Faecal haemoglobin concentration is related to detection of advanced colorectal neoplasia in the next screening round

Abstract

Objective

To examine associations between faecal haemoglobin concentrations below the cut-off used in colorectal cancer screening and outcomes in the next screening round.

Methods

In the Scottish Bowel Screening Programme, faecal haemoglobin concentrations and diagnostic outcomes were investigated for participants with a negative result (faecal haemoglobin concentrations < 80.0 µg Hb/g faeces), followed by a positive result within two years.

Results

Of 37,780 participants with negative results, at the next screening round, 556 (1.5%) screened positive and 30,293 (80.2%) negative. Initial median faecal haemoglobin concentrations (2.1 µg Hb/g faeces, IQR: 0.0–13.2) were higher in those with subsequent positive results than those with subsequent negative results (0.0 µg Hb/g faeces, IQR: 0.0–1.4; p < 0.0001). Using faecal haemoglobin concentrations 0.0–19.9 µg Hb/g faeces as reference, logistic regression analysis showed high adjusted odds ratios for advanced neoplasia (advanced neoplasia: colorectal cancer or higher risk adenoma) detection at the next round of 14.3 (95% CI: 8.9–23.1) in those with initial faecal haemoglobin concentrations 20.0–39.9 µg Hb/g faeces, and 38.0 (95% CI: 20.2–71.2) with 60.0–79.9 µg Hb/g faeces.

Conclusions

A higher proportion of participants with faecal haemoglobin concentrations of ≥ 20 µg Hb/g faeces had advanced neoplasia detected at the next round than participants with lower faecal haemoglobin concentrations. Although most relevant when using high faecal haemoglobin concentrations cut-offs, studies of faecal haemoglobin concentrations and outcomes over screening rounds may provide strategies to direct available colonoscopy towards those at highest risk.

Keywords

Advanced neoplasia colorectal cancer faecal haemoglobin faecal immunochemical test faecal occult blood test risk screening

Introduction

In colorectal cancer (CRC) screening programmes, participants with faecal haemoglobin concentrations (f-Hb) below the selected cut-off for positivity and referral for follow-up colonoscopy, as determined by a quantitative faecal immunochemical test (FIT) for haemoglobin (Hb), are more likely to be diagnosed with an interval cancer (IC) than those with undetectable f-Hb.¹ Establishing the relationship between a detectable f-Hb below the programme selected cut-off, and diagnostic outcome in the longer term could assist in determining the role of f-Hb as a predictor of the future risk of advanced neoplasia (AN), defined as CRC plus higher risk adenoma (HRA).

Two previous studies from Taiwan are relevant. Firstly,² a cohort of 44,324 CRC screening participants aged 40–69, with f-Hb below the cut-off of ≥ 20.0 µg Hb/g faeces, was followed for a median of 4.39 years. The incidence of AN rose from 1.75/1000 person-years for those with f-Hb of 0.2–3.9 µg Hb/g faeces to 7.08/1000 in those with f-Hb of 16.0–19.9 µg Hb/g faeces. Relative to those with f-Hb of 0.2–3.9 µg Hb/g faeces, adjusted hazard ratios (HR) for AN were calculated as 3.41 for those with f-Hb in the category lower than but closest to the cut-off f-Hb used (16.0–19.9 µg Hb/g faeces). The second study used regression modelling to further establish the value of f-Hb as a predictor for colorectal neoplasia³; 54,921 participants invited between 2001 and 2007 were followed up in the Taiwanese annual FIT-based CRC screening programme to identify those diagnosed with IC, screen-detected CRC and adenoma. A trend of increasing HR with increasing baseline f-Hb was demonstrated for colorectal neoplasia, including for those with f-Hb above the traditionally used cut-off of ≥ 20 µg Hb/g faeces.

