Journal of Breast Imaging最新文献

Objective: Breast MRI affords high sensitivity with intermediate specificity for cancer detection. Ultrafast dynamic contrast-enhanced (DCE) MRI assesses early contrast inflow with potential to supplement or replace conventional DCE-MRI kinetic features. We sought to determine whether radiologist's evaluation of ultrafast DCE-MRI can increase specificity of a clinical MRI protocol.

Methods: In this IRB-approved, HIPAA-compliant study, breast MRIs from March 2019 to August 2020 with a BI-RADS category 3, 4, or 5 lesion were identified. Ultrafast DCE-MRI was acquired during the first 40 seconds after contrast injection and before conventional DCE-MRI postcontrast acquisitions in the clinical breast MRI protocol. Three radiologists masked to outcomes retrospectively determined lesion time to enhancement (TTE) on ultrafast DCE-MRI. Interreader agreement, differences between benign and malignant lesion TTE, and TTE diagnostic performance were evaluated.

Results: Ninety-five lesions (20 malignant, 75 benign) were included. Interreader agreement in TTE was moderate to substantial for both ultrafast source images and subtraction maximum intensity projections (overall κ = 0.63). Time to enhancement was greater across benign lesions compared with malignancies (P <.05), and all lesions demonstrating no enhancement during the ultrafast series were benign. With a threshold TTE ≥40 seconds, ultrafast DCE-MRI yielded an average 40% specificity (95% CI, 30%-48%) and 92% sensitivity (95% CI, 81%-100%), yielding a potential reduction in 31% (95% CI, 23%-39%) of benign follow-ups based on conventional DCE-MRI.

Conclusion: Ultrafast imaging can be added to conventional DCE-MRI to increase diagnostic accuracy while adding minimal scan time. Future work to standardize evaluation criteria may improve interreader agreement and allow for more robust ultrafast DCE-MRI assessment.

The clinical approach, differential diagnosis, diagnostic imaging, and management of breast masses differs between adult and pediatric patients. Breast symptoms in pediatric populations are likely to be normal, variants of normal development, or classically benign. When evaluating pediatric patients, a thorough history and physical examination is key and may include US imaging if clinically indicated. Currently, there are no consensus guidelines from professional societies specific to the imaging evaluation, reporting, or management of the pediatric breast. Inappropriate utilization of additional imaging or biopsy recommendations can cause stress to patients and caregivers, added financial costs, and potential damage to the developing breast. As such, it is increasingly important for radiologists to have a clear understanding of the expected physical and imaging findings for developing breast tissue, developmental abnormalities, benign conditions, and rare malignancies. This review summarizes the expected normal developmental breast changes in the pediatric population as well as common anatomic variants. Both benign and rare malignant breast pathologies are reviewed with a discussion about clinical presentation and management to guide breast imaging trainees and practicing radiologists.

Objective: To identify imaging, patient, and radiologist factors associated with malignant pathology for mammographic asymmetries referred for image-guided biopsy.

Methods: This is an IRB-approved study of consecutive image-guided core-needle biopsy reports from an academic center and affiliated imaging centers January 1, 2015 to June 30, 2022 with lesion type reported by biopsy radiologist as "asymmetry" on biopsy requisition form retrieved from the pathology database. Imaging features (asymmetry type, associated mammographic features, US correlate, lesion size) and patient demographics (age at biopsy, breast density, family or personal history of breast cancer) were extracted from imaging reports and electronic health record. Cases were excluded if the diagnostic or biopsy radiologist reported the lesion other than "asymmetry" or if imaging review identified the finding as predominantly mass or architectural distortion. Multiple logistic regression was performed to calculate the odds ratio (OR) of malignant outcome.

Results: Image-guided core-needle biopsy was requested of 326 asymmetries; 79 were excluded, yielding 247 asymmetries (9 asymmetry, 218 developing asymmetry, 20 focal asymmetry). Overall, 40/247 (16.2%) were malignant and 207/247 (83.8%) were benign. Presence of associated distortion (OR 14.78; 95% CI, 4.87-44.83; P <.001) or calcifications (OR 9.86; 95% CI, 2.74-35.53; P <.001), personal history of breast cancer (OR 2.65; 95% CI, 1.04-6.77; P = .041), and increasing patient age at biopsy (OR 1.08; 95% CI, 1.04-1.12; P <.001) were associated with likelihood of malignancy.

Conclusion: Malignancy of mammographic asymmetries is more likely in the presence of associated distortion or calcifications, personal history of breast cancer, and increasing patient age. These results may be useful in diagnostic management of mammographic asymmetries.

