[WITHDRAWN] Predicting breast cancer with AI for individual risk-adjusted MRI screening and early detection.

This paper has been withdrawn by Lukas Hirsch. Major revisions and rewriting in progress.

Population-Level Access to Breast Cancer Early Detection and Diagnosis in Nigeria.

PURPOSE: Mammography, breast ultrasound (US), and US-guided breast biopsy are essential services for breast cancer early detection and diagnosis. This study undertook a comprehensive evaluation to determine population-level access to these services for breast cancer early detection and diagnosis in Nigeria using a previously validated geographic information system (GIS) model. METHODS: A comprehensive list of public and private facilities offering mammography, breast US, and US-guided breast biopsy was compiled using publicly available facility data and a survey administered nationally to Nigerian radiologists. All facilities were geolocated. A cost-distance model using open-source population density (GeoData Institute) and road network data (OpenStreetMap) was used to estimate population-level travel time to the nearest facility for mammography, breast US, and US-guided biopsy using GIS software (ArcMAP). RESULTS: In total, 1,336 facilities in Nigeria provide breast US, of which 47.8% (639 of 1,336) are public facilities, and 218 provide mammography, of which 45.4% (99 of 218) are public facilities. Of the facilities that provide breast US, only 2.5% (33 of 1,336) also provide US-guided breast biopsy. At the national level, 83.1% have access to either US or mammography and 61.7% have access to US-guided breast biopsy within 120 minutes of a continuous one-way travel. There are differences in access to mammography (64.8% v 80.6% with access at 120 minutes) and US-guided breast biopsy (49.0% v 77.1% with access at 120 minutes) between the northern and southern Nigeria and between geopolitical zones. CONCLUSION: To our knowledge, this is the first comprehensive evaluation of breast cancer detection and diagnostic services in Nigeria, which demonstrates geospatial inequalities in access to mammography and US-guided biopsy. Targeted investment is needed to improve access to these essential cancer care services in the northern region and the North East geopolitical zone.

Automated Triage of Screening Breast MRI Examinations in High-Risk Women Using an Ensemble Deep Learning Model.

OBJECTIVES: The aim of the study is to develop and evaluate the performance of a deep learning (DL) model to triage breast magnetic resonance imaging (MRI) findings in high-risk patients without missing any cancers. MATERIALS AND METHODS: In this retrospective study, 16,535 consecutive contrast-enhanced MRIs performed in 8354 women from January 2013 to January 2019 were collected. From 3 New York imaging sites, 14,768 MRIs were used for the training and validation data set, and 80 randomly selected MRIs were used for a reader study test data set. From 3 New Jersey imaging sites, 1687 MRIs (1441 screening MRIs and 246 MRIs performed in recently diagnosed breast cancer patients) were used for an external validation data set. The DL model was trained to classify maximum intensity projection images as "extremely low suspicion" or "possibly suspicious." Deep learning model evaluation (workload reduction, sensitivity, specificity) was performed on the external validation data set, using a histopathology reference standard. A reader study was performed to compare DL model performance to fellowship-trained breast imaging radiologists. RESULTS: In the external validation data set, the DL model triaged 159/1441 of screening MRIs as "extremely low suspicion" without missing a single cancer, yielding a workload reduction of 11%, a specificity of 11.5%, and a sensitivity of 100%. The model correctly triaged 246/246 (100% sensitivity) of MRIs in recently diagnosed patients as "possibly suspicious." In the reader study, 2 readers classified MRIs with a specificity of 93.62% and 91.49%, respectively, and missed 0 and 1 cancer, respectively. On the other hand, the DL model classified MRIs with a specificity of 19.15% and missed 0 cancers, highlighting its potential use not as an independent reader but as a triage tool. CONCLUSIONS: Our automated DL model triages a subset of screening breast MRIs as "extremely low suspicion" without misclassifying any cancer cases. This tool may be used to reduce workload in standalone mode, to shunt low suspicion cases to designated radiologists or to the end of the workday, or to serve as base model for other downstream AI tools.

Increased tumor glycolysis is associated with decreased immune infiltration across human solid tumors.

