that patients with a PSA level of greater than 1.55 ng/ml comprised the at-risk group. Within this at-risk group, CART analysis further identified subgroups as high-risk for a positive biopsy. Overall, the model was able to reduce the number of unnecessary biopsies by 38 percent while still detecting 95.5 percent of all cancers in the study, the same as the standard screening.
"This model struck a nice balance between specificity and sensitivity," said Garzotto. "CART makes it possible to better ascertain which patients don't need biopsies while maintaining the same level of prostate cancer detection as current screening methods."
Researchers created the initial CART model with data prospectively collected from 1,436 men with a PSA level of 10 ng/ml or below who underwent a prostate biopsy. The data were put into the CART database to create a prostate cancer diagnosis decision tree. Demographic data such as age, race, reason for referral, history of vasectomy and family health history, DRE results, PSA level, PSA density (PSAD), and transverse rectal ultrasound (TRUS) volume and lesions were included in the data base. Prostate cancer was detected in 24 percent of the patients.
Twenty percent of the data were reserved randomly for study validation. CART analysis was carried out on the remaining 80 percent of the data. First PSA level and DRE results were examined independently. Then remaining variables were examined in men found to be at risk based on the initial CART model.
Prostate cancer is the most common malignancy among men and the second leading cause of cancer death in men in the United States. Overall, roughly one in six American men will develop prostate cancer during his lifetime.
PSA is a protein made only by prostate cells. Certain prostate conditions, including prostate cancer, can cause high levels of PSA in the blood. PSA blood levels are monitored to help predict the presence and progression of prostate cancer.
Mark Garzotto, M.D., is an assistant professor of surgery (urology) in the OHSU School of Medicine, director of urologic oncology at the Portland Veterans Affairs Medical Center, and a member of the OHSU Cancer Institute.
"While current prostate cancer screening practices are good at helping us find patients with cancer, they unfortunately also identify many patients who don't have cancer. In fact, three out of four men who undergo a prostate biopsy do not have cancer at all," said Mark Garzotto, M.D., lead study investigator and member of the OHSU Cancer Institute. "Until now most patients with abnormal screening results were counseled to have prostate biopsies because physicians were unable to discriminate between those with cancer and those without cancer."
Current screening practice standards include a digital rectal exam (DRE) and a prostate specific antigen (PSA) blood test. Up to 25 percent of men who receive prostate screenings have an abnormal DRE, an elevated PSA or both. The majority of these men undergo prostate biopsies to determine if the abnormalities are cancerous.
"Seventy-five percent of men who have biopsies don't have prostate cancer," Garzotto said. "Oncologists currently don't have tools to limit the number of these procedures."
Prostate biopsy is an outpatient procedure that requires local anesthesia. It can cause patient discomfort, bleeding and infection, and can burden the health care system with extra costs.
In this study, researchers used Classification and Regression Tree (CART) analyses to predict whether a biopsy would find cancer. CART is a computer program that can build a decision tree or predictive model.
"While CART itself is not new, before this study it had not been examined as a tool for prostate cancer diagnosis," Garzotto said. "CART told us that there are certain patients who don't need biopsies."
CART analyses indicated that men with a PSA level below 1.55 ng/ml did not necessarily need to be biopsied. Using this cutoff, nearly 97 percent of tumors were detected, regardless of DRE results. CART analyses determined that patients with a PSA level of greater than 1.55 ng/ml comprised the at-risk group. Within this at-risk group, CART analysis further identified subgroups as high-risk for a positive biopsy. Overall, the model was able to reduce the number of unnecessary biopsies by 38 percent while still detecting 95.5 percent of all cancers in the study, the same as the standard screening.
"This model struck a nice balance between specificity and sensitivity," said Garzotto. "CART makes it possible to better ascertain which patients don't need biopsies while maintaining the same level of prostate cancer detection as current screening methods."
Researchers created the initial CART model with data prospectively collected from 1,436 men with a PSA level of 10 ng/ml or below who underwent a prostate biopsy. The data were put into the CART database to create a prostate cancer diagnosis decision tree. Demographic data such as age, race, reason for referral, history of vasectomy and family health history, DRE results, PSA level, PSA density (PSAD), and transverse rectal ultrasound (TRUS) volume and lesions were included in the data base. Prostate cancer was detected in 24 percent of the patients.
Twenty percent of the data were reserved randomly for study validation. CART analysis was carried out on the remaining 80 percent of the data. First PSA level and DRE results were examined independently. Then remaining variables were examined in men found to be at risk based on the initial CART model.
Prostate cancer is the most common malignancy among men and the second leading cause of cancer death in men in the United States. Overall, roughly one in six American men will develop prostate cancer during his lifetime.
PSA is a protein made only by prostate cells. Certain prostate conditions, including prostate cancer, can cause high levels of PSA in the blood. PSA blood levels are monitored to help predict the presence and progression of prostate cancer.
Mark Garzotto, M.D., is an assistant professor of surgery (urology) in the OHSU School of Medicine, director of urologic oncology at the Portland Veterans Affairs Medical Center, and a member of the OHSU Cancer Institute.
###