Mammography screening in the Netherlands: delay in the diagnosis of breast cancer after breast cancer screening

Impact of clinical breast examination-based screening program on care pathway, stage at diagnosis, nature of treatment, and overall survival among breast cancer patients in Morocco

BACKGROUND

This study aimed to indirectly examine whether the implementation of clinical breast examination-based screening program in Morocco has been successful in downstaging and improving survival rates. Breast cancer patients detected through the screening pathway were compared with those detected through self-referral over the same period in terms of cancer stage at diagnosis, tumor characteristics, care delays, and survival.

METHODS

A prospective observational study was conducted between April 2019 and August 2020 at two major public oncology centers.

RESULTS

A total of 896 women with confirmed breast cancer were recruited (483 were program-referred and 413 were self-referred). The authors did not report any significant difference between the two groups in terms of stage at diagnosis, molecular profile, or histopathological grade. Early-stage cancer (stage I-II) was detected in 55.7% of self-referred participants compared to 55.5% of program-referred participants. Median intervals between symptom recognition, pathological diagnosis, and treatment initiation were not significantly different between the two groups. Similarly, survival after treatment showed no significant difference between patients screened by the program and self-referred patients. The 3-year survival rate after treatment was 94.5% for patients referred through the program and 88.6% for patients not referred through the program (p = .16).

CONCLUSIONS

This study highlights the importance of equitable and timely access to high-quality diagnosis and treatment facilities, leading to substantial downstaging and enhanced survival rates. Continued efforts to improve quality and expand coverage to include asymptomatic women will consolidate the health infrastructure gains achieved by the Moroccan breast cancer screening program.

Causes and consequences of delayed diagnosis in breast cancer screening with a focus on mammographic features and tumour characteristics

PURPOSE

To study the prevalence, causes and consequences of delayed breast cancer diagnosis in the screening population.

METHODS

This retrospective study was performed in women who underwent biennial screening mammography between January 1, 2009 and June 30, 2019. Patients were divided into 3 groups; screen-detectedbreast cancer (SDC) without a diagnostic delay, a primary diagnostic delay(i.e. missed cancer at previous screening round)and a delay in diagnostic work-up after recall. Women with a true interval cancer (IC; i.e. not visible on prior examinations) were excluded. Outcome parameters included mammographic and tumour characteristics, lymph node status and surgical treatment.

RESULTS

In our sample of 4491 women with breast cancer (4292 SDC and 199 'missed' IC), respectively, a total of 1112 women experienced a diagnostic delay of ≥ 4 months. Compared to women without a diagnostic delay (n = 2720), the 176 women with a delay in diagnostic work-up showed overall similar mammographic abnormalities (P = 0.052). These groups show similar distributions in invasive tumours, tumour stage and lymph node status (P = 0.25, P = 0.95 and P = 0.93, respectively). Women with a primary diagnostic delay (n = 936) showed less calcifications (P < 0.001), and more masses with calcifications and architectural distortions on mammography (P = 0.01 and P = 0.04, respectively). Moreover, this group comprised larger tumours (P < 0.001) and lymph node metastases (P < 0.001), and more often underwent mastectomy (P < 0.001).

CONCLUSIONS

A primary diagnostic delay in breast cancer diagnosis results in less favourable tumour characteristics and relatively more mastectomies compared to no delay in breast cancer diagnosis and a delay in diagnostic work-up after recall.

Effects of delayed callback from screening mammography due to the COVID-19 pandemic

Objective

To determine the frequency and distinguishing imaging characteristics of breast cancers detected on screening mammography which was initially evaluated as a probably benign lesion and the workup was delayed due to the COVID-19 pandemic.

Materials and methods

REB-approved multicenter retrospective screening mammography studies and patient's chart review carried out between February 2020 and March 2020. According to an institutional decision, the frequency and imaging findings deemed probably benign on screening mammography after review by a breast fellowship-trained radiologist with workup deferred until after the first pandemic wave plateau in late July 2020 were recorded. Results were correlated with histopathology if tissue sample performed or an uneventful 2 years follow-up. Descriptive statistical analysis was used to describe the retrieved data set.

Results

Out of 1816 mammography screening between February 2020 and March 2020, 99 women, median age 58 years (range 35–84), 99 mammography had possibly benign findings with workup delayed, and two patients, age 49 and 56, had cancers (2.02%), misinterpreted as benign findings. Both malignant cases were focal asymmetries, with pathology of invasive ductal carcinoma, 12 mm and 9 mm in size. No in-situ carcinoma was detected.

Conclusion

The low rate of cancer detected suggests that a delay callback may be a reasonable option for some likely benign findings when immediate callback is not an option, such as during a pandemic. Larger studies would be helpful to support our findings and may allow us to translate the adoption of such a model during potential future pandemic.

Clinical relevance

The results of this study may be helpful for a future situation when delaying a call back from screening mammography is again required.

Delayed breast cancer diagnosis after repeated recall at biennial screening mammography: an observational follow-up study from the Netherlands

Background

Delay in detection of breast cancer may worsen tumour characteristics, with progression of tumour size and a higher risk of metastatic lymph nodes. The purpose of this study was to investigate delayed breast cancer diagnosis after repeated recall for the same mammographic abnormality at screening.

Methods

This was a retrospective study performed in two cohorts of women enrolled in a mammography screening programme in the Netherlands. All women aged 50−75 who underwent biennial screening mammography either between January 1, 1997 and December 31, 2006 (cohort 1) or between January 1, 2007 and December 31, 2016 (cohort 2) were included.

Results

The cohorts showed no difference in proportions of women with delayed breast cancer diagnosis of at least 2 years (2.2% versus 2.8%, P = 0.29). Most delays were caused by incorrect BI-RADS classifications after recall (74.2%). An increase in mean tumour size was seen when comparing sizes at initial false-negative recall and at diagnosis of breast cancer (P < 0.001).

