How we provided appropriate breast imaging practices in the epicentre of the COVID-19 outbreak in Italy

Updates on Breast Cancer

This collection of 18 articles, comprising 12 original studies, 1 systematic review, and 5 reviews, is a collaborative effort by distinguished experts in breast cancer research, and it has been edited by Dr [...].

Deep learning performance for detection and classification of microcalcifications on mammography

BACKGROUND

Breast cancer screening through mammography is crucial for early detection, yet the demand for mammography services surpasses the capacity of radiologists. Artificial intelligence (AI) can assist in evaluating microcalcifications on mammography. We developed and tested an AI model for localizing and characterizing microcalcifications.

METHODS

Three expert radiologists annotated a dataset of mammograms using histology-based ground truth. The dataset was partitioned for training, validation, and testing. Three neural networks (AlexNet, ResNet18, and ResNet34) were trained and evaluated using specific metrics including receiver operating characteristics area under the curve (AUC), sensitivity, and specificity. The reported metrics were computed on the test set (10% of the whole dataset).

RESULTS

The dataset included 1,000 patients aged 21-73 years and 1,986 mammograms (180 density A, 220 density B, 380 density C, and 220 density D), with 389 malignant and 611 benign groups of microcalcifications. AlexNet achieved the best performance with 0.98 sensitivity, 0.89 specificity of, and 0.98 AUC for microcalcifications detection and 0.85 sensitivity, 0.89 specificity, and 0.94 AUC of for microcalcifications classification. For microcalcifications detection, ResNet18 and ResNet34 achieved 0.96 and 0.97 sensitivity, 0.91 and 0.90 specificity and 0.98 and 0.98 AUC, retrospectively. For microcalcifications classification, ResNet18 and ResNet34 exhibited 0.75 and 0.84 sensitivity, 0.85 and 0.84 specificity, and 0.88 and 0.92 AUC, respectively.

CONCLUSIONS

The developed AI models accurately detect and characterize microcalcifications on mammography.

RELEVANCE STATEMENT

AI-based systems have the potential to assist radiologists in interpreting microcalcifications on mammograms. The study highlights the importance of developing reliable deep learning models possibly applied to breast cancer screening.

KEY POINTS

• A novel AI tool was developed and tested to aid radiologists in the interpretation of mammography by accurately detecting and characterizing microcalcifications. • Three neural networks (AlexNet, ResNet18, and ResNet34) were trained, validated, and tested using an annotated dataset of 1,000 patients and 1,986 mammograms. • The AI tool demonstrated high accuracy in detecting/localizing and characterizing microcalcifications on mammography, highlighting the potential of AI-based systems to assist radiologists in the interpretation of mammograms.

An Unenhanced Breast MRI Protocol Based on Diffusion-Weighted Imaging: A Retrospective Single-Center Study on High-Risk Population for Breast Cancer

PURPOSE

This study aimed to investigate the use of contrast-free magnetic resonance imaging (MRI) as an innovative screening method for detecting breast cancer in high-risk asymptomatic women. Specifically, the researchers evaluated the diagnostic performance of diffusion-weighted imaging (DWI) in this population.

METHODS

MR images from asymptomatic women, carriers of a germline mutation in either the BRCA1 or BRCA2 gene, collected in a single center from January 2019 to December 2021 were retrospectively evaluated. A radiologist with experience in breast imaging (R1) and a radiology resident (R2) independently evaluated DWI/ADC maps and, in case of doubts, T2-WI. The standard of reference was the pathological diagnosis through biopsy or surgery, or ≥1 year of clinical and radiological follow-up. Diagnostic performances were calculated for both readers with a 95% confidence interval (CI). The agreement was assessed using Cohen's kappa (κ) statistics.

RESULTS

Out of 313 women, 145 women were included (49.5 ± 12 years), totaling 344 breast MRIs with DWI/ADC maps. The per-exam cancer prevalence was 11/344 (3.2%). The sensitivity was 8/11 (73%; 95% CI: 46-99%) for R1 and 7/11 (64%; 95% CI: 35-92%) for R2. The specificity was 301/333 (90%; 95% CI: 87-94%) for both readers. The diagnostic accuracy was 90% for both readers. R1 recalled 40/344 exams (11.6%) and R2 recalled 39/344 exams (11.3%). Inter-reader reproducibility between readers was in moderate agreement (κ = 0.43).

