This website is intended for healthcare professionals only.
Take a look at a selection of our recent media coverage:
23rd September 2021
Globally, in 2020, there were an estimated 2.3 million women diagnosed with breast cancer leading to 685,000 deaths. Fortunately, improvements in survival over recent decades have been attributed to population-based breast cancer screening with mammography. In fact, a recent UK study suggested that screening reduces cancer mortality by 38% among women screened at least once.
The use of artificial intelligence (AI) systems for image recognition in breast cancer screening could lead to improvements in the detection of cases, either as a standalone system or as an aid to radiologists. Indeed, there is some evidence to support the value of AI with one retrospective analysis of an AI screening algorithm concluding that it showed better diagnostic performance than a radiologist. Nevertheless, in a 2019 review, it was concluded that while AI systems have good accuracy for breast cancer detection, methodological concerns and evidence gaps exist that limit translation into clinical breast cancer screening settings.
In light of these concerns, a team from the Division of Health Sciences, University of Warwick, UK, were commissioned by the UK National Screening Committee to undertake a systematic review to determine whether there was sufficient evidence to support the introduction of AI for mammographic image analysis in breast screening. They conducted literature searches up to May 2021 and included studies that reported the test accuracy of AI algorithms either alone or in combination with radiologists, to detect breast cancer in digital mammograms in screening practice or in test sets. The team included cancer confirmed by histological analysis of biopsy samples in cases where women were referred for further tests after screening as the reference standard or from symptomatic presentation during follow-up.
The review identified a total of 12 studies including 131,822 women undergoing breast cancer screening. In studies with a standalone AI system, the algorithm calculated a cancer risk score, categorising women at either high (recall) or low (no recall) risk. When used to assist the radiologist, the AI system simply provided a level of suspicion. In two large retrospective studies including 76,813 women, that compared the AI system with the clinical decisions of a radiologist, 96% of systems were less accurate than a single radiologist and all were less accurate than a double read.
Overall, the authors reported considerably heterogeneity in study methodology, some of which resulted in high concerns over the risk of bias and applicability. In their study, they commented that “evidence is insufficient on the accuracy or clinical effect of introducing AI to examine mammograms anywhere on the screening pathway”.
In the conclusion, the authors noted how AI systems for breast cancer screening are a long way from having the quality and quantity required for implementation into clinical practice.
Freeman K et al. Use of artificial intelligence for image analysis in breast cancer screening programmes: systematic review of test accuracy. BMJ 2021
21st September 2021
Detection of breast cancer requires that mammography scans are interpreted correctly and thus guidelines on screening require that robust quality control and quality assurance are in place. Although an assessment of performance against standards can be achieved through audits, such audits offer little insight of the diagnostic skills of individual radiologists. The use of on-line self-assessment, using pre-diagnosed mammograms, offers a means of identifying any knowledge gaps and could serve as an important educational tool.
An on-line self-assessment tool was adopted by the Italian Society of Medical Radiology in 2018 as part of a national on-line self-assessment program for interpretation of mammograms and the results have published. The test set was a collection of 24 pathology-confirmed cancers combined with 108 mammograms which were reported on as negative after double reading. All cases were acquired with full-field digital mammography and came from women aged 50 – 69 years, who represent those at average risk of breast cancer. For the assessment, the society determined a pass threshold for sensitivity of 62% (amounting to at least 15/24 cancers correctly identified) and 86% specificity (93/108 negative mammograms correctly identified). The on-line self-assessment was posted on a dedicated website and upon registration, radiologists provided information on demographics such as age, gender, place of work (public or private sector), years of breast imaging experience and a qualitative self-judgement of mammography interpretative skills as “beginner”, “average” and “expert”. As with usual mammography practice, the radiologists were required to submit a dichotomous diagnosis (i.e., positive/negative) for every case. In addition to reporting on the radiologist’s success in the test, the authors employed regression analysis and odds ratios to identify the most important predictor variables for diagnostic accuracy.
A total of 685 radiologists registered for the on-line self-assessment and 49.9% (342) with a mean age of 46 years (69% female), completed test. Among those completing, two-thirds (64%) self-judged their interpretative skills as “average” and respondents had a median of 8 years breast imaging experience although 38.3% reporting having more than 10 years’ experience. Participants reported a median number of 1501 mammographic interpretations per year and the majority (68.7%) worked in the public rather than private sector.
When examining the proportion who successfully completed the assessment, only 28.7% of radiologists (98/342) passed on their first attempt. After an initial failure, 138 of the remaining 244 radiologists, re-took the test of whom, only 35.5% (49/138) passed. In fact, the authors reported that overall, only 44.2% (151/342) of radiologists who completed the on-line self-assessment were successful.
Using regression analysis, a significant association for diagnostic accuracy was found only for the assessment of > 3,000 mammograms per year compared to < 1,000 (Odds ratio, OR = 3.88, 95% CI 1.07 – 14.14, p = 0.04) and working the in public rather than private sector (OR = 1.65). Other variables such as age, self-judged interpretative skills and years of breast imaging experience had no significant effect.
The authors concluded that their study suggested that breast imaging experience does not guarantee diagnostic accuracy in screening reading. They also noted that their on-line self-assessment test could be included as a criterion for the accreditation process of breast units.
Brancato B et al. Mammography self?evaluation online test for screening readers: an Italian Society of Medical Radiology (SIRM) initiative. Eur Radiol 2021
9th July 2021
Breast cancer is the most common cancer in women worldwide and in 2020, there were 2.2 million cases and nearly 685,000 deaths. Breast cancer is a heterogenous disease with various types and different sensitivities to treatment, e.g., oestrogen receptor (ER) positive, progesterone receptor (PR) positive and hormone receptor (HR) negative. Among those with ER positive tumours, treatment with adjunctive therapy such as tamoxifen for five years is recommended. Nevertheless, even after treatment with tamoxifen, breast cancer can return over the intervening years, metastasise and lead to death. The use of clinical breast cancer markers such as tumour size or grade can be used to provide estimates of survival for up to 10 years and both increased tumour size and grade are associated with a reduced short-term survival. However, what is far less clear, is whether any of these markers are associated with longer term survival.
Given that both ER-positive and ERBB2-negative (i.e., HER2) disease are associated with a continuous risk of recurrent disease after the primary diagnosis, a team from the Department of Oncology and Pathology, Karolinska Institute, Stockholm, Sweden, sought to determine whether both clinically used markers and treatment with tamoxifen were associated with long-term survival in patients with breast cancer. The team turned to data from the Stockholm tamoxifen (STO-3) trial, which enrolled postmenopausal women with lymph-node negative breast cancer and tumours less than 30mm in diameter, between 1976 and 1990 and who were randomised to tamoxifen 40mg daily or no endocrine therapy. In 1983, women with no recurrence after 2 years on tamoxifen, were given the drug for a further 3 years. The team undertook an analysis of tissue samples that were available from patients in the study. The tumour grade (1 to 3) had been assessed in 2014 by a pathologist and tumour size was categorised as T1a/b if 10mm or less, T1c if 11–20mm and T2 if larger than 20mm. All patients had a unique national registration number and which was linked with a cancer registry, allowing the team to obtain long-term follow-up data after 25 years.
The sample contained 565 postmenopausal women with a mean age of 62 years. Just over half (52.2%) of tumours were graded as T1c and nearly a third (30%), T1a/b. Patients with T1a/b tumours had the best long-term survival (88%) compared to those with T1c tumours (76%) or T2 tumours (63%). Similarly, the highest level of survival (81%) occurred in those with grade 1 tumours compared with grade 3 tumours (65%). Using this data, the team estimated a 69% reduced risk of disease recurrence for those with T1a/b sized tumours (hazard ratio, HR = 0.31, 95% CI 0.17–55). In terms of treatment, those given tamoxifen with grade 1–2 tumours also experienced a significantly reduced risk of disease recurrence (HR = 0.24, 95% CI 0.07 – 0.82) compared with those who did receive the drug. However, use of tamoxifen also produced a significant survival benefit in those with T2 tumours (HR = 0.34).
The authors concluded that tumour size, followed by grade and use of tamoxifen were independently associated with long-term survival in breast cancer.
Dar H et al. Assessment of 25-Year Survival of Women with Oestrogen Receptor–Positive/ERBB2-Negative Breast Cancer Treated with and Without Tamoxifen Therapy. A Secondary Analysis of Data from the Stockholm Tamoxifen Randomised Clinical Trial. JAMA Netw Open 2021.
25th May 2021
Cancer of the breast is the most common form of cancer in women although with an early diagnosis, the 5-year survival prognosis ranges from 86 to 99%. Nevertheless, women who survive breast cancer have a 17% increased risk for a second cancer compared to the general population. One factor known to be associated with cancer is obesity with one US study estimating that 40% of all cancer diagnoses occurred in people who were either overweight or obese. However, while much attention has been paid to the effect of obesity on the development of an initial cancer, far less is known about how obesity impacts on the development of a second cancer. As a result, a team from Kaiser Permanente, Denver, US, sought to examine the association between body mass index (BMI) and a second cancer among women who survived invasive breast cancer. Data were extracted from an electronic database and a surveillance tumour registry which provided information on the incidence and type of secondary cancers that occurred. Height and weight measurements within two years prior through one year after the date of the initial breast cancer diagnosis were used to calculate the BMI. All women included had surgery as part of their initial breast cancer and had no evidence of a second cancer one year later. The study outcomes included all second cancers, cancers for which there was a known association with obesity (e.g., oesophageal adenocarcinoma), and ER-positive second breast cancers.
A total of 6481 women were included in the analysis with a mean age of 60.2 years, of whom 33.4% were classed as overweight or obese (33.8%) at the time of their initial breast cancer diagnosis. During a median follow-up of 88 months, 822 (12.7%) women developed a second cancer, of which 508 (61.8%) were obesity-related and 333 (40.5%) were breast cancer, the majority of which (69.4%) were ER-positive. The authors calculated that every 5 unit increase in BMI was associated with a 7% increased risk of developing any second cancer (relative risk, RR = 1.07, 95% CL 1.01–1.14), a 13% increased for an obesity-related cancer and by 15% for a second ER-positive breast cancer.
The authors calculated that the risk of a second cancer was increased by 5% for every 5 unit increase in BMI. They concluded that these data had important public health implications given the prevalence of obesity and underscored the need for effective preventative strategies.
Feigelson HS et al. Body Mass Index and Risk of Second Cancer Among Women with Breast Cancer. J Natl Cancer Inst 2021