Sage Journals: Discover world-class research

Abstract

Objective

To compare interval breast cancer rates (ICR) between a biennial organized screening programme in Norway and annual opportunistic screening in North Carolina (NC) for different conceptualizations of interval cancer.

Setting

Two regions with different screening practices and performance.

Methods

620,145 subsequent screens (1996–2002) performed in women aged 50–69 and 1280 interval cancers were analysed. Various definitions and quantification methods for interval cancers were compared.

Results

ICR for one year follow-up were lower in Norway compared with NC both when the rate was based on all screens (0.54 versus 1.29 per 1000 screens), negative final assessments (0.54 versus 1.29 per 1000 screens), and negative screening assessments (0.53 versus 1.28 per 1000 screens). The rate of ductal carcinoma in situ was significantly lower in Norway than in NC for cases diagnosed in both the first and second year after screening. The distributions of histopathological tumour size and lymph node involvement in invasive cases did not differ between the two regions for interval cancers diagnosed during the first year after screening. In contrast, in the second year after screening, tumour characteristics remained stable in Norway but became prognostically more favorable in NC.

Conclusion

Even when applying a common set of definitions of interval cancer, the ICR was lower in Norway than in NC. Different definitions of interval cancer did not influence the ICR within Norway or NC. Organization of screening and screening performance might be major contributors to the differences in ICR between Norway and NC.

Introduction

Publications from the International Breast Cancer Screening Network have shown that different ways of organizing and conducting screening, including varying definitions and procedures for data collection, create difficulties in international comparisons of screening performance.^1–3 Yankaskas et al.¹ compared recall rates across 13 countries with emphasis on data availability and definitions used for evaluating screening performance. Considerable variability was documented and a call for data standardization was made to promote comparability. Törnberg et al.⁴ compared the interval cancer rate (ICR) in four European countries with organized, population-based screening mammography. The rate ranged from 4.3 to 23.8 per 10,000 screens. Differences in completeness of cancer reporting were suggested as the main reason for the observed variation.⁴ Other studies have aimed to explain variations in ICR,^3–6 and a recent publication by Bulliard et al.³ considered the effects of different identification, definition, and quantification methods, recommending further work to illuminate the relationship between screening practices and programme performance.

The ICR is an important indicator of screening performance.^7,8 An interval breast cancer is defined as breast cancer diagnosed within a defined period following a negative screening examination, most often two years in Europe and one year in the U.S.^7,8 Interval cancers are considered a shortcoming in a screening programme due to prognostically unfavourable tumour characteristics compared with screen-detected cancers^9–11 and consequently less favourable survival from the disease.^9,10,12

Screening mammography programmes can vary in their invitation procedures, recommended screening intervals, number of views performed, reading procedures, and performance. There is also considerable variation in the general health care systems within Europe,^4,13 and also between Europe and the U.S.^1,6,13,14 With these challenges in mind, we investigated the ICR in Norway and North Carolina (NC), USA, two places that have population registration for documentation of cancer following screening mammography. The main objective of the study was to test the effects of various definitions of interval cancer as delineated by Bulliard et al.³ on the ICR between and within the organized screening programme in Norway and the self-referring, opportunistic screening in NC.

Material and Methods

Data were obtained from the Norwegian Breast Cancer Screening Program (NBC SP) and the Carolina Mammography Registry (CMR) for the time period 1996–2002. The NBCSP and CMR both collect prospective data on mammography activity (for all women invited to the screening programme in Norway, and for women seen in facilities that participate in CMR in NC), and the mammography data are linked to population-based breast cancer outcomes.

Women aged 50–69 at a subsequent screening without prior breast cancer were included in the study. Prevalent screens were excluded since performance at prevalent and subsequent screens tends to differ.^4,11,15 Subsequent screens from 151,678 Norwegian and 153,219 NC women were thus the basis for the study. Over the six-year study period, the Norwegian women had an average of 1.8 screens resulting in 269,816 screens, while the NC women had an average of 2.3 screens totalling 350,599 screens. The average age was 59.7 years (SD 5.1) for Norwegian and 58.8 years (SD 5.7) for NC women. Further, 52% (141,182/269,816) of the Norwegian women were aged 50–59 at screening, compared with 59% (207,062/350,599) of the NC women.