Response to immunotherapy across multiple cancer types is approximately 25%, with some tumor types showing increased response rates compared to others (i.e. response rates in melanoma and non-small cell lung cancer (NSCLC) are typically 30-60%). Patients whose tumors are resistant to immunotherapy often lack high levels of pre-existing inflammation in the tumor microenvironment. Increased tumor glycolysis, acting through glucose deprivation and lactic acid accumulation, has been shown to have pleiotropic immune suppressive effects using in-vitro and in-vivo models of disease. To determine whether the immune suppressive effect of tumor glycolysis is observed across human solid tumors, we analyzed glycolytic and immune gene expression patterns in multiple solid malignancies. We found that increased expression of a glycolytic signature was associated with decreased immune infiltration and a more aggressive disease across multiple tumor types. Radiologic and pathologic analysis of untreated estrogen receptor (ER)-negative breast cancers corroborated these observations, and demonstrated that protein expression of glycolytic enzymes correlates positively with glucose uptake and negatively with infiltration of CD3+ and CD8+ lymphocytes. This study reveals an inverse relationship between tumor glycolysis and immune infiltration in a large cohort of multiple solid tumor types.

Radiologist-Level Performance by Using Deep Learning for Segmentation of Breast Cancers on MRI Scans.

PURPOSE: To develop a deep network architecture that would achieve fully automated radiologist-level segmentation of cancers at breast MRI. MATERIALS AND METHODS: In this retrospective study, 38 229 examinations (composed of 64 063 individual breast scans from 14 475 patients) were performed in female patients (age range, 12-94 years; mean age, 52 years ± 10 [standard deviation]) who presented between 2002 and 2014 at a single clinical site. A total of 2555 breast cancers were selected that had been segmented on two-dimensional (2D) images by radiologists, as well as 60 108 benign breasts that served as examples of noncancerous tissue; all these were used for model training. For testing, an additional 250 breast cancers were segmented independently on 2D images by four radiologists. Authors selected among several three-dimensional (3D) deep convolutional neural network architectures, input modalities, and harmonization methods. The outcome measure was the Dice score for 2D segmentation, which was compared between the network and radiologists by using the Wilcoxon signed rank test and the two one-sided test procedure. RESULTS: The highest-performing network on the training set was a 3D U-Net with dynamic contrast-enhanced MRI as input and with intensity normalized for each examination. In the test set, the median Dice score of this network was 0.77 (interquartile range, 0.26). The performance of the network was equivalent to that of the radiologists (two one-sided test procedures with radiologist performance of 0.69-0.84 as equivalence bounds, P < .001 for both; n = 250). CONCLUSION: When trained on a sufficiently large dataset, the developed 3D U-Net performed as well as fellowship-trained radiologists in detailed 2D segmentation of breast cancers at routine clinical MRI.Keywords: MRI, Breast, Segmentation, Supervised Learning, Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning AlgorithmsPublished under a CC BY 4.0 license. Supplemental material is available for this article.

Breast MRI Background Parenchymal Enhancement Categorization Using Deep Learning: Outperforming the Radiologist.

BACKGROUND: Background parenchymal enhancement (BPE) is assessed on breast MRI reports as mandated by the Breast Imaging Reporting and Data System (BI-RADS) but is prone to inter and intrareader variation. Semiautomated and fully automated BPE assessment tools have been developed but none has surpassed radiologist BPE designations. PURPOSE: To develop a deep learning model for automated BPE classification and to compare its performance with current standard-of-care radiology report BPE designations. STUDY TYPE: Retrospective. POPULATION: Consecutive high-risk patients (i.e. >20% lifetime risk of breast cancer) who underwent contrast-enhanced screening breast MRI from October 2013 to January 2019. The study included 5224 breast MRIs, divided into 3998 training, 444 validation, and 782 testing exams. On radiology reports, 1286 exams were categorized as high BPE (i.e., marked or moderate) and 3938 as low BPE (i.e., mild or minimal). FIELD STRENGTH/SEQUENCE: A 1.5 T or 3 T system; one precontrast and three postcontrast phases of fat-saturated T1-weighted dynamic contrast-enhanced imaging. ASSESSMENT: Breast MRIs were used to develop two deep learning models (Slab artificial intelligence (AI); maximum intensity projection [MIP] AI) for BPE categorization using radiology report BPE labels. Models were tested on a heldout test sets using radiology report BPE and three-reader averaged consensus as the reference standards. STATISTICAL TESTS: Model performance was assessed using receiver operating characteristic curve analysis. Associations between high BPE and BI-RADS assessments were evaluated using McNemar's chi-square test (α* = 0.025). RESULTS: The Slab AI model significantly outperformed the MIP AI model across the full test set (area under the curve of 0.84 vs. 0.79) using the radiology report reference standard. Using three-reader consensus BPE labels reference standard, our AI model significantly outperformed radiology report BPE labels. Finally, the AI model was significantly more likely than the radiologist to assign "high BPE" to suspicious breast MRIs and significantly less likely than the radiologist to assign "high BPE" to negative breast MRIs. DATA CONCLUSION: Fully automated BPE assessments for breast MRIs could be more accurate than BPE assessments from radiology reports. LEVEL OF EVIDENCE: 4 TECHNICAL EFFICACY STAGE: 3.

