Utilizing populationbased cancer registry data in the study
Utilizing population-based cancer registry data in the study of earlyonset CRC Jordan Karlitz, M. D. Associate Professor of Clinical Medicine Director, GI Hereditary Cancer and Genetics Program Tulane University School of Medicine Division of Gastroenterology
Overview Review of population-based cancer registry data systems available in U. S. n Example of utilizing registry data to identify CRC geographic “hot spots” w/ focus on early-onset CRC. n Example of utilizing registry data to understand Lynch syndrome management in early-onset CRC. n
SEER (Surveillance, Epidemiology, and End Results) SEER collects cancer incidence and survival data from population-based cancer registries covering approximately 28% of the population of the U. S. n Patient demographics, primary tumor site, tumor morphology, stage at diagnosis, first course of treatment. n NCI and CDC funding n
SEER States/Regions
CDC CER Data n n Comprehensive effectiveness research data-not typically collected by state registries, 2011 only. - Tumor markers (MSI, KRAS), treatment regimens, comorbidities, smoking, BMI, outcomes etc. -Population based--minimize bias, look at all patients in a region, state etc. Only 10 states participating in CER data collection. http: //www. cdc. gov/cancer/npcr/cer_data_collection. htm
Research Questions-SEER How can SEER incidence data be used to identify areas in the U. S. with disproportionately higher rates of CRC? n Formulate a hypothesis based on knowledge of risk factors, SES, heritage etc. n Look for cancer clustering based on hypothesis. n
Acadian Region Study-Background n n n Acadian parishes (S. Louisiana)--population 1. 2 million, approximately 500, 000 -700, 000 Cajuns. Cajun population-- established in the 1700 s when French settlers (Acadians) left Nova Scotia, Canada. Joined by others who emigrated from France. Compared with other US populationsn n n strong geographic ties to specific region maintained distinct community preserved language, Cajun French.
Cajuns- A Founder Population n Founder effect- occurs when a new population originates from a small subgroup of a larger population. Results in decreased genetic variation in the new population, which may have a nonrandom sample of genes from the original population. The Cajun population is the largest regional founder population in the U. S.
Founder Population. Continued n If mutation associated with disease introduced into the founder population, leads to unusually high frequency of inherited diseases, including certain malignancies. n Founder effects in Cajun population for Tay. Sachs, Oculopharyngeal Muscular Dystrophy, and Usher syndrome type 1 C.
Founder population= n Founder populations are like a magnifying glass for the study of genetic diseases as mutations can be found that are also present in other populations but are “diluted out” due to heterogeneity.
Founder population examples Newfoundland, Canada n Amongst the highest CRC rates in the world. n Founder mutations for Lynch Syndrome discovered. n Woods MO, Hyde AJ, Curtis FK et al. High frequency of hereditary colorectal cancer in Newfoundland likely involves novel susceptibility genes. Clin Cancer Res 2005; 11(19 Pt 1): 6853– 6861
Acadian Parishes Epidemiologic Study. Objective n Given the establishment of the Cajuns as a founder population, determine whether rates of colorectal cancer (CRC) are higher in the Acadian region compared with statewide and US rates as this may imply an increased frequency of hereditary CRC. Kartliz JJ et al. CRC incidence rates in Louisiana Acadian Parishes demonstrated to be among the highest in U. S. Clinical and Translational Gastro, 2014.
Study Design (Based at Louisiana Tumor Registry—SEER registry) n n n French language census data from 2000 used to identify an 18 -parish region (“Acadian 18”) with increased rates of French speakers at home as a marker for the Cajun population. 9 -parish subgroup (“Acadian 9”) of the 18 parishes with higher numbers of French speakers. Analyzed CRC incidence rates in the 9 -parishes, the 18 -parishes, LA and the U. S (stratified by race and gender).
18 Acadian Parish Region Language Map
CRC incidence n Acadian 9 (higher proportion of French speakers) whites: n n CRC 13% higher than LA and 23% higher than U. S Acadian 9, white males: n CRC 19% higher than LA and 37% higher than U. S. Strong trend of disproportionately increased CRC in white patients <50.