Conclusions

The proportion of women with a long delay in breast cancer confirmation following repeated recall at screening mammography has not decreased during 20 years of screening. These delays lead to larger tumour size at detection and may negatively influence prognosis.

Frequency and characteristics of additionally detected ipsilateral breast lesions following recall at screening mammography

PURPOSE

To determine the frequency and outcome of additionally detected ipsilateral breast abnormalities following recall at screening mammography.

METHODS AND MATERIALS

We included a consecutive series of 130,338 screening mammograms obtained between January 1, 2014 and January 1, 2016. During 2-year follow-up, clinical data were collected of all recalls. Women with a bilateral recall (115) and women recalled for multiple lesions in one breast (165) were excluded from the analyses. Screening outcome parameters were determined for recalled women with or without evaluation of additional ipsilateral breast abnormalities following recall.

RESULTS

A total of 3995 women were recalled (recall rate, 3.1%). In 258 (6.4%) of these women, another lesion was detected in the ipsilateral breast than the one for which she had been recalled. Biopsy was more frequently performed of additionally detected ipsilateral lesions than of recalled lesions (55.8% (144/258)) versus 39.7% (1375/3457), (p < 0.001)). The proportion of malignancy in recalled lesions and additionally detected lesions was comparable (21.5% (743/3457) versus 19.0% (49/258), p = 0.34). Of all 144 biopsies of additionally detected ipsilateral lesions, 9 revealed a synchronous tumour in addition to a malignant recalled lesion, and 33 biopsies revealed multicentric or multifocal tumours. In 5 women, the recalled lesion turned out to be benign, whereas the additional lesion in a different quadrant was malignant at biopsy. A total of 97 biopsies showed benign findings.

CONCLUSION

A substantial proportion of women are analyzed for additional ipsilateral breast lesions following recall. These lesions are more frequently biopsied than recalled lesions, but have a comparable probability of being malignant. The majority of additionally detected cancerous lesions are part of multifocal or multicentric malignancies.

Breast cancer risk is increased in the years following false-positive breast cancer screening

Supplemental Digital Content is available in the text.

A small number of studies have investigated breast cancer (BC) risk among women with a history of false-positive recall (FPR) in BC screening, but none of them has used time-to-event analysis while at the same time quantifying the effect of false-negative diagnostic assessment (FNDA). FNDA occurs when screening detects BC, but this BC is missed on diagnostic assessment (DA). As a result of FNDA, screenings that detected cancer are incorrectly classified as FPR. Our study linked data recorded in the Flemish BC screening program (women aged 50–69 years) to data from the national cancer registry. We used Cox proportional hazards models on a retrospective cohort of 298 738 women to assess the association between FPR and subsequent BC, while adjusting for potential confounders. The mean follow-up was 6.9 years. Compared with women without recall, women with a history of FPR were at an increased risk of developing BC [hazard ratio=2.10 (95% confidence interval: 1.92–2.31)]. However, 22% of BC after FPR was due to FNDA. The hazard ratio dropped to 1.69 (95% confidence interval: 1.52–1.87) when FNDA was excluded. Women with FPR have a subsequently increased BC risk compared with women without recall. The risk is higher for women who have a FPR BI-RADS 4 or 5 compared with FPR BI-RADS 3. There is room for improvement of diagnostic assessment: 41% of the excess risk is explained by FNDA after baseline screening.

Frequency and characteristics of contralateral breast abnormalities following recall at screening mammography

PURPOSE

To determine the frequency and characteristics of contralateral, non-recalled breast abnormalities following recall at screening mammography.

METHODS

We included a series of 130,338 screening mammograms performed between 1 January 2014 and 1 January 2016. During the 1-year follow-up, clinical data were collected for all recalls. Screening outcome was determined for recalled women with or without evaluation of contralateral breast abnormalities.

RESULTS

Of 3,995 recalls (recall rate 3.1%), 129 women (3.2%) underwent assessment of a contralateral, non-recalled breast abnormality. Most lesions were detected at clinical mammography and/or breast tomosynthesis (101 women, 78.3%). The biopsy rate was similar for recalled lesions and contralateral, non-recalled lesions, but the positive predictive value of biopsy was higher for recalled lesions (p = 0.01). A comparable proportion of the recalled lesions and contralateral, non-recalled lesions were malignant (p = 0.1). The proportion of ductal carcinoma in situ was similar for both groups, as well as invasive cancer characteristics and type of surgical treatment.

CONCLUSIONS

About 3% of recalled women underwent evaluation of contralateral, non-recalled breast lesions. Evaluation of the contralateral breast after recall is important as we found that 15.5% of contralateral, non-recalled lesions were malignant. Contralateral cancers and screen-detected cancers show similar characteristics, stage and surgical treatment.

KEY POINTS

• 3% of recalled women underwent evaluation of contralateral, non-recalled lesions • One out of seven contralateral, non-recalled lesions was malignant • A contralateral cancer was diagnosed in 0.5% of recalls • Screen-detected cancers and non-recalled, contralateral cancers showed similar histological characteristics • Tumour stage and surgical treatment were similar for both groups.

BI-RADS 3: Current and Future Use of Probably Benign

PURPOSE OF REVIEW

Probably benign (BI-RADS 3) causes confusion for interpreting physicians and referring physicians and can induce significant patient anxiety. The best uses and evidence for using this assessment category in mammography, breast ultrasound, and breast MRI will be reviewed; the reader will have a better understanding of how and when to use BI-RADS 3.

RECENT FINDINGS

Interobserver variability in the use of BI-RADS 3 has been documented. The 5th edition of the BI-RADS atlas details the appropriate use of BI-RADS 3 for diagnostic mammography, ultrasound, and MRI, and discourages its use in screening mammography. Data mining, elastography, and diffusion weighted MRI have been evaluated to maximize the accuracy of BI-RADS 3.

SUMMARY

BI-RADS 3 is an evolving assessment category. When used properly, it reduces the number of benign biopsies while allowing the breast imager to maintain a high sensitivity for the detection of early stage breast cancer.

Trends in incidence and tumour grade in screen-detected ductal carcinoma in situ and invasive breast cancer

PURPOSE

In a biennial screening mammography programme, we analysed the trends in incidence of screen-detected DCIS and invasive breast cancers in the era of screen-film mammography (SFM) screening, the period of the transition to full-field digital mammography (FFDM) screening and the period of FFDM screening. We also investigated a possible association between the incidence and grading of screen-detected DCIS and invasive breast cancer.

METHODS

In the southern part of the Netherlands, FFDM screening gradually replaced SFM screening between May 2009 and April 2010. We included a consecutive series of 484, 422 screens obtained between July 2005 and July 2015 and divided these screens into three groups; SFM-only cohort, transition cohort and FFDM-only cohort.

RESULTS

A total of 3059 referred women were diagnosed with DCIS (n = 623) or invasive breast cancer (n = 2436). The majority of DCIS were high-grade (48.2%), whereas the majority of the invasive breast cancers were low-grade (45.4%) or intermediate-grade (41.6%). The cancer detection rate (CDR) per 1000 screened women showed the same distribution by grade in both groups. The transition to FFDM was characterised by an increased overall detection rate of invasive cancers.

CONCLUSIONS

Screening mammography detects mostly high-grade DCIS and low- or intermediate-grade invasive cancers. The grade distribution as well as the CDR in the years after the introduction of FFDM remained stable compared to the era of SFM screening. By diagnosing and treating high-grade DCIS, which otherwise may develop into high-grade invasive carcinoma, our findings provide new evidence for the beneficial value of screening mammography programmes.

A clinical model for identifying the short-term risk of breast cancer

BACKGROUND

Most mammography screening programs are not individualized. To efficiently screen for breast cancer, the individual risk of the disease should be determined. We describe a model that could be used at most mammography screening units without adding substantial cost.

METHODS

The study was based on the Karma cohort, which included 70,877 participants. Mammograms were collected up to 3 years following the baseline mammogram. A prediction protocol was developed using mammographic density, computer-aided detection of microcalcifications and masses, use of hormone replacement therapy (HRT), family history of breast cancer, menopausal status, age, and body mass index. Relative risks were calculated using conditional logistic regression. Absolute risks were calculated using the iCARE protocol.

RESULTS

Comparing women at highest and lowest mammographic density yielded a fivefold higher risk of breast cancer for women at highest density. When adding microcalcifications and masses to the model, high-risk women had a nearly ninefold higher risk of breast cancer than those at lowest risk. In the full model, taking HRT use, family history of breast cancer, and menopausal status into consideration, the AUC reached 0.71.

CONCLUSIONS

Measures of mammographic features and information on HRT use, family history of breast cancer, and menopausal status enabled early identification of women within the mammography screening program at such a high risk of breast cancer that additional examinations are warranted. In contrast, women at low risk could probably be screened less intensively.

Trends in surgery for screen-detected and interval breast cancers in a national screening programme

Background

This population-based study aimed to evaluate trends in surgical approach for screen-detected cancer versus interval breast cancer, and to determine the factors associated with positive resection margins.

Methods

Screening mammograms of women aged 50–75 years, who underwent biennial screening in a Dutch breast-screening region between 1997 and 2011, were included. Patient and tumour characteristics were compared between women who underwent mastectomy or breast-conserving surgery (BCS) for screen-detected or interval cancer, and women with a negative or positive resection margin after BCS.

Results

Some 417 013 consecutive screening mammograms were included. A total of 2224 screen-detected and 825 interval cancers were diagnosed. The BCS rate remained stable (mean 6·1 per 1000 screened women; P = 0·099), whereas mastectomy rates increased significantly during the study from 0·9 (1997–1998) to 1·9 (2009–2010) per 1000 screened women (P &lt; 0·001). The proportion of positive resection margins for invasive cancer was 19·6 and 7·6 per cent in 1997–1998 and 2009–2010 respectively (P &lt; 0·001), with significant variation between hospitals. Dense breasts, preoperative magnetic resonance imaging, microcalcifications, architectural distortion, tumour size over 20 mm, axillary lymph node metastasis and treating hospital were independent risk factors for mastectomy. Interval cancer, image-guided tumour localization, microcalcifications, breast parenchyma asymmetry, tumour size greater than 20 mm, lobular tumour histology, low tumour grade, extensive invasive component and treating hospital were independent risk factors for positive resection margins.

Conclusion

Mastectomy rates doubled during a 14-year period of screening mammography and the proportion of positive resection margins decreased, with variation among hospitals. The latter observation stresses the importance of quality control programmes for hospitals treating women with breast cancer.

Variations in screening outcome among pairs of screening radiologists at non-blinded double reading of screening mammograms: a population-based study

OBJECTIVES

Substantial inter-observer variability in screening mammography interpretation has been reported at single reading. However, screening results of pairs of screening radiologists have not yet been published. We determined variations in screening performances among pairs of screening radiologists at non-blinded double reading.

METHODS

We included pairs of screening radiologists with at least 7,500 screening examinations per pair, obtained between 1997 and 2011. During 2-year follow-up, breast imaging reports, surgical reports and pathology results were collected of all referred women and interval cancers. Referral rate, cancer detection rate, positive predictive value and sensitivity were calculated for each pair.

RESULTS

A total of 310,906 screening mammograms, read by 26 pairs of screening radiologists, were included for analysis. The referral rate ranged from 1.0 % (95 % CI 0.8 %-1.2 %) to 1.5 % (95 % CI 1.3 %-1.8 %), the cancer detection rate from 4.0 (95 % CI 2.8-5.2) to 6.3 (95 % CI 4.5-8.0) per 1,000 screens. The programme sensitivity and positive predictive value of referral ranged from 55.1 % (95 % CI 45.1 %-65.1 %) to 81.5 % (95 % CI 73.4 %-89.6 %) and from 28.7 % (95 % CI 20.8 %-36.6 %) to 49.5 % (95 % CI 39.7 %-59.3 %), respectively.

CONCLUSION

We found significant variations in screening outcomes among pairs of screening radiologists at non-blinded double reading. This stresses the importance of monitoring screening results on a local scale.

KEY POINTS

• Substantial inter-observer variability in screening mammography interpretation is known at single reading • Population-based study showed significant variations in outcomes among pairs of screening radiologists • Local monitoring and regular feedback are important to optimise screening outcome.

Impact of the transition from screen-film to digital screening mammography on interval cancer characteristics and treatment - a population based study from the Netherlands

INTRODUCTION

In most breast screening programmes screen-film mammography (SFM) has been replaced by full-field digital mammography (FFDM). We compared interval cancer characteristics at SFM and FFDM screening mammography.

PATIENTS AND METHODS

We included all 297 screen-detected and 104 interval cancers in 60,770 SFM examinations and 427 screen-detected and 124 interval cancers in 63,182 FFDM examinations, in women screened in the period 2008-2010. Breast imaging reports, biopsy results and surgical reports of all cancers were collected. Two radiologists reviewed prior and diagnostic mammograms of all interval cancers. They determined breast density, described mammographic abnormalities and classified interval cancers as missed, showing a minimal sign abnormality or true negative.

RESULTS

The referral rate and cancer detection at SFM were 1.5% and 4.9‰ respectively, compared to 3.0% (p<0.001) and 6.6‰ (p<0.001) at FFDM. Screening sensitivity was 74.1% at SFM (297/401, 95% confidence interval (CI)=69.8-78.4%) and 77.5% at FFDM (427/551, 95% CI=74.0-81.0%). Significantly more interval cancers were true negative at prior FFDM than at prior SFM screening mammography (65.3% (81/124) versus 47.1% (49/104), p=0.02). For interval cancers following SFM or FFDM screening mammography, no significant differences were observed in breast density or mammographic abnormalities at the prior screen, tumour size, lymph node status, receptor status, Nottingham tumour grade or surgical treatment (mastectomy versus breast conserving therapy).

CONCLUSION

FFDM resulted in a significantly higher cancer detection rate, but sensitivity was similar for SFM and FFDM. Interval cancers are more likely to be true negative at prior FFDM than at prior SFM screening mammography, whereas their tumour characteristics and type of surgical treatment are comparable.

Re-attendance after false-positive screening mammography: a population-based study in the Netherlands

Background:

In the current study, mammography adherence of women who had experienced a false-positive referral is evaluated, with emphasis on the probability of receiving surveillance mammography outside the national screening programme.

Methods:

We included 424 703 consecutive screens and collected imaging, biopsy and surgery reports of 3463 women who experienced a false-positive referral. Adherence to screening, both in and outside the screening programme, was evaluated.

Results:

Two years after the false-positive referral, overall screening adherence was 94.6%, with 64.7% of women returning to the national screening programme, compared with 94.9% of women re-attending the screening programme after a negative screen (P<0.0001). Four years after the false-positive screen, the overall adherence had decreased to 85.2% (P<0.0001) with a similar proportion of the women re-attending the screening programme (64.4%) and a lower proportion (20.8%) having clinical surveillance mammography. Women who had experienced a false-positive screen at their first screening round were less likely to adhere to mammography than women with an abnormal finding at one of the following screening rounds (92.4% vs 95.5%, P<0.0001).

Conclusion:

Overall screening adherence after previous false-positive referral was comparable to the re-attendance rate of women with a negative screen at 2-year follow-up. Overall adherence decreased 4 years after previous false-positive referral from 94.6% to 85.2%, with a relatively high estimate of women who continue with clinical surveillance mammography (20.8%). Women with false-positive screens should be made aware of the importance to re-attend future screening rounds, as a way to improve the effectiveness of the screening programme.

Characteristics and screening outcome of women referred twice at screening mammography

OBJECTIVES

To determine the characteristics and screening outcome of women referred twice at screening mammography.

METHODS

We included 424,703 consecutive screening mammograms and collected imaging, biopsy and surgery reports of women with screen-detected breast cancer. Review of screening mammograms was performed to determine whether or not an initial and second referral comprised the same lesion.

RESULTS

The overall positive predictive value of referral for cancer was 38.6% (95% CI 37.3-39.8%). Of 147 (2.6%) women referred twice, 86 had been referred for a different lesion at second referral and 32 of these proved malignant (37.2%, 95% CI 27.0-47.4%). Sixty-one women had been referred twice for the same lesion, of which 22 proved malignant (36.1%, 95% CI 24.1-48.0%). Characteristics of these women were comparable to women with cancer diagnosed after first referral. Compared with women without cancer at second referral for the same lesion, women with cancer more frequently showed suspicious densities at screening mammography (86.4% vs 53.8%, P = 0.02) and work-up at first referral had less frequently included biopsy (22.7% vs 61.5%, P = 0.004).

CONCLUSIONS

Cancer risk in women referred twice for the same lesion is similar to that observed in women referred once, or referred for a second time but for a different lesion.

KEY POINTS

Cancer risk was 36% for lesions referred twice at screening mammography. The cancer risk was similar for lesions referred only once at screening. Densities at first referral were associated with increased cancer risk at second referral. No biopsy at first referral was associated with increased cancer risk at second referral. Patient and tumour characteristics were similar for women with and without diagnostic delay.

Malpractice claims following screening mammography in The Netherlands

Although malpractice lawsuits are frequently related to a delayed breast cancer diagnosis in symptomatic patients, information on claims at European screening mammography programs is lacking. We determined the type and frequency of malpractice claims at a Dutch breast cancer screening region. We included all 85,274 women (351,009 screens) who underwent biennial screening mammography at a southern breast screening region in The Netherlands between 1997 and 2009. Two screening radiologists reviewed the screening mammograms of all screen detected cancers and interval cancers and determined whether the cancer had been missed at the previous screen or at the latest screen, respectively. We analyzed all correspondence between the screening organization, clinicians and screened women, and collected complaints and claims until September 2011. At review, 20.9% (308/1,475) of screen detected cancers and 24.3% (163/670) of interval cancers were considered to be missed at a previous screen. A total of 19 women (of which 2, 6 and 11 women had been screened between 1997 and 2001 (102,439 screens), 2001 and 2005 (114,740 screens) and 2005 and 2009 (133,830 screens), respectively) had contacted the screening organization for additional information about their screen detected cancer or interval cancer, but filed no claim. Three other women directly initiated an insurance claim for financial compensation of their interval cancer without previously having contacted the screening organization. We conclude that screening-related claims were rarely encountered, although many screen detected cancers and interval cancers had been missed at a previous screen. A small but increasing proportion of women sought additional information about their breast cancer from the screening organization.

Trends in incidence and detection of advanced breast cancer at biennial screening mammography in The Netherlands: a population based study

Introduction

The aims of this study were to determine trends in the incidence of advanced breast cancer at screening mammography and the potential of screening to reduce it.

Methods

We included a consecutive series of 351,009 screening mammograms of 85,274 women aged 50-75 years, who underwent biennial screening at a Dutch breast screening region in the period 1997-2008. Two screening radiologists reviewed the screening mammograms of all advanced screen detected and advanced interval cancers and determined whether the advanced cancer (tumor > 20 mm and/or lymph node positive tumor) had been visible at a previous screen. Interval cancers were breast cancers diagnosed in women after a negative screening examination (defined as no recommendation for referral) and before any subsequent screen. Patient and tumor characteristics were compared between women with advanced cancer and women with non-advanced cancer, including ductal carcinoma in situ.

Results

A total of 1,771 screen detected cancers and 669 interval cancers were diagnosed in 2,440 women. Rates of advanced cancer remained stable over the 12-year period; the incidence of advanced screen-detected cancers fluctuated between 1.5 - 1.9 per 1,000 screened women (mean 1.6 per 1,000) and of advanced interval cancers between 0.8 - 1.6 per 1,000 screened women (mean 1.2 per 1,000). Of the 570 advanced screen-detected cancers, 106 (18.6%) were detected at initial screening; 265 (46.5%) cancers detected at subsequent screening had been radiologically occult at the previous screening mammogram, 88 (15.4%) had shown a minimal sign, and 111 (19.5%) had been missed. Corresponding figures for advanced interval cancers were 50.9% (216/424), 24.3% (103/424) and 25.1% (105/424), respectively. At multivariate analysis, women with a ≥ 30 months interval between the latest two screens had an increased risk of screen-detected advanced breast cancer (OR 1.63, 95%CI: 1.07-2.48) and hormone replacement therapy increased the risk of advanced disease among interval cancers (OR 3.04, 95%CI: 1.22-7.53).

Conclusion

We observed no decline in the risk of advanced breast cancer during 12 years of biennial screening mammography. The majority of these cancers could not have been prevented through earlier detection at screening.

A modified artificial immune system based pattern recognition approach--an application to clinical diagnostics

OBJECTIVE

This paper introduces a modified artificial immune system (AIS)-based pattern recognition method to enhance the recognition ability of the existing conventional AIS-based classification approach and demonstrates the superiority of the proposed new AIS-based method via two case studies of breast cancer diagnosis.

METHODS AND MATERIALS

Conventionally, the AIS approach is often coupled with the k nearest neighbor (k-NN) algorithm to form a classification method called AIS-kNN. In this paper we discuss the basic principle and possible problems of this conventional approach, and propose a new approach where AIS is integrated with the radial basis function--partial least square regression (AIS-RBFPLS). Additionally, both the two AIS-based approaches are compared with two classical and powerful machine learning methods, back-propagation neural network (BPNN) and orthogonal radial basis function network (Ortho-RBF network).

RESULTS

The diagnosis results show that: (1) both the AIS-kNN and the AIS-RBFPLS proved to be a good machine leaning method for clinical diagnosis, but the proposed AIS-RBFPLS generated an even lower misclassification ratio, especially in the cases where the conventional AIS-kNN approach generated poor classification results because of possible improper AIS parameters. For example, based upon the AIS memory cells of "replacement threshold=0.3", the average misclassification ratios of two approaches for study 1 are 3.36% (AIS-RBFPLS) and 9.07% (AIS-kNN), and the misclassification ratios for study 2 are 19.18% (AIS-RBFPLS) and 28.36% (AIS-kNN); (2) the proposed AIS-RBFPLS presented its robustness in terms of the AIS-created memory cells, showing a smaller standard deviation of the results from the multiple trials than AIS-kNN. For example, using the result from the first set of AIS memory cells as an example, the standard deviations of the misclassification ratios for study 1 are 0.45% (AIS-RBFPLS) and 8.71% (AIS-kNN) and those for study 2 are 0.49% (AIS-RBFPLS) and 6.61% (AIS-kNN); and (3) the proposed AIS-RBFPLS classification approaches also yielded better diagnosis results than two classical neural network approaches of BPNN and Ortho-RBF network.

