Sage Journals: Discover world-class research

Abstract

Objective

To evaluate interval cancers in the population-based colorectal cancer screening programme of Stockholm/Gotland, Sweden.

Methods

From 2008, individuals aged 60–69 were invited to colorectal cancer screening using biennial guaiac-based faecal occult blood test (Hemoccult®). Interval cancers, defined as colorectal cancer among participants not diagnosed by the screening programme but registered in the Swedish cancer register, were evaluated by cross-checking the screening histories for all cancers in the region 2008–2012.

Results

Of 203,848 individuals from nine different birth cohorts who participated (∼60%), 4530 (2.2%) tested positive. All invited individuals were followed up for 24 months after invitation. The cancer register reported 557 colorectal cancer, 219 (39.3%) screen-detected cancers and 338 (60.7%) interval cancers, generating both test- and episode sensitivities of approximately 40% and an interval cancer-rate of 17.1/10,000 tests. Among individuals with positive tests without colorectal cancer diagnosed at work-up colonoscopy, 37 interval cancers (10.9%) occurred. There was statistically significant lower sensitivity in women, ranging 22.4–32.2%, compared with 43.2–52.0% in men. Age-group and tumour location were not strongly correlated to screen-detected cancer rates. The programme sensitivity increased by year (20.3–25.0%), with successively more colorectal cancers diagnosed within the expanding programme (11.6–16.2%).

Conclusion

Interval cancer is a quality indicator of a screening programme. As the interval cancer-rate determined in a well-organized population-based screening programme was actually higher than the screen-detected cancer rate, a change to a more sensitive screening test is indicated. The lower screen-detected cancers among women, and compliance and quality of work-up colonoscopies also need attention.

Keywords

Colorectal screening population faecal occult blood test interval cancer sensitivity

Introduction

Following European Union recommendations to screen for colorectal cancer (CRC) with faecal occult blood test (FOBT) in 2003,¹ many European countries introduced national and regional programmes.² The recommendations were based on the results from four randomized controlled trials with guaiac-based FOBT (gFOBT), initiated in the 1980s, demonstrating a CRC-mortality reduction of around 16%.^3–6 The 2010 European Guidelines for Quality Assurance in Colorectal Cancer Screening and Diagnosis (EG) provide further details.⁷

The Stockholm/Gotland (Sweden) CRC screening programme with gFOBT using Hemoccult® started in 2008. To evaluate effectiveness, the implementation followed a model with random invitation of different birth cohorts aged between 60 and 69.⁸ Over five years, more than 200,000 individuals from nine different birth cohorts were invited, with an approximate 60% compliance rate, that increased with age, female gender and subsequent screening round.⁹

We used the EG definition of interval cancer (IC), i.e. a CRC that occurs in a screening participant after a negative FOBT-test result or after a positive test with a negative work-up colonoscopy, before the next invitation round,⁷ which in the Stockholm/Gotland screening programme is fixed 24 months after the latest invitation. Hence, an IC is a CRC not diagnosed by the screening programme. This definition is also in line with that used in breast cancer screening by European guidelines for quality assurance in breast cancer screening and diagnosis.¹⁰

The number of ICs is an intermediate measure of the effectiveness of the programme in decreasing CRC mortality. The frequency of ICs, together with the participation rate, determines the ability of the screening programme to detect cancer in the target population (programme sensitivity) and has to be monitored regularly. Hence, the programme sensitivity is used to evaluate the effect of the screening programme in the population. In a CRC screening programme with gFOBT, the programme sensitivity depends on the number of cancers not detected by the test (test sensitivity), but also on the number of cancers missed at the work-up colonoscopies after a positive test (episode sensitivity).^11,12 For simplicity, we have not used the proportional incidence method, but defined test sensitivity as the screen-detected cancers (SDC)/(SDC + IC with negative gFOBT) and the episode sensitivity (screening process sensitivity) as SDC/(SDC + all IC). Similar simple measurements have been validated by Andersen et al.¹³ in breast cancer screening. The IC-rate is an important quality indicator of the screening programme, and to enable comparison of ICs between different screening tests and CRC screening programmes, a standardized nomenclature has been proposed by the Working Group on interval CRC of the World Endoscopy Organization.¹⁴

This study aimed to prospectively evaluate the sensitivities of the Stockholm/Gotland CRC screening programme using gFOBT, by identifying the ICs, defined as the number of IC/10,000 negative tests and negative outcome of assessing a positive test, and to retrospectively evaluate the effect of the programme on the cancers occurring in the target population. Secondarily, we evaluated the differences in ICs and SDCs with respect to age-group, gender, and localization in the colorectum.

Methods

From 2008 in the Stockholm/Gotland screening programme, individuals aged between 60 and 69 were invited by randomly selected age-cohorts to expand the screening programme. By 2012, four cohorts had been invited once, four cohorts twice, and one cohort three times.⁹ In 2014, all age-groups were invited. A kit with three test-cards, instructions, and a prepaid return-envelope was sent to the individuals’ homes. People with negative test results were informed by a standard letter and further re-invited after 24 months. Individuals with positive tests (at least one out of three test-cards positive) were referred electronically to an assessment work-up colonoscopy. All test results, findings during the colonoscopy and histopathological findings were registered in the programme data administrative IT-system which also includes a database on all results, i.e. the screening register. Table 1 gives the screening performance indicators of the programme.⁹

Table 1.

Screening performance indicators of the guaiac-based programme of Stockholm/Gotland.

Year	Invited	Tested	Compliance Rate^a(%)	Positivity rate (%)	No. screening detected cancer (SDC)	PPV^b(%)	Colonoscopy compliance rate^c(%)
2008	44,930	25,049	55.8	2.2	26	4.8	87.9
2009	63,299	38,097	60.2	1.8	45	6.5	90.7
2010	68,392	40,602	59.4	2.3	45	4.9	90.2
2011	69,402	43,439	62.6	1.7	45	6.0	89.3
2012	94,537	56,661	59.9	2.9	58	3.6	86.0

No. of individuals invited and tested that particular year. Participation occurs up to six months after invitation.

Positive predicted value.

Individuals invited in that year having had a colonoscopy⁹.

Each individual in Sweden is assigned a unique identification number that is used in all population statistics. The number allows for matching with relevant health registers. The screening register was matched with the Swedish cancer register maintained by the National Board of Health and Welfare.¹⁵ The cancer register has a 98% coverage of all cancers diagnosed in Sweden.¹⁶ Information on CRC diagnosed during 2008–2012 in all individuals aged 60–69 living in the Stockholm/Gotland region was retrieved and stratified into CRC among screening participants, non-participants and non-invited.

Prospective analysis

The aim of the prospective study was to evaluate the sensitivity of the gFOBT (test sensitivity), i.e. the ability of gFOBT to identify CRC in the preclinical phase.¹² From the screening register we collected all test results among the screening participants, and all individuals were followed for 24 months after the latest invitation, i.e. the time frame of the study between 1 January 2008 and 31 December 2014. We prospectively categorized the CRC into SDCs or ICs, the latter only registered in the Swedish cancer register and not in the screening register (Figure 1). As an indirect estimate of how well the work-up colonoscopy confirms the CRC diagnosis after a positive test (episode sensitivity),¹² we categorized the ICs after a positive test result as an IC after a negative colonoscopy, an incomplete colonoscopy, or non-colonoscopy, i.e. in an individual not compliant with the work-up colonoscopy after the positive test.

Figure 1.

Prospective and retrospective analyses with time line and events.

Retrospective analysis – Mode of detection

To study the effectiveness of the screening programme to detect CRC in the target population (programme sensitivity),¹² we retrieved all CRC diagnoses between 2010 and 2012 from the Swedish cancer register. The rational for the time interval was that one screening interval of 24 months must have passed from start of the programme in 2008. We cross-checked with the screening register to evaluate all the CRC in the birth cohorts invited to the screening programme. There is a steady state of invitations from start of the programme in 2008, but due to the successive implementation, four cohorts had had prevalence screening rounds and five cohorts had undergone incidence rounds as well.⁹ For each individual with a CRC, we then retrospectively evaluated if they, within a screening interval (24 months), were invited, participated or did not participate, were diagnosed with a SDC, or participated and had a negative screening result (i.e. IC (negative test or positive test without CRC diagnosed with work-up colonoscopy)). The CRCs were categorized according to the mode by which they were detected (mode of detection); CRCs among non-invited (not yet covered by the programme), CRCs among invitees detected by the programme (SDCs), or not detected by the screening programme (ICs), CRCs among non-participants, or CRC among individuals having an incorrect test.

Statistical analysis

Each participant was prospectively cross-checked between the years 2008 and 2012 in the Swedish cancer register to analyse SDC-rate, IC-rate, and sensitivities. The screening episode sensitivity was defined as the number of SDCs divided by the sum of SDCs plus all ICs among participants (episode sensitivity = SDC/(SDC + all IC)). The test sensitivity only considers the IC after a false negative test (test sensitivity = SDC/(SDC + IC with negative gFOBT)). IC-rate has been presented as number of IC/10,000 negative tests, or a positive test without an assessment colonoscopy diagnosing the CRC. The IC-rate by year of screening invitation was then compared with the predicted CRC incidence per 10,000 individuals in the absence of screening (expected incidence rate (EIR)) according to the proportional incidence method.¹⁷ The EIR is estimated by a 10-year extrapolation of the actual age-standardized incidence of CRC in the region between 1998 and 2007. The difference between the IC-rate and the predicted CRC incidence without screening is an indirect estimate of the ability of gFOBT to detect CRC in the population targeted for screening.

Comparative analyses of differences in SDC, IC, and episode sensitivity by age-groups 60–64 vs. 65–71, and by gender, and tumour localization in colorectum (right- vs. left-sided) were then performed using two-tailed Fisher’s exact test. The CRC of the right colon was defined according the International Classification of Diseases 9th revision as C18.0–3 and C18.4 (from transverse colon and proximal) and the CRC of the left colon and rectum as C18.5–7 and C19–20 (splenic flexure and distal). A p-value of < 0.05 was considered statistically significant.

Results

Prospective analysis

In total, 203,848 individuals from nine different birth cohorts participated (∼60%) and sent in gFOBT between 2008 and 2012 (Table 2). Four cohorts were invited to one screening round, four to two rounds and one to three rounds. All those invited were followed for 24 months after the last invitation, to evaluate SDC and IC. Of the 557 diagnosed CRCs, 219 (39.3%) were SDCs and 338 (60.7%) were ICs, generating an IC-rate of 17.1/10,000 negative tests or positive tests without CRC diagnosed at colonoscopy (338 × 10,000/[193,690 + 4530−219]). The EIR ranged from 11.5–11.8 per 10,000 individuals, with an actual IC-rate from 5.0–10.6 per 10,000 at 0–12 months, and 7.8–12.9 at 13–24 months. Hence, all IC-rates were lower than the EIR except for the IC-rate at 13–24 months in 2011 (Table 3).

Table 2.

Prospective analysis of test result, screen-detected cancer (SDC) and interval cancer (IC).

Year	Number of tested individuals				No. SDC	SDC-rate (SDC/10,000 tested)	Interval cancer (IC)	IC-rate (IC/10,000 neg. tests^a)	Test sensitivity (%) SDC/SDC+IC
Year	Total	No. negative tests	No. positive tests	No. incorrect tests	No. SDC	SDC-rate (SDC/10,000 tested)	Interval cancer (IC)	IC-rate (IC/10,000 neg. tests^a)	Test sensitivity (%) SDC/SDC+IC
2008	25,049	23,586	539	924	26	10.4	45	18.7	36.6
2009	38,097	36,407	694	996	45	11.8	72	19.4	38.5
2010	40,602	38,511	914	1177	45	11.1	58	14.7	43.7
2011	43,439	41,490	755	1194	45	10.4	87	20.6	34.1
2012	56,661	53,696	1628	1337	58	10.2	76	13.8	43.3
Total	203,848	193,690	4530	5628	219	10.7	338	17.1	39.3

IC/10,000 tested negative or tested positive with a negative colonoscopy.

Table 3.

Prospective analysis of interval cancer (IC) by negative test result, colonoscopy outcome after a positive test, and time period when the IC was diagnosed outside the Hemoccult® CRC screening programme. The rate of IC is correlated to the experienced incidence of Stockholm/Gotland, Sweden, by year.

Year	Total no. IC					IC 0–12 months			IC 13–24 months
Year	Neg. test	Pos. test + neg. colo.	Pos. test + incompl.colo.	Pos. test + non-colo.	EIR^a	No.	IC-rate^b	IC-rate^b/ EIR^a(%)	No.	IC-rate^b	IC-rate^b/ EIR^a(%)
2008	37	1	3	4	11.5	25	10.6	92.2	20	8.5	73.7
2009	63	3	2	4	11.6	31	8.5	73.4	41	11.3	97.1
2010	55	1	1	1	11.7	27	7.0	59.9	30^c	7.8	66.6
2011	79	3	2	3	11.7	36	8.7	74.2	51	12.9	105.0
2012	67	1	3	5	11.8	27	5.0	42.6	49	9.1	77.4

Expected mean incidence per 10,000 individuals for the period 2008–2012 in absence of screening estimated by extrapolating the incidence 10 years prior to screening start (1998–2007).

The number of IC/10,000 negative tests or a positive test with a negative colonoscopy.

One IC > 24 months.

Approximately 11% (37/338) of the ICs were among individuals having had a positive test but negative follow-up colonoscopy (9/37, 24%), incomplete colonoscopy (11/37, 30%), or not complying with the colonoscopy work-up (17/37, 46%) (Table 3) The screening test sensitivity, only considering the CRC missed due to a negative gFOBT (false negative test), was 42% (219/[219 + 338−37]) and the episode sensitivity was just above 39% (219/557).

There was a non-significant higher sensitivity in the younger age-group 60–64 years (range 36.4–47.1%) compared with 65–71 years (30.2–41.3%), except for 2009 (Table 4). On the other hand, except for 2009, there was statistically significant higher sensitivity among males, with a more than double rate of SDC. Overall, as expected, there were more SDC localized in the left side of the colon.

Table 4.

Participating individuals followed for 24 months after a test with screening-detected cancer (SDC), and interval cancer (IC) by age-group, gender and tumour site stratified as right (dx) or left (sin) colorectum, estimation of differences in sensitivity, and by year of invitation.

Year	Age-group							Gender							Tumour site^a
	60–64			65–71			Sens. diff. (p-value^b)	Male			Female			Sens. diff. (p-value^b)	Dx			Sin			Sens. diff. (p-value^b)
	SDC	IC	Sens. %	SDC	IC	Sens. %	Sens. diff. (p-value^b)	SDC	IC	Sens. %	SDC	IC	Sens. %	Sens. diff. (p-value^b)	SDC	IC	Sens. %	SDC	IC	Sens. %	Sens. diff. (p-value^b)
2008	13	15	46.4	13	30	30.2	16.2 (0.211)	19	21	47.5	7	24	22.6	24.9 (0.047)	5	14	26.3	21	31	40.4	14.1 (0.405)
2009	12	21	36.4	33	51	39.3	2.9 (0.835)	27	34	44.3	18	38	32.1	12.2 (0.190)	12	25	32.4	31	47	39.7	7.3 (0.538)
2010	24	27	47.1	21	31	40.4	6.7 (0.554)	33	31	51.6	12	27	30.8	20.8 (0.043)	13	28	31.7	26	30	46.4	14.7 (0.208)
2011	11	18	37.9	34	69	33.0	4.9 (0.661)	32	42	43.2	13	45	22.4	20.8 (0.016)	16	28	36.4	27	59	31.4	5.0 (0.694)
2012	32	39	45.1	26	37	41.3	3.8 (0.728)	39	36	52.0	19	40	32.2	19.8 (0.024)	17	36	32.1	40	40	50.0	17.9 (0.050)

SDCs not specified as dx or sin; 2 in 2009, 6 in 2010, 2 in 2011 and 1 in 2012.

Fisher’s exact test.

Retrospective analysis – Mode of detection

The programme sensitivity increased by year (20.3–25.0%) (Table 5). Dividing the cancers by mode of detection, there were successively more CRCs diagnosed within the programme as it was expanding. Both SDC and IC frequencies increased by year (11.6–16.2% and 14.8–22.5%, respectively), but the CRC-rate among non-participants decreased. Just over one-third of CRCs in the target population were among non-invited in 2012.

Table 5.

Mode of colorectal cancer detection. Retrospective analysis of screening history for all cases of CRC of the screening-population registered in the Swedish cancer register for the years 2010, 2011 and 2012.

	2010	%	2011	%	2012	%
Total colorectal cancer	344	100	351	100	333	100
Not invited (not covered by screening programme)	147	42.7	135	38.5	117	35.1
Invited (covered by screening programme)	197	57.3	216	61.5	216	64.9
- Tested, cancer detected (SDC)	40	11.6	50	14.2	54	16.2
- Tested, no cancer detected (IC)	51	14.8	71	20.2	75	22.5
- Tested incorrect	9	2.6	4	1.1	9	2.7
- Not tested (non-participants)	97	28.2	91	25.9	78	23.4
Programme sensitivity	40/197	20.3	50/216	23.1	54/216	25.0

Discussion

In the Stockholm/Gotland population-based gFOBT CRC screening programme, the number of ICs was actually larger than the number of SDCs between 2008 and 2012, generating a test- and episode sensitivity of just above and below 40%, respectively. The episode sensitivity was in the lower limit compared with other population-based gFOBT screening programmes, predominantly in Europe, reporting sensitivities around 40–60%.^18–22 In Scotland and in the Catalonian region of Spain, increasing frequencies of IC were reported over the screening rounds from 31–59% to 32–46%, respectively.^19,22 The lower sensitivity in the last round in Scotland was explained by a decrease of SDCs rather than an increase in ICs. There could be many explanations for the IC differences.

First, compared with the Spanish programme using hospital data to identify CRC,²² we used Swedish cancer registry data. The completeness of the cancer register is around 98%, so close to all of IC-cases in Stockholm/Gotland are registered and detected in retrospective analysis of mode of detection.¹⁶ Underestimation of the IC-frequency could generate a falsely high sensitivity.

Second, the sensitivity of the gFOBT seems to increase with lower age, and the target groups of both the Scottish and Spanish programmes were aged 50–69,^19,22 compared with age 60–69 in the Stockholm/Gotland programme. Older people have a higher incidence of adenomas and CRC, but the unspecific bleeding of other causes dominates as the positive predictive value for cancer of the positive test in the cohort aged 60–69 in the Stockholm/Gotland programme is only 5.6%.⁹

Third, 11% of ICs in Stockholm/Gotland occurred among individuals with positive tests but without CRC diagnosed at work-up colonoscopy. The episode sensitivity, combining the ability of the screening test and the work-up colonoscopy to detect CRC, should be addressed.¹² The programme sensitivity could then be increased through higher compliance with work-up colonoscopies after a positive test, and higher quality examinations. High quality work-up colonoscopies are essential both to ensure that screening has a positive effect for the individual, and also for the overall incidence- and mortality reduction potential of the screening programme.^23–27 A programme intervention is required when close to half of individuals with IC after a positive test simply do not comply with the diagnostic work-up colonoscopy. The intervention could be more thorough information on the necessity of a work-up colonoscopy following a positive test, and frequent re-calls as reminders.