Race and gender stratified CRC incidence rates per 100, 000 popu in the 9 Acadian parishes and 18 Acadian parishes compared with LA and US SEER rates
CRC incidence rate/100, 000, Age 4049: Acadian 9 compared with Acadian 18, LA, U. S. SEER Patient Acadian 9 Demographic Acadian 18 Louisiana US SEER Ages 40 -49 White male/female 31. 3 P < 0. 21 26. 0 P<0. 21 25. 2 P<0. 10 21. 5 P<0. 01 White male 34. 4 27. 4 P<0. 27 26. 7 P<0. 16 22. 9 P<0. 01 White female 28. 1 24. 6 P<0. 58 23. 8 P<0. 41 19. 9 P<0. 07 P-values reflect comparison of the Acadian 18, LA, or US SEER with the Acadian 9
CRC incidence rate/100, 000, Age < 50: Acadian 9 compared with Acadian 18, LA, U. S. SEER Patient Demographic Acadian 9 Acadian 18 LA US SEER White male/female 8. 6 7. 5 P<0. 30 7. 4 P<0. 18 6. 2 P<0. 01 White male 9. 9 7. 5 P<0. 13 7. 8 P<0. 13 6. 6 P<0. 01 White female 7. 3 7. 4 P<1. 00 6. 9 P<0. 80 5. 9 P<0. 21 Age < 50 P-values reflect comparison of the Acadian 18, LA, or US SEER with the Acadian
n If the Acadian 9 were considered a “state, ” white males have highest CRC incidence in the U. S. by 11% (P<0. 0175) compared with other white male populations (Kentucky).
Colorectal cancer (CRC) incidence rates per 100, 000 population in the overall white population and white male population of the 18 Acadian parishes compared with the highest ranking states
Non-CRC in Acadian region Other cancers sharing environmental risk factors with CRC (breast, lung, pancreatic etc) examined- not disproportionately higher than statewide rates. n Argues against shared environmental risk factor as etiology of increased CRC in Acadian region. n Smoking, obesity and health insurance rates--no difference in these variables in comparison to Louisiana. n
Northern Maine n n n Aroostook County, Maine has the highest regional percentage of French speakers (22. 4%) in the U. S. outside of LA. Similar to the Acadian parishes of LAestablished Acadian population descended from Nova Scotia. CRC incidence in Aroostook county in whites is 26. 9% higher than statewide rates (P <. 05). CRC in white males is 30. 5% higher than statewide rates (P <. 05). Markedly different environment compared with LA
Aroostook county, Maine
Acadian Study Conclusions n n n Something unique about the white population, and particularly white male population, in the Acadian parishes, where French is more frequently spoken. Concerning trend of increased early-onset CRC Known founder population- ? founder effect for a hereditary form of CRC. ? Environmental risk factor (diet, obesity, etc. ) disproportionately effecting white population. Further study underway.
Lynch Syndrome Accounts for 3 -5% of CRC. n CRC < 50 years old= LS in 17%. n Autosomal Dominant with up to 80% lifetime risk of CRC. n Critical to screen—risk of other cancers, screen family members. n
Lynch Syndrome n Associated with mutations in mismatch repair genes MLH 1, MSH 2, MSH 6, PMS 2). Leads to microsatellite instability (MSI). n Leads to deficient corresponding MMR proteins (MLH 1, MSH 2, MSH 6, PMS 2) on immunohistochemical staining of tumor tissue. n
HNPCC/LS Testing Overview n Tumor Tissue Microsatellite Instability Testing (MSI) n Immunohistochemistry (IHC) n n Germline Testing- Blood Sample (if abnormal tumor tissue and/or strong suspicion—PREMM score).
Using CER data to assess Lynch Syndrome management in early-onset CRC n How effectively are we screening young patients with CRC for HNPCC/LS?
LS Screening by MSI and/or IHC in Early-Onset CRC n n First population-based study examining LS screening practices in U. S. All patients ≤ age 50 (highest risk) in Louisiana diagnosed with CRC in 2011 (274 patients). 61 healthcare facilities (public, teaching, comprehensive cancer centers etc. ). LS screening rate with MSI and/or IHC only 23% (should be 100%). Karlitz JJ et al: Population-based Lynch Syndrome screening by microsatellite instability in patients ≤ 50: prevalence, testing determinants, and result availability prior to colon surgery. Am J Gastroenterol 2015.
Variables Associated with the Ordering (no/yes) of MSI and/or IHC Testing-Multivariate Analysis
Table: Variables Associated with the Ordering (no/yes) of MSI and/or IHC Testing
Talk Conclusions n n n Tumor registries are an important source of population-based data that can help us better understand CRC/early-onset CRC. SEER cancer incidence data can provide descriptive information on cancer in regions--identify geographic hot spots (Acadian region). CER data allows more detailed assessment of populations that allow understanding of management (Lynch syndrome in early-onset CRC).
- Slides: 33