CONCLUSION

In summary, this paper proposed a new machine learning method for complex systems by integrating the AIS system with RBFPLS. This new method demonstrates its satisfactory effect on classification accuracy for clinical diagnosis, and also indicates its wide potential applications to other diagnosis and detection problems.

Inter-observer variability in mammography screening and effect of type and number of readers on screening outcome

We prospectively determined the variability in radiologists' interpretation of screening mammograms and assessed the influence of type and number of readers on screening outcome. Twenty-one screening mammography radiographers and eight screening radiologists participated. A total of 106 093 screening mammograms were double-read by two radiographers and, in turn, by two radiologists. Initially, radiologists were blinded to the referral opinion of the radiographers. A woman was referred if she was considered positive at radiologist double-reading with consensus interpretation or referred after radiologist review of positive cases at radiographer double-reading. During 2-year follow-up, clinical data, breast imaging reports, biopsy results and breast surgery reports were collected of all women with a positive screening result from any reader. Single radiologist reading (I) resulted in a mean cancer detection rate of 4.64 per 1000 screens (95% confidence intervals (CI)=4.23–5.05) with individual variations from 3.44 (95% CI=2.30–4.58) to 5.04 (95% CI=3.81–6.27), and a sensitivity of 63.9% (95% CI=60.5–67.3), ranging from 51.5% (95% CI=39.6–63.3) to 75.0% (95% CI=65.3–84.7). Sensitivity at non-blinded, radiologist double-reading (II), radiologist double-reading followed by radiologist review of positive cases at radiographer double-reading (III), triple reading by one radiologist and two radiographers with referral of all positive readings (IV) and quadruple reading by two radiologists and two radiographers with referral of all positive readings (V) were as follows: 68.6% (95% CI=65.3–71.9) (II); 73.2% (95% CI=70.1–76.4) (III); 75.2% (95% CI=72.1–78.2) (IV), and 76.9% (95% CI=73.9–79.9) (V). We conclude that screener performance significantly varied at single-reading. Double-reading increased sensitivity by a relative 7.3%. When there is a shortage of screening radiologists, triple reading by one radiologist and two radiographers may replace radiologist double-reading.

Facilitated "fast track" referral reduces time from abnormal screening mammogram to diagnosis

BACKGROUND

The Screening Mammography Program of British Columbia (SMPBC) implemented voluntary, facilitated referral to diagnostic imaging ("Fast Track") after testing 5 interventions to reduce time from an abnormal screening mammogram to diagnosis. The purpose of this study was to compare time intervals for patients evaluated through the Fast Track process with patients who were not.

METHODS

Data were extracted from the SMPBC database for women with abnormal screens conducted from January 1, 2003 to June 30, 2005 (N = 40,292). After exclusions, 39,607 screens were analyzed. Median and 90th percentile times were calculated from abnormal screen to diagnosis and for three subintervals: abnormal screen to notification, notification to first assessment, and first assessment to diagnosis.

RESULTS

One third of abnormal screens were investigated through Fast Track imaging facilities. Overall, the median time from abnormal screen to diagnosis was 8 days faster for Fast Track compared with non-Fast Track. There was no clinically significant difference in time from abnormal screen to notification. The median time from notification to first assessment was 1.1 weeks (Fast Track) compared with 2.4 weeks (non-Fast Track), a reduction of 9 days or 54% in the interval targeted by the Fast Track strategy. The time interval distribution from first assessment to diagnosis was significantly different only for those having a core biopsy (average 3 days faster for Fast Track).

INTERPRETATION

Facilitated referral to diagnostic imaging reduces average time from notification of abnormal screen to first assessment by more than half. Additional strategies are needed to address diagnostic investigation beyond initial imaging procedures.

Classification of breast masses using selected shape, edge-sharpness, and texture features with linear and kernel-based classifiers

Breast masses due to benign disease and malignant tumors related to breast cancer differ in terms of shape, edge-sharpness, and texture characteristics. In this study, we evaluate a set of 22 features including 5 shape factors, 3 edge-sharpness measures, and 14 texture features computed from 111 regions in mammograms, with 46 regions related to malignant tumors and 65 to benign masses. Feature selection is performed by a genetic algorithm based on several criteria, such as alignment of the kernel with the target function, class separability, and normalized distance. Fisher's linear discriminant analysis, the support vector machine (SVM), and our strict two-surface proximal (S2SP) classifier, as well as their corresponding kernel-based nonlinear versions, are used in the classification task with the selected features. The nonlinear classification performance of kernel Fisher's discriminant analysis, SVM, and S2SP, with the Gaussian kernel, reached 0.95 in terms of the area under the receiver operating characteristics curve. The results indicate that improvement in classification accuracy may be gained by using selected combinations of shape, edge-sharpness, and texture features.

Additional Double Reading of Screening Mammograms by Radiologic Technologists: Impact on Screening Performance Parameters

BACKGROUND

Studies have shown that having mammography technologists review screening mammograms in addition to radiologist review may increase the number of breast cancers that are detected at screening mammography. We prospectively examined the effects on screening performance of adding independent double reading of screening mammograms by technologists to standard double reading by radiologists.

