Sage Journals: Discover world-class research

Abstract

Objective

To compare interval cancer rates from the Irish breast screening programme, BreastCheck, for the period 2000–2007 with those from other European countries.

Methods

Data from BreastCheck was linked to National Cancer Registry breast cancer registrations, to calculate numbers of women screened, screen-detected cancers, and interval cancers, by year of screening, in the first and second years after screening, and by initial or subsequent screen. Estimated underlying cancer incidence from the period 1996–1999 inclusive was used to calculate proportionate incidence. We calculated the interval cancer ratio as an alternative measure of the burden of interval cancers.

Results

There were 372,658 screening records for 178,147 women in the period 2000–2007. The overall interval rate was 9.6 per 10,000 screens. In the first year after screening, the interval cancer rate was 5.8 per 10,000 screens and this increased to 13.4 in the second year after screening. The screen detection rate for the period was 53.6 per 10,000 screened for all screens combined. Initial screens produced a higher detection rate at 66.9 per 10,000 screened compared with subsequent screens with a screen-detected rate of 41.4 per 10,000 screens.

Conclusion

Interval breast cancer rates for the first years of the programme are within acceptable limits and are comparable with those in other European programmes. Nationwide roll-out together with the adoption of digital mammography may have an impact on interval cancer rates in future years.

Keywords

Interval cancer Breast Cancer Screening

Introduction

Mammographic screening programmes, introduced in many countries with the aim of reducing breast cancer mortality,^1–3 must balance the costs and benefits of screening. While these are still the subject of debate,^4,5 the consensus is that well-organized and systematically conducted breast screening, with rigorous internal and external quality control, is effective at the population level.^6,7 Population-based screening programmes usually have no internal mechanism for monitoring the incidence of breast cancer in the interval between screens, which is generally done by following up screening attendees through population-based cancer registries.⁷ Interval cancer rates can provide evidence on the sensitivity of the screening test, and on appropriate screening intervals, and may provide further information on the natural history of the disease.⁸ If interval cancer rates are high, further investigation may be required into why this is happening and to ascertain the nature of the interval cancer. Most programmes with at least one full round of screening have reported on interval cancer rates.⁹

Following a pilot programme in the Dublin region,¹⁰ the Irish breast cancer screening programme, BreastCheck, commenced in 2000 in the east of the country, covering about 50% of the eligible population, and was extended to the rest of the country beginning in 2007.¹¹ A full round of screening of the entire national eligible population was achieved in 2013. Women aged 50–64 are invited for free screening every two years. Two-view mammography and double reading of mammograms were used from the outset of the programme, and digital mammography was adopted in 2007.^12,13

We aimed to evaluate the incidence of interval cancers in the early years of the Irish programme and to compare this with recent data from population-based screening programmes in other European countries.

Methods

Ireland has a mixed public-private health system; screening mammography is available through the public national BreastCheck programme. Some opportunistic screening is performed in private hospitals and clinics, but we do not have access to this data. A linked dataset suggests that there were 765 in-situ cancer diagnoses within two years of a screening (with no further invasive breast cancer in the period 2000–2009) of which 176 (23%) were interval cancers. This paper describes invasive cancers detected in women attending the organized programme, BreastCheck.

Data was provided by BreastCheck on all women screened from 2000–2007. On registration, each woman was asked to tick a box by BreastCheck to consent to her data being shared with the National Cancer Registry. After the exclusion of a small number of duplicate observations and records relating to women outside the screening age group, there were 372,658 screening records for 178,147 women. Using probabilistic matching methods, these records were linked to breast cancer registrations, both invasive (ICD10 C50) and in situ (ICD10 D05), from the National Cancer Registry diagnosed in the period 2000–2009. The National Cancer Registry has registered all new cancer cases, including in situ disease, since 1994, and case ascertainment for breast cancer is at least 97–98% complete.¹⁴

