A Deep Learning Decision Support Tool to Improve Risk Stratification and Reduce Unnecessary Biopsies in BI-RADS 4 Mammograms.

Purpose: To evaluate the performance of a biopsy decision support algorithmic model, the intelligent-augmented breast cancer risk calculator (iBRISK), on a multicenter patient dataset. Materials and Methods: iBRISK was previously developed by applying deep learning to clinical risk factors and mammographic descriptors from 9700 patient records at the primary institution and validated using another 1078 patients. All patients were seen from March 2006 to December 2016. In this multicenter study, iBRISK was further assessed on an independent, retrospective dataset (January 2015-June 2019) from three major health care institutions in Texas, with Breast Imaging Reporting and Data System (BI-RADS) category 4 lesions. Data were dichotomized and trichotomized to measure precision in risk stratification and probability of malignancy (POM) estimation. iBRISK score was also evaluated as a continuous predictor of malignancy, and cost savings analysis was performed. Results: The iBRISK model's accuracy was 89.5%, area under the receiver operating characteristic curve (AUC) was 0.93 (95% CI: 0.92, 0.95), sensitivity was 100%, and specificity was 81%. A total of 4209 women (median age, 56 years [IQR, 45-65 years]) were included in the multicenter dataset. Only two of 1228 patients (0.16%) in the "low" POM group had malignant lesions, while in the "high" POM group, the malignancy rate was 85.9%. iBRISK score as a continuous predictor of malignancy yielded an AUC of 0.97 (95% CI: 0.97, 0.98). Estimated potential cost savings were more than $420 million. Conclusion: iBRISK demonstrated high sensitivity in the malignancy prediction of BI-RADS 4 lesions. iBRISK may safely obviate biopsies in up to 50% of patients in low or moderate POM groups and reduce biopsy-associated costs.Keywords: Mammography, Breast, Oncology, Biopsy/Needle Aspiration, Radiomics, Precision Mammography, AI-augmented Biopsy Decision Support Tool, Breast Cancer Risk Calculator, BI-RADS 4 Mammography Risk Stratification, Overbiopsy Reduction, Probability of Malignancy (POM) Assessment, Biopsy-based Positive Predictive Value (PPV3) Supplemental material is available for this article. Published under a CC BY 4.0 license.See also the commentary by McDonald and Conant in this issue.

Efficacy and Safety of Half-Dose Gadopiclenol versus Full-Dose Gadobutrol for Contrast-enhanced Body MRI.

Background Gadopiclenol is a macrocyclic gadolinium-based contrast agent (GBCA) with higher relaxivity compared with standard GBCAs, potentially allowing gadolinium dose reduction without decreasing efficacy. Purpose To investigate whether gadopiclenol at 0.05 mmol/kg is noninferior to gadobutrol at 0.1 mmol/kg for lesion visualization in body MRI. Materials and Methods A randomized, double-blind, crossover, phase 3 study was conducted between August 2019 and December 2020 at 33 centers in 11 countries. Adults with at least one suspected focal lesion in one of three different body regions (head and neck; breast, thorax, abdomen, or pelvis; or musculoskeletal system) underwent two contrast-enhanced MRI examinations, randomized to start with either gadopiclenol or gadobutrol. MRI examinations were read by three blinded expert readers for each respective body region. Readers rated border delineation, internal morphologic characteristics, and visual contrast enhancement. Three additional blinded readers assessed reader preference. For safety analysis, adverse events were recorded. The differences between gadopiclenol- and gadobutrol-enhanced MRI in terms of lesion visualization were analyzed with a generalized linear mixed model using a two-sided paired t test. Results Among 273 participants (mean age, 57 years ± 13 [SD]; 162 women) who underwent both gadopiclenol- and gadobutrol-enhanced MRI and had at least one correlating lesion, 260 participants without major protocol deviations were analyzed for noninferiority. Gadopiclenol was noninferior to gadobutrol for all qualitative visualization parameters and for all readers (lower limit 95% CI of the difference of at least -0.10, which was above the noninferiority margin [-0.35]; P < .001). For most participants (75%-83% [206-228 of 276]), readers reported no preference between gadopiclenol- and gadobutrol-enhanced images. Adverse events did not differ in frequency, intensity, type, or association with GBCA injection (12 of 288 participants receiving gadopiclenol and 16 of 290 receiving gadobutrol). Conclusion Gadopiclenol at 0.05 mmol/kg was comparable with gadobutrol at 0.1 mmol/kg for lesion evaluation at contrast-enhanced body MRI and had a similar safety profile. Clinical trial registration no. NCT03986138 Published under a CC BY 4.0 license. Supplemental material is available for this article. See also the editorial by Bashir and Thomas in this issue.

