COLON CANCER SCREENING: AN UPDATE

Overview COLON CANCER SCREENING: AN UPDATE Siddharth Verma, DO, JD Rutgers New Jersey Medical School Background Screening Updates in Specific Populations African Americans CRC in the younger age USPSTF Update Screening Modalities Cost of Screening Background CRC is 2 nd leading cause of cancer death in US 8% of all new cancer cases in US annually Adults in US have a lifetime CRC risk of ~5% Most frequently diagnosed: 65-74 years Median age at death: 68 years Estimated 135,000 new cases of CRC Estimated 50,000 deaths related to CRC Background 5 year survival: 90% if localized to colon 68% for regional disease (local LN) only 10% if distant mets CRC incidence and mortality rates declining over past 2 decades Attributed to CRC screening /early detection Between 2004-2013, CRC incidence declined at average rate of 3% per/yr CRC mortality declined at average rate of 2.7% per/yr As of 2012, ~28% of eligible adults still not screened. Who to Screen? Average Risk : 50 75 (USPSTF, 2008, 2016) Age 50+: AGA (2008); ACG (2008); ASGE (2006) https://www.cancer.gov/colorectalcancerrisk/ Calculates person s 5 yr, 10 yr, and lifetime risk Takes into consideration age, gender (M>F), race (AA> white), diet (red meat, low veg), lifestyle (tobacco, obesity, exercise, etc) in addition to family history (2-6X increased risk) Screening in African Americans 1

Screening in African Americans AGA (2008): screening beginning @ age 45 Based on ACG Committee on Minority Affairs and Cultural Diversity from 2005 ASGE followed suit in 2010 ACS, USMSTF, and USPSTF postponed due to lack of evidence on effect on screening/survival Joinpoint regression analysis demonstrated that incidence of CRC begins to increase at younger age in AA vs Caucasians. Paquette et al. (Gastrointest Endosc. 2015;82:878) Retrospective SEER database (2000-2011) review Calculate age-specific incidence in AA and calculate joinpoint. CRC incidence in AA younger than Caucasians is irrespective of SES. Screening in African Americans Cross sectional observational study of pts 50-75 yoa California Health Interview Survey: nonadherence to 2008 USPSTF CRC screening guidelines While age of screening changed, are clinicians actually referring for CRC screening? May et al (Am J Gastro. 2015;110:1388) examined association between patient race and provider recommendation for CRC screening No difference in: Education level, marital status, clinical settting (clinic v hospital), residency (urban/rural). Screening in the Young CRC: ~ 15% of patients are < 50 yoa. Screening in the Young Guidelines not yet suggesting screening < 50 (exception: @ 45 for AA) USPSTF 2016: due to modest increase in life-years gained by starting screening earlier and lack of empirical evidence in younger populations, no need to reduce screening age. Based on evidence from CISNET models CRC rates and all cause mortality declining due to improved screening, except for those < 50 2

Evidence of CRC in young O Connell J et al (Am Surg. 2003;69:866): SEER database (1973-1999) compared incidence rates in young patients (20-40 years, n = 5383) v older patients (60+ years, n = 256,401) Young: colon Ca incidence increased by 17% (P < 0.05), and rectal Ca incidence increased by 75% (P < 0.05) Younger patients also had less localized CRC and higher rates of poorly differentiated CRC Siegel et al (JNCI. 2017;109(8): djw322): Retrospective cohort study (SEER database) of patients age 20+ diagnosed with invasive CRC from 1974-2013 JNCI. 2017;109(8) USPSTF Guidelines on CRC Screening CRC Screening Guideline Update 2008 guidelines: colonoscopy q10 years, annual FIT/gFOBT, or flex sig q5 years with gfobt q3 years. No mention of CTC or DNA due to lack of evidence Changes to 2016 edition: Does not state any specific modality. No one size fits all approach More concerned about access to screening Added recommendations on CTC and DNA testing USPSTF Update: 2016 Screening Modalities 3

Modalities In general, 2 broad categories: Stool based (FIT, gfobt, DNA) gfobt Detects presence of blood in stool through peroxidase dependent chemical reaction with heme Has shown to decrease CRC mortality rates by 15%-33% Structural based (C/S, FS, DCBE, CTC) Sensitivity: 61% - 79% Specificity: 87% - 96% Annual screening better than biennial Minnesota Colon Cancer Control Study: 30 year follow up of pts randomly assigned to annual or biennial gfobt vs usual care showed a 32% decrease (annual testing) and 22% decrease (biennial testing) in CRC mortality (Shaukat et al, NEJM 2013:369;12) gfobt In office testing (ie stool with DRE)? Collin et al (Ann Int Med. 2005;142): specificity 97.5%, but sensitivity 4.9%! Limitations: High false-positive due to dietary intake of foods w/ peroxidase Need to stay away from beef, pork, poultry, fish, etc Ascorbic acid also can decrease test sensitivity ASA/NSAIDS/AC can lower PPV Requires 3 different stool samples FIT testing Directly measures human Hgb in stool, using monoclonal or polyclonal antibodies against globin moiety. Hassan et al (Aliment Pharmacol Ther 2012;36:929) meta analysis: FIT superior to gfobt both for CRC detection (RR, 1.96) and advanced neoplasia (RR, 2.28) Lee et al (Ann Intern Med 2014;160:171) 2014 meta-analysis (19 studies): pooled sensitivity 79% and specificity 94% for CRC FIT Testing At least 4 RCTs and 2 meta analyses showed improved adherence of FIT to gfobt 20% improved adherence Due to need for fewer samples and no dietary restrictions FIT and ASA/NSAID/anticoagulant use: 2 prospective studies (Am J Gastroenterol 2009;104:933; JAMA 2010;304:2513): improved sensitivity with no change in specificity No need to stop these meds if FIT offered ACG guidelines and updated ASGE recommendations (2017) have essentially replaced gfobt in favor of FIT 4