For each woman the follow-up time from screening to the end of 2009, or age 66, or date of breast cancer diagnosis, whichever occurred first, was calculated to give woman years at risk (wyar) overall, for initial (prevalent) and subsequent (incident) screen and by year of screening. If a woman had more than one tumour, only the earliest invasive tumour was retained in the dataset. If two invasive tumours were diagnosed at the same time, the one with worse prognosis was retained. Cancers of unknown type and those coded as Paget’s disease and phyllodes tumour were excluded from the dataset. Each successfully linked cancer was classified as screen-detected or interval (diagnosed within two years of a negative screen). It was not possible to tell if 527 cancers diagnosed more than two years after a screen were due to a missed screening episode or delay by either the programme or the woman, so these were excluded from the analysis. These women were considered lapsed attenders¹⁵ and considered a separate group from those with screen detected or interval cancers. Numbers and rates are given here for invasive breast cancer only.

Screen detected (interval) cancer rates per 10,000 screens are calculated by dividing the number of screen detected (interval) cancers by the number of women screened and multiplying by 10,000. Several methods exist to describe the incidence rate of interval cancers: as a rate for a specified period (the number of interval cancers divided by the number of wyar for that period); as a proportion of the expected rate for that period (proportionate incidence);⁷ or a proportion of total cancers in the screened population (number of interval cancers divided by the total number of screen-detected and interval cancers for that period). This last method, the interval cancer ratio has been argued by Andersen et al.¹⁶ to be a good estimate of the burden of interval cancers, particularly when the estimation of background incidence is problematic. Published data from other screening programmes are not consistent in their use of these methods, and few give all of these statistics. In this analysis, we calculated the rate per wyar and per woman screened in addition to the proportionate incidence (observed/expected rate ratio) and interval cancer ratio. The background incidence was defined as the pre-screening incidence rates in the four years 1996–1999 overall and by 5-year age group for women aged 50–64. An upward trend in incidence during 1994–1999 was not statistically significant, and so we did not project this trend for the years 2000–2007. Rates were not age-standardized. The 95% confidence interval for proportionate incidence (PrI) was approximated by $e^{\ln (Prl) \pm 1.96 \sqrt{\frac{1}{O} + \frac{1}{IC}}}$ , where O is the number of cancers for which the expected incidence rate was calculated while IC is the numbers of interval cancers for the relevant group.

We used Joinpoint,¹⁷ a piecewise log-linear model with limited constraints on the positions of the joinpoints, to examine whether there were trends in screen-detected or interval cancer rates over the time period and to estimate any annual percent changes in the rates. Finally, we abstracted relevant information from papers identified on PubMed to report interval cancer rates from other population-based breast cancer screening programmes in Europe.

Results

BreastCheck carried out 372,658 screening episodes in 2000–2007, of which 48% were initial screens. The number of screens increased annually throughout the period, from 30,288 in 2001, the first complete year of screening, to 65,948 in 2007. The proportion of initial screens declined from 99.9% in 2001 to 23.3% in 2007. Women were screened up to four times in the period 2000–2007 (table 1), and 1996 invasive cancers were detected, 59% of them at the initial screening episode. The screen detection rate (invasive cancers only) was 66.9 per 10,000 screened for initial screens and 41.4 per 10,000 for subsequent screens (table 2). The screen detection rate decreased from 36.8 per 10,000 wyar in 2000 to 24.8 per 10,000 wyar in 2007, with an annual percentage change of −7.7% (95% confidence interval (−10.6%, −4.8%)). However this trend appears to be levelling off at the end of the period.

Table 1.

Number of screen-detected and interval cancers by year and screening episode, 2000–2007.

	Screening episode
Year of screening	Initial	Subsequent				Number of women screened
Year of screening	1st	2nd	3rd	4th	all	Number of women screened
Screen-detected cancers
2000	92	0	0	0	92	12,597
2001	218	0	0	0	218	30,288
2002	240	22	0	0	262	38,806
2003	182	112	0	0	294	49,671
2004	130	134	1	0	265	51,949
2005	114	82	71	0	267	59,099
2006	119	59	95	4	277	64,300
2007	96	74	94	57	320	65,948
2000–2007	1,191	483	261	61	1,996	372,658
Interval cancers
2000	20	0	0	0	20	12,597
2001	66	0	0	0	66	30,288
2002	69	9	0	0	78	38,806
2003	61	43	0	0	104	49,671
2004	28	45	0	0	73	51,949
2005	32	31	34	0	97	59,099
2006	41	46	58	1	146	64,300
2007	32	26	45	30	133	65,948
2000–2007	349	200	137	31	717	372,658

Table 2.

