Answers are grounded in this paper and cite their sources.
The National Cancer Institute's Surveillance Epidemiology, and End Results (SEER) program has presented statistics on cancer since it was first established in 1973, and program leaders are now working on multiple changes to better address the evolving complexity of cancer science. Researchers across the country and around the world rely on SEER for the most accurate cancer statistics. Over the past 40 years, SEER's population-based registries have expanded to cover 28% of the US population. These data have enabled the program to provide comprehensive information concerning cancer survival as well as key demographics, stage at diagnosis, and tumor characteristics. However, SEER leaders now are working to broaden the breadth and depth of the information they gather. In the early days of SEER, patients with cancer were seen at a hospital, but as more patients have started to be treated in free-standing radiation facilities or medical oncology offices, some never see a hospital, says Lynne Penberthy, MD, MPH, associate director of SEER in the National Cancer Institute's division of cancer control and population sciences. “We're working to capture information from a broader diversity of providers,” she says. One major method SEER researchers have developed over the past decade is real-time electronic capture of approximately 80% of pathology reports from 350 pathology laboratories nationwide for approximately 450,000 new cancer cases annually. “Many of our registries, such as Kentucky and Louisiana, have almost 100% of their pathology reports received electronically, but our goal is to get close to 100% nationwide,” Dr. Penberthy says. These electronic pathology (or “E-path”) reports have the potential to provide detailed information regarding biomarkers and specimens from patients, which enables SEER registries to work with researchers through real-time reporting. The data can help researchers quickly identify patients for clinical trials and other types of studies. “There is a tremendous amount of very rich information in those pathology reports, and we're working on using natural language processing [NLP] to reduce the manual burden on hospital registrars and others to collect the information,” Dr. Penberthy says. NLP enables computers to capture relevant data from unstructured medical text, such as a physician's dictated notes. Essentially, it places the information from the reports into context. SEER leaders are working with both academic and commercial partners to determine how well software can extract the data elements they would like to capture. “We're very early in the process, but we would like to develop a more automated, efficient, and agile system,” Dr. Penberthy says. Ahmedin Jemal, DVM, PhD, vice president of surveillance research for the American Cancer Society, is enthusiastic about these new goals for SEER. “It will be an amazing resource for cancer researchers,” he says. He and his colleagues not only rely on the data for research but are key partners, helping to decide on priorities for the program. “Clinical trials, as good as they are, have their limitations,” Dr. Jemal says. “Only 3% of the population participates in clinical trials, and these patients are more likely to be younger, healthier, and wealthier. We don't know if the results from trials will apply to the general population.” New data from the SEER database will enable researchers to generate similar data in the general population to support the findings from clinical trials. Researchers also will have access to more detailed treatment data and costs of treatment. “This will not only generate new knowledge on patterns of care and outcomes, but also [add] knowledge that will translate into policy changes in areas such as quality of care and cancer control,” says Dr. Jemal. In SEER's first strategic priority area, treatment, the program already has good data regarding surgery and radiation, but not on chemotherapy and hormonal therapy. Dr. Penberthy and her colleagues aim to gather detailed information concerning infusion and targeted therapy as well as orally administered chemotherapy. The former is easier to obtain through insurance claims information. SEER already acquires that information for patients aged 65 years and older through Medicare, and is working with community oncology providers to collect the same data for younger patients. Capturing information regarding oral chemotherapy is more of a challenge because patients are given prescriptions and take the drugs at home. Furthermore, because many of these treatments are so expensive, physicians have reported anecdotally that some patients are either not filling their prescriptions or not taking them as often as prescribed to save money. Because many more patients are being treated with oral chemotherapy, capturing data regarding these individuals is vitally important, particularly for understanding disparities and differences in outcomes associated with the various drugs, Dr. Penberthy says. SEER leaders are exploring ways to capture information from pharmacies or other organizations that gather pharmacy data to obtain these additional data. They also are examining alternative ways to gather the information, including tests given to patients receiving targeted therapy drugs, which would indicate whether they are taking their medications. The second priority, seeking more data on biomarkers and other tests, will enable researchers to have a better overall picture of these biomarkers in the general population. One example of how this information can be used by researchers is linking biomarkers and test results of mutations with cancer registry data to document whether treatments are administered according to guidelines in routine clinical practice, says Dr. Jemal. For example, investigators could examine chemotherapy (cetuximab and panitumumab) patterns and survival in patients with colon cancer who either have or do not have KRAS gene mutations (the treatment is only recommended for patients without the mutation). “We currently capture data on about 32 biomarkers, and we're looking at starting to obtain more through the E-path reports, such as EGFR and ALK for lung cancer and BRAF for colon cancer,” Dr. Penberthy says. And even with current SEER biomarker data, she notes there is room for improvement, adding that not every biomarker is recorded. For example, SEER researchers found that from 2004 to 2013, approximately 34% of these biomarker tests were not reported through the hospital. As a result, SEER leaders will seek additional ways to capture the data, Dr. Penberthy says. Through NLP and by working with organizations such as Genomic Health Inc, they hope to enhance information regarding the predictive and prognostic markers that characterize tumors. The program's third priority of capturing additional outcomes data beyond survival and death focuses on gathering detailed information regarding disease recurrence. By assessing biopsy and imaging results, leaders hope to develop a more comprehensive picture of cancer progression in patients. Dr. Penberthy and her team are also developing new ways to present their data that are clinically relevant, and she notes that more information will be included on incidence by cancer molecular subtype. For example, featured data might include specific incidence trends such as a decline in squamous cell esophageal cancer along with an increase in adenocarcinoma of the esophagus. “Patients [now] are more sophisticated, and we have an obligation to present data in a more relevant manner,” she says. This will not only generate new knowledge on patterns of care and outcomes, but also [add] knowledge that will translate into policy changes.—Ahmedin Jemal, DVM, PhD
No takes yet. Share an insight, caveat, or question.
Carrie Printz (2015) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: