An analysis of screen-detected invasive cancers by grade in the English breast cancer screening programme: are we failing to detect sufficient small grade 3 cancers?

OBJECTIVE: Randomised controlled trials have shown a reduction in breast cancer mortality from mammography screening and it is the detection of high-grade invasive cancers that is responsible for much of this effect. We determined the detection rates of invasive cancers by grade, size and type of screen and estimated relative sensitivities with emphasis on grade 3 detection. METHODS: This observational study analysed data from over 11 million screening episodes (67,681 invasive cancers) from the English NHS breast screening programme over seven screening years 2009/2010 to 2015/2016 for women aged 45-70. RESULTS: At prevalent (first) screens (which are unaffected by screening interval), the detection rate of small (< 15 mm) invasive cancers was 0.95 per 1000 for grade 1, but for grade 3 only 0.30 per 1000. The ratio of small (< 15 mm) to large (≥ 15 mm) cancers was 1.8:1 for grade 1 but reversed to 0.5:1 for grade 3. We estimated that the relative sensitivity for grade 3 invasive cancers was 52% of that for grade 1 and the relative sensitivity for small (< 15 mm) grade 3 only 26% of that for small (< 15 mm) grade 1 invasive cancers. CONCLUSIONS: Sensitivity for small grade 3 invasive cancers is poor compared with that for grade 1 and 2 invasive cancers and larger grade 3 malignancies. This observation is likely a limitation of the current technology related to the absence of identifiable mammographic features for small high-grade cancers. Future work should focus on technologies and strategies to improve detection of these clinically most significant cancers. KEY POINTS: • The detection of small high-grade invasive cancers is vital to reduce breast cancer mortality. • We estimate the sensitivity for small grade 3 invasive cancers may be only 26% of that of small grade 1 invasive cancers. This is likely to be associated with the non-specific mammographic features for these cancers. • New technologies and appropriate strategies using current technology are required to maximise the detection of small grade 3 invasive cancers.

Correction to: An analysis of 11.3 million screening tests examining the association between recall and cancer detection rates in the English NHS breast cancer screening programme.

The original version of this article, published on 04 February 2019, unfortunately contained a mistake.

An analysis of 11.3 million screening tests examining the association between needle biopsy rates and cancer detection rates in the English NHS Breast Cancer Screening Programme.

AIM: To examine the association between recall, needle biopsy, and cancer detection rates to inform the setting of target ranges to optimise the benefit to harm ratio of breast screening programmes. MATERIALS AND METHODS: Annual screening programme information from 2009/10 to 2015/16 for the 80 screening units of the English National Health Service Breast Screening Programme (totalling 11.3 million screening tests) was obtained from annual (KC62) returns. Linear regression models were used to examine the association between needle biopsy rates and recall rates and non-linear regression models to examine the association between cancer detection rates and needle biopsy rates. RESULTS: The models show and quantify the diminishing returns for prevalent screens with increasing biopsy rates. A biopsy rate increase from 10 to 20 per 1,000 increases the cancer detection rate by 2.13 per 1,000 with four extra biopsies per extra cancer detected. Increasing the biopsy rate from 40 to 50 per 1,000, increases the cancer detection rate by only 0.25 per 1,000, with 40 extra biopsies per extra cancer detected. Although diminishing returns are also seen at incident screens, screening is generally more efficient. CONCLUSIONS: Increasing needle biopsy rates leads to rapidly diminishing returns in cancer detection and a marked increase in non-malignant/benign needle biopsies. Much of the harms associated with screening in terms of false-positive recall rates and non-cancer biopsies occur at prevalent screens with much lower rates at incident screens. Needle biopsy rate targets should be considered together with recall rate targets to maximise benefit and minimise harm.

An analysis of 11.3 million screening tests examining the association between recall and cancer detection rates in the English NHS breast cancer screening programme.

OBJECTIVE: To develop methods to model the relationship between cancer detection and recall rates to inform professional standards. METHODS: Annual screening programme information for each of the 80 English NHSBSP units (totalling 11.3 million screening tests) for the seven screening years from 1 April 2009 to 31 March 2016 and some Dutch screening programme information were used to produce linear and non-linear models. The non-linear models estimated the modelled maximum values (MMV) for cancers detected at different grades and estimated how rapidly the MMV was reached (the modelled 'slope' (MS)). Main outcomes include the detection rate for combined invasive/micro-invasive and high-grade DCIS (IHG) detection rate and the low/intermediate grade DCIS (LIG) detection rate. RESULTS: At prevalent screens for IHG cancers, 99% of the MMV was reached at a recall rate of 7.0%. The LIG detection rate had no discernible plateau, increasing linearly at a rate of 0.12 per 1000 for every 1% increase in recall rate. At incident screens, 99% of the MMV for IHG cancer detection was 4.0%. LIG DCIS increased linearly at a rate of 0.18 per 1000 per 1% increase in recall rate. CONCLUSIONS: Our models demonstrate the diminishing returns associated with increasing recall rates. The screening programme in England could use the models to set recall rate ranges, and other countries could explore similar methodology. KEY POINTS: • Question: How can we determine optimum recall rates in breast cancer screening? • Findings: In this large observational study, we show that increases in recall rates above defined levels are almost exclusively associated with false positive recalls and a very small increase in low/intermediate grade DCIS. • Meaning: High recall rates are not associated with increases in detection of life-threatening cancers. The models developed in this paper can be used to help set recall rate ranges that maximise benefit and minimise harm.

Image processing can cause some malignant soft-tissue lesions to be missed in digital mammography images.

AIM: To investigate the effect of image processing on cancer detection in mammography. METHODS AND MATERIALS: An observer study was performed using 349 digital mammography images of women with normal breasts, calcification clusters, or soft-tissue lesions including 191 subtle cancers. Images underwent two types of processing: FlavourA (standard) and FlavourB (added enhancement). Six observers located features in the breast they suspected to be cancerous (4,188 observations). Data were analysed using jackknife alternative free-response receiver operating characteristic (JAFROC) analysis. Characteristics of the cancers detected with each image processing type were investigated. RESULTS: For calcifications, the JAFROC figure of merit (FOM) was equal to 0.86 for both types of image processing. For soft-tissue lesions, the JAFROC FOM were better for FlavourA (0.81) than FlavourB (0.78); this difference was significant (p=0.001). Using FlavourA a greater number of cancers of all grades and sizes were detected than with FlavourB. FlavourA improved soft-tissue lesion detection in denser breasts (p=0.04 when volumetric density was over 7.5%) CONCLUSIONS: The detection of malignant soft-tissue lesions (which were primarily invasive) was significantly better with FlavourA than FlavourB image processing. This is despite FlavourB having a higher contrast appearance often preferred by radiologists. It is important that clinical choice of image processing is based on objective measures.

Diagnostic quality of 50 and 100 µm computed radiography compared with screen-film mammography in operative breast specimens.

OBJECTIVE: To compare reader ratings of the clinical diagnostic quality of 50 and 100 µm computed radiography (CR) systems with screen-film mammography (SFM) in operative specimens. METHODS: Mammograms of 57 fresh operative breast specimens were analysed by 10 readers. Exposures were made with identical position and compression with three mammographic systems (Fuji 100CR, 50CR and SFM). Images were anonymised and readers blinded to the CR system used. A five-point comparative scoring system (-2 to +2) was used to assess seven quality criteria and overall diagnostic value. Statistical analysis was subsequently performed of reader ratings (n = 16,925). RESULTS: For most quality criteria, both CR systems were rated as equivalent to or better than SFM. The CR systems were significantly better at demonstrating skin edge and background tissue (p < 1 × 10(-5)). Microcalcification was best demonstrated on the CR50 system (p < 1 × 10(-5)). The overall diagnostic value of both CR systems was rated as being as good as or better than SFM (p < 1 × 10(-5)). CONCLUSION: In this clinical setting, the overall diagnostic performance of both CR systems was as good as or better than SFM, with the CR50 system performing better than the CR100.

Early detection of breast cancer: overview of the evidence on computer-aided detection in mammography screening.

We review the evidence on computer-aided detection (CAD) as an adjunct to mammography interpretation, and discuss the complexity of its impact on decision-making and potential medico-legal aspects. CAD prompts the reader to review lesions on the mammogram and re-evaluate the decision on whether to recall CAD-prompted findings. Studies show that CAD can improve the sensitivity of a single reader, with an incremental cancer detection rate (from adding CAD to a single read) ranging between 1 and 19%. However, CAD will also substantially increase the recall rate (decrease the reader's specificity) causing additional recall in approximately 6-35% of women. Evidence indicates that CAD does not perform as well as double (human) reading in the context of organized breast screening where double reading is the standard of care. Although CAD can identify and prompt readers to missed cancers, the high number of false-positive prompts (1.5-4 false prompts per case) can have an adverse effect on clinical decision-making, and detracts from CAD's application in screening practice. Refinement in CAD algorithms, in combination with increasing implementation of digital mammography, may improve the potential use of CAD in mammography reading, but will require prospective evaluation.

Evaluation of computer-aided detection (CAD) devices.