Screening in Norway

The NBCSP is a government-sponsored, population-based screening programme administered by the Cancer Registry of Norway. Administration of the programme includes screening operations, data collection, and quality assurance. The NBCSP is mandated in the Cancer Registry regulation,¹⁶ which permits collection of data and subsequent evaluation. All women aged 50–69 are invited to participate in the screening programme by a personal letter stating the time and place for examination. The screening interval is 24 months (±6 months); 97% of women receive invitations at that interval. An unknown, but presumably small, number of women have an additional (‘opportunistic’) screening at private clinics in between two screening examinations in the NBCSP. Data on opportunistic screening are not available in Norway.

The programme performs independent double reading with consensus.¹⁷ A five-point rating scale indicating the probability of a cancer is used to interpret screening mammograms. All cases having a score of two or higher by one or both of the radiologists are discussed at a consensus meeting, where the final decision whether to recall women is made. Further assessment is performed at centralized breast clinics and includes additional imaging, ultrasound, clinical examination and/or invasive procedures, if necessary. Usually, all the procedures related to further assessment are performed in one day. The programme does not recommend short-term follow-up.

The participation rate was 78%, and the coverage 84% during the first ten years of the programme.¹¹ The recall rate among subsequent screens was 2.7% and the screening detection rate 0.51% in the study population (Figure 1).

Figure 1

Numbers of screens, rescreens and interval cancers detected 1–12 and 13–24 months after screening in Norway and North Carolina

Screening in North Carolina

The CMR is based in the state of NC and funded by the National Cancer Institute as a member of the multi-site Breast Cancer Surveillance Consortium (BCSC).¹⁸ CMR collects prospective data from community-based mammography practices in NC. There are standardized quality control procedures for data collection and analysis. CMR is reviewed and approved by the University of North Carolina Institutional Review Board annually, and holds a US Public Health Service Certificate of Confidentiality.¹⁹

Women in NC self-refer for screening or are screened on recommendation from their primary care physician, considered opportunistic screening in contrast to the organized screening programme in Norway. Guidelines recommend that women aged 40 and older be screened every 12–24 months.^20,21 Most physicians recommend a screening interval of 12 months.

NC uses the BI-RADS^® coding system for reporting the assessment of the mammography exam.²² The assessment is coded as 1 = normal, 2 = benign, 3 = probably benign 4 = suspicious abnormality, 5 = suspicious for cancer and 0 = needs further evaluation. Assessments with a code of 0, 4 or 5 are recalled for further diagnostic work-up. The ‘3’ in BI-RADS is associated with a recommendation for short-term rescreen, usually 6 months, and is not considered a recall. Mammograms with a score of 3 were excluded from this study (N = 5,131), as there is no short-term rescreen in Norway. Although single reading is most usual, approximately 20% of mammograms underwent independent double reading.

The facilities that participate in CMR cover approximately 25% of the women being screened in NC. The recall rate in subsequently screened women aged 50–69 was 6.8% and the cancer detection rate 0.37% (Figure 1).

Interval cancers

The general definition of an interval cancer is a breast cancer diagnosed within a defined period (the formal screening interval for the programme mentioned above) following a negative screening examination.^7,8 A screening examination is considered negative if the screening assessment or final assessment is negative. A screening assessment is based on the screening mammograms only, while the final assessment is also based on additional imaging and histology reports. A histological diagnosis of breast cancer is mandatory in the identification process of interval cancers.

Both invasive cancers and ductal carcinomas in situ (DCIS) were included as interval cancers, while lobular in situ cancers were excluded.

Analysis

An interval cancer can only be diagnosed after a negative screening examination. Women were followed to a diagnosis of interval cancer, to the next screening examination, or at least two years after the last mammogram. ICR are presented by time since screening: one and two year follow-up (1–12 months and 13–24 months). We considered the number of emigrations and deaths negligible in the study groups, so that no censuring for these events was performed.