Artificial intelligence (AI) in breast imaging has garnered significant attention given the numerous reports of improved efficiency, accuracy, and the potential to bridge the gap of expanded volume in the face of limited physician resources. While AI models are developed with specific data points, on specific equipment, and in specific populations, the real-world clinical environment is dynamic, and patient populations are diverse, which can impact generalizability and widespread adoption of AI in clinical practice. Implementation of AI models into clinical practice requires focused attention on the potential of AI bias impacting outcomes. The following review presents the concept, sources, and types of AI bias to be considered when implementing AI models and offers suggestions on strategies to mitigate AI bias in practice.

Over the last 30 years, axillary surgery in patients with breast cancer has undergone a consistent patient-centered, evidence-based de-escalation of surgical intervention, including considerable practice changes over the last decade. As surgical approaches change, the role of breast imaging radiologists and axillary imaging must also evolve. Axillary imaging remains variable across radiology practices, with implementation of new protocols occurring at various speeds. Breast radiologists must be aware of recent trials and remain agile in adopting or responding to changing treatment paradigms. Breakthrough studies, such as the SOUND (Sentinel Node Vs Observation After Axillary Ultra-souND) trial, must be vetted and evaluated by individual practices before adoption. Breast radiologists should play a key role in multidisciplinary collaboration with colleagues involved in breast cancer care to assist with strategic planning and appropriate resource utilization.

Objective: This study assessed opportunities for graduated autonomy in fellowship programs registered with the Society of Breast Imaging (SBI) Fellowship Match.

Methods: A 16-question survey developed by the SBI Fellowship Match Committee was distributed electronically to fellowship program directors registered with the SBI. Responses were analyzed, with subgroup comparison using Fisher's exact test.

Results: The response rate was 51.5% (52/101). Most respondents (63.5%, 33/52) do not allow fellows to final sign reports. Of programs that do offer this practice, 36.8% (7/19) have done so for <3 years, 21.1% (4/19) for 3 to 5 years, and 42.1% (8/19) for >5 years. There was no association between fellowship class size or length of fellowship training and final-sign opportunities. Fellow education (84.2%, 16/19) and fellow interest (73.7%, 14/19) were the most common reasons for offering final-sign privileges. Faculty consensus was the main criterion for assessing fellow readiness for graduated autonomy. Of examination types, independent interpretation was most common for diagnostic mammogram and US examinations (36.5%, 19/52), initiated before the last 2 months of fellowship. Approximately 30% (16/52) of respondents allow fellows to perform and final sign procedures, most commonly 5 to 10 months into fellowship training. In 52.6% (10/19) of programs allowing independent reads, no additional compensation is provided.

Conclusion: Most breast imaging fellowship programs do not allow fellows to independently render examination interpretations or perform breast procedures. However, more than half of programs offering fellow autonomy have done so for ≤5 years, suggesting a potential shift in final-sign opportunities.

Receiving feedback can sometimes be difficult and uncomfortable but is an essential component of professional development in breast imaging. Trainees have an opportunity to leverage feedback in breast imaging by incorporating self-assessments, real-world patient outcomes, procedural feedback, patient interactions, and available audit data to build confidence and competency in residency and fellowship. We present strategies for seeking and receiving feedback with a growth mindset, including specific scenarios in breast imaging where trainees can incorporate feedback and maximize learning potential.

Objective: The purpose of this study was to describe the safety and efficacy of percutaneous drain placement for postoperative fluid collections in the breast.

Methods: A retrospective review was conducted of the patient characteristics, intervention data, and clinical outcomes of the 43 adult patients who underwent percutaneous drain placement for fluid collections at a tertiary care hospital over a 13-year period ending February 28, 2023.

Results: Most fluid collections treated with percutaneous drain placement were secondary to ipsilateral breast surgery (92%, 44/48), most commonly breast reduction (23%, 10/44) and mastectomy with immediate tissue expander reconstruction (16%, 7/44). Additional patients had fluid collections without prior ipsilateral breast surgery (8%, 4/48) and were excluded from further analysis. The fluid cultures from 39% of the cultured postsurgical fluid collections were positive (16/41), and Staphylococcus aureus was the most commonly cultured organism (15%, 6/41). The only immediate complication was the rupture of a tissue expander during drain placement in 1 patient. The median duration of percutaneous drainage for postsurgical fluid collections was 12 days (range: 1 to 49 days). Percutaneous drain placement achieved clinical success without any subsequent treatments in 73% (32/44) of patients. An additional 7% (3/44) of patients required subsequent image-guided aspiration procedures and/or percutaneous drain placements but avoided surgical treatment for a persistent fluid collection.

Conclusion: Postsurgical fluid collections after diverse breast surgeries represented the vast majority of the fluid collections referred to our academic practice for percutaneous drain placement. Percutaneous drain placement was a safe and effective treatment in this patient population.