Keywords::
“Knowing the risk factors provides opportunities for intervention.”
Ann Graham Zauber is an epidemiologist and biostatistician at the Memorial Sloan Kettering Cancer Center (USA), and she works primarily in the field of colorectal cancer (CRC). Here, she tells us about her career to-date.
Please tell us about yourself
I entered the field of CRC prevention in 1983 through the National Polyp Study (NPS), a randomized clinical trial assessing the timing of surveillance intervals for patients with adenomas, the precursor lesion to CRC, following an initial colonoscopic polypectomy. I am a biostatistician with a background in cancer epidemiology and health policy. With the advent of colonoscopy in the 1970s, it was possible to detect and remove adenomas. However, given that adenomas could progress to cancer, there was concern as to how frequently adenoma patients should undergo subsequent colonoscopy looking for additional adenomas. Initially, patients were being brought back for colonoscopy every 6–12 months. The NPS assessed whether adenoma patients could wait 3 years for first surveillance or whether a colonoscopy at year 1 and again at year 3 was required. The percent of patients with advanced adenomas at surveillance colonoscopy was 3%, regardless of follow-up interval. These findings provided evidence of favorable risk benefit to waiting 3 years for the first surveillance if the initial colonoscopy was of high quality. Given this evidence, colonoscopy resources could be utilized to reduce the number of surveillance colonoscopies and to increase the number of patients having an initial colonoscopy. These results were used as evidence for surveillance guidelines for the national gastroenterology societies.
The high-quality data of the NPS yielded other important findings, including that colonoscopic polypectomy reduced CRC incidence by 76% over 10 years, and CRC mortality by 53% over 20 years, in comparison with the general population. These results demonstrating a long-term benefit of colonoscopy to reduce CRC mortality have contributed to the increase in colonoscopy screening, with 90% of current CRC screening being colonoscopy.
I have also served as biostatisician for several other randomized clinical trials for chemoprevention and screening interventions to decrease the risk of CRC. Given this background, in 2002, I had the opportunity to lead a team of epidemiologists, statisticians and health economists to develop microsimulation modeling for CRC. Our group is part of the National Cancer Institute (NCI; MD, USA) Cancer Intervention and Surveillance Modeling Network (CISNET), which conducts microsimulation modeling of risks and benefits of possible screening strategies. These results have been used to inform health policy. One of our major projects was to provide a decision analysis for age to begin, age to end, intervals of screening and different screening tests for the US Preventive Services Task Force in their recommendations for CRC screening in 2008 and again in 2016. Our analysis suggested that continuing screening past age 75 had more risk than benefit. Additionally, our analysis suggested beginning screening at age 50 provided a reasonable benefit in life years saved relative to potential risks of screening.
What made you choose a career in your field?
Each year the NCI publishes an annual report to the nation on the status of cancer. When I joined the NPS in the early 1980s, I could see that the US rates of CRC were increasing rapidly with an accelerated number of CRC cases and CRC deaths. However the incidence rates began falling as we entered the 2000s. There was consensus that a rise in CRC screening was most likely contributing to the decline in CRC incidence. However, in 2017, Rebecca Siegel of the American Cancer Society (ACS; GA, USA) provided an age-period-cohort analysis of the trends in CRC incidence, which demonstrated decreasing rates of CRC for those 50 years and older but, increasing incidence in younger ages.
In 2017, the ACS was considering its guidelines for CRC screening and whether screening should start at age 45 rather than at 50 for the average risk population. ACS requested our CISNET colon group to provide a decision analysis on age to begin screening comparing starting at age 45 rather than age 50. Furthermore, ACS requested the decision analysis provide results by race and sex.
It was necessary to update our models for the natural history of the adenoma carcinoma pathway to have increased risk of CRC occurring for the young adult population. Siegel’s work demonstrated that the current CRC incidence rate for age 45 is almost as high as the incidence rate at age 50 in 1997, when the guidelines for CRC screening recommended starting at age 50. Given this observation, our modeling showed that the balance of life years gained by beginning screening at age 45, relative to the risk associated with screening, was higher for beginning screening at age 45 rather than at age 50. Furthermore, beginning at age 45 was more favorable than age 50 for men and women, blacks and whites. The ACS published their guidelines with a qualified guideline to begin CRC screening at age 45 for the average risk population. They also revised their web sites and information packets to include the qualified guideline to start at age 45.
Screening has made a large impact on decreasing CRC incidence and deaths. Nonetheless, consensus is that universal screening below age 45 would have to be more focused.
What is the most difficult challenge you have encountered in your work?
The seminal work of Rebecca Siegal provided the message that we have a problem, and the extent of the problem, of rising rates for younger cohorts for CRC. Now we need to know what are the risk factors for early age CRC. Knowing the risk factors provides opportunities for intervention. High risk individuals would be considered for screening. Early age CRC’s are being diagnosed at late stage of disease. Informing medical professionals to consider CRC as a possible diagnosis could result in earlier stage cancer being detected. There is great interest in determining whether the early age cancers are different in their molecular genetic profile, compared with CRC at older ages. Epidemiologists are assessing risk factors such as obesity. Familial risk other than the established syndromes are also being considered.
What are the main highlights of your career so far?
In 1993, the NPS demonstrated that detection and removal of the adenoma, the precursor for CRC, led to reduction in CRC incidence. Thus, CRC can be prevented. Furthermore, in 2012 the NPS also demonstrated that removing adenomas led to reduction in CRC mortality. Thus, the CRC’s prevented were those that had the potential to cause CRC deaths. These papers have been used to recommend colonoscopy for screening. Today in the USA, colonoscopy is the main screening test for preventing CRC and CRC deaths.
My work with the NPS led to my leading a microsimulation modeling consortium for the CISNET group funded by the NCI. In 2008, the US Preventive Services Task Force asked our CISNET models for a decision analysis of CRC screening strategies of age to begin, age to end and intervals between subsequent tests that provided the optimal balance between life years gained and burden. This analysis was to inform their CRC screening recommendations in conjunction with a comprehensive literature review from an evidence based practice center. This was the first time the USPSTF had used a microsimulation modeling decision analysis as part of the background for their decision. This led USPSTF to include CISNET modeling analysis when deciding on screening recommendations for breast, cervix and lung cancer as well as for new CRC recommendations in 2016. This also led to the American Cancer Society using a CISNET modeling decision analysis when ACS provided a qualified recommendation to begin CRC screening at age 45 for the average population.
What are your main aims for the future?
Personalized or precision medicine has the potential to more precisely evaluate a person’s risk for CRC, and what screening strategies provide the greatest benefit relative to the risks of screening. We are currently working with a young investigator from the Harvard Nurses’ Health Study II to develop a risk prediction model incorporating secular changes of relevant lifestyle factors for early age onset of CRC. We will focus on modeling risk factor modifications to reduce CRC in the younger age cohorts. We are also collaborating with a consortium of multiple CRC case–control and cohort studies in building a risk prediction model based on genetic epidemiology risk factors.
We are collaborating with the New Hampshire Colonoscopy registry to inform our modeling of the serrated polyp pathway which is an alternate pathway to CRC. We will assess cost–effectiveness, as well as effectiveness, of CRC prevention strategies for early age of onset.
We are planning to use our CISNET microsimulation modeling to investigate past trends in risk factors and screening patterns to inform hypotheses for causes for the increase in CRC in young adults.
Financial & competing interests disclosure
The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
Disclaimer
The opinions expressed in this interview are those of the interviewee and do not necessarily reflect the views of Future Medicine Ltd.