Fourth, intra-observer differences in the subjective assessment of a positive Hemoccult®-test result in a variation of the positivity rate that could be avoided by changing gFOBT to the quantitative faecal immunochemistry test (FIT). As FIT analyses the quantity of human occult blood, all that is required is the quantity of blood that would be considered a positive test, i.e. the cut-off level. FIT is more sensitive than gFOBT in detecting CRC and adenomatous polyps (adenoma) and generates lower IC-rates in population-based screening. There is also the advantage of increases in screening participation rates,^28–31 probably due to easier faeces sample collection and because it does not require dietary restrictions. In observational studies, FIT has also demonstrated promising effectiveness on CRC incidence and mortality when used in population-based programmes.^32–35

We did not find any statistically significant differences in sensitivity between age-groups nor, as had been reported previously, more ICs on the right side of colon.^19,36 The proportion of IC was the same in both the right and left sides of the colon, but because the expected incidence is higher on the left side (approximately two thirds of all CRC³⁷), the denominator is higher, and test sensitivity is also higher for left-sided CRC.

On the other hand, even though there were similar numbers of ICs between men and women, we found more than twice as many SDCs in absolute numbers diagnosed among men during the time period. Women have an up to 10% higher compliance rate in the Stockholm/Gotland screening programme,⁹ so the rate of IC in this programme is actually relatively higher among women, as has been reported elsewhere.^19,21,36,38 To equalize the SDCs and ICs between genders, a change to FIT screening could also be adequate. The amount of blood in the stool in screening is associated with CRC and adenomas, but also with male gender.^39–41 In changing to FIT, the cut-off for a positive test could be set to a more sensitive lower level for women, to generate the same positivity rate and volume of assessment colonoscopies as in men, although, because more women comply with the programme, the potential decrease in ICs among women with FIT at lower cut-off would be at the expense of an increased number of negative work-up colonoscopies after a positive test.

In general, all IC-rates except one were lower then the predicted incidence rates of CRC in the absence of screening. We could not adjust for potential selection bias from individuals with lower risk of CRC participating in the screening programme, which could generate an overestimation of the effect of gFOBT in detecting CRC (low IC due to few CRC in the population participating). On the other hand, by definition, the non-attenders could not have had an IC.

As the screening programme expanded, both the SDC and IC-rates increased by year, compared with CRCs among the non-invited. The CRC-rate among non-participants decreased by year, possibly due to screening invitations to younger age-groups, with both a lower participation-rate and lower prevalence of CRC, compared with the older age-groups.⁹

The four randomized studies that constitute the basis for the evidence that screening reduces CRC mortality were all guaiac-based.^3–6 It is reasonable to believe that, despite the low sensitivity in the Stockholm/Gotland routine screening programme, a similar effect is achievable. Changing to a more sensitive test (e.g. FIT) would eventually contribute to an improved effect on disease specific mortality reduction.

Conclusion

The IC-rate indicates the sensitivity of a screening programme to detect CRC and is a surrogate measure of programme effectiveness. In the organized Stockholm/Gotland gFOBT screening programme, there are actually more ICs than SDCs. A change to a more sensitive screening test with the potential to alter cut-off levels to equalize work-up colonoscopies between genders may increase programme sensitivity. Moreover, interventions to increase compliance with and quality of work-up colonoscopies have to be evaluated.

Footnotes

Acknowledgement

We thank Sini Kilpeläinen, Regional Cancer Centre Stockholm Gotland, Sweden for statistical work.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

European Council. COUNCIL RECOMMENDATION of 2 December 2003 on cancer screening (2003/878/EC). Off J Eur Union 2003; 327: 34–37.