The impact of scaling up access to treatment and imaging modalities on global disparities in breast cancer survival: a simulation-based analysis.

BACKGROUND: Female breast cancer is the most commonly diagnosed cancer in the world, with wide variations in reported survival by country. Women in low-income and middle-income countries (LMICs) in particular face several barriers to breast cancer services, including diagnostics and treatment. We aimed to estimate the potential impact of scaling up the availability of treatment and imaging modalities on breast cancer survival globally, together with improvements in quality of care. METHODS: For this simulation-based analysis, we used a microsimulation model of global cancer survival, which accounts for the availability and stage-specific survival impact of specific treatment modalities (chemotherapy, radiotherapy, surgery, and targeted therapy), imaging modalities (ultrasound, x-ray, CT, MRI, PET, and single-photon emission computed tomography [SPECT]), and quality of cancer care, to simulate 5-year net survival for women with newly diagnosed breast cancer in 200 countries and territories in 2018. We calibrated the model to empirical data on 5-year net breast cancer survival in 2010-14 from CONCORD-3. We evaluated the potential impact of scaling up specific imaging and treatment modalities and quality of care to the mean level of high-income countries, individually and in combination. We ran 1000 simulations for each policy intervention and report the means and 95% uncertainty intervals (UIs) for all model outcomes. FINDINGS: We estimate that global 5-year net survival for women diagnosed with breast cancer in 2018 was 67·9% (95% UI 62·9-73·4) overall, with an almost 25-times difference between low-income (3·5% [0·4-10·0]) and high-income (87·0% [85·6-88·4]) countries. Among individual treatment modalities, scaling up access to surgery alone was estimated to yield the largest survival gains globally (2·7% [95% UI 0·4-8·3]), and scaling up CT alone would have the largest global impact among imaging modalities (0·5% [0·0-2·0]). Scaling up a package of traditional modalities (surgery, chemotherapy, radiotherapy, ultrasound, and x-ray) could improve global 5-year net survival to 75·6% (95% UI 70·6-79·4), with survival in low-income countries improving from 3·5% (0·4-10·0) to 28·6% (4·9-60·1). Adding concurrent improvements in quality of care could further improve global 5-year net survival to 78·2% (95% UI 74·9-80·4), with a substantial impact in low-income countries, improving net survival to 55·3% (42·2-67·8). Comprehensive scale-up of access to all modalities and improvements in quality of care could improve global 5-year net survival to 82·3% (95% UI 79·3-85·0). INTERPRETATION: Comprehensive scale-up of treatment and imaging modalities, and improvements in quality of care could improve global 5-year net breast cancer survival by nearly 15 percentage points. Scale-up of traditional modalities and quality-of-care improvements could achieve 70% of these total potential gains, with substantial impact in LMICs, providing a more feasible pathway to improving breast cancer survival in these settings even without the benefits of future investments in targeted therapy and advanced imaging. FUNDING: Harvard T H Chan School of Public Health, and National Cancer Institute P30 Cancer Center Support Grant to Memorial Sloan Kettering Cancer Center.

Tumor-Nipple Distance of ≥ 1 cm Predicts Negative Nipple Pathology After Neoadjuvant Chemotherapy.

BACKGROUND: As neoadjuvant chemotherapy (NAC) for breast cancer has become more widely used, so has nipple-sparing mastectomy. A common criterion for eligibility is a 1 cm tumor-to-nipple distance (TND), but its suitability after NAC is unclear. In this study, we examined factors predictive of negative nipple pathologic status (NS-) in women undergoing total mastectomy after NAC. METHODS: Women with invasive breast cancer treated with NAC and total mastectomy from August 2014 to April 2018 at our institution were retrospectively identified. Following review of pre- and post-NAC magnetic resonance imaging (MRI) and mammograms, the association of clinicopathologic and imaging variables with NS- was examined and the accuracy of 1 cm TND on imaging for predicting NS- was determined. RESULTS: Among 175 women undergoing 179 mastectomies, 74% of tumors were cT1-T2 and 67% were cN+ on pre-NAC staging; 10% (18/179) had invasive or in situ carcinoma in the nipple on final pathology. On multivariable analysis, after adjusting for age, grade, and tumor stage, three factors, namely number of positive nodes, pre-NAC nipple-areolar complex retraction, and decreasing TND, were significant predictors of nipple involvement (p < 0.05). The likelihood of NS- was higher with increasing TND on pre- and post-NAC imaging (p < 0.05). TND ≥ 1 cm predicted NS- in 97% and 95% of breasts on pre- and post-NAC imaging, respectively. CONCLUSIONS: Increasing TND was associated with a higher likelihood of NS-. A TND ≥ 1 cm on pre- or post-NAC imaging is highly predictive of NS- and could be used to determine eligibility for nipple-sparing mastectomy after NAC.

