Population-based service mammography screening: the Icelandic experience

Figures & data

Table 1 Mammography screening in Iceland

Screening outcomes per average 2-year screening round in 1998–2010	Age 40–49		Age 50–69		Age 40–69
	N	%	N	%	N	%
Population	20,725		27,605		48,330
Participants	12,138	58.6	16,836	61.0	28,974	60.0
Recalls	572	4.7	606	3.6	1,178	4.1
Needle biopsies	153	1.3	217	1.3	370	1.3
Surgery (including surgical biopsies)	64	0.5	124	0.7	188	0.6
Invasive cancer	22	0.2	85	0.5	107	0.4
≤10 mm cancer of invasive cancer	7/22	31.8	24/85	28.2	31/107	29.0
≤15 mm cancer of invasive cancer	11/22	50.0	47/85	55.3	58/107	54.2
DCIS	6	<0.1	11	<0.1	17	<0.1
Total cancer (invasive and DCIS)	28	0.2	97	0.6	125	0.4
Referral rate of recalls for needle biopsy	153/572	26.7	217/606	35.8	370/1,178	31.4
Referral rate of recalls for surgery	64/572	11.2	124/606	20.5	188/1,178	16.0
Total cancer-PPV of recalls	28/572	4.9	97/606	16.0	125/1,178	10.6
Total cancer-PPV of surgery	28/64	43.8	97/124	78.2	125/188	66.5
Number of surgeries per total cancer	64/28	2.3*	124/97	1.3*	188/125	1.5*
Benign to malignant biopsy ratio	36/28	1.3*	27/97	0.3*	63/125	0.5*

Notes: Performance parameters in 1998–2010. Average numbers (N) of women to be screened per 2-year screening round, participants, recalls, referrals for needle biopsy and surgery (including surgical biopsies), cancer and DCIS detected. Average referral rates and positive predictive values.

* Proportions.

Abbreviations: DCIS, ductal carcinoma in situ; PPV, positive predictive value.

Figure 1 Invasive breast cancer in Iceland.

Notes: All ages 1969–2010, age-world-standardized incidence, and mortality rates per 100,000 per year, with 5-year moving average rates. Linear regression lines for incidence in 1969–1987 and 1992–2010 and for mortality in 1969–1995 and from that point to 2010.

Figure 2 Invasive breast cancer in Iceland. (A) Age-specific incidence and mortality rates per 100,000 at age 40–49, with 5-year moving average rates. Linear regression lines for the incidence trends between 1969–1987 and 1992–2010 and for the mortality trends between 1969–1995 and from that point to 2010 in the respective age groups. (B) Age-specific incidence and mortality rates per 100,000 at age 50–69 with 5-year moving average rates. Linear regression lines for the incidence trends between 1969–1987 and 1992–2010 and for the mortality trends between 1969–1995 and from that point to 2010 in the respective age groups. (C) Age-specific incidence and mortality rates per 100,000 at age 70–79 with 5-year moving average rates.

Notes: Linear regression lines for the incidence trends between 1969–1987 and 1992–2010 and for the mortality trends in 1969–1995 and from that point to 2010 in the respective age groups.

Table 2 Distribution of risk factors and disease specific death rates according to mode of cancer detection 1988–2009

	Screening cases (N = 954)	Interval cases (N = 568)	Refuser cases (N = 624)	P-values*
				Sc versus Int	Sc versus Ref	Int versus Ref
Disease specific death (%)	9%	17%	26%	<0.001	<0.001	<0.001
Follow-up time in years (mean)	8.6 (0.0–21.8)	7.2 (0.0–21.1)	6.8 (0.0–21.8)	<0.001	<0.001	0.197
Year of birth (mean)	1942 (1919–1969)	1947 (1919–1966)	1944 (1919–1969)	<0.001	0.019	<0.001
Year of diagnosis (mean)	2000 (1988–2009)	2001 (1988–2009)	1999 (1988–2009)	0.006	0.406	<0.001
Age at diagnosis (mean)	57 (40–69)	54 (40–69)	55 (40–69)	<0.001	<0.001	<0.001
Tumor size (mm) mean	16.4 (1–120)	23.5 (1–100)	28.0 (1–150)	<0.001	<0.001	<0.001
Missing (N)	9	12	30
Axillary metastatis
No (%)	69%	52%	47%	<0.001	<0.001	0.076
Yes (%)	31%	48%	53%
Missing (N)	34	33	64
Tumor grade
Grade 1 (%)	33%	16%	16%	<0.001	<0.001	0.99
Grade 2 (%)	43%	45%	45%
Grade 3 (%)	25%	39%	39%
Missing (N)	167	145	117
Estrogen receptors
ER pos (%)	83%	68%	73%	<0.001	<0.001	0.092
ER neg (%)	17%	32%	27%
Missing (N)	232	119	118
Progesterone receptors
PR pos (%)	75%	62%	65%	<0.001	<0.001	0.451
PR neg (%)	25%	38%	35%
Missing (N)	249	124	125

Notes:

* Chi-square test for proportion; t-test for mean.

Abbreviations: ER, estrogen receptor; Int, interval cases; N, number of cases; neg, negative; pos, positive; PR, progesterone receptor; Ref, refuser cases; Sc, screening cases.

Table 3 The likelihood of dying from breast cancer

Risk factors	Screening		Refusers		Univariate analysis		Multivariate analysis (Cox)*
	N	Deaths	N	Deaths	HR	CI	HR	95% CI
Screening attendance
Refusers			618	164	1.00
Attenders	950	82			0.24	0.18–0.31	0.46	0.31–0.69
Year of birth**	950	82	618	164	0.97	0.96–0.98	0.97	0.94–0.99
Year of diagnosis**	950	82	618	164	0.95	0.93–0.98	0.98	0.94–1.02
Tumor size
<15 mm	507	16	124	10	1.00
≥15 mm	338	65	470	139	4.73	3.10–7.22	4.53	1.92–10.66
Axillary metastasis
Not present	633	30	261	33	1.00
Present	286	47	298	103	3.55	2.62–4.81	3.24	2.10–5.02
Tumor grade
Grade 1 + 2	594	26	311	45	1.00
Grade 3	192	32	195	67	2.83	1.89–4.26	2.11	1.37–3.25
Estrogen receptors
Not present	123	22	135	45	1.00
Present	597	44	369	91	0.56	0.42–0.76	0.85	0.52–1.37
Progesterone receptors
Not present	172	30	176	56	1.00
Present	531	35	321	72	0.57	0.43–0.76	0.80	0.50–1.29

Notes: Women diagnosed with breast cancer at age 40–69 in 1988–2009 after start of mammography screening. Results of Cox regression: HR and 95% CI for known risk factors. N: number of women with information on risk factors in the Cancer Registry. Interval cases were excluded from this analysis.

* Cases with information on all variables (556 screening cases and 365 refuser cases);

** continuous variables.

Abbreviations: CI, confidence interval; HR, hazard ratio; N, number.