We present a review of three major UK studies of computer-aided detection (CAD) for mammography. A short account of the motivation, methods and results is given for each of the three. A number of conclusions are drawn, particularly about the merits and difficulties of research in the field. The first two studies measured the impact of CAD on the sensitivity and specificity of film readers interpreting cases with known outcomes displayed on rollers with an artificially high frequency of cancers. In the first study 50 film readers each read 180 cases, including 60 cancers (40 screen-detected and 20 interval). In the second study 35 film readers viewed 120 cases including 44 cancers, of which 40 were selected to be difficult cases that CAD prompted correctly. The third study was carried out prospectively. 6111 films were independently double read by film readers who recorded a judgement before and after viewing CAD prompts. In addition to this, intraobserver measure of the impact of CAD, we compared the cancer detection rate in these cases with that in 1339 cases read over the same period without the benefit of CAD. None of the three studies showed a statistically significant effect attributable to CAD. There is evidence that a high proportion of missed cancers are prompted and that "emphasised" prompts, which have a greater positive predictive value, have a stronger impact on decision-making that other prompts.

Increasing the diagnosis of multifocal primary breast cancer by the use of bilateral whole-breast ultrasound.

AIM: The aim of this study was to evaluate the contribution of bilateral whole-breast ultrasound (BBUS) to the diagnosis and management of women with newly diagnosed breast cancer. METHODS: Over a period of 6 months, 102 women presenting with breast cancer underwent BBUS. Data were collected on clinical findings, radiology, histology and surgical outcome. These women were compared with a control group of 124 women presenting over a similar 6-month period 1 year previously, who had undergone targeted breast ultrasound. RESULTS: Multicentric/multifocal tumours were demonstrated in 35 (34%) of the 102 participants and in 18 (15%) of the 124 controls, a statistically significant difference (Fisher's exact test, p = 0.001). Multiple tumours were diagnosed preoperatively in 18% of the study population compared with 8% of the controls, and BBUS identified invasive multifocal/multicentric tumours in significantly more women in the study population (11 versus 1 control) (Fisher's exact test, p = 0.019). Contralateral cancer was diagnosed in 4 women in the study population and none in the control population (Fisher's exact test, p = 0.040). Surgical review showed that the surgical management changed significantly in 8% (95% confidence interval 4 to 14%) of cases in the study population following BBUS. The increase in the number of women undergoing benign biopsies in the study population (10 versus 5 controls) was not statistically significant (Fisher's exact test, p = 0.11). CONCLUSION: BBUS increased the preoperative diagnosis of multiple tumours in women presenting with primary breast cancer, resulting in a management change in 8% of cases.

Impact of computer-aided detection prompts on the sensitivity and specificity of screening mammography.

OBJECTIVES: To determine the value of computer-aided detection (CAD) for breast cancer screening. DESIGN: Two sets of mammograms with known outcomes were used in two studies. Participants in both studies read the films with and without the benefit of a computer aid. In both studies, the order of reading sessions was randomised separately for each reader. The first set of 180 films, used in study 1, included 20 false-negative interval cancers and 40 screen-detected cancers. The second set of 120 films, used in study 2, was designed to be favourable to CAD: all 44 cancer cases had previously been missed by a film reader and cancers prompted by CAD were preferentially included. SETTING: The studies were conducted at five UK screening centres between January 2001 and April 2003. PARTICIPANTS: Thirty radiologists, five breast clinicians and 15 radiographers participated. INTERVENTIONS: All cases in the trial were digitised and analysed using the R2 ImageChecker version 2.2. Participants all received training on the use of CAD. In the intervention condition, participants interpreted cases with a prompt sheet on which regions of potential abnormality were indicated. MAIN OUTCOME MEASURES: The sensitivity and specificity of participants were measured in both intervention and control conditions. RESULTS: No significant difference was found for readers' sensitivity or specificity between the prompted and unprompted conditions in study 1 [95% confidence index (CI) for sensitivity with and without CAD is 0.76 to 0.80, for specificity it is 0.81 to 0.86 without CAD and 0.81 to 0.87 with CAD]. No statistically significant difference was found between the sensitivity and specificity of different groups of film reader (95% CI for unprompted sensitivity of radiologists was 0.75 to 0.81, for radiographers it was 0.71 to 0.81, prompted sensitivity was 0.76 to 0.81 for radiologists and 0.69 to 0.79 for radiographers). Thirty-five readers participated in study 2. Sensitivity was improved in the prompted condition (0.81 from 0.78) but the difference was slightly below the threshold for statistical significance (95% CI for the difference -0.003 to 0.064). Specificity also improved (0.87 from 0.86); again, the difference was not significant at 0.05 (95% CI -0.003 to 0.034). A cost-effectiveness analysis showed that computer prompting increases cost. CONCLUSIONS: No significant improvement in film readers' sensitivity or specificity or gain in cost-effectiveness was established in either study. This may be due to the system's low specificity, its relatively poor sensitivity for subtle cancers or the fact the prompts cannot serve as aids to decision-making. Readers may have been better able to make use of the prompts after becoming more accustomed to working with them. Prompts may have an impact in routine use that is not detectable in an experimental setting. Although the case for CAD as an element of the NHS Breast Screening Programme is not made here, further research is required. Evaluations of new CAD tools in routine use are underway and their results should be given careful attention.