Disparities in Cancer Genetic Testing and Variants of Uncertain Significance in the Hispanic Population of South Texas.

PURPOSE: Racial and ethnic disparities have included a lack of access to both genetic testing and research, resulting in poor understanding of the genomic architecture in under-represented populations. The South Texas population is primarily of Hispanic background and has been largely devoid of genetic services. We extended access to this underserved population and uncovered genetic variants previously not observed, emphasizing the need to continually improve both genomic databases and clarification of variant significance to provide meaningful patient counseling. METHODS: This study consisted of a retrospective cohort review of patients seen through a cancer genetics education and service program across 24 counties in South Texas. In total, 1,595 individuals were identified as appropriate for cancer genetic counseling and 1,377 completed genetic testing. RESULTS: Eighty percent of those receiving genetic counseling self-identified as Hispanic, 16% as non-Hispanic White (NHW), 3% as African American, and 1% as other race/ethnicity. Of reported variants, 18.8% were pathogenic and 13.7% were reported as a variant of uncertain significance (VUS). VUS was reported in 17.2% of the Hispanic individuals compared with 9% NHW (P = .005). CONCLUSION: Individuals of Hispanic ethnicity were significantly more likely to harbor a VUS compared with NHW. The extended reach into our regional communities revealed a gap in the ability to accurately interpret genomic variation with implications for advising patients on screening, prevention, and management strategies. A higher percentage of VUS also emphasizes the challenge of continued follow-up amid existing barriers that led to disparities in access. As understanding of the variants develops, hopefully gaps in knowledge of the genomic landscape will be lessened with increased clarity to provide accurate cancer risk assessment and recommendations for implementing prevention initiatives.

Optoacoustic Imaging and Gray-Scale US Features of Breast Cancers: Correlation with Molecular Subtypes.

Background Optoacoustic imaging can assess tumor hypoxia coregistered with US gray-scale images. The combination of optoacoustic imaging and US may have a role in distinguishing breast cancer molecular subtypes. Purpose To investigate whether optoacoustic US feature scores correlate with breast cancer molecular subtypes. Materials and Methods A total of 1972 women (with a total of 2055 breast masses) underwent prebiopsy optoacoustic US in a prospective multi-institutional study between December 2012 and September 2015. Seven readers blinded to pathologic diagnosis scored gray-scale US and optoacoustic US features of the known cancers. Optoacoustic US features within (internal) and outside of the tumor boundary (external) were scored. Immunohistochemistry findings were obtained from pathology reports. Multinomial logistic regression analysis was used to fit the US scores, adding optoacoustic US features to the model to investigate the incremental benefit of each feature. Kruskal-Wallis tests were used to analyze the relationship between molecular subtypes and feature scores. Results Among 653 invasive cancers identified in 629 women, a total of 532 cancers in 519 women, all of which had molecular markers available, were included in the analysis. Mean age ± standard deviation was 57.9 years ± 12.6. Mean total external optoacoustic US feature scores of luminal (A and B) breast cancers were higher (9.9 vs 8.8; P < .05) and total internal scores were lower (6.8 vs 7.7; P < .001) than those of triple-negative and human epidermal growth factor receptor 2-positive (HER2+) cancers. A multinomial logistic regression model showed that optoacoustic internal vessel (odds ratio [OR], 0.6; 95% confidence interval [CI]: 0.5, 0.8; P = .002), optoacoustic internal blush (OR, 0.7; 95% CI: 0.5, 0.9; P = .02), and optoacoustic internal hemoglobin (OR, 0.6; 95% CI: 0.5, 0.8; P = .001) were associated with classification of luminal versus triple-negative and HER2+ cancer subtypes. Conclusion Combined optoacoustic US imaging and gray-scale US features may help distinguish luminal breast cancers from triple-negative and human epidermal growth factor receptor 2-positive cancers. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Mann in this issue.

Controversies surrounding screening mammography.

Screening for breast cancer with mammography has been shown in multiple randomized control trials to decrease breast cancer mortality. Despite this, there are ongoing controversies regarding the benefits and risks of mammographic screening. Different professional groups have published conflicting guidelines regarding when to start screening, how often to screen, and when to stop screening which have added to the confusion. In this article we describe the evidence supporting and challenging the efficacy of mammography and review our real world approach to breast cancer screening.

Controversies in breast MRI.