Fecal DNA Testing Has been studied since 1990s (Sidransky, et al identified ras mutations in stool, 1992) 1 st generation DNA testing: Better than gfobt (Imperiale et al, 2004) Limitations: DNA degradation, lack of buffers, limited marker panels. 2 nd generation DNA testing: 4 methylated genes, KRAS, alpha-actin DNA panel Fecal DNA Testing: Cologuard Imperiale et al (NEJM;2014): Asymptomatic persons b/w 50-84 yoa considered to be at average risk for CRC Exclusion: personal history of colorectal neoplasia or digestive cancer IBD or colon resection (except for sigmoid diverticula) Colonoscopy within the previous 9 years or a barium enema CT colangiography or sigmoidoscopy in the previous 5 years Positive FIT or gfobt in previous 6 months Overt rectal bleeding within the previous 30 days Personal or family history of CRC Cologuard vs. FIT Fecal DNA Testing Limitations Cost! Does not test all markers for CRC No guidelines on how often to screen Per manufacturer, testing q1-3 years ACG: q3 years USPSTF 2016: q1-3 years What to do with a + DNA and (-) colonoscopy? 5

CT Colonography Structure Based Testing CT images obtained after bowel preparation Followed by CO2 insufflation in rectum Colonoscopy indicated for lesions 10 mm or more than 3 lesions > 5 mm El Zoghbi et al, 2016 CT Colonography Accurately detect 90% of lesions > 10 mm in diameter (NEJM. 2008;359:1207) Heresbach D et al (Gut 2011;60: 658): Multicenter trial (845 patients) screening with CTC followed by colonoscopy. Sensitivity: 69% sensitivity and 91% specificity in detecting polyps > 6 mm Advantages: Quick: On average 10 minutes to complete No need for sedation/no risk of sedation related adverse events Low risk of perforation CT Colonography Limitations Operator dependent are radiologists skilled/trained Software availability? Only diagnostic: Still need colonoscopy if anything abnormal Radiation exposure though minimal Extracolonic findings: are common 40% to 70% of screening tests DCBE Flexible Sigmoidoscopy Sensitivity (38%) and specificity (86%) for polyps of any size. Still need to drink bowel prep Needs 20-40 minutes to complete Still need a colonoscopy if abnormal Therefore, ACG no longer recommends DCBE USPSTF: insufficient evidence to recommend DCBE 6

Flexible Sigmoidoscopy Aim: examine efficacy of a single flex sig in reducing CRC incidence and mortality rates after a 17 year follow-up 170,034 people: 112,936 in the control group 57,098 in the intervention group: 40,621 (71%) were screened Colonoscopy No RCTs have shown effect of colonoscopy on CRC mortality Several underway: COLONPREV (Spain) 1 time C/S v FIT q2 yr follow up 10 yrs Interim data (2012): FIT = C/S in CRC detection, but C/S > FIT in advanced adenoma (OR, 2.30) and non-advanced adenomas (OR, 9.80) detection rates (NEJM.2012:23) SCREESCO (Sweden) - ClinicalTrials.gov NCT00883792 CONFIRM (US) - ClinicalTrials.gov NCT01239082, NCT02078804 Observational studies: colonoscopies reduce CRC incidence and mortality Colonoscopy Pan et al (Am J Gastroenterol 2016; 11): Meta analysis (11 studies, 1.5 milion pts) Aim: evaluate CRC incidence/mortality in patients w/ nonmalignant findings on initial C/S Non-malignant: negative findings, hyperplastic polyps, adenomas, serrated lesions 11 studies: 5 cohort, 3 prospective, 2 retrospective, and 6 case control Follow up period for CRC/mortality: 3 studies >10 years, 7 studies 5 10 years, and 1 study of <5 years Incidence/Mortality: 6 studies CRC incidence only, 4 on mortality only, and 1 on both incidence and mortality All 11 studies: colonoscopy at baseline compared with no colonoscopy Forest plot of reduction in CRC incidence Pooled RR: 0.39 61% RR reduction in CRC incidence after C/S w/ non malignant findings Cost Considerations Forest plot of reduction in CRC mortality Pooled RR: 0.39 61% RR reduction in CRC mortality after C/S w/ non malignant findings 7

Cost Considerations Given burden of CRC, multiple economic studies indicate that CRC screening is highly cost effective But which test is the most cost effective? All strategies except for COLO were both more effective and less costly compared with no screening Sharaf. Am J Gastro 2013; 108 Yearly FIT yielded the highest mean QALYs per person, followed by screening C/S, MT sdna Every 3 years, and FIT every 2 years Ladabaum, Gastro 2016;151 Take Home Points Screening making a difference in incidence/mortality in most Rising among the younger patient population Room for improvement in getting people screened. USPSTF changes reflect need for access to screening gfobt has limitations; don t do in the office setting FIT may be slightly better (adherence, limited restrictions vs. cost) Stool DNA: Cost effective? Intervals? +test/(-) C/S? DCBE out! CTC pretty good, but still need prep C/S? RCT on C/S and mortality coming soon (FSig data assuring) 8