METHODS

Twenty-one screening mammography technologists and eight certified screening radiologists participated in this study. From January 1, 2003, to January 1, 2005, all 61,251 screening mammograms obtained at two mammography screening units in The Netherlands were independently read (although the second reader was not blinded to the first reader's interpretation) by two technologists and, in turn, by two radiologists. Radiologists were blinded to the referral opinion of the technologists. During a 2-year follow-up period, we collected clinical data, breast imaging reports, biopsy results, and breast surgery reports of all women with a positive screening result (i.e., those that required additional imaging) from any reader. The distributions of categorical variables between subgroups were compared using chi-square or Fisher's exact tests. Differences in referral and detection by radiologists and technologists were assessed using McNemar's test. All statistical tests were two-sided.

RESULTS

The radiologists referred 905 women (referral rate = 1.48%, 95% confidence interval [CI] = 1.38% to 1.57%), of whom 323 had breast cancer, corresponding to a positive predictive value of referral (PPV) of 35.7% (95% CI = 32.6% to 38.8%). Review of 446 additional technologist-positive readings led to another 80 referrals, resulting in the detection of 22 additional cancers. These extra referrals increased the initial referral rate from 1.48% to 1.61% (difference = 0.13%; 95% CI = 0.10% to 0.16%) and the cancer detection rate (CDR) from 5.27 to 5.63 cancers per 1000 women screened (difference = 0.36 cancers per 1000 women screened; 95% CI = 0.24 to 0.55). With technologist double reading only, 829 women would have been referred (referral rate = 1.35%, 95% CI = 1.26% to 1.45%); among these women, 286 cancers were diagnosed (PPV = 34.5%, 95% CI = 31.3% to 37.7%; CDR = 4.67 cancers per 1000 women screened, 95% CI = 4.13 to 5.21). Referral of all 1351 radiologist- and/or technologist-positive readings would have led to 362 cancers found at screening. The cancer detection rate for radiologist double reading would have increased from 5.27 to 5.91 cancers per 1000 women screened (relative increase = 12.1%, 95% CI = 8.8% to 16.5%; difference = 0.64 cancers per 1000 women screened, 95% CI = 0.47 to 0.87).

CONCLUSION

A referral strategy that includes all technologist-positive readings, which would have increased the cancer detection rate while maintaining a low referral rate, should be considered.

Classification of breast masses via nonlinear transformation of features based on a kernel matrix

We propose methods to perform a certain nonlinear transformation of features based on a kernel matrix, before the classification step, aiming to improve the discriminating power of the comparatively weak edge-sharpness and texture features of breast masses in mammograms, and seek better incorporation of features representing different radiological characteristics than shape features only. Kernel principal component analysis (KPCA) is applied to improve the discriminating power of each single feature in an expanded feature space and the discriminating capability of different feature combinations in other transformed, more informative, lower-dimensional feature spaces. A kernel partial least squares (KPLS) method is developed to derive score vectors for a shape feature set, and an edge-sharpness and texture feature set, respectively, with minimal covariance between each other, to help in achieving improved diagnosis using multiple radiological characteristics of breast masses. Fisher's linear discriminant analysis (FLDA) is employed to evaluate the classification capability of the transformed features. The methods were tested with a set of 57 regions in mammograms, of which 20 are related to malignant tumors and 37 to benign masses, represented using five shape features, three edge-sharpness features, and 14 texture features. The classification performance of the edge-sharpness and texture features, via KPCA transformation, was significantly improved from 0.75 to 0.85 in terms of the area under the receiver operating characteristics curve (Az). The classification performance of all of the shape, edge-sharpness, and texture features, via KPLS transformation, was improved from 0.95 to 1.0 in Az value.

A decade of breast cancer screening in The Netherlands: trends in the preoperative diagnosis of breast cancer

PURPOSE

In a prospective, multi-institutional follow-up study we describe the trends in the preoperative pathologic confirmation of breast cancer of women who underwent breast cancer screening between 1995 and 2005.

METHODS

We included all women aged 50-75 years who underwent biennial screening mammography in the southern breast cancer screening region of the Netherlands between February 13, 1995 and December 22, 2004. Clinical data, breast imaging reports, biopsy results and breast surgery reports were collected of all women with a positive screening result. Follow-up lasted through the next biennial screening examination and was approximately two years for all referred women.

RESULTS

Of 258,900 mammographic screening examinations, 3,064 (1.2%) were positive screens. The majority of women (92%) were analyzed in four regional hospitals and workup yielded breast cancer in 1,332 women. From 1995 to 2005, the percentage of breast cancer cases that underwent percutaneous biopsy prior to surgery, increased from 42.4 to 100%. The proportion of cancers with a preoperative diagnosis of malignancy by percutaneous biopsy, increased from 27.1% in 1995 to 92.7% in 2004. Preoperative breast cancer confirmation by fine needle aspiration cytology (FNAC) gradually decreased from 91.3% to 14.5%, whereas preoperative confirmation by ultrasound guided core biopsy (USCB) or stereotactic core needle biopsy (SCNB) increased from 8.7% to 69.1% and from 0 to 17.4% respectively.

CONCLUSIONS

A preoperative diagnosis of breast cancer is currently obtained in more than 90% of breast cancer patients. The increase in preoperative breast cancer diagnosis through 1995-2004 is correlated with the introduction of SNCB and increased use of USCB at the expense of FNAC.

Minority report – false negative breast assessment in women recalled for suspicious screening mammography: imaging and pathological features, and associated delay in diagnosis

AIM

We studied imaging, pathology and diagnostic aspects of false negative assessment (FNA) in women recalled for suspicious screening mammography.

METHOD

Subjects were women aged 50-69 years undergoing biennial screening mammography within the Florence District screening programme from January 1992-December 2001 (339,953 consecutive screens). We identified all cancers occurring in women recalled to assessment and ascertained, and reviewed, all cases considered as negative on assessment and subsequently diagnosed with breast cancer. We compared imaging features, tumour histology and stage, and diagnostic testing on assessment for all women with cancer, and presentation and length of delay in women falsely negative on assessment.