When used for appropriate indications, breast magnetic resonance imaging (MRI) is a powerful diagnostic tool. However, breast MRI has its share of controversies. These controversies can be a source of confusion for the radiologist or referring physician. This paper addresses 4 breast MRI controversies that we frequently encounter at our university hospital practice: (1) what are the appropriate indications for screening breast MRI? (2) what are the appropriate indications for the use of breast MRI as a problems-solving modality? (3) how does one interpret MRI imaging features that have substantial overlap between benign and malignant conditions? and (4) what are the appropriate indications for preoperative breast MRI? Illustrative case examples are provided.

The satisfaction of Latina breast cancer survivors with their healthcare and health-related quality of life.

OBJECTIVE: The aim of this study was to evaluate the relationship of satisfaction with the cancer care doctor and health-related quality of life (HRQOL) among Latina breast cancer survivors (BCS) by (1) assessing whether satisfaction would be positively correlated with HRQOL and (2) assessing whether satisfaction would significantly influence HRQOL while controlling for covariates. METHODS: The cross-sectional study used self-report data from 117 Latina BCS. Satisfaction was measured with the Hall Satisfaction Index, and HRQOL was measured with the Functional Assessment of Cancer Therapy-General (FACT-G). Analyses included calculation of descriptive statistics, t tests, bivariate correlations, analyses of variance (ANOVAs), and multivariate analyses. RESULTS: Latina BCS had high satisfaction and generally good HRQOL. The Hall Satisfaction Index total score was positively associated with FACT-G functional well-being (r=0.265, p=0.004). Multivariate analyses showed that the Hall Satisfaction Index total score was a significant predictor of FACT-G functional well-being (p=0.012). Employment status was also a significant predictor, where being employed or retired resulted in better functional well-being than being unemployed. CONCLUSIONS: Latina BCS were quite satisfied with their cancer care doctors, and high levels of satisfaction with the cancer care doctor influenced functional well-being when confounding variables were controlled. Despite reportedly high satisfaction, Latina BCS did report barriers to satisfaction that could be considered cultural. Implications are discussed.

Radiobiological characterization of post-lumpectomy focal brachytherapy with lipid nanoparticle-carried radionuclides.

Post-operative radiotherapy has commonly been used for early stage breast cancer to treat residual disease. The primary objective of this work was to characterize, through dosimetric and radiobiological modeling, a novel focal brachytherapy technique which uses direct intracavitary infusion of ß-emitting radionuclides (186Re/188Re) carried by lipid nanoparticles (liposomes). Absorbed dose calculations were performed for a spherical lumpectomy cavity with a uniformly injected activity distribution using a dose point kernel convolution technique. Radiobiological indices were used to relate predicted therapy outcome and normal tissue complication of this technique with equivalent external beam radiotherapy treatment regimens. Modeled stromal damage was used as a measure of the inhibition of the stimulatory effect on tumor growth driven by the wound healing response. A sample treatment plan delivering 50 Gy at a therapeutic range of 2.0 mm for 186Re-liposomes and 5.0 mm for 188Re-liposomes takes advantage of the dose delivery characteristics of the ß-emissions, providing significant EUD (58.2 Gy and 72.5 Gy for 186Re and 188Re, respectively) with a minimal NTCP (0.046%) of the healthy ipsilateral breast. Modeling of kidney BED and ipsilateral breast NTCP showed that large injected activity concentrations of both radionuclides could be safely administered without significant complications.

Postlumpectomy focal brachytherapy for simultaneous treatment of surgical cavity and draining lymph nodes.

PURPOSE: The primary objective was to investigate a novel focal brachytherapy technique using lipid nanoparticle (liposome)-carried ß-emitting radionuclides (rhenium-186 [(186)Re]/rhenium-188 [(188)Re]) to simultaneously treat the postlumpectomy surgical cavity and draining lymph nodes. METHODS AND MATERIALS: Cumulative activity distributions in the lumpectomy cavity and lymph nodes were extrapolated from small animal imaging and human lymphoscintigraphy data. Absorbed dose calculations were performed for lumpectomy cavities with spherical and ellipsoidal shapes and lymph nodes within human subjects by use of the dose point kernel convolution method. RESULTS: Dose calculations showed that therapeutic dose levels within the lumpectomy cavity wall can cover 2- and 5-mm depths for (186)Re and (188)Re liposomes, respectively. The absorbed doses at 1 cm sharply decreased to only 1.3% to 3.7% of the doses at 2 mm for (186)Re liposomes and 5 mm for (188)Re liposomes. Concurrently, the draining sentinel lymph nodes would receive a high focal therapeutic absorbed dose, whereas the average dose to 1 cm of surrounding tissue received less than 1% of that within the nodes. CONCLUSIONS: Focal brachytherapy by use of (186)Re/(188)Re liposomes was theoretically shown to be capable of simultaneously treating the lumpectomy cavity wall and draining sentinel lymph nodes with high absorbed doses while significantly lowering dose to surrounding healthy tissue. In turn, this allows for dose escalation to regions of higher probability of containing residual tumor cells after lumpectomy while reducing normal tissue complications.