Multispectral Imaging for Metallic Biopsy Marker Detection During MRI-Guided Breast Biopsy: A Feasibility Study for Clinical Translation.

PURPOSE: To assess the feasibility and diagnostic accuracy of multispectral MRI (MSI) in the detection and localization of biopsy markers during MRI-guided breast biopsy. METHODS: This prospective study included 20 patients undergoing MR-guided breast biopsy. In 10 patients (Group 1), MSI was acquired following tissue sampling and biopsy marker deployment. In the other 10 patients (Group 2), MSI was acquired following tissue sampling but before biopsy marker deployment (to simulate deployment failure). All patients received post-procedure mammograms. Group 1 and Group 2 designations, in combination with the post-procedure mammogram, served as the reference standard. The diagnostic performance of MSI for biopsy marker identification was independently evaluated by two readers using two-spectral-bin MR and one-spectral-bin MR. The κ statistic was used to assess inter-rater agreement for biopsy marker identification. RESULTS: The sensitivity, specificity, and accuracy of biopsy marker detection for readers 1 and 2 using 2-bin MSI were 90.0% (9/10) and 90.0% (9/10), 100.0% (10/10) and 100.0% (10/10), 95.0% (19/20) and 95.0% (19/20); and using 1-bin MSI were 70.0% (7/10) and 80.0% (8/10), 100.0% (8/8) and 100.0% (10/10), 85.0% (17/20) and 90.0% (18/20). Positive predictive value was 100% for both readers for all numbers of bins. Inter-rater agreement was excellent: κ was 1.0 for 2-bin MSI and 0.81 for 1-bin MSI. CONCLUSION: MSI is a feasible, diagnostically accurate technique for identifying metallic biopsy markers during MRI-guided breast biopsy and may eliminate the need for a post-procedure mammogram.

Accuracy of Magnetic Resonance Imaging-Guided Biopsy to Verify Breast Cancer Pathologic Complete Response After Neoadjuvant Chemotherapy: A Nonrandomized Controlled Trial.

Importance: After neoadjuvant chemotherapy (NAC), pathologic complete response (pCR) is an optimal outcome and a surrogate end point for improved disease-free and overall survival. To date, surgical resection remains the only reliable method for diagnosing pCR. Objective: To evaluate the accuracy of magnetic resonance imaging (MRI)-guided biopsy for diagnosing a pCR after NAC compared with reference-standard surgical resection. Design, Setting, and Participants: Single-arm, phase 1, nonrandomized controlled trial in a single tertiary care cancer center from September 26, 2017, to July 29, 2019. The median follow-up was 1.26 years (interquartile range, 0.85-1.59 years). Data analysis was performed in November 2019. Eligible patients had (1) stage IA to IIIC biopsy-proven operable invasive breast cancer; (2) standard-of-care NAC; (3) MRI before and after NAC, with imaging complete response defined as no residual enhancement on post-NAC MRI; and (4) definitive surgery. Patients were excluded if they were younger than 18 years, had a medical reason precluding study participation, or had a prior history of breast cancer. Interventions: Post-NAC MRI-guided biopsy without the use of intravenous contrast of the tumor bed before definitive surgery. Main Outcomes and Measures: The primary end point was the negative predictive value of MRI-guided biopsy, with true-negative defined as negative results of the biopsy (ie, no residual cancer) corresponding to a surgical pCR. Accuracy, sensitivity, positive predictive value, and specificity were also calculated. Two clinical definitions of pCR were independently evaluated: definition 1 was no residual invasive cancer; definition 2, no residual invasive or in situ cancer. Results: Twenty of 23 patients (87%) had evaluable data (median [interquartile range] age, 51.5 [39.0-57.5] years; 20 women [100%]; 13 White patients [65%]). Of the 20 patients, pre-NAC median tumor size on MRI was 3.0 cm (interquartile range, 2.0-5.0 cm). Nineteen of 20 patients (95%) had invasive ductal carcinoma; 15 of 20 (75%) had stage II cancer; 11 of 20 (55%) had ERBB2 (formerly HER2 or HER2/neu)-positive cancer; and 6 of 20 (30%) had triple-negative cancer. Surgical pathology demonstrated a pCR in 13 of 20 (65%) patients and no pCR in 7 of 20 patients (35%) when pCR definition 1 was used. Results of MRI-guided biopsy had a negative predictive value of 92.8% (95% CI, 66.2%-99.8%), with accuracy of 95% (95% CI, 75.1%-99.9%), sensitivity of 85.8% (95% CI, 42.0%-99.6%), positive predictive value of 100%, and specificity of 100% for pCR definition 1. Only 1 patient had a false-negative MRI-guided biopsy result (surgical pathology showed <0.02 cm of residual invasive cancer). Conclusions and Relevance: This study's results suggest that the accuracy of MRI-guided biopsy to diagnose a post-NAC pCR approaches that of reference-standard surgical resection. MRI-guided biopsy may be a viable alternative to surgical resection for this population after NAC, which supports the need for further investigation. Trial Registration: ClinicalTrials.gov Identifier: NCT03289195.

Designing Participatory Needs Assessments to Support Global Health Interventions in Time-Limited Settings: A Case Study From Nigeria.

Social scientists have advocated for the use of participatory research methods for Global Health project design and planning. However, community-engaged approaches can be time and resource-intensive. This article proposes a feasible framework for conducting a participatory needs assessment in time-limited settings using multiple, triangulated qualitative methods. This framework is outlined through a case study: a participatory needs assessment to inform the design of an ultrasound-guided biopsy training program in Nigeria. Breast cancer is the leading cause of death for Nigerian women and most cases in Nigeria are diagnosed at an advanced stage; timely diagnosis is impeded by fractious referral pathways, costly imaging equipment, and limited access outside urban centers. The project involved participant observation, surveys, and focus groups at the African Research Group for Oncology (ARGO) in Ile-Ife, Nigeria. Through this timely research and engagement, participants spoke about diagnostic challenges, institutional power dynamics, and infrastructure considerations for program implementation.

Using Deep Learning to Improve Nonsystematic Viewing of Breast Cancer on MRI.

OBJECTIVE: To investigate the feasibility of using deep learning to identify tumor-containing axial slices on breast MRI images. METHODS: This IRB-approved retrospective study included consecutive patients with operable invasive breast cancer undergoing pretreatment breast MRI between January 1, 2014, and December 31, 2017. Axial tumor-containing slices from the first postcontrast phase were extracted. Each axial image was subdivided into two subimages: one of the ipsilateral cancer-containing breast and one of the contralateral healthy breast. Cases were randomly divided into training, validation, and testing sets. A convolutional neural network was trained to classify subimages into "cancer" and "no cancer" categories. Accuracy, sensitivity, and specificity of the classification system were determined using pathology as the reference standard. A two-reader study was performed to measure the time savings of the deep learning algorithm using descriptive statistics. RESULTS: Two hundred and seventy-three patients with unilateral breast cancer met study criteria. On the held-out test set, accuracy of the deep learning system for tumor detection was 92.8% (648/706; 95% confidence interval: 89.7%-93.8%). Sensitivity and specificity were 89.5% and 94.3%, respectively. Readers spent 3 to 45 seconds to scroll to the tumor-containing slices without use of the deep learning algorithm. CONCLUSION: In breast MR exams containing breast cancer, deep learning can be used to identify the tumor-containing slices. This technology may be integrated into the picture archiving and communication system to bypass scrolling when viewing stacked images, which can be helpful during nonsystematic image viewing, such as during interdisciplinary tumor board meetings.

Multistakeholder Needs Assessment to Inform the Development of an mHealth-Based Ultrasound-Guided Breast Biopsy Training Program in Nigeria.

PURPOSE: The incidence of breast cancer is rising in Nigeria, and one major barrier to care is the lack of affordable and appropriate breast cancer diagnosis by ultrasound (US)-guided biopsy. The prohibitive cost of US devices limits their availability in low- and middle-income countries. The emergence of mobile health (mHealth) imaging devices may offer an acceptable low-cost alternative. The purpose of this research was to perform a comprehensive needs assessment to understand knowledge, use, training needs, and attitudes as regards image-guided biopsy in Nigeria to inform the development of an mHealth-based US-guided biopsy training program. METHODS: A multistakeholder needs assessment was conducted at the Sixth Annual African Research Group for Oncology Symposium. Voluntary anonymous surveys were administered to all attendees. A subset of attendees (ie, surgeons, radiologists, pathologists, and nurses) participated in six focus groups. Survey items and interview guides were developed collaboratively with local and international input. RESULTS: Surveys focusing on use, training needs, and attitudes regarding US-guided biopsies were completed with a 55% response rate (n = 54 of 98) among participants from 22 hospitals across Nigeria. Respondents expressed dissatisfaction with the way breast biopsies were currently performed at their hospitals and high interest in having their institution participate in a US-guided biopsy training program. Focus group participants (n = 37) identified challenges to performing US-guided procedures, including equipment functionality and cost, staff training, and access to consumables. Groups brainstormed the design of an mHealth US-guided biopsy training program, preferring a train-the-trainer format combining in-person teaching with independent modules. CONCLUSION: A multidisciplinary needs assessment of local stakeholders identified a need for and acceptability of an mHealth-based US-guided biopsy training program in Nigeria.