The available limited evidence indicates that the higher the baseline f-Hb, the greater the likelihood of a future finding of AN. If verified, this could have implications for the design and execution of CRC screening programmes. For example, those with a negative screening test result, who are at high risk of having AN by virtue of a detectable f-Hb, despite being lower than the cut-off f-Hb applied, could be prioritised for future repeat screening at a shorter interval than the current usual annual or biennial invitation. Conversely, those with undetectable f-Hb, and therefore deemed very unlikely to have a future finding of AN, could require less frequent invitations to screening. Such a strategy could be an important consideration in countries such as Scotland where, in view of demands on the colonoscopy resource, adoption of a cut-off f-Hb of ≥ 80.0 µg Hb/g faeces (equivalent to ≥ 400 ng Hb/ml buffer) for the planned FIT-based programme is likely, given that this was used in the pilot evaluation required to inform future adoption of FIT as a first-line test.⁴ This f-Hb cut-off was selected to achieve approximately 2% positivity, to mimic the current screening algorithm based on a two-tier reflex guaiac faecal occult blood test (gFOBT)/qualitative FIT strategy.⁴ As ICs are common at this f-Hb cut-off,¹ it is reasonable to postulate that a proportion of participants with f-Hb < 80.0 µg Hb/g faeces will have AN detected at the next screening round. We therefore aimed to investigate the relationship between f-Hb reported as a negative screening test result in our FIT evaluation and diagnostic outcomes identified over the two-year period of the next screening round, to test the hypothesis that those with f-Hb lower than but closer to the cut-off applied are more likely to be subsequently diagnosed with AN.

Methods

Outside the setting of the FIT evaluation, the Scottish Bowel Screening Programme (SBoSP) uses a two-tier reflex gFOBT/FIT screening algorithm (previously described).⁴ The pilot evaluation using FIT as a first-line screening test in Scotland has also been previously described.⁵ All participants with f-Hb ≥ 80.0 µg Hb/g faeces were reported as positive and offered colonoscopy.

Following completion of the FIT evaluation, the SBoSP returned the two NHS Boards involved (NHS Tayside or NHS Ayrshire and Arran) to the two-tier reflex gFOBT/FIT screening algorithm. The screening test results for all those eligible to take part in the two years of the next screening round after the FIT evaluation (13 January 2011 to 12 January 2013) and resident in either NHS area were examined. The f-Hb from the previous FIT-based pilot evaluation round for those with a positive gFOBT/FIT screening test result was retrieved from the Scottish Bowel Screening IT System. Data for colonoscopy outcomes and any subsequent pathology for those with a positive test result were downloaded from the clinical IT systems in two NHS Board areas. Data on colonoscopy findings, including number, size, Dukes’ stage and localisation of CRC and adenomas were collected. As in the SBoSP, HRAs were defined as adenomas > 10 mm in diameter, or when there were three or more adenoma present.⁶ AN was defined as CRC and HRA.

MedCalc (MedCalc Software, Mariakerke, Belgium) statistical software was used for all calculations. The Mann-Whitney U test was used for comparison of median f-Hb between classes. Probability of p < 0.05 was considered significant. Logistic regression analysis was performed to calculate odds ratios (OR) for diagnosis of AN at the next screening round amongst those in different f-Hb classes and different demographic groups, both adjusted and unadjusted for age and gender as known confounding variables.

This study was approved by the Programme Board, SBoSP. Ethical approval was not required.

Results

The majority of participants (37,780) in the FIT pilot evaluation had an initial negative test result.⁵ Of those, 92.7% were eligible for an invitation to further screening with the gFOBT/FIT algorithm. The majority of those not invited were above the age limit (75) applied in the SBoSP; others had died or were no longer resident in Scotland. There were 30,849 participants in the FIT evaluation and subsequent screening round with gFOBT/FIT. Table 1 details the outcomes and cut-off used in the FIT evaluation, by age and gender, of all participants with f-Hb < 80 µg Hb/g faeces.

Table 1.

Number of participants with a negative FIT screening test result and their screening test results in two years in the next gFOBT/FIT screening round.