Zavoral

Suchanek

Zavada

. Colorectal cancer screening in Europe. World J Gastroenterol 2009; 15: 5907–5915.

Hardcastle

Chamberlain

Robinson

. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996; 348: 1472–1477.

Kronborg

Fenger

Olsen

. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996; 348: 1467–1471.

Mandel

Church

Ederer

. Colorectal cancer mortality: effectiveness of biennial screening for fecal occult blood. J Natl Cancer Inst 1999; 91: 434–437.

Lindholm

Brevinge

Haglind

. Survival benefit in a randomized clinical trial of faecal occult blood screening for colorectal cancer. Br J Surg 2008; 95: 1029–1036.

European Colorectal Cancer Screening Guidelines Working Group, von Karsa

Patnick

. European guidelines for quality assurance in colorectal cancer screening and diagnosis: overview and introduction to the full supplement publication. Endoscopy 2013; 45: 51–59.

Malila

Oivanen

Hakama

. Implementation of colorectal cancer screening in Finland: experiences from the first three years of a public health programme. Z Gastroenterol 2008. April (46 Suppl 1): S25–S28.

Blom

Kilpeläinen

Hultcrantz

. Five-year experience of organized colorectal cancer screening in a Swedish population – increased compliance with age, female gender and subsequent screening round. J Med Screen 2014; 21: 144–150.

10.

Perry

Broeders

de Wolf

. European guidelines for quality assurance in breast cancer screening and diagnosis. Fourth edition – summary document. Ann Oncol 2008; 19: 614–622.

11.

Hakama

Auvinen

Day

. Sensitivity in cancer screening. J Med Screen 2007; 14: 174–177.

12.

Malila

Oivanen

Malminiemi

. Test, episode, and programme sensitivities of screening for colorectal cancer as a public health policy in Finland: experimental design. BMJ 2008; 337: a2261–a2261.

13.

Andersen

Törnberg

Lynge

. A simple way to measure the burden of interval cancers in breast cancer screening. BMC Cancer 2014; 14: 782–782.

14.

Sanduleanu

le Clercq

Dekker

. Definition and taxonomy of interval colorectal cancers: a proposal for standardising nomenclature. Gut 2015; 64: 1257–1267.

15.

The National Board of Health and Welfare, Swedish cancer register, www.socialstyrelsen.se/register/halsodataregister/cancerregistret/inenglish (22 November 2016).

16.

Barlow

Westergren

Holmberg

. The completeness of the Swedish Cancer Register: a sample survey for year 1998. Acta Oncol 2009; 48: 27–33.

17.

Day

. Estimating the sensitivity of a screening test. J Epidemiol Community Health 1985; 39: 1440–1448.

18.

Tazi

Faivre

Lejeune

. Interval cancers in a community-based programme of colorectal cancer screening with faecal occult blood test. Eur J Cancer Prev 1999; 8: 131–135.

19.

Steele

McClements

Watling

. Interval cancers in a FOBT-based colorectal cancer population screening programme: implications for stage, gender and tumour site. Gut 2012; 61: 576–581.

20.

Gill

Bramble

Rees

. Comparison of screen-detected and interval colorectal cancers in the Bowel Cancer Screening Programme. Br J Cancer 2012; 107: 417–421.

21.

Moss

Campbell

Melia

. Performance measures in three rounds of the English bowel cancer screening pilot. Gut 2012; 61: 101–107.

22.

Garcia

Domenech

Vidal

. Interval cancers in a population-based screening program for colorectal cancer in catalonia, Spain. Gastroenterol Res Pract 2015; 2015: 672410–672410.

23.

Bretthauer

. Colorectal cancer screening. J Intern Med 2011; 270: 87–98.

24.

Brouwers

De Vito

Bahirathan

. What implementation interventions increase cancer screening rates? A systematic review. Implement Sci 2011; 6: 111–111.