Numbers and rates of screen-detected and interval cancers, overall and for initial and subsequent round, by age at screening.

		age group
		50–54	55–59	60–64	all ages
	Expected incidence rate (based on 1994–1999)	22.4	25.8	24.9	24.2
Initial screens	number of screens	93,091	49,008	36,047	178,147
	woman years at risk	183,486	96,324	70,998	350,809
	number of screen detected cancers	479	376	336	1,191
	number of interval cancers	180	91	78	349
	screen detected cancers per 10,000 screens	51.5	76.7	93.2	66.9
	interval cancers per 10,000 screens	19.3	18.6	21.6	19.6
	interval cancers per 10,000 wyar	9.8	9.4	11.0	9.9
	Proportionate incidence (95% CI)	43%	36%	43%	40%
		(37%,51%)	(30%, 46%)	(34%, 55%)	(37%, 46%)
	Interval cancer ratio	0.27	0.19	0.19	0.23
Subsequent screens	number of screens	39,092	87,193	68,226	194,511
	woman years at risk	76,901	171,432	134,502	382,835
	number of screen detected cancers	125	330	350	805
	number of interval cancers	58	165	145	368
	screen detected cancers per 10,000 screens	32.0	37.8	51.3	41.4
	interval cancers per 10,000 screens	14.8	18.9	21.3	18.9
	interval cancers per 10,000 wyar	7.5	9.6	10.8	9.6
	Proportionate incidence (95% CI)	33%	37%	43%	39%
		(26%, 44%)	(31%, 44%)	(37%, 52%)	(36%, 44%)
	Interval cancer ratio	0.32	0.33	0.29	0.31
All screens	number of screens	132,183	136,201	104,273	372,658
	woman years at risk	260,387	267,756	205,500	733,644
	number of screen detected cancers	604	706	686	1,996
	number of interval cancers	238	256	223	717
	screen detected cancers per 10,000 screens	45.7	51.8	65.8	53.6
	interval cancers per 10,000 screens	18.0	18.8	21.4	19.2
	interval cancers per 10,000 wyar	9.1	9.6	10.9	9.8
	Proportionate incidence (95% CI)	40%	36%	43%	40%
		(35%, 47%)	(32%, 43%)	(38%, 51%)	(37%, 44%)
	Interval cancer ratio	0.28	0.27	0.25	0.26

There were 717 invasive interval cancers diagnosed in women aged 50–64 at screening and aged 50–66 at diagnosis (table 2). The interval cancer rate was 19.6 per 10,000 screens for initial screens and 18.9 per 10,000 screens for subsequent screens (table 2). Proportional incidence of interval cancers and interval cancer ratio are presented for three age-groups and overall in table 2.

The incidence of invasive interval cancer was 9.8 per 10,000 wyar overall (Table 3), 5.8 per 10,000 wyar in the first 12 months after screening, 13.8 per 10,000 wyar in the second year after a negative screening. The interval cancer rate per 10,000 wyar (9.8) was almost identical to the rate per woman (per 10,000) screened (9.6). Over the period 2000–2007, interval cancer rates were consistently higher in the 12–23 months after screening compared with 0–11 months afterwards. There was no noticeable trend in overall interval cancer rates by calendar year in the period 2000–2007.

Table 3.

Woman years at risk, numbers and rates of interval cancers, by time since last negative screen and screening year.