RESULTS

Eleven thousand six hundred and twenty four women were recalled to diagnostic assessment (recall rate = 3.4%) predominantly for suspicious mammography (9,216 positive screens). Breast cancer was missed in 57 cases: a FNA rate of 0.50% (0.37-0.62%) and comprising 4.1% (3.0-5.1%) of cancers occurring in women recalled after a positive screen. Two types of abnormalities were significantly more frequent in FNA cases than cancers detected at assessment: mass with regular borders (21.1 vs. 5.6%, p = 10(-5)), and asymmetrical density (22.8 vs. 5.4%, p = 10(-5)). On review 56% of FNAs were benign or probably benign BI-RADS categories. FNA occurred in 1.4% of early recalls and in 0.4% of initial assessment (p=0.0001). Significantly fewer tests were performed when assessing missed cancers than detected cancers with the most significant difference noted for FNAC (29.8 vs. 96.0%, p=10(-6)); mammography as the only evaluation on assessment was more frequent in missed cancers (31.5% vs 0.2%, p = 10(-6)). The 57 missed cases were subsequently diagnosed at early recall (2 cases), next biennial screen (11 cases), or as interval breast cancers (44 cases) with a mean delay in diagnosis of 628 days. Tumour histology, size and nodal status did not significantly differ between cancers missed and cancers diagnosed on assessment.

CONCLUSION

False negatives on assessment represent a minority group in whom screening has failed. They might be reduced by adopting a more intensive diagnostic approach to assessment. Although there was no evidence of a worse prognosis in cancers missed at assessment, the delay in diagnosis is substantial and may impact long-term outcomes.

Effect of screening result on waiting times to assessment and breast cancer diagnosis: results from the Ontario Breast Screening Program

BACKGROUND

The effect of severity of screening result on delays to diagnosis has primarily been examined for mammographic abnormalities. This study will examine delays to assessment and diagnosis for women with an abnormal mammogram compared to women with an abnormal clinical breast examination (CBE) or abnormal CBE and mammogram.

METHODS

Using data routinely collected by Ontario Breast Screening Program (OBSP), 12,675 women aged 50 to 69 with an abnormal screening result between January 1, 2000 and December 31, 2000 were followed prospectively to the completion of their assessment process. Median waiting times from abnormal screen to first assessment procedure and diagnosis were compared by modality of referral and among women with a breast cancer diagnosis by prognostic features.

RESULTS

The median waiting time to first assessment and to diagnosis was significantly longer for women with only a clinical abnormality compared to women with a mammographic abnormality. In addition, women diagnosed with cancers of larger size had longer delays when the abnormality was detected only clinically. However, women referred by both modalities had significantly shorter waiting times to first assessment procedure and to diagnosis of poor prognosis cancers compared to women referred by mammography alone.

INTERPRETATION

Women with an abnormal CBE and mammogram are assessed more promptly and have shorter diagnostic times. However, women with only a CBE abnormality had delays to diagnosis as a result of longer waiting times to first assessment procedure. Integration of the OBSP with assessment centres should improve times to diagnosis irrespective of modality of referral.

Methodological issues in international comparison of interval breast cancers

International comparisons of interval cancers (IC) are important to better understand the relationship between programmes' performance and screening practices. In this respect, differences in (i) definition, (ii) identification and (iii) quantification of IC have received little attention. To examine these 3 comparability issues and activities involving IC, an assessment was conducted among member countries of the International Breast Cancer Screening Network, and the impact of accuracy of identification and quantification practices was estimated using 1996-98 data from the Dutch breast cancer screening programme. Information was obtained from 19 screening programmes in 18 countries, 16 of which acknowledged the coexistence of opportunistic screening. IC data were collected to evaluate performance of the screening programme (100% of programmes) and the radiologists (89%); 53% of programmes had a designated review process for IC. Most programmes (84%) agreed with the European Guidelines definition of IC, but a case situation exercise evidenced substantial discrepancy in classification of cancers that occurred after a positive screen. Completeness of identification of IC appears to contribute most to international variation, and cannot be easily controlled for in methodologically rigorous comparisons. Statistically significant differences of about 4% were measured between quantification methods for IC. An operational definition of IC is proposed to enhance international comparability. Valid comparisons of IC are possible with careful attention to the definition but true differences in IC frequency across screening programmes should exceed 10% to be possibly indicative of real differences between programmes.

Episode sensitivity in association with process indicators in the Finnish breast cancer screening program

We assessed the episode sensitivity of the Finnish mammography program for women aged 50-59 (partly 60-64) years and explored associations between the episode sensitivity and the screening process indicators from 1991 to 1999. For the study period, data were available from 10 screening centers. Records of 721,000 screening visits were linked to the files of the nationwide Finnish Cancer Registry from 1991 to 2001 at an individual level. The rates of screen-detected and interval breast cancers were calculated at the first and at the subsequent screens. The episode sensitivity was determined by contrasting the incidence of interval cancers with the expected population incidence rate without screening (incidence method) and as a proportion of interval cancers out of all cancers detected (detection method). At the subsequent screens, the episode sensitivity determined by the incidence method was 54% and by the detection method 65%. The sensitivity 0-11 and 12-23 months after the screening was 70% and 38%, respectively. The episode sensitivity decreased toward the end of the study period and increased with age. The center-specific sensitivity increased 13% per 1% absolute increase in the recall rate. In general, our study provides further information on the effectiveness of screening programs. The sensitivity estimates were comparable with those from other European service screening programs. The variability in the episode sensitivity suggests potential for variations in the future screening outcome.