Regional Lymph Node Involvement Among Patients With De Novo Metastatic Breast Cancer.

Importance: Regional nodal irradiation (RNI) for node-positive breast cancer reduces distant metastases and improves survival, albeit with limited reduction in regional nodal recurrences. The mechanism by which RNI robustly reduces distant metastases while modestly influencing nodal recurrences (ie, the presumed target of RNI) remains unclear. Objective: To determine whether some distant metastases putatively arise from occult regional nodal disease and whether regional recurrences otherwise remain largely undetected until an advanced cancer presentation. Design, Setting, and Participants: This cohort study examined patients presenting with de novo stage IV breast cancer to the Memorial Sloan Kettering Cancer Center in New York, New York, from 2006 to 2018. Medical records were reviewed to ascertain clinicopathological parameters, including estrogen receptor status and survival. Pretreatment positron emission tomography-computed tomography (PET-CT) imaging was reviewed to ascertain the extent of regional nodal involvement at metastatic diagnosis using standard nodal assessment criteria. A subset underwent regional lymph node biopsy for diagnostic confirmation and served to validate the radiographic nodal assessment. Data analysis was performed from October 2019 to February 2020. Exposures: Untreated metastatic breast cancer. Main Outcome and Measures: The primary outcome was the likelihood of regional nodal involvement at the time of metastatic breast cancer presentation and was determined by reviewing pretreatment PET-CT imaging and lymph node biopsy findings. Results: Among 597 women (median [interquartile range] age, 53 [44-65] years) with untreated metastatic breast cancer, 512 (85.8%) exhibited regional lymph node involvement by PET-CT or nodal biopsy, 509 (85%) had involvement of axillary level I, 328 (55%) had involvement in axillary level II, 136 (23%) had involvement in axillary level III, 101 (17%) had involvement in the supraclavicular fossa, and 96 (16%) had involvement in the internal mammary chain. Lymph node involvement was more prevalent among estrogen receptor-negative tumors (92.4%) than estrogen receptor-positive tumors (83.6%). Nodal involvement at the time of metastatic diagnosis was not associated with overall survival. Conclusions and Relevance: These findings suggest that a majority of patients with de novo metastatic breast cancer harbor regional lymph node disease at presentation, consistent with the hypothesis that regional involvement may precede metastatic dissemination. This is in alignment with the findings of landmark trials suggesting that RNI reduces distant recurrences. It is possible that this distant effect of RNI may act via eradication of occult regional disease prior to systemic seeding. The challenges inherent in detecting isolated nodal disease (which is typically asymptomatic) may account for the more modest observed benefit of RNI on regional recurrences. Alternative explanations of nodal involvement that arises concurrently or after metastatic dissemination remain possible, but do not otherwise explain the association of RNI with distant recurrence.

Harmonization of Quantitative Parenchymal Enhancement in T1 -Weighted Breast MRI.