			Year of last routine screen
			2000	2001	2002	2003	2004	2005	2006	2007	2000–2007
		Number of women screened	12,597	30,288	38,806	49,671	51,949	59,099	64,300	65,948	372,658
0–11 months		woman years at risk (wyar)	12,511	30,079	38,555	49,376	51,685	58,837	64,021	65,624	370,687
	Interval cancers	number	4	15	29	32	25	21	47	43	216
		per 10,000 screens*	3.2	5.0	7.5	6.4	4.8	3.6	7.3	6.5	5.8
		per 10,000 wyar	3.2	5.0	7.5	6.5	4.8	3.6	7.3	6.6	5.8
12–23 months		woman years at risk	12,459	29,930	37,576	48,775	50,394	57,299	62,626	63,899	362,957
	Interval cancers	number	16	51	49	72	48	76	99	90	501
		per 10,000 screens*	12.7	16.8	12.6	14.5	9.2	12.9	15.4	13.6	13.4
		per 10,000 wyar	12.8	17.0	13.0	14.8	9.5	13.3	15.8	14.1	13.8
0–23 months		woman years at risk	24,970	60,009	76,131	98,151	102,078	116,136	126,647	129,523	733,644
	Interval cancers	number	20	66	78	104	73	97	146	133	717
		per 10,000 screens*	7.9	10.9	10.0	10.5	7.0	8.2	11.4	10.1	9.6
		per 10,000 wyar	8.0	11.0	10.2	10.6	7.2	8.4	11.5	10.3	9.8
	Screen-detected	number	92	218	262	294	265	267	277	321	1,996
	cancers	rate per 10000*	36.5	36.0	33.8	29.6	25.5	22.6	21.5	24.3	26.8
		per 10,000 wyar	36.8	36.3	34.4	30.0	26.0	23.0	21.9	24.8	27.2

(average per 12 month period)

Table 4 shows these results compared with those published by other European population-based screening programmes and the pilot programme in Ireland, ranked by overall proportionate incidence. The incidence of interval cancers (2000–2007), relative to the background incidence rate, in the first 12 months following screening in Ireland was the fifth lowest of 13. The proportionate incidence of interval cancer improved in Ireland with the rollout of the national screening programme compared with the pilot programme.¹⁰

Table 4.

Comparison of interval cancer rates and proportionate incidence in Ireland and other European countries, ordered by proportionate incidence (0–23 months).

					T time since last screening episode
					0 – 11 months			12 – 23 months			0 – 23 months
Area/country	Dates of screening	EIR	Age group invited	No of tests	IC	IC rate per 10000 screens	Proportionate incidence	IC	IC rate per 10000 screens	Proportionate incidence	IC	IC rate per 10000 screens	Proportionate incidence
Navarra (Spain)⁹	1990–1996	16.2	45–65	126,318	27	2.1	13.0%	79	6.3	38.6%	106	8.4	25.9%
Isère (France)¹⁸	2002–2004	24.5²¹	50–69	31,377	10	3.2	13.0%	†	†	†	48	15.3	31.2%
Norrbotten (Sweden)¹⁹ *	1989–2002	16	40–74	261,190	132	5.1	31.6%	147	5.6	35.2%	279	10.7	33.4%
Firenze (Italy)⁹	1990–1994	22.2	50–69	49,148	19	3.9	17.4%	56	11.4	51.3%	75	15.3	34.4%
Torino (Italy)⁹	1992–1996	20	50–59	41911	23	5.5	25.0%	42	10	45.0%	65	15.5	35.0%
England/Wales/ Northern Ireland⁸	1997–2003	21.6²²	50–64	7,320,648	4,048	5.9	25.6%	8,289	11.3	52.4%	12,337	16.9	39.0%
Netherlands²⁰	1990–1993	23.2	50–69	756,283	464	6.1	26.4%	915	12.1	52.1%	1,379	18.2	39.3%
Ireland	2000–2007	24.3	50–64	372,658	216	5.8	23.9%	501	13.6	55.1%	717	19.5	39.5%
Marseille (France)⁹	1993–1998	20.1	50–69	140,086	75	5.4	26.6%	169	12.1	60.0%	244	17.4	43.3%
Strasbourg (France⁹	1989–1997	22.6	50–65	168,186	116	6.9	30.5%	243	14.4	63.9%	359	21.3	47.2%
Isère (France)¹⁸	1994–1999	24.5	50–69	66,232	45	6.8	27.7%	†	†	†	158	23.9	48.7%
Ireland (pilot programme)¹⁰	1989–1994	22.9	50–64	35159	†	†	†	†	†	†	77	21.9	47.8%
Four counties (Norway)⁹	1996–1997	20	50–69	126,779	57	4.5	22.5%	190	15.0	74.9%	247	19.5	48.7%
Pirkanmaa (Finland)⁹	1988–1999	13.1	50–59	64803	42	6.5	49.6%	71	11	84.0%	113	17.4	66.4%

EIR: expected incidence rate (per 100,000 women)

IC: number of interval cancers

ICR: interval cancer rate per 10,000 screens

Isère is listed twice in the above table, as the study reported on two separate periods when single view (1994–1999) and two view mammography (2002–2004) was performed.