	All		Men		Women
	n	%	n	%	n	%
Total with negative FIT screening test result < 80.0 µg Hb/g faeces	37,780		17,525		20,255
Test result in next gFOBT/FIT round:
Positive	556	1.5	339	1.9	217	1.1
Negative	30,293	80.2	13,910	79.4	16,383	80.9
Non-responder	4,165	11.0	1976	11.3	2189	10.8
Excluded	2,766	7.3	1300	7.4	1466	7.2

There were 556 (1.5%) participants with a positive result and 30,293 (80.2%) with negative results in the subsequent screening round. The initial median f-Hb was statistically significantly higher in those who then had a positive test result in the next round than in those who remained negative (2.1 µg Hb/g faeces, IQR 0.0–13.2 vs. 0.0 µg Hb/g faeces, IQR 0.0–1.4; p < 0.0001).

Table 2 shows the diagnostic outcomes, with median f-Hb, of participants who had a positive test result in the next screening round. The differences in median f-Hb between men and women were only statistically significant in those diagnosed at colonoscopy with either non-neoplastic pathology or normal findings (both p < 0.05).

Table 2.

Diagnostic outcomes and median faecal haemoglobin concentrations (f-Hb) at the initial FIT screening round in participants then having a positive screening test result at the next gFOBT/FIT screening round.

	All				Men				Women
	n	%	Initial median f-Hb (µg Hb/g faeces)	IQR^a	n	%	Initial median f-Hb (µg Hb/g faeces)	IQR	n	%	Initial median f-Hb (µg Hb/g faeces)	IQR
Total positive	556		2.1	0.0–13.0	339		1.6	1.2–3.0	217		2.4	0.0–12.9
Colorectal cancer (CRC)	26	4.7	16.7	1.2–31.6	19	5.6	21.0	0.5–36.3	7	3.2	10.2	2.5–24.0
Higher risk adenoma (HRA)	85	15.3	13.6	1.2–38.5	63	18.6	9.6	0.9–38.8	22	10.1	16.6	1.8–35.2
Advanced neoplasia (AN: CRC and HRA)	111	20.0	13.6	1.2–37.6	82	24.2	14.7	0.8–38.8	29	13.4	13.6	1.8–35.2
Low risk adenoma	65	11.7	1.7	0.0–8.5	42	12.4	1.6	0.0–10.9	23	10.6	1.8	0.0–7.0
Non-neoplastic pathology	131	23.6	1.6	0.0–7.4	72	21.2	0.7	0.0–5.8	59	27.2	3.4	0.1–15.1
Normal findings	169	30.4	1.4	0.0–6.2	93	27.4	0.3	0.0–5.6	76	35.0	2.0	0.0–7.3

IQR is the interquartile range (difference between the upper quartile and the lower quartile); non-neoplastic pathology comprises conditions including diverticular disease, haemorrhoids and inflammatory bowel disease and haemorrhoids.

A numerical f-Hb result was available for 30,823/30,849 participants in both screening rounds. In the majority of these (96.6%), f-Hb in the FIT round was in the lowest class of 0.0–19.9 µg Hb/g faeces; 87.4% of participants with a positive gFOBT/FIT result who did not have AN detected were in this low f-Hb class, compared with only 56.8% of those with AN. Table 3 shows the proportions of participants with different diagnostic outcomes according to class of f-Hb at the initial screening round with FIT. A lower proportion of participants with previously low f-Hb had AN detected than less severe outcomes, but this trend reversed with rising f-Hb (see Figure 1). A commonly used f-Hb cut-off for screening programmes outside Scotland is ≥ 20 µg Hb/g faeces. Table 4 shows further analysis of the class with f-Hb below this concentration, dissected into two f-Hb classes: those with undetectable f-Hb at the initial FIT screening round and those with f-Hb 1.0–19.9 µg Hb/g faeces.

Figure 1.

Percentages of FIT positive screening participants completing follow-up who had advanced neoplasia or less severe outcomes in the next gFOBT/FIT screening round, according to faecal haemoglobin concentration (f-Hb) class.

Table 3.

Outcomes of participants in both screening rounds according to faecal haemoglobin concentration class at the initial FIT screening round.