ICR were calculated for all cancers, DCIS alone, and invasive cancers alone using different denominators: a) all screening examinations, b) negative screening assessments with negative final assessments and c) negative screening assessments only (Table 1). Rates are presented without any age adjustment since age adjustment, using the combined Norway and NC study populations as the standard, showed negligible differences. Statistical differences in rates both between and within Norway and NC were explored using a chi-square test. The significance level was set at 5%. The analyses were conducted using SPSS (Version 14.0.1 for Windows, SPSS Inc, Chicago, Illinois), R Statistical Computing (Version 2.0.1), and SAS (Version 8, Cary, NC).

Table 1

Numerators and denominators used for calculations of interval cancer rates in this study

Numerator	Denominator	Explanation
Invasive cancers and DCIS	All screens^*	Screen-positives: TP + FP
Invasive cancers DCIS		Screen-negatives: TN + FN
Invasive cancers and DCIS	Negative final assessment^*	Screen-positives: FP only
Invasive cancers DCIS		Screen-negatives: TN + FN
Invasive cancers and DCIS	Negative screening assessment^*	Screen-negatives only: TN + FN
Invasive cancers
DCIS

Each screening examination is followed one and two years for an interval cancer; DCIS, ductal carcinoma in situ; TP, true-positive; FP, false-positive; TN, true-negative; FN, false-negative

Results

A total of 269,816 subsequent screening examinations in Norway and 350,599 in NC were performed during the study period (Figure 1). A total of 1387 screen-detected (5.14 per 1000 screens) and 486 interval cancers (1.80 per 1000 screens) were diagnosed in Norway, and 1282 screen-detected (3.66 per 1000 screens) and 794 interval cancers (2.26 per 1000 screens) in NC. During the first 12 months after screening, 145 interval cancers were diagnosed in Norway (0.54 per 1000 screened) and 451 in NC (1.29 per 1000 screens). In NC, 239,106 screen examinations were followed by a rescreen within one year (68%, 239,106/349,317; Figure 1). These women were thus not at risk for an interval cancer 13–24 months after screening. The numbers of interval cancers 13–24 months after screening were 341 (1.27 per 1000 screens) for Norway and 343 (3.13 per 1000 screens) for NC. All of these differences were statistically significant (P < 0.05).

The ICR per 1000 screens (DCIS and invasive) 1–12 months after screening were lower in Norway compared with NC for all screens (0.54 versus 1.29), negative final assessment (0.54 versus 1.29), and negative screening assessment (0.53 versus 1.28, Table 2). The rates were 1.27 versus 3.13, 1.27 versus 3.13, and 1.23 versus 3.06 respectively for all screens, negative final assessment and negative screening assessment, 13–24 months after screening. The differences between Norway and NC were most pronounced for DCIS in all categories. The ICR of DCIS for all screens were 0.03 and 0.09 per 1000 screens in Norway, and 0.12 and 0.46 per 1000 screens in NC, for 1–12 and 13–24 months after screening. All of these differences were statistically significant (P < 0.05).

Table 2

Number and rates of interval breast cancers in women aged 50–69, screened in Norway and North Carolina, 1996–2002, stratified by all screens, negative final assessment, negative screening assessment, and follow-up time

	1–12 months after screening				13–24 months after screening				1–24 months after screening
	Norway		NC		Norway		NC		Norway		NC
	n	Rate^*	n	Rate^*	n	Rate^*	n	Rate^*	n	Rate^*	n	Rate^*
All screens (TP + FP + TIM + FN)
Screening examinations (n)	269,816		350,599		268,284^**		109,760^***		269,816		350,599
All interval cancers	145	0.54	451	1.29^†	341	1.27^‡	343	3.13^†^‡	486	1.80	794	2.26^†
DCIS	7	0.03	43	0.12^†	24	0.09^‡	50	0.46^†^‡	31	0.11	93	0.27^†
Invasive	138	0.51	408	1.16^†	317	1.18^‡	293	2.67^†^‡	455	1.69	701	2.00^†
Negative final assessment (FP + TN + FN)
Screening examinations (n)	268,429		349,317		268,284^**		109,760^***		268,429		349,317
All interval cancers	145	0.54	451	1.29^†	341	1.27^‡	343	3.13^†^‡	486	1.81	794	2.27^†
DCIS	7	0.03	43	0.12^†	24	0.09^‡	50	0.46^†^‡	31	0.12	93	0.27^†
Invasive	138	0.51	408	1.17^†	317	1.18^‡	293	2.67^†^‡	455	1.70	701	2.01^†
Negative screening assessment (TN + FN)
Screening examinations (n)	262,617		326,804		262,478^§		108,975^¶		262,617		326,804
All interval cancers	139	0.53	417	1.28^†	324	1.23^‡	334	3.06^†^‡	463	1.76	751	2.30^†
DCIS	5	0.02	37	0.13^†	21	0.08^‡	48	0.44^†^‡	26	0.10	85	0.26^†
Invasive	134	0.51	380	1.16^†	317	1.15^‡	286	2.62^†^‡	437	1.66	666	2.04^†