†

the figure was not reported in the relevant paper

Norrbotten is the most northern and most remote region in Sweden with an EIR of 16. Stockholm represents a capital area and has an EIR of 25.8.

Table 5 ranks countries by their overall interval cancer ratio. With the exception of Isere (2002–2004, two-view mammography) the programmes had higher interval cancer ratio for subsequent screens than for initial screens. Ireland ranked mid-way in this list.

Table 5.

Comparison of interval cancer ratios in Ireland and other European countries, ordered by overall interval cancer ratio.

				Initial screens			Subsequent screens			Overall
Area/country	Dates of screening	Age-group	No of tests	Screen detected	Interval cancers	Interval cancer ratio	Screen detected	Interval cancers	Interval cancer ratio	Screen detected	Interval cancers	Interval cancer ratio
Torino (Italy)⁹	1992–1996	50–59	41911	248	40	0.14	82	25	0.23	330	65	0.16
Firenze (Italy)⁹	1990–1994	50–69	49148	325	47	0.13	54	28	0.34	379	75	0.17
Navarra (Spain)⁹	1990–1996	45–65	126318	256	29	0.10	268	77	0.22	524	106	0.17
Isère (France)¹⁸	2002–2004	50–69	31377	91	25	0.22	130	23	0.15	221	48	0.18
Four counties (Norway)⁹	1996–1997	50–69	126779	†	†	†	†	†	†	852	247	0.22
Netherlands¹⁹	1990–1993	50–69	756283	3635	1002	0.22	701	377	0.35	4336	1379	0.24
Ireland	2000–2007	50–64	372658	1191	349	0.23	805	368	0.31	1996	717	0.26
Norrbotten (Sweden)¹⁹	1989–2002	40–74	261190	†	†	†	†	†	†	768	279	0.27
Marseille (France)⁹	1993–1998	50–69	140086	483	179	0.27	171	65	0.28	654	244	0.27
Isère (France)¹⁸	1994–1999	50–69	66232	167	72	0.30	184	86	0.32	351	158	0.31
Pirkanmaa (Finland)⁹	1988–1999	50–59	64803	†	†	†	†	†	†	235	113	0.32
Strasbourg (France⁹	1989–1997	50–65	168186	328	129	0.28	390	230	0.37	718	359	0.33
England/Wales/ Northern Ireland⁸	1997–2003	50–64	7320648	†	†	†	†	†	†	34369	21281	0.38*
Ireland (pilot programme)¹⁰	1989–1994	50–64	35159	133	45	0.34	58	32	0.55	191	77	0.40

this is based on a three-year screening interval

†

the figure was not reported in the relevant paper

Discussion

A strength of this study is the high level of completeness of the National Cancer Registry data in Ireland, in which information on cancer diagnoses for women who reside in one part of the country but are treated for cancer in another part (common in rural populations) is captured. With 372,658 women participating in the programme in the period 2000–2007, there were sufficient numbers to compare incident and subsequent screens in addition to the number of early (0–11 months) and late (12–23 months) interval cancers. The main limitation is the difficulty in calculating the background incidence, which makes inter-country comparison problematic.

BreastCheck sets a minimum standard of no more than 7.5 interval cancers per 10,000 women screened in the first year, and 12.5 per 10,000 in the second year, following a screening episode.²³ The “achievable” standard was set as 5.0 per 10,000 women screened in the first year, and 7.5 per 10,000 in the second year. For 2000–2007, the programme comfortably met the minimum standard. The rate was slightly above the “achievable” standard in the first year after screening, however, in the second year after screening episode, the rate was 13.4 per 10,000 - above the minimum standard of 12.5.