BACKGROUND: Differences in imaging parameters influence computer-extracted parenchymal enhancement measures from breast MRI. PURPOSE: To investigate the effect of differences in dynamic contrast-enhanced MRI acquisition parameter settings on quantitative parenchymal enhancement of the breast, and to evaluate harmonization of contrast-enhancement values with respect to flip angle and repetition time. STUDY TYPE: Retrospective. PHANTOM/POPULATIONS: We modeled parenchymal enhancement using simulations, a phantom, and two cohorts (N = 398 and N = 302) from independent cancer centers. SEQUENCE FIELD/STRENGTH: 1.5T dynamic contrast-enhanced T1 -weighted spoiled gradient echo MRI. Vendors: Philips, Siemens, General Electric Medical Systems. ASSESSMENT: We assessed harmonization of parenchymal enhancement in simulations and phantom by varying the MR parameters that influence the amount of T1 -weighting: flip angle (8°-25°) and repetition time (4-12 msec). We calculated the median and interquartile range (IQR) of the enhancement values before and after harmonization. In vivo, we assessed overlap of quantitative parenchymal enhancement in the cohorts before and after harmonization using kernel density estimations. Cohort 1 was scanned with flip angle 20° and repetition time 8 msec; cohort 2 with flip angle 10° and repetition time 6 msec. STATISTICAL TESTS: Paired Wilcoxon signed-rank-test of bootstrapped kernel density estimations. RESULTS: Before harmonization, simulated enhancement values had a median (IQR) of 0.46 (0.34-0.49). After harmonization, the IQR was reduced: median (IQR): 0.44 (0.44-0.45). In the phantom, the IQR also decreased, median (IQR): 0.96 (0.59-1.22) before harmonization, 0.96 (0.91-1.02) after harmonization. Harmonization yielded significantly (P < 0.001) better overlap in parenchymal enhancement between the cohorts: median (IQR) was 0.46 (0.37-0.58) for cohort 1 vs. 0.37 (0.30-0.44) for cohort 2 before harmonization (57% overlap); and 0.35 (0.28-0.43) vs. .0.37 (0.30-0.44) after harmonization (85% overlap). DATA CONCLUSION: The proposed practical harmonization method enables an accurate comparison between patients scanned with differences in imaging parameters. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY STAGE: 4.

A machine learning model that classifies breast cancer pathologic complete response on MRI post-neoadjuvant chemotherapy.

BACKGROUND: For breast cancer patients undergoing neoadjuvant chemotherapy (NAC), pathologic complete response (pCR; no invasive or in situ) cannot be assessed non-invasively so all patients undergo surgery. The aim of our study was to develop and validate a radiomics classifier that classifies breast cancer pCR post-NAC on MRI prior to surgery. METHODS: This retrospective study included women treated with NAC for breast cancer from 2014 to 2016 with (1) pre- and post-NAC breast MRI and (2) post-NAC surgical pathology report assessing response. Automated radiomics analysis of pre- and post-NAC breast MRI involved image segmentation, radiomics feature extraction, feature pre-filtering, and classifier building through recursive feature elimination random forest (RFE-RF) machine learning. The RFE-RF classifier was trained with nested five-fold cross-validation using (a) radiomics only (model 1) and (b) radiomics and molecular subtype (model 2). Class imbalance was addressed using the synthetic minority oversampling technique. RESULTS: Two hundred seventy-three women with 278 invasive breast cancers were included; the training set consisted of 222 cancers (61 pCR, 161 no-pCR; mean age 51.8 years, SD 11.8), and the independent test set consisted of 56 cancers (13 pCR, 43 no-pCR; mean age 51.3 years, SD 11.8). There was no significant difference in pCR or molecular subtype between the training and test sets. Model 1 achieved a cross-validation AUROC of 0.72 (95% CI 0.64, 0.79) and a similarly accurate (P = 0.1) AUROC of 0.83 (95% CI 0.71, 0.94) in both the training and test sets. Model 2 achieved a cross-validation AUROC of 0.80 (95% CI 0.72, 0.87) and a similar (P = 0.9) AUROC of 0.78 (95% CI 0.62, 0.94) in both the training and test sets. CONCLUSIONS: This study validated a radiomics classifier combining radiomics with molecular subtypes that accurately classifies pCR on MRI post-NAC.

Ultrafast dynamic contrast-enhanced breast MRI may generate prognostic imaging markers of breast cancer.

BACKGROUND: Ultrafast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI)-derived kinetic parameters have demonstrated at least equivalent accuracy to standard DCE-MRI in differentiating malignant from benign breast lesions. However, it is unclear if they have any efficacy as prognostic imaging markers. The aim of this study was to investigate the relationship between ultrafast DCE-MRI-derived kinetic parameters and breast cancer characteristics. METHODS: Consecutive breast MRI examinations between February 2017 and January 2018 were retrospectively reviewed to determine those examinations that meet the following inclusion criteria: (1) BI-RADS 4-6 MRI performed on a 3T scanner with a 16-channel breast coil and (2) a hybrid clinical protocol with 15 phases of ultrafast DCE-MRI (temporal resolution of 2.7-4.6 s) followed by early and delayed phases of standard DCE-MRI. The study included 125 examinations with 142 biopsy-proven breast cancer lesions. Ultrafast DCE-MRI-derived kinetic parameters (maximum slope [MS] and bolus arrival time [BAT]) were calculated for the entire volume of each lesion. Comparisons of these parameters between different cancer characteristics were made using generalized estimating equations, accounting for the presence of multiple lesions per patient. All comparisons were exploratory and adjustment for multiple comparisons was not performed; P values < 0.05 were considered statistically significant. RESULTS: Significantly larger MS and shorter BAT were observed for invasive carcinoma than ductal carcinoma in situ (DCIS) (P < 0.001 and P = 0.008, respectively). Significantly shorter BAT was observed for invasive carcinomas with more aggressive characteristics than those with less aggressive characteristics: grade 3 vs. grades 1-2 (P = 0.025), invasive ductal carcinoma vs. invasive lobular carcinoma (P = 0.002), and triple negative or HER2 type vs. luminal type (P < 0.001). CONCLUSIONS: Ultrafast DCE-MRI-derived parameters showed a strong relationship with some breast cancer characteristics, especially histopathology and molecular subtype.