Rate per 1000 screens;

All screens minus 1387 screen-detected and 145 IC diagnosed 1–12 months after screening;

***

All screens minus 1282 screen-detected and 451 IC diagnosed 1–12 months after screening and 239,106 rescreens;

†

Statistically significant different from Norway (P < 0.05);

‡

Statistically significant different from 1–12 months follow-up (P < 0.05);

All screens with negative screening assessment 1–12 months after screening (n = 326,804), minus rescreens (n = 217,41 2) and interval cancers (n = 417) among screens with a negative screening assessment 1–12 months after screening;

All screens with negative screening assessment 1–12 months after screening (n = 262,617), minus interval cancers (n = 139) diagnosed among screens with negative screening assessment 1–12 months after screening; DCIS, ductal carcinoma in situ; TP, true-positive; FP, false-positive; TN, true-negative; FN, false-negative

ICR within Norway and NC

In both Norway and NC, the rates differed little by denominator chosen or quantification method applied. In Norway, limiting follow-up to 1–12 months after screening resulted in an ICR of 0.54 per 1000 screens, 0.54 per 1000 negative final assessments, and 0.53 per 1000 negative screening assessments (Table 2). Corresponding values for NC were 1.29, 1.29 and 1.28. For all screens, the DCIS rate increased three-fold between the first and second year after screening in Norway and increased four-fold (from 0.12 to 0.46 per 1000 screens) in NC. All differences were statistically significant (P < 0.05).

Histopathological characteristics

The proportion of DCIS among all interval cancers was lower in Norway (6.4%, 31/486) than in NC (11.7%, 93/794, P < 0.001) during the entire study period (Table 3). The percentage of DCIS tended to be higher 13–24 months after screening compared with 1–12 months after screening: 4.8% versus 7.0% (P = 0.478) in Norway and 9.5% versus 14.6% (P = 0.038) in NC (Table 3). The percentage of invasive ductal carcinoma, invasive lobular carcinoma, and other invasive cancers did not differ between Norway and NC for cancers diagnosed 1–12 months after screening (P = 0.411, 0.565, and 0.688, respectively), while there was a higher proportion of both invasive ductal and lobular carcinoma in Norway 13–24 months after screening (73.9% and 14.7%) compared with NC (63.3% and 8.7%, P < 0.001 for both). This may be explained by the proportion of ‘other’ invasive cancers being three-fold greater in NC compared with Norway.

Table 3

Distribution of histopathological types and tumour characteristics in invasive interval cancers diagnosed 1–12 and 13–24 months after screening examination in Norway and North Carolina, 1996–2002