The interval cancer rate in Ireland is in the mid-range compared with rates published for a number of European screening programmes, although there is little information on interval cancer rates covering recent periods (table 4 and^18–20,24), and some of the published European data precede the routine use of double view mammography, which is standard in the Irish programme.

Interval cancers include those potentially detectable at the time of screening, and true negatives (including occult cancers). The incidence of true interval cancers depends on the background incidence of cancer and on the proportion of rapidly growing tumours in this background incidence. A high background incidence of breast cancer will lead to a higher interval cancer rate. The estimated background incidence in Ireland (based on incidence during 1995–1999) was relatively high, second only to one of the other European programmes listed, and may have been an under-estimate of the true background incidence during 2000–2007 if the upward trend in incidence rates observed prior to 2001²⁴ was maintained. However, the incidence of interval cancers relative to the expected rate (proportionate) was comparable with the other regions and countries described here, as was the interval cancer ratio. The calculation of background incidence outside a trial situation is difficult. In Ireland, due to the pilot programme,¹⁰ the entire population was not screening naïve and thus background rates may have been slightly lower than presented. However, the upwards trend in incidence rates in the period 1994–1997 was not carried forward, so these may be underestimated. The interval cancer ratio is a useful alternative measure for regions that have mature screening programmes, unstable incidence rates in the period prior to the programme, or where there is no population based cancer registry to supply such information.

Detection through mammography carried out privately, between BreastCheck rounds, may have inflated the number of cancers classified as intervals. We have indirect evidence of private mammography via the number of in-situ interval cancers found in this group.

Interval rate estimation depends on complete case ascertainment and successful case linkage to the screening register. The published registry data cited here was from areas covered by population-based cancer registries, linked to screening registers. Incomplete linkage of registered cancers to the screening register will result in under-estimation of interval cancer rates, while incomplete linkage to screen-detected cancers will lead to overestimation. In Ireland, almost all screen-detected breast cancers reported to the Registry by BreastCheck were successfully linked to registrations, using probabilistic matching software and identifiers including name, address, date of birth, and hospital medical record number. Only two cancers detected by BreastCheck could not be linked to a National Cancer Registry registration.

In breast cancer screening programmes choices must be made between high sensitivity, which will reduce the number of interval cancers but increase false positive rates, or high specificity, which will reduce recall rates at the expense of increasing the number of interval cancers. While avoiding interval cancers must be a priority, false positive results cause distress to the women involved and may reduce future participation rates;²⁵ for a programme to be effective, participants need to be retained and attend for multiple screens. Digital mammography, which may have better performance characteristics than conventional mammography,^26,27 was introduced by the Irish programme in 2007, and this could reduce interval cancer rates.

Conclusion

The interval cancer rates over the first eight years of BreastCheck were centred in the range reported from other population-based screening programmes in Europe. Although the programme achieved the minimum standard it had set itself in the first 12 months after screening, it did not attain the “achievable” standard, especially in the second year after screening. It is important for all population-based screening programmes to undergo periodic external evaluation, including estimation of interval cancer rates. The challenges inherent in maintaining quality during nationwide programme roll-out mean that assessing interval cancer rates will remain important for the Irish programme in coming years.

Footnotes

Acknowledgements

We thank staff at BreastCheck for providing data and for their assistance during this project, and the Tumour Registration Officers of the National Cancer Registry.

The Authors have no conflicting interests to declare.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Schopper

de Wolf

. How effective are breast cancer screening programmes by mammography? Review of the current evidence. Eur J Cancer 2009; 45(11): 1916–23.

Moller

Weedon-Fekjaer

. The influence of mammographic screening on national trends in breast cancer incidence. Eur J Cancer Prev 2005; 14(2): 117–28.

Moss

. The impact of mammographic screening on breast cancer mortality in Europe: a review of trend studies. J Med Screen 2012; 19(Suppl 1): 26–32.

Jørgensen

Gøtzsche

. Who evaluates public health programmes? A review of the NHS Breast Screening Programme. J R Soc Med 2010 Jan; 103(1): 14–20.