Outcome in next FOBT/FIT screening round	Faecal haemoglobin concentration class in the initial FIT screening round (µg Hb/g faeces)
	0.0–19.9		20.0–39.9		40.0–59.9		60.0–79.9
	n	%	n	%	n	%	n	%
Negative screening test result	29,049	96.6	661	2.2	227	0.8	125	0.4
Positive screening test result	449	80.8	59	10.6	26	4.7	22	4.0
Follow-up complete	383		52		22		20
Colorectal cancer (CRC)	14	3.7	7	13.5	1	4.5	4	20.0
Higher risk adenoma (HRA)	49	12.8	17	32.7	10	45.5	9	45.0
Advanced neoplasia (AN: CRC and HRA)	63	16.4	24	46.2	11	50.0	13	65.0
Low-risk adenoma	58	15.1	8	15.4	2	9.1	0	0.0
Non-neoplastic pathology^a	115	30.0	6	11.5	7	31.8	3	15.0
No pathology detected	147	38.4	14	26.9	2	9.1	4	20.0

Non-neoplastic pathology comprises conditions including diverticular disease, haemorrhoids and inflammatory bowel disease and haemorrhoids.

Table 4.

Outcomes of participants of both screening rounds with faecal haemoglobin concentration < 20 µg Hb/g faeces at the initial FIT screening round, dissected into two classes.

Outcome in subsequent screening round	Faecal haemoglobin concentration class in the initial FIT screening round (µg Hb/g faeces)
	Undetectable		1.0–19.9
	n	%	n	%
Negative screening test result	16,621	56.8	12,633	43.2
Positive screening test result	181	40.3	268	59.7
Follow-up complete	151		232
Colorectal cancer (CRC)	5	3.3	9	3.8
Higher risk adenoma (HRA)	14	9.3	35	15.1
Advanced neoplasia (AN: CRC and HRA)	19	12.6	44	19.0
Lower risk adenoma (LRA)	24	15.9	34	14.7
Non-neoplastic pathology^a	43	28.5	72	31.0
No pathology detected	65	43.0	82	35.3

Non-neoplastic pathology comprises conditions including diverticular disease, haemorrhoids and inflammatory bowel disease and haemorrhoids.

Logistic regression analysis showed very high age and gender-adjusted OR for AN, even in those with f-Hb 20.0–39.9 µg Hb/g faeces, using those with f-Hb 0.0–19.9 µg Hb/g faeces as the reference group (adjusted OR = 14.3, 95% CI 8.9–23.1) (see Table 5). Almost one-tenth of all participants who had f-Hb within the highest f-Hb class examined (60.0–79.9 µg Hb/g faeces), and participated in the subsequent round, had AN detected at follow-up for a positive test result. AN was over 40 times more prevalent in this class than among those with f-Hb previously 0.0–19.9 µg/g faeces, of whom 0.21% had AN. The adjusted OR for AN in the highest f-Hb class was 38.0 (95% CI 20.2–71.2).

Table 5.

Odds ratios (both unadjusted and adjusted for age and gender) for advanced neoplasia (AN) found at the next gFOBT/FIT round after a positive screening test result, according to faecal haemoglobin concentration (f-Hb) class in the previous FIT screening round.

f-Hb class (µg Hb/g faeces)	% with AN^a detected, %	Odds ratio
		Unadjusted	Adjusted
		(95% CI)	(95% CI)
0.0–19.9	0.2	1.0	1.0
20.0–39.9	3.4	16.2 (10.1–26.2)	14.3 (8.9–23.1)
40.0–59.9	4.4	21.5 (11.2–41.4)	17.7 (9.2–34.2)
60.0–79.9	9.0	45.9 (24.7–85.4)	38.0 (20.2–71.2)

Advanced neoplasia (AN) includes colorectal cancer and higher risk adenoma.

Discussion

A higher proportion of participants with a previous f-Hb lower than, but approaching, the cut-off used in our FIT pilot evaluation had AN detected at the next round than participants with low f-Hb. This provides further evidence that f-Hb is a strong predictor of future risk.