The clinical importance of axillary lymphadenopathy detected on screening mammography: revisited.

AIM: The aim of this study was to re-evaluate our protocol for the management of isolated axillary lymphadenopathy (ALP) on mammographic screening. METHODS: In a retrospective review of 200,716 women screened at the South West London Breast Screening Service (SWLBSS) over 7 years, 72 women with ALP with an otherwise normal mammogram were identified. Thirteen patients were not recalled, nine of who had a known underlying diagnosis and the remainder had longstanding unchanged mammograms. Fifty-nine patients were recalled for further clinical assessment and investigations, including ultrasound, further mammographic views, fine-needle aspiration cytology (FNAC), blood tests and a chest radiograph. Those with a definite diagnosis were referred for appropriate management and those with benign reactive cytology on FNAC reviewed at 6 weeks with subsequent referral for excision of persisting abnormal nodes. RESULTS: The ultimate diagnosis was benign in 45 cases: 26 benign reactive changes, 11 arthritides, five with dermatological and viral conditions and three with tuberculosis. Malignancy was diagnosed in 13 cases: four with metastatic breast carcinoma and nine with lymphoma/leukaemia. The total number of newly diagnosed malignancies was 20% of women recalled. Another 5% of patients had active tuberculosis. Of the 22 patients with benign reactive cytology, one had significant pathology on excision biopsy: tuberculosis. Over 95% of the results from excision biopsy in these patients did not alter management. CONCLUSION: In the majority of patients, the FNAC results were representative of the final excision pathology. The present study suggests that excision biopsy could be omitted for those patients whose FNAC and culture are negative.

Assessing the impact of CAD on the sensitivity and specificity of film readers.

AIM: To assess the impact of computer-aided detection (CAD) prompts on film readers' sensitivity and specificity. MATERIAL AND METHODS: Thirty-five readers read 120 films, including 44 cancers, 40 of which were prompted. All readers looked at all cases with and without prompts. The sensitivity and specificity were calculated for each reader under each condition. RESULTS: The sensitivity improved when CAD prompts were used (0.80 from 0.77). The difference was slightly below the threshold for statistical significance (95% CI for the difference is -0.0027-0.064). The specificity also improved (0.86 from 0.85), but not significantly. There was a significant improvement in sensitivity when readers' judgements were combined to simulate double reading, from 0.77 to 0.81. (95% CI for the difference is 0.014-0.077). CONCLUSIONS: Analysis of prompted cancers that readers did and did not recall, found that cases were more likely to be correctly recalled if there were emphasized prompts, more prompts or if the case was harder. There was no statistically significant effect for type of abnormality or tumour size or for the performance, attitude or experience of the reader.

An evaluation of the impact of computer-based prompts on screen readers' interpretation of mammograms.

The NHS Breast Screening Programme is changing working practices in response to increased workload and manpower shortages. Radiographers are being trained as film readers. Computer aids are being considered. We have evaluated the impact of prompts placed by the R2 ImageChecker 1000 on the sensitivity and specificity of film readers, including non-radiologist film readers. 30 radiologists, 5 breast clinicians and 15 radiographers each read 180 films, including 60 cancers (20 false negative interval cases and 40 screen detected cancers). Each reader read each case twice, once with and once without computer prompts. The order in which the reading sessions were carried out was randomized separately for each reader. 36 out of 40 screen-detected cases were prompted by the ImageChecker, a sensitivity of 90%. Eight out of 20 interval cases were prompted, a sensitivity of 40%. No significant difference was found for readers' sensitivity or specificity between the prompted and unprompted conditions. No significant difference was found between the sensitivity and specificity of the different groups of film reader. No difference in impact of prompts was found for well or poorly performing film readers. The result suggests that this version of the ImageChecker would not have a significant impact on the UK screening programme.

Mammographic casting-type calcification associated with small screen-detected invasive breast cancers: is this a reliable prognostic indicator?

