Society: AGA

Introduction: The ultimate goal of colorectal cancer (CRC) screening is to prevent CRC by detecting advanced adenomas that potentially progress to CRC. Previous studies demonstrated that the Asia-Pacific Colorectal Screening (APCS) score could select more advanced neoplasia (AN) subjects and reduce the number needed to screen (NNS). Using the recommended cutoff FIT for the average-risk individual at 100 ng/ml could have missed AN in the high-risk individuals. Hypothetically, using the lower cutoff of FIT in those with high risk may improve the sensitivity for AN detection. Therefore, we aimed to evaluate the performance of the cutoff FIT of 25 ng/ml in the high-risk population and the cutoff FIT of 100 ng/ml in the non-high-risk population.
Methods: We conducted an organized CRC screening program in four districts in Thailand. Asymptomatic subjects aged 50-75 were invited and interviewed by the village health volunteer. Using the APCS score (i.e., age, male, smoking, and a first-degree relative with CRC), subjects were stratified into the high-risk or non-high-risk group. The positive FIT was called in the high-risk group with quantitative FIT using the cutoff ≥25 ng/ml (FIT25) and ≥100 ng/ml (FIT100) in the non-high-risk group. Positive FIT subjects were invited for colonoscopy. We assessed the positive predictive value (PPV) and NNS to detect AN—AN comprised of advanced adenoma and CRC. We compared the two cutoffs (FIT25/100) in high and non-high-risk individuals and the single cutoff in all individuals (FIT100). The analysis population was intention-to-treat.
Results: Of 12,000 invited subjects, 10262 subjects (85.5%) participated in the FIT program. The mean age was 60 years, and 33.7% were male. Of those, 8578 (83.6%) screenees were classified as non-high-risk, and 1684 (16.4%) were high-risk. The return rate of FIT was 99.6% (n=10221). The acceptance rate for colonoscopy was 68.3% (n=915). The positivity rate of FIT25/100 was 13.1% (n=1340) in the FIT25/100 group and 9.4% (n=961) in the FIT100 group (p<0.001). In the FIT25/100 and FIT100 groups, AN and CRC were detected in 158 and 34 screenees and 133 and 33 screenees, respectively. The PPV for AN (11.8% vs. 13.8%, p=0.17) and CRC (2.5% vs. 3.4%, p=0.21) were not significantly different between the FIT25/100 and FIT100 groups. The NNS to detect one AN in FIT25/100 was 64.9 (95%CI, 63.7 - 66.2) and significantly lower than that of in FIT100 (77.2 (95%CI, 75.7 - 78.7); p<0.001) (Table 1). This strategy reduced 15.8% in NNS to detect AN compared to FIT100 for all individuals (p<0.001).
Conclusion: Using a lower FIT cutoff in the high-risk individuals showed better performance with lower NNS while maintaining high PPV for AN and CRC detection compared to using a high cutoff for all individuals. This strategy could reduce the workload in CRC screening in limited-resource countries.

INTRODUCTION
Colorectal cancer (CRC) is the 2nd leading cause of cancer deaths, and screening is crucial to improve outcomes by early detection. Fecal immunochemical test (FIT) is the most common CRC screening method worldwide, and recommended annually for those at average risk. FIT-based screening programs vary in quality and efficacy, and rely on providers correctly ordering FIT, reliable patient participation, and timely diagnostic follow-up after abnormal results. We aimed to assess the frequency of and reasons behind inappropriate FIT utilization at the Phoenix Veterans Affairs Healthcare System.

METHODS
A retrospective review of FITs ordered from January to February 2019 was performed. For each FIT, the patient chart was reviewed to identify patient demographics, date of prior screening, and presence of high-risk features i.e., personal history of advanced polyps or CRC, IBD, or family history of CRC. FIT was deemed inappropriate if it was performed for any indication other than CRC screening, in a high-risk patient, or within the recommended interval from prior screening. Among positive FITs, performance of follow-up colonoscopy and reasons for lack of diagnostic follow-up was assessed.

RESULTS
A total of 546 patients underwent FIT during the observed period. Of them, 52 tested positive (9.5%), 490 tested negative (89.7%), and 4 tests were canceled (0.7%). Of all FITs, 312 of 546 were obtained for CRC screening (57%). The remainder (234 tests, 42.8%) were completed for inappropriate indications i.e., evaluation of anemia (9.5%), gastrointestinal bleeding (2.9%), other symptoms (2.9%), or unknown indication (27.5%). In those who underwent FIT for CRC screening, 88 of 312 (28.2%) had high-risk features including history of CRC (1%) or adenoma (25%), 1st degree relative with CRC (5.1%), or IBD (0.6%). In addition, 69 patients (22.1%) had recent FIT within 1 year or adequate colonoscopy within 5 years. FIT for CRC screening was appropriate in only 192 (35.2%) of all tests, and inappropriate in 354 (64.8%). Lastly, only 53.8% of patients who had a positive FIT underwent colonoscopy, due to patient refusal (13.5%), lack of referral for colonoscopy (30.8%), or GI advising against colonoscopy (5.8%).