Positivity in participants with a previous round f-Hb below the applied cut-off of ≥ 80.0 µg Hb/g faeces was 1.5%. This was lower than either the overall positivity seen in the Scottish FIT pilot evaluation⁷ or the current gFOBT/FIT algorithm at ca.2%.⁴ This suggests that participants with a previously negative screening test result, even with a relatively high cut-off f-Hb, are less likely to have a positive test result in the subsequent screening round than the overall population (including first-time participants and participants with a previous positive test result) invited for screening.

Our findings further support the concept of f-Hb being a valuable predictor of risk of AN, as f-Hb was statistically significantly elevated, not only in those who would have a positive test result in the next round compared with those who again had a negative result but also in those who had AN detected at follow-up investigations compared with those with no pathology (both p < 0.0001). One in five participants with f-Hb in the f-Hb range nearest to the cut-off applied who were then positive in the next screening round were diagnosed with CRC. This compares with 1 in 27 participants in the lowest f-Hb class examined. Those with elevated f-Hb who go on to have a positive screening test result in the next round are over five times more likely to have CRC detected. When also taking into account detection of HRA, 65% of participants with a positive screening test result who previously had f-Hb in the highest f-Hb class examined had any AN diagnosed, compared with just 16.4% in the lowest. There is also a trend of a higher proportion of AN in participants from each of the four classes of increasing f-Hb. From these data, the odds of having AN detected following a positive screening result are even in those with previous f-Hb between 40.0 and 59.9 µg/g faeces. After this point, it is more likely than not that AN will be present at follow-up investigations in those with a positive screening test result in the next round. Almost one-tenth of all participants with f-Hb approaching the cut-off f-Hb applied went on to have AN detected, representing a 40-fold increase in the risk compared with those in the lowest f-Hb class. This finding is important when considering f-Hb as a predictor of future risk of AN at the time of a negative screening test result. Strong evidence is provided not only for the value of f-Hb as a risk factor for AN being detected at follow-up investigations of participants going on to have a positive test result but also for predicting future positivity and the subsequent detection of AN. As it has been suggested that f-Hb is a predictor not only of CRC mortality but also of all-cause death,⁸ there may be other benefits in the more general application of f-Hb as a predictor of ill health.

Figure 1 shows a graph that could be used to determine the effect of cut-off f-Hb on yield of AN for screening programmes collecting data such as these, irrespective of the faecal test used. The f-Hb, where the percentage of participants with positive screening results completing follow-up who had AN crosses that of those less severe outcomes, can indicate the cut-off f-Hb associated with the greatest benefit in terms of yield of AN, although this only includes participants with a positive result at the next screening round, not those with negative subsequent results. CRC FIT screening programmes use a wide range of f-Hb cut-offs, and that used by the SBoSP FIT evaluation was much higher than most. Our data allow some comparison of strategies using lower f-Hb cut-offs. Combining the data presented here with those in our recent publication on IC arising in the pilot evaluation of FIT screening,¹ if Scotland used a cut-off of ≥ 60.0 µg Hb/g faeces rather than the ≥ 80.0 µg used in the FIT evaluation,⁴ then 25.6% additional colonoscopy would yield 20.0% more CRC, or at least possible precursor lesions, presuming these were present at the time of the FIT-based screening. However, in addition to the reduction in the number of false-negative screening test results, the number of false-positive results will also increase, requiring assessment of the associated harms, such as overdiagnosis of lesions that would never have become symptomatic, and the risk of complications of colonoscopy.

Further analysis was performed on clinical outcomes at the subsequent screening round of those with an initial f-Hb below the commonly used cut-off f-Hb ≥ 20 µg Hb/g faeces. With the prevalence of AN higher in those with f-Hb between 1.0 and 19.9 µg Hb/g faeces than in those with undetectable f-Hb, evidence is provided that f-Hb is a predictor of future risk of colorectal neoplasia, even at a lower f-Hb cut-off than that used in Scotland.