	1–12 months after screening				13–24 months after screening				1–24 months after screening
	Norway		North Carolina		Norway		North Carolina		Norway		North Carolina
	n = 145		n = 451		n = 341		n = 343		n = 486		n = 794
	n	%	n	%	n	%	n	%	n	%	n	%
Histopathological type
Ductal carcinoma in situ	7	4.8	43	9.5	24	7.0	50	14.6^,^*	31	6.4	93	11.7
Invasive ductal carcinoma	99	68.3	289	64.1	252	73.9	217	63.3^*	351	72.2	506	63.7
Invasive lobular carcinoma	21	14.5	55	12.2	50	14.7	30	8.7^*	71	14.6	85	10.7
Other invasive	18	12.4	64	14.2	15	4.4^***	46	13.4^*	33	6.8	110	13.9
	N = 138		N = 408		N = 317		N = 293		N = 455		N = 701
Characteristics invasive cases
Tumour size
Average size (mm)	21		23		22		22		22		22
Median size (mm)	18		19		18		17		18		18
≤ 15 mm	50	40.0	135	37.3	111	39.4	141	52.2^,^*	161	39.6	276	43.7
16–20 mm	29	23.2	70	19.3	57	20.2	40	14.8	86	21.1	110	17.4
>20 mm	46	36.8	157	43.4	114	40.4	89	33.0^**	160	39.9	246	38.9
Missing/no information	13		46		35		23		48		69
Lymph node involvement
Yes	53	43.1	156	38.8	116	41.0	84	29.0^,^*	169	41.6	240	34.7
No	70	56.9	246	61.2	167	59.0	206	71.0^,^*	237	58.4	452	65.3
Missing/no information	15		6		34		3		49		9

P ≤ 0.005 between Norway versus NC for the period 13–24 months after screening examination;

P ≤ 0.005 between interval cancers diagnosed in NC 1–12 versus 13–24 months after screening examination;

***

P ≤ 0.005 between interval cancers diagnosed in Norway 1–12 versus 13–24 months after screening examination

During the entire study period, the largest proportion of tumours were >20 mm, both in Norway and NC (Table 3). The percentages of tumours ≤15 mm, 16–20 mm and >20 mm diagnosed 1–12 months after screening did not differ between Norway and NC, in contrast to the distribution 13–24 months after screening; the percentage of cancers ≤15 mm was stable in Norway (40.0% and 39.4%, P = 0.991) while it increased in NC (from 37.3% to 52.2%, P < 0.001). The proportion of tumours >20 mm remained stable in Norway (36.8% and 40.4%, P = 0.561), while it decreased in NC (from 43.4% to 33.0%, P = 0.010). The percentage of lymph node positive tumours was 43.1% in invasive interval cancers diagnosed 1–12 months after screening examination in Norway, and 38.8% in NC (P = 0.457). The percentage remained stable by time after screening in Norway (from 43.1% to 41.0%, P = 0.776), while it decreased significantly in NC (from 38.8% to 29.0%, P = 0.009).

Discussion

Despite the uniform definition and quantification of interval cancer, the rates were lower in Norway compared with NC. The ICR within the two locations did not differ markedly by varying the denominator – all screens, negative screening assessment, and negative final assessment. A substantially lower rate and proportion of DCIS were found in Norway than in NC. Histopathological tumour characteristics did not differ between Norway and NC in invasive cancers diagnosed 1–12 months after screening examination. Tumour characteristics remained stable during the follow-up time in Norway, but tended to have more prognostically favourable characteristics in NC 13–24 months after screening.

As common definitions and quantification of interval cancer were applied in Norway and NC, the observed differences in the ICR between the two places are likely due to other reasons, such as programme organization, recommended screening intervals, the individual screening history, reading procedures, and screening performance.

Screening organization

In the organized screening programme in Norway, women are invited by a personal letter to a fixed place and time for screening examination biennially. Due to programme logistics, no woman is screened in the programme sooner than in the two-year period. If women choose to do so, they are referred to private clinics, and an eventual cancer diagnosed is defined as an interval cancer. The nationwide screening database in combination with the 11-digit personal identification number and the completeness of the Cancer Registry of Norway ensure accurate surveillance.²³

CMR covers 25% of the screening facilities in NC and the opportunistic screening in NC is more flexible compared with Norway. In contrast to the Norwegian invitation system, screening in NC is defined by the radiologist, who records the visit as a screening or diagnostic examination at the time of the procedure. All women are reminded to return for their next screening exam, although each facility determines its own procedures for reminders. Most NC women were rescreened within 12 months (66%, 239,106/349,317), and 77% (307,285/349,317) re-attended within 18 months. Coverage policies of insurance companies are one important reason for choosing an interval of more than 12 months. Many women re-attend in the 13th month because of insurance policies that will only pay for one mammogram per year. This would be annual screening.

Other possible reasons for a delayed screening were not investigated in this study, but the topic warrants further research. A delayed screening examination has been shown to increase the risk of breast cancer in Norway;²⁴ this finding may also apply to NC, but should be interpreted with care in the absence of NC-specific data.