Post-lumpectomy intracavitary retention and lymph node targeting of (99m)Tc-encapsulated liposomes in nude rats with breast cancer xenograft.

Liposomes are recognized drug delivery systems with tumor-targeting capability. In addition, therapeutic or diagnostic radionuclides can be efficiently loaded into liposomes. This study investigated the feasibility of utilizing radiotherapeutic liposomes as a new post-lumpectomy radiotherapy for early-stage breast cancer by determining the locoregional retention and systemic distribution of liposomes radiolabeled with technetium-99m ((99m)Tc) in an orthotopic MDA-MB-231 breast cancer xenograft nude rat model. To test this new brachytherapy approach, a positive surgical margin lumpectomy model was set up by surgically removing the xenograft and deliberately leaving a small tumor remnant in the surgical cavity. Neutral, anionic, and cationic surface-charged fluorescent liposomes of 100 and 400 nm diameter were manufactured and labeled with (99m)Tc-BMEDA. Locoregional retention and systemic distribution of (99m)Tc-liposomes injected into the post-lumpectomy cavity were determined using non-invasive nuclear imaging, ex vivo tissue gamma counting and fluorescent stereomicroscopic imaging. The results indicated that (99)Tc-liposomes were effectively retained in the surgical cavity (average retention was 55.7 ± 24.2% of injected dose for all rats at 44 h post-injection) and also accumulated in the tumor remnant (66.9 ± 100.4%/g for all rats). The majority of cleared (99m)Tc was metabolized quickly and excreted into feces and urine, exerting low radiation burden on vital organs. In certain animals (99m)Tc-liposomes significantly accumulated in the peripheral lymph nodes, especially 100 nm liposomes with anionic surface charge. The results suggest that post-lumpectomy intracavitary administration of therapeutic radionuclides delivered by 100-nm anionic liposome carrier is a potential therapy for the simultaneous treatment of the surgical cavity and the draining lymph nodes of early-stage breast cancer.

Computed tomography-generated templates: a new approach to frontal sinus osteoplastic flap surgery.

BACKGROUND: The aim of this study was to present a new technique for frontal sinus template creation used in osteoplastic flap surgery, obviating the need for the traditional 6-ft Caldwell radiograph. METHODS: Seventeen human cadaver skulls were studied prospectively. Using digital addition algorithms of sequential coronal computed tomography (CT) images of the frontal sinus to create a composite image, an image is printed to 1 cm = 1 cm scale. A frontal sinus template was created using this processed image and compared with a control 6-ft Caldwell radiograph. RESULTS: The mean variance between corresponding vertical points on the templates was 0.71 mm (95% confidence interval, 0.56-0.87 mm). The mean variance between horizontal points was 1.11 mm (95% confidence interval, 0.97-1.28 mm). The mean difference in width was 2.46 mm. The plain film templates were always wider than the CT-generated templates. CONCLUSION: The CT-generated frontal sinus templates are virtually identical to the Caldwell radiograph-derived templates. They may obviate the need for additional imaging and minimize the potential errors commonly associated with 6-ft Caldwell templates.

Analysis of sentinel lymph node mapping with immediate pathologic review in patients receiving preoperative chemotherapy for breast carcinoma.

BACKGROUND: Sentinel lymph node mapping (SLNM) and neoadjuvant chemotherapy are becoming established components of therapy for selected patients with breast carcinoma. However, neoadjuvant therapy has been considered a relative contraindication to SLNM. In an effort to learn whether patients who have received preoperative chemotherapy can undergo accurate SLNM, we evaluated our experience with this technique. METHODS: From January 1997 to June 2000, SLNM and axillary lymph node dissection were concurrently performed in 35 patients who received preoperative chemotherapy. Mapping was performed with (99m)Tc sulfur colloid only in one patient and Lymphazurin dye only in 15 patients, and the two methods were combined in the remainder. RESULTS: SLNM successfully identified a sentinel lymph node in 30 (86%) patients. Metastatic disease was identified in the sentinel lymph nodes of four patients during surgery. The intraoperative pathologic diagnosis proved to be correct in 19 (79%) of 24 patients. The final pathologic diagnosis of the sentinel lymph node reflected the status of the axillary contents in all patients in whom it was identified. CONCLUSIONS: These results demonstrate that SLNM can be consistently performed in patients receiving preoperative chemotherapy for breast cancer, suggesting the utility of this technique in this patient population.