As it is well accepted that men have higher median f-Hb than women,^9,10 it is surprising that the median initial f-Hb of participants with a positive screening test result at the next round was higher (although not statistically significantly so) in women than in men. Significantly higher median initial f-Hb was detected in women compared with men in those with non-neoplastic pathology, or normal findings at colonoscopy triggered by a positive screening test result at the next screening round. Men who had a false-positive screening test result appeared to be more likely to have had a low f-Hb in the previous screening round before experiencing short-term colonic blood loss to trigger their subsequent positive screening test result. Women, however, may have had longer term colonic bleeding, rising over time. The reasons for this are unclear and may be worthy of further study. Although not statistically significant, median f-Hb of participants with CRC detected at the next screening round was higher in men than in women, but the opposite was apparent for HRA. The previous median f-Hb for women with CRC was very similar to previous median f-Hb of men with HRA, further demonstrating the variation in f-Hb by gender as documented previously,^9,10 which should be taken into account in setting cut-off f-Hb for any screening programme.

An obvious weakness of this analysis is that different tests were employed in the two screening rounds examined. However, the cut-off used in the quantitative FIT pilot evaluation was equivalent, in terms of positivity, to the current gFOBT/FIT algorithm, to which the programme reverted. It would be of interest to conduct a similar study in which quantitative FIT was used in consecutive screening rounds, to allow investigation of the variation in f-Hb over time according to diagnostic outcomes. In addition, the number of CRC cases is low, limiting detailed analysis by sub-groups such as age, gender, and lesion site.

Two recent studies have compared clinical outcomes between two consecutive screening rounds with very large cohorts, allowing sub-analyses by gender and lesion site. A Spanish study showed that the second screening round was associated with lower median f-Hb and an increased proportion of proximal CRC.¹¹ The authors suggested various strategies to improve diagnosis of proximal CRC, such as using a lower cut-off f-Hb at the first compared with the second round. A French study has demonstrated that subsequent screening was associated with increased detection of proximal AN, particularly in women, and CRC was more often later stage.¹² It was concluded that, as the late stage CRC was likely to be previously missed lesions, the findings stressed the importance of repeated screening after a previous negative test result. These studies demonstrate the scope that exists for further work into the relationship between f-Hb and outcomes in subsequent screening rounds according to various sub groups, if working with a large enough sample size.

Our data could have an impact on CRC screening strategies. Risk scoring models are becoming increasingly developed, with an escalation in published studies in recent years.^13–22 We have confirmed that f-Hb predicts the presence of colorectal disease, through analysis of positivity in the next screening round. Screening programmes using FIT and considering the application of risk-scoring should incorporate f-Hb into such models, along with age and gender, and perhaps deprivation,²³ to improve their predictive power. The fact that f-Hb is related, not only to colorectal disease severity,²⁴ but also to future risk, has important implications for FIT-based CRC screening strategies.

Footnotes

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: CGF undertook consultancy with Immunostics Inc., Ocean, NJ, USA, Mode Diagnostics, Glasgow, Scotland, and Kyowa-Medex, Tokyo, Japan.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the Chief Scientist Office of the Scottish Government Health Department. Project: Establishing a Bowel Screening Research Unit (Grant Ref: CZH/6/4). JD was supported from a grant from Tenovus Tayside.

References

Digby J, Fraser CG, Carey FA, et al. Interval cancers using a quantitative faecal immunochemical test for haemoglobin (FIT) when colonoscopy capacity is limited. J Med Screen. Epub ahead of print 19 November 2015. DOI: 10.1177/0969141315609634.

Chen

Yen

Chiu

. Baseline immunochemical faecal occult blood test (iFOBT) concentration as a predictor of incident colorectal neoplasia: longitudinal follow-up of a Taiwanese population-based colorectal cancer screening cohort. Lancet Oncol 2011; 12: 551–8.

Yen

Chen

Chiu

. A new insight into fecal haemoglobin concentration-dependent predictor for colorectal neoplasia. Int J Cancer 2014; 135: 1203–12.

Fraser

Digby

McDonald

. Experience with a two-tier reflex gFOBT/FIT strategy in a national bowel screening programme. J Med Screen 2012; 19: 8–13.