CONCLUSIONS
In our review, inappropriate FIT use was substantial, with 64.8% of the 546 patient tests analyzed deemed inappropriate. The most common reason of inappropriate FIT was its use outside of CRC screening (42.8% of tests). When performed for CRC screening, 38.5% tests were inappropriately performed in patients with at least 1 high-risk feature (28.2%) or within the recommended surveillance interval from prior testing (22.1%). This review highlights the need for optimizing adherence to CRC screening guidelines by promoting accurate identification of eligible patients for screening with FIT and avoiding the use of FIT when a colonoscopy is indicated.

<b>Table 1. Factors leading to inappropriate FIT use determination.</b> <i>Patients who underwent FIT were deemed to have had an inappropriate test if they underwent FIT for any other indication than CRC screening. Among those who underwent FIT for CRC screening, FIT was deemed inappropriate if the patient had any of the listed historical risk factors (either confirmed per review of the patient’s chart, or unknown and therefore unable to confirm that patient did not have the specific risk factor).</i>

Table 1. Factors leading to inappropriate FIT use determination. Patients who underwent FIT were deemed to have had an inappropriate test if they underwent FIT for any other indication than CRC screening. Among those who underwent FIT for CRC screening, FIT was deemed inappropriate if the patient had any of the listed historical risk factors (either confirmed per review of the patient’s chart, or unknown and therefore unable to confirm that patient did not have the specific risk factor).

<b>Table 2. Follow-up diagnostics after abnormal (positive) FIT.</b>

Table 2. Follow-up diagnostics after abnormal (positive) FIT.

Introduction: Fecal immunochemical test (FIT) effectiveness for colorectal cancer (CRC) prevention and early detection is contingent on timely colonoscopy after abnormal results. Receipt of follow-up colonoscopy is low in safety-net settings, including Federally Qualified Health Centers (FQHCs), which provide primary care services to 30 million Americans annually. We aimed to determine patient factors associated with receipt of, and time to, follow-up colonoscopy in one of the largest FQHCs in California.

Methods: We identified patients ages 50-75 with an abnormal FIT result between 1/1/2016 and 8/13/2019 at Northeast Valley Health Corporation (NEVHC) in Los Angeles County. Data were obtained from electronic health records. Outcomes were 1) receipt of follow-up colonoscopy at 6 mos. and 2) number of days between FIT result and follow-up colonoscopy. We used multivariable logistic regression to assess correlates of colonoscopy at 6 months after abnormal FIT and Cox regression to identify patient characteristics associated with time to colonoscopy.

Results: Our study included 1,475 patients (abnormal FIT rate 12.4%) and the overall follow-up colonoscopy completion rate was 19.3% at 6 months. The mean (sd) time to colonoscopy completion was 8.5 (7.5) months, or 259.6 (228.6) days. Median age at FIT completion was 58.6 years, 62.4% were female, 80.5% were Latino/a, 74.0% reported Spanish language preference, and 49.5% were Medicaid insured (Table). Patients of other race/ethnicity (aOR 0.12, 95%CI 0.02-0.97) and with Spanish preference (aOR 0.55, 95%CI 0.35-0.85) were less likely to have completed colonoscopy within 6 months compared to White patients and English speakers (Table). Patients with health insurance or with an abnormal FIT in later study years were more likely to have completed colonoscopy within 6 months (Table). Black (aHR 0.41, 95%CI 0.18-0.95) and other race/ethnicity (aHR 0.36, 95%CI 0.13-0.99) patients were more likely to have longer time to colonoscopy compared to White patients. Patients with Medicare had shorter time to colonoscopy compared to uninsured patients (HR 1.45, 95%CI 1.05-2.01), and patients with an abnormal FIT in later study years also had shorter time to colonoscopy (Figure).

Discussion: FQHCs face considerable challenges to obtaining colonoscopies for patients with abnormal FIT results. In this large FQHC, patients with abnormal FIT were less likely to undergo colonoscopy within 6 months if their primary language was Spanish or if they were uninsured and were more likely to have a long delay to colonoscopy if they were Black or uninsured. We also observed improvement in FIT follow-up over time, likely reflecting the impact of system-level interventions at NEVHC, including tracking and navigation for patients with abnormal FIT results. Additional policies, resources, and interventions are urgently needed.

Catherine Crespi

Appearances