CONCLUSIONS

In female carriers of a BRCA1/2 mutation, breast DWI supplemented with T2-WI allowed breast cancer detection with high sensitivity and specificity by a radiologist with extensive experience in breast imaging, which is comparable to other screening tests. The findings suggest that DWI and T2-WI have the potential to serve as a stand-alone method for unenhanced breast MRI screening in a selected population, opening up new perspectives for prospective trials.

Assessment and Response to Neoadjuvant Treatments in Breast Cancer: Current Practice, Response Monitoring, Future Approaches and Perspectives

Neoadjuvant treatments (NAT) for breast cancer (BC) consist in the administration of chemotherapy-more rarely endocrine therapy-before surgery. Firstly, it was introduced 50 years ago to downsize locally advanced (inoperable) BCs. NAT are now widespread and so effective to be used also at the early stage of the disease. NAT are heterogeneous in terms of therapeutic patterns, class of used drugs, dosage, and duration. The poly-chemotherapy regimen and administration schedule are established by a multi-disciplinary team, according to the stage of disease, the tumor subtype and the age, the physical status, and the drug sensitivity of BC patients. Consequently, an accurate monitoring of treatment response can provide significant clinical advantages, such as the treatment de-escalation in case of early recognition of complete response or, on the contrary, the switch to an alternative treatment path in case of early detection of resistance to the ongoing therapy. Future is going toward increasingly personalized therapies and the prediction of individual response to treatment is the key to practice customized care pathways, preserving oncological safety and effectiveness. To gain such goal, the development of an accurate monitoring system, reproducible and reliable alone or as part of more complex diagnostic algorithms, will be promising.

Women’s perceptions and attitudes to the use of AI in breast cancer screening: a survey in a cancer referral centre

Objective:

Although breast cancer screening can benefit from Artificial Intelligence (AI), it is still unknown whether, to which extent or under which conditions, the use of AI is going to be accepted by the general population. The aim of our study is to evaluate what the females who are eligible for breast cancer screening know about AI and how they perceive such innovation.

Methods:

We used a prospective survey consisting of a 11-multiple-choice questionnaire evaluating statistical associations with Chi-Square-test or Fisher-exact-test. Multinomial-logistic-regression was performed on items with more than two response categories. Odds ratio (OR) with 95% CI were computed to estimate the probability of a specific response according to patient’s characteristics.

Results:

In the 800 analysed questionnaires, 51% of respondents confirmed to have knowledge of AI. Of these, 88% expressed a positive opinion about its use in medicine. Non-Italian respondents were associated with the belief of having a deep awareness about AI more often than Italian respondents (OR = 1.91;95% CI[1.10–3.33]). Higher education level was associated with better opinions on the use of AI in medicine (OR = 4.69;95% CI[1.36–16.12]). According to 94% of respondents, the radiologists should always produce their own report on mammograms, whilst 77% agreed that AI should be used as a second reader. Most respondents (52%) considered that both the software developer and the radiologist should be held accountable for AI errors.

Conclusions:

Most of the females undergoing screening in our Institute approve the introduction of AI, although only as a support to radiologist, and not in substitution thereof. Yet, accountability in case of AI errors is still unsolved. advances in knowledge: This survey may be considered as a pilot-study for the development of large-scale studies to understand females’s demands and concerns about AI applications in breast cancer screening.

Legal and Regulatory Framework for AI Solutions in Healthcare in EU, US, China, and Russia: New Scenarios after a Pandemic

The COVID-19 crisis has exposed some of the most pressing challenges affecting healthcare and highlighted the benefits that robust integration of digital and AI technologies in the healthcare setting may bring. Although medical solutions based on AI are growing rapidly, regulatory issues and policy initiatives including ownership and control of data, data sharing, privacy protection, telemedicine, and accountability need to be carefully and continually addressed as AI research requires robust and ethical guidelines, demanding an update of the legal and regulatory framework all over the world. Several recently proposed regulatory frameworks provide a solid foundation but do not address a number of issues that may prevent algorithms from being fully trusted. A global effort is needed for an open, mature conversation about the best possible way to guard against and mitigate possible harms to realize the potential of AI across health systems in a respectful and ethical way. This conversation must include national and international policymakers, physicians, digital health and machine learning leaders from industry and academia. If this is done properly and in a timely fashion, the potential of AI in healthcare will be realized.