AIM: The aim of the present study was to establish whether mammographic casting-type calcification associated with small screen-detected invasive breast cancers is a reliable prognostic indicator. METHODS AND MATERIALS: We retrospectively identified 50 consecutive women diagnosed with an invasive cancer less than 15 mm who showed associated casting calcification on their screening mammograms. Controls were identified that showed no microcalcification and were matched for tumour size, histological type and lymph node status. A minimum of 5 years follow-up was obtained, noting recurrence and outcome. Conditional and unconditional logistic regression, depending on the outcome variable, were used to analyse the data, taking the matched design into account in both cases. Where small numbers prohibited the use of logistic regression, Fisher's exact test was used. RESULTS: Five deaths from breast cancer occurred out of the 50 cases, of which three were lymph node positive, two were lymph node negative and none were grade 3. None of the 78 control cases died from breast cancer. The difference in breast cancer death rates was significant by Fisher's exact test (p=0.02). Risk of recurrence was also significantly increased in the casting cases (OR=3.55, 95% CI 1.02-12.33, p=0.046). CONCLUSION: Although the overall outcome for small screen-detected breast cancers is good, our study suggests that casting calcification is a poorer prognostic factor. The advantage of a mammographic feature as an independent prognostic indicator lies in early identification of high-risk patients, allowing optimization of management.

Reproducibility of prompts in computer-aided detection (CAD) of breast cancer.

AIM: We evaluated the reproducibility of prompts using the R2 ImageChecker M2000 computer-aided detection (CAD) system. MATERIALS AND METHODS: Forty selected two-view mammograms of women with breast cancer were digitized and analysed using the ImageChecker on 10 separate occasions. The mammograms were chosen to provide both straightforward and subtle signs of malignancy. Data analysed included mammographic abnormality, pathology, and whether the cancer was prompted or given an emphasized prompt. RESULTS: Correct prompts were generated in 86 out of 100 occasions for screen-detected cancers. Reproducibility was less in the other categories of more subtle cancers: 21% for cancers previously missed by CAD, a group that contained more grade 1 and small (<10 mm) tumours. Prompts for calcifications were more reproducible than those for masses (76% versus 53%) and these cancers were more likely to have an emphasized prompt. CONCLUSIONS: Probably the most important cause of variability of prompts is shifts in film position between sequential digitizations. Consequently subtle lesions that are only just above the threshold for display may not be prompted on repeat scanning. However, users of CAD should be aware that even emphasized prompts are not consistently reproducible.

The diagnostic contribution of the frontal lumbar spine radiograph in community referred low back pain--a prospective study of 1030 patients.

AIM: The diagnostic contribution of the anteroposterior (AP) view was studied to assess whether this view could be omitted safely, thus reducing the radiation burden received by patients undergoing lumbar spine radiography. MATERIALS AND METHODS: Prospective analysis of 1030 consecutive referrals for lumbar spine radiography from general practice. RESULTS: In the majority of cases (90.5%) the AP view was non-contributory. In 4.2% the diagnosis was strengthened and it was altered in 4.6%. However, in the latter group only 1.3% of the total were considered significant alterations. These were cases of possible, but not definite, pars defects and sacroiliitis. Specific important conditions such as infection, malignancy and benign tumours were not missed on the lateral view alone, in our study population. The radiation burden is reduced by 75% by omitting the AP view. CONCLUSION: A single lateral view is an adequate examination, with the proviso that sacroiliac joint disease is not assessed on this view and some pars defects and facet joint degenerative changes may be overlooked. The radiation protection gains are considerable. A single lateral lumbar view is now our routine practice unless sacroiliitis is a specific clinical concern.

Vitamin D receptor gene polymorphisms are associated with breast cancer risk in a UK Caucasian population.

There is increasing evidence that vitamin D can protect against breast cancer. The actions of vitamin D are mediated via the vitamin D receptor (VDR). We have investigated whether polymorphisms in the VDR gene are associated with altered breast cancer risk in a UK Caucasian population. We recruited 241 women following a negative screening mammogram and 181 women with known breast cancer. The VDR polymorphism Bsm I, an intronic 3' gene variant, was significantly associated with increased breast cancer risk: odds ratio bb vs BB genotype = 2.32 (95% CI, 1.23-4.39). The Bsm I polymorphism was in linkage disequilibrium with a candidate translational control site, the variable length poly (A) sequence in the 3' untranslated region. Thus, the 'L' poly (A) variant was also associated with a similar breast cancer risk. A 5' VDR gene variant, Fok I, was not associated with breast cancer risk. Further investigations into the mechanisms of interactions of the VDR with other environmental and/or genetic influences to alter breast cancer risk may lead to a new understanding of the role of vitamin D in the control of cellular and developmental pathways.