Screening interval

The different screening intervals make comparison of ICR difficult both with respect to screening history prior to the actual interval cancer and according to follow-up time. This study includes only subsequent screens, and does not take into account the individual screening history in terms of total number of previous screening examinations and their regularity. It appears that little is known about ICR in women having annual screens for more than 10 years (NC) compared with biannually screened women who had their first invitation when they were aged 50 (Norway). Considering women's screening history and its influence on the ICR is an important area for future research. Moreover, an ideal follow-up time cannot be suggested from this study since one year follow-up is artificial for Norway as is two years for NC.

The different screening intervals could be dealt with by using women-years in the denominator.⁶ Using women-years would have given a larger increase in the denominator for Norway compared with NC, but lowered the ICR in both places. Considering a cancer appearing a certain time after a screening examination regardless of the chosen screening interval by following each screen until the diagnosis of a breast cancer (screen-detected or interval), the next screen or one/two years after the last screen is a possible way of addressing the challenge. Analysis stratified by time after screening as a simple proxy for estimating women-years at risk is another analytic approach. There are several methods for calculating ICR, including as a rate of the background incidence.⁴ This method presents some weaknesses as obtaining an accurate estimate of the incidence before screening started can be difficult and the background incidence may have changed independently from screening after initiation of the screening programme. ICR could also be presented as a proportion of the total number of cancers detected within a given time frame.

Reading procedures

Use of independent double reading versus single reading is shown to yield a higher screening detection rate.^6,14 Less is known about its influence on the ICR, though this study documents a favorable effect of independent double reading on the ICR. In addition, recommendations for minimum reading volume differ in Norway and NC: 5000 screens annually for radiologists in Norway and 960 screens biennially for radiologists in NC. Most of the Norwegian radiologists are general radiologists, and only about one-third actually read the annual recommended volume (Hofvind et al., in preparation). Studies show conflicting results regarding the influence of volume on the sensitivity of cancer detection by reading procedure.^25,26

The recall rate and the positive predictive value are two other performance measures that should be taken into account when analysing the ICR. The US is known to have recall rates that are two to three times higher than those observed in Europe.^6,13,14,27 The recall rate in this study was 2.7% for Norway and 6.8% for NC. The higher recall rate and lower cancer detection rate in NC are associated with a higher ICR in our study. All these performance indicators are closely linked and should be interpreted together.

Definition of interval cancer

Bulliard et al. defined interval breast cancers as ‘those tumors that are diagnosed in the time period between two screening rounds’.⁷ In NC and the US, further assessment ends with the imaging assessment and a recommendation for biopsy, and does not include the biopsy results.^8,22 The transformation of Norway's data into the US definition resulted in five fewer interval cancers. These five cases had a benign biopsy following a positive screening assessment and additional imaging. We consider this an insignificant number (1.03%, 5/486).

Another definition issue is whether or not to include DCIS. We chose to analyse all cancers, DCIS, and invasive cancers separately. We obtained different results for DCIS and invasive cancers. Thus, whether DCIS is included or not in the ICR appears to be an important factor and should be clearly explained in all presentations.

Histopathological type and tumour characteristics

The higher rate of DCIS in NC compared with Norway could be due to differences in the pathologists’ perception and interpretation. The use of digital mammography and computer-aided detection (CAD) cannot explain the difference in DCIS rates since digital mammography was infrequently used in Norway and NC during the study period.

A proportion of DCIS that is twice as high in NC as in Norway is striking. However, the same trend was also found in a recent study comparing ICR in Norway and Vermont.⁶ Usually, the DCIS proportion of interval cancers is low because it rarely causes symptoms and does not differ much from that found in breast cancers of non-screened women.^28,29 A high proportion of DCIS among interval cancers could be due to misclassification or could be indicative of women undergoing (‘opportunistic’) screening mammography in the period between programme invitations.