Gotzsche

. Relation between breast cancer mortality and screening effectiveness: systematic review of the mammography trials. Dan Med Bull 2011 Mar; 58(3): A4246–A4246.

Independent UK Panel on Breast Cancer Screening. The benefits and harms of breast cancer screening: an independent review. Lancet 2012; 380(9855): 1778–1786.

Perry

Broeders

. European guidelines for quality assurance in breast cancer screening and diagnosis. Fourth edition–summary document. Ann Oncol 2008; 19(4): 614–22.

Bennett

Sellars

Moss

. Interval cancers in the NHS breast cancer screening programme in England, Wales and Northern Ireland. Br J Cancer 2011; 104(4): 571–7.

Törnberg

. A pooled analysis of interval cancer rates in six European countries. Eur J Cancer Prev 2010 Mar; 19(2): 87–93.

10.

Törnberg

Codd

Rodrigues

Segnan

Ponti

. Ascertainment and evaluation of interval cancers in population-based mammography screening programmes: a collaborative study in four European centres. Journal of medical screening 2005; 12(1): 43–49.

11.

BreastCheck Programme Report 2011/12. 2013.

12.

McCullagh

Baldelli

Phelan

. Clinical dose performance of full field digital mammography in a breast screening programme. Br J Radiol 2011 Nov; 84(1007): 1027–33.

13.

Keavey

. Comparison of the clinical performance of three digital mammography systems in a breast cancer screening programme. Br J Radiol 2012 Aug; 85(1016): 1123–7.

14.

O'Brien

Comber

Sharp

. Completeness of case ascertainment at the Irish National Cancer Registry. Ir J Med Sci 2014 Jun; 183(2): 219–24.

15.

Threlfall

Collins

Woodman

. Impact of NHS breast screening on advanced disease and mortality from breast cancer in the North West of England. Br J Cancer 2003 Jul 7; 89(1): 77–80.

16.

Andersen

Törnberg

Lynge

Njor

. A simple way to measure the burden of interval cancers in breast cancer screening. BMC cancer 2014; 14(1): 782–782.

17.

Kim

Fay

Feuer

Midthune

. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000; 19(3): 335–51.

18.

Seigneurin

. Comparison of interval breast cancer rates for two-versus single-view screening mammography: a population-based study. Breast 2009; 18(5): 284–8.

19.

Bordás

Jonsson

Nyström

Lenner

. Interval cancer incidence and episode sensitivity in the Norrbotten Mammography Screening Programme, Sweden. Journal of medical screening 2009; 16(1): 39–45.

20.

Fracheboud

De Koning

. Interval cancers in the Dutch breast cancer screening programme. British journal of cancer 1999; 81(5): 912–912.

21.

Junod

Zahl

Kaplan

Olsen

Greenland

. An investigation of the apparent breast cancer epidemic in France: screening and incidence trends in birth cohorts. BMC Cancer 2011 Sep 21; 11: 401–401.

22.

Fielder

Rogers

Gower-Thomas

Monypenny

Dallimore

Brook

Greening

. Results from 10 years of breast screening in Wales. J Med Screen 2001; 8: 21–2.

23.

National Cancer Screening Service Board. Guidelines for Quality Assurance in Mammography Screening, Third ed. Dublin: The National Cancer Screening Service Board, 2008.

24.

National Cancer Registry. Breast cancer incidence, mortality, treatment and survival in Ireland: 1994–2009, Cork, Ireland: National Cancer Registry, 2012.

25.

McCann

Stockton

Godward

. Impact of false-positive mammography on subsequent screening attendance and risk of cancer. Breast Cancer Res 2002; 4(5): R11–R11.

26.

Vinnicombe

. Full-field digital versus screen-film mammography: comparison within the UK breast screening program and systematic review of published data. Radiology 2009; 251(2): 347–58.

27.

Skaane

. Studies comparing screen-film mammography and full-field digital mammography in breast cancer screening: updated review. Acta Radiol 2009; 50(1): 3–14.

Interval cancer rates in the Irish national breast screening programme

Abstract

Objective

Methods

Results

Conclusion

Keywords

Introduction

Methods

Results

Discussion

Conclusion

Footnotes

Acknowledgements

References