Steele

McDonald

Digby

. Clinical outcomes using a faecal immunochemical test for haemoglobin as a first-line test in a national programme constrained by colonoscopy capacity. United European Gastroenterol J 2013; 1: 198–205.

Atkin

Saunders

. Surveillance guidelines after removal of colorectal adenomatous polyps. Gut 2002; 51(Suppl v): v6–vv9.

Steele

McClements

Libby

. Results from the first three rounds of the Scottish demonstration pilot of FOBT screening for colorectal cancer. Gut 2009; 58: 530–5.

Chen

Yen

Fraser

. Impact of faecal haemoglobin concentration on colorectal cancer mortality and all-cause death. BMJ Open 2013; 3: e003740–e003740.

McDonald

Strachan

Digby

. Faecal haemoglobin concentrations by gender and age: implications for population-based screening for colorectal cancer. Clin Chem Lab Med 2012; 50: 935–40.

10.

Fraser

Rubeca

Rapi

. Faecal haemoglobin concentrations vary with sex and age, but data are not transferable across geography for colorectal cancer screening. Clin Chem Lab Med 2014; 52: 1211–6.

11.

Bujanda

Sarasqueta

Castells

. Colorectal cancer in a second round after a negative faecal immunochemical test. Eur J Gastroenterol Hepatol 2015; 27: 813–8.

12.

Azimoufoussé Assogba

Jezewski-Serra

Lastier

. Impact of subsequent screening episodes on the positive predictive value for advanced neoplasia and on the distribution of anatomic subsites of colorectal cancer: a population-based study on behalf of the French colorectal cancer screening program. Cancer Epidemiol 2015; 39: 964–71.

13.

Aniwan

Rerknimitr

Kongkam

. A combination of clinical risk stratification and fecal immunochemical test results to prioritize colonoscopy screening in asymptomatic participants. Gastrointest Endosc 2015; 81: 719–27.

14.

Stegeman

de Wijkerslooth

Stoop

. Combining risk factors with faecal immunochemical test outcome for selecting CRC screenees for colonoscopy. Gut 2014; 62: 466–71.

15.

Auge

Pellise

Escudero

. Risk stratification for advanced colorectal neoplasia according to fecal haemoglobin concentration in a colorectal cancer screening program. Gastroenterology 2014; 147: 628–36.

16.

Chen

Mao

Pan

. Prediction rule for estimating advanced colorectal neoplasm risk in average-risk populations in Southern Jiangsu Province. Chin J Cancer Res 2014; 26: 4–11.

17.

Kaminski

Polkowski

Kraszewska

. A score to estimate the likelihood of detecting advanced colorectal neoplasia at colonoscopy. Gut 2014; 63: 1112–29.

18.

Wong

Lam

Tsoi

. A validated tool to predict colorectal neoplasia and inform screening choice for asymptomatic subjects. Gut 2014; 63: 1130–6.

19.

Chong

Roslani

Law

. Colonoscopic prioritization in colorectal carcinoma screening using quantitative immunochemical faecal occult blood test: a pilot study. Med J Malaysia 2013; 68: 30–3.

20.

Yeoh

Chiu

. Asia-Pacific Working Group on Colorectal Cancer. The Asia-Pacific Colorectal Screening score: a validated tool that stratifies risk for colorectal advanced neoplasia in asymptomatic Asian subjects. Gut 2011; 60: 1236–41.

21.

Omata

Shintani

Isozaki

. Diagnostic performance of quantitative fecal immunochemical test and multivariate prediction model for colorectal neoplasms in asymptomatic individuals. Eur J Gastroenterol Hepatol 2011; 23: 1036–41.

22.

Driver

Gaziano

Gelber

. Development of a risk score for colorectal cancer in men. Am J Med 2007; 120: 257–63.

23.

Digby

McDonald

Strachan

. Deprivation and faecal haemoglobin: implications for bowel cancer screening. J Med Screen 2014; 21: 95–7.

24.

Digby

Fraser

Carey

. Faecal haemoglobin concentration is related to severity of colorectal neoplasia. J Clin Pathol 2013; 66: 415–9.