25.

Tinmouth

Ritvo

McGregor

. A qualitative evaluation of strategies to increase colorectal cancer screening uptake. Can Fam Physician 2011; 57: e7–e15.

26.

Blom

Liden

Jeppsson

. Compliance and findings in a Swedish population screened for colorectal cancer with sigmoidoscopy. Eur J Surg Oncol 2002; 28: 827–831.

27.

Blom

Liden

Nilsson

. Colorectal cancer screening with flexible sigmoidoscopy-participants' experiences and technical feasibility. Eur J Surg Oncol 2004; 30: 362–369.

28.

Zappa

Castiglione

Paci

. Measuring interval cancers in population-based screening using different assays of fecal occult blood testing: the District of Florence experience. Int J Cancer 2001; 92: 151–154.

29.

Zorzi

Fedato

Grazzini

. High sensitivity of five colorectal screening programmes with faecal immunochemical test in the Veneto Region, Italy. Gut 2011; 60: 944–949.

30.

Rabeneck

Rumble

Thompson

. Fecal immunochemical tests compared with guaiac fecal occult blood tests for population-based colorectal cancer screening. Can J Gastroenterol 2012; 26: 131–147.

31.

Crotta

Segnan

Paganin

. High rate of advanced adenoma detection in 4 rounds of colorectal cancer screening with the fecal immunochemical test. Clin Gastroenterol Hepatol 2012; 10: 633–638.

32.

Ventura

Mantellini

Grazzini

. The impact of immunochemical faecal occult blood testing on colorectal cancer incidence. Dig Liver Dis 2014; 46: 82–86.

33.

Giorgi Rossi

Vicentini

Sacchettini

. Impact of screening program on incidence of colorectal cancer: a cohort study in Italy. Am J Gastroenterol 2015; 110: 1359–1366.

34.

Zorzi

Fedeli

Schievano

. Impact on colorectal cancer mortality of screening programmes based on the faecal immunochemical test. Gut 2015; 64: 784–790.

35.

Chiu

Chen

Yen

. Effectiveness of fecal immunochemical testing in reducing colorectal cancer mortality from the One Million Taiwanese Screening Program. Cancer 2015; 121: 3221–3229.

36.

Morris

Whitehouse

Farrell

. A retrospective observational study examining the characteristics and outcomes of tumours diagnosed within and without of the English NHS Bowel Cancer Screening Programme. Br J Cancer 2012; 107: 757–764.

37.

Shinya

Wolff

. Morphology, anatomic distribution and cancer potential of colonic polyps. Ann Surg 1979; 190: 679–683.

38.

Paimela

Malila

Palva

. Early detection of colorectal cancer with faecal occult blood test screening. Br J Surg 2010; 97: 1567–1571.

39.

Kapidzic

van der Meulen

Hol

. Gender differences in fecal immunochemical test performance for early detection of colorectal neoplasia. Clin Gastroenterol Hepatol 2015; 13: 1464–1471.e4.

40.

Rozen

Levi

Hazazi

. Identification of colorectal adenomas by a quantitative immunochemical faecal occult blood screening test depends on adenoma characteristics, development threshold used and number of tests performed. Aliment Pharmacol Ther 2009; 29: 906–917.

41.

Ciatto

Martinelli

Castiglione

. Association of FOBT-assessed faecal Hb content with colonic lesions detected in the Florence screening programme. Br J Cancer 2007; 96: 218–221.

Interval cancers in a guaiac-based colorectal cancer screening programme: Consequences on sensitivity

Abstract

Objective

Methods

Results

Conclusion

Keywords

Introduction

Methods

Prospective analysis

Retrospective analysis – Mode of detection

Statistical analysis

Results

Prospective analysis

Retrospective analysis – Mode of detection

Discussion

Conclusion

Footnotes

Acknowledgement

Declaration of conflicting interests

Funding

References