There may be an expectation of increased tumour size and a higher proportion of lymph node positive tumours in interval cancers following longer screening intervals, due to increased delay in diagnosis. These expectations have been disproved recently,^30,31 and this study showed no differences for Norway in tumour size or proportion of lymph node positive tumours 1–12 and 13–24 months after screening. Interestingly, interval cancers diagnosed 13–24 months after screening examination in NC had a smaller tumour size and the percentage of tumours with lymph node involvement was lower than for those diagnosed 1–12 months after screening. It is possible that these mammograms were misclassified as diagnostic, when they were indeed screening, but we have no data at this time to support this. We do not have a ready explanation for the better prognostic characteristics of interval cancers in NC.

Conclusion

The ICR were higher in NC compared with Norway when applying uniform quantification for interval cancer. The rates did not differ statistically within Norway and NC by varying the denominator – all screens, negative screening assessment, and negative final assessment. Using common definitions cannot outweigh differences in the health care systems, organization of mammography screening, and screening performance in the two regions.

Footnotes

Acknowledgements

We thank the International Cancer Screening Network (ICSN) for patience and support during many years of work. Bob Allis has done a great job with the extracting of data from the Carolina Mammography DataBase. We hope this work will stimulate discussions about the topic of IC in breast cancer screening.

The CMR is supported by a grant from the National Cancer Institute, U01-CA70040. The collection of cancer incidence data was in part, supported by the North Carolina Central Cancer Registry which is partially supported by the Centers for Disease Control and Prevention under cooperative agreement NC U58 DP000123–02. The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the Centers for Disease Control and Prevention.

References

Yankaskas

, Klabunde

, Ancelle-Park

. International comparison of performance measures for screening mammography: can it be done? J Med Screen 2004; 11: 187–93.

Klabunde

, Sancho-Garnier

, Taplin

, Thoresen

, Ohuchi

, Ballard-Barbash

. Quality assurance in follow-up and initial treatment for screening mammography programs in 22 countries. Int J Qual Health Care 2002; 14: 449–61.

Bulliard

, Sasieni

, Klabunde

, De Landtsheer

, Yankaskas

, Fracheboud

. Methodological issues in international comparison of interval breast cancer. Int J Cancer 2006; 119: 1158–63.

Törnberg

, Codd

, Rodrigues

, Segnan

, Ponti

. Ascertainment and evaluation of interval cancers in population-based mammography screening programmes: a collaborative study in four European centres. J Med Screen 2005; 12: 43–9.

Klabunde

, Bouchard

, Taplin

, Scharpantgen

, Ballard-Barbash

. Quality assurance for screening mammography: an international comparison. J Epidemiol Community Health 2001; 55: 204–12.

Hofvind

, Vacek

, Skelly

, Weaver

, Geller

. Comparing screening mammography for early breast cancer detection in Vermont and Norway. J Natl Cancer Inst 2008; 100: 1082–91.

Perry

, Broeders

, de Wolf

, Tornberg

, Holland

, von Karsa

. European guidelines for quality assurance in breast cancer screening and diagnosis. Printed in Belgium. European Communities; 2006. http://bookshop.eu.int ISBN: 92-79-01258-4. Assessed November 2008

Rosenberg

, Yankaskas

, Hunt

. Effect of variations in operational definitions on performance estimates for screening mammography. Acad Radiol 2000; 7: 1058–68.

Bordas

, Jonsson

, Nystrom

, Lenner

. Survival from invasive breast cancer among interval cases in the mammography screening programmes of northern Sweden. Breast 2007; 16: 47–54.

10.

Vitak

, Olsen

, Manson

, Arnesson

, Stal

. Tumour characteristics and survival in patients with invasive interval breast cancer classified according to mammographic findings at the latest screening: a comparison of true interval and missed interval cancers. Eur Radiol 1999; 9: 460–9.

11.

Hofvind

, Geller

, Vacek

, Thoresen

, Skaane

. Using the European Guidelines to evaluate the Norwegian Breast Cancer Screening Program. Eur J Epidemiol 2007; 22: 447–55.

12.

Zackrisson

, Janzon

, Manjer

, Andersson

. Improved survival rate for women with interval breast cancer - results from the breast cancer screening programme in Malmo, Sweden 1976-1999. J Med Screen 2007; 14: 138–43.

13.

Broeders

, Scharpantgen

, Ascunce

. Comparison of early performance indicators for screening projects within the European Breast Cancer Network: 1989-2000. Eur J Cancer Prev 2005; 14: 107–16.

14.

Smith-Bindman

, Ballard-Barbash

, Miglioretti

, Patnick

, Kerlikowske

. Comparing the performance of mammography screening in the USA and the UK. J Med Screen 2005; 12: 50–4.

15.

Yankaskas

, Taplin

, Ichikawa

. Association between mammography timing and measures of screening performance in the United States. Radiology 2005; 234: 363–73.

16.

The Ministry of Health, and Social Affairs. Regulations on the collection and processing of personal health data in the Cancer Registry of Norway (Cancer Registry Regulations). Oslo, Norway, the Ministry of Health and Social Affairs, 2001

17.

The Cancer Registry of Norway. Kvalitetsikring i Mammografiprogrammet. A Quality Assurance Manual for the Norwegian Breast Cancer Screening Program (in Norwegian). The Cancer Registry of Norway, Oslo, Norway, 2003

18.

SEER (Surveillance Epidemiology and End Results). http://seer.cancer.gov/. Assessed November 2008

19.

Yankaskas

, Jones

, Aldrich

. The Carolina Mammography Registry: A population-based mammography and cancer surveillance project. Journal Registry Management 1996; 23: 4

20.

Department of Health and Human Services. Quality Mammography Standards. Final Rules-21 CRF Parts 16 and 900 (docket No. 95 N-0192) RIN 0910-AA24 edition. Washington DC, Department of Health and Human Services 1997; 55925–6

21.

U.S. Preventive Task Force.

Screening for breast cancer: recommendations and rationale.

Ann Intern Med2002; 137(5 Part 1): 344–6

22.

American College of Radiology. Illustrated Breast Imaging Reporting and Data System (BI-RADS), 4th edn. American College of Radiology: Reston, VA, 2003

23.

Larsen

, Småstuen

, Johannesen

. Data quality at the Cancer Registry of Norway: an overview of comparability, completeness, validity and timeliness. Eur J Cancer 2009: 1218–31. Epub 2008 Dec 16

24.

Hofvind

. Breast cancer screening-prevalence of disease in women who only respond after an invitation reminder. J Med Screen 2007; 14: 21–2.

25.

Harvey

, Geller

, Oppenheimer

, Pinet

, Riddell

, Garra

. Increase in cancer detection and recall rates with independent double interpretation of screening mammography. AJR Am J Roentgenol 2003; 180: 1461–7.

26.

Taplin

, Rutter

, Elmore

, Seger

, White

, Brenner

. Accuracy of screening mammography using single versus independent double interpretation. AJR Am J Roentgenol 2000; 174: 1257–62.

27.

Botha

, Bray

, Sankila

, Parkin

. Breast cancer incidence and mortality trends in 16 European countries. Eur J Cancer 2003; 39: 1718–29.

28.

Fracheboud

, de Koning

, Beemsterboer

. Interval cancers in the Dutch breast cancer screening programme. Br J Cancer 1999; 81: 912–7.

29.

Hofvind

, Sørum

, Thoresen

. Incidence and tumor characteristics of breast cancer diagnosed before and after implementation of a population-based screening-program. Acta Oncol 2007; 47: 225–31.

30.

Hofvind

, Bjurstam

, Sørum

, Bjorndal

, Thoresen

, Skaane

. Number and characteristics of breast cancer diagnosed in four periods in the screening interval of a biennial population based screening program. J Med Screen 2006; 13: 192–6.

31.

Porter

, Evans

, Burrell

, Lee

, Ellis

, Chakrabarti

. Interval breast cancers: prognostic features and survival by subtype and time since screening. J Med Screen 2006; 13: 115–22.

Comparing Interval Breast Cancer Rates in Norway and North Carolina: Results and Challenges

Abstract

Objective

Setting

Methods

Results

Conclusion

Introduction

Material and Methods

Screening in Norway

Screening in North Carolina

Interval cancers

Analysis

Results

ICR within Norway and NC

Histopathological characteristics

Discussion

Screening organization

Screening interval

Reading procedures

Definition of interval cancer

Histopathological type and tumour characteristics

Conclusion

Footnotes

Acknowledgements

References