Survival Outcomes of Screening with Breast MRI in Women at Elevated Risk of Breast Cancer.

OBJECTIVE: To determine survival outcomes in women with breast cancer detected at combined screening with breast MRI and mammography versus screening mammography alone. METHODS: This is an institutional review board-approved retrospective study, and the need for informed consent was waived. A total of 3002 women with an increased risk of breast cancer were screened between 2001 and 2004. Of the 3002 women, 1534 (51.1%) had 2780 combined screenings (MRI and mammography) and 1468 (48.9%) had 4811 mammography-only screenings. The Χ 2 test and the Kaplan-Meier method were used to compare cancer detection rates and survival rates. RESULTS: The overall cancer detection rate was significantly higher in the MRI plus mammography group compared with the mammography-only group (1.4% [40 of 2780] vs 0.5% [23 of 4811]; P < 0.001). No interval cancers occurred in the MRI plus mammography group, whereas 9 interval cancers were found in the mammography-only group. During a median follow-up of 10.9 years (range: 0.7 to 15.2), a total of 11 recurrences and 5 deaths occurred. Of the 11 recurrences, 6 were in the MRI plus mammography group and 5 were in the mammography-only group. All five deaths occurred in the mammography-only group. Disease-free survival showed no statistically significant difference between the two groups (P = 0.32). However, overall survival was significantly improved in the MRI plus mammography group (P = 0.002). CONCLUSION: Combined screening with MRI and mammography in women at elevated risk of breast cancer improves cancer detection and overall survival.

Background Parenchymal Enhancement on Breast MRI as a Prognostic Surrogate: Correlation With Breast Cancer Oncotype Dx Score.

PURPOSE: Breast MRI background parenchymal enhancement (BPE) can potentially serve as a prognostic marker, by possible correlation with molecular subtype. Oncotype Dx, a gene assay, is a prognostic and predictive surrogate for tumor aggressiveness and treatment response. The purpose of this study was to investigate the association between contralateral non-tumor breast magnetic resonance imaging (MRI) background parenchymal enhancement and tumor oncotype score. METHODS: In this retrospective study, patients with ER+ and HER2- early stage invasive ductal carcinoma who underwent preoperative breast MRI, oncotype risk scoring, and breast conservation surgery from 2008-2010 were identified. After registration, BPE from the pre and three post-contrast phases was automatically extracted using a k-means clustering algorithm. Four metrics were calculated: initial enhancement (IE) relative to the pre-contrast signal, late enhancement, overall enhancement (OE), and area under the enhancement curve (AUC). Histogram analysis was performed to determine first order metrics which were compared to oncotype risk score groups using Mann-Whitney tests and Spearman rank correlation analysis. RESULTS: This study included 80 women (mean age = 51.1 ± 10.3 years); 46 women were categorized as low risk (≤17) and 34 women were categorized as intermediate/high risk (≥18) according to Oncotype Dx. For the mean of the top 10% pixels, significant differences were noted for IE (p = 0.032), OE (p = 0.049), and AUC (p = 0.044). Using the risk score as a continuous variable, correlation analysis revealed a weak but significant correlation with the mean of the top 10% pixels for IE (r = 0.26, p = 0.02), OE (r = 0.25, p = 0.02), and AUC (r = 0.27, p = 0.02). CONCLUSION: BPE metrics of enhancement in the non-tumor breast are associated with tumor Oncotype Dx recurrence score, suggesting that the breast microenvironment may relate to likelihood of recurrence and magnitude of chemotherapy benefit.

Differentiation between subcentimeter carcinomas and benign lesions using kinetic parameters derived from ultrafast dynamic contrast-enhanced breast MRI.

OBJECTIVES: This study aims to evaluate ultrafast DCE-MRI-derived kinetic parameters that reflect contrast agent inflow effects in differentiating between subcentimeter BI-RADS 4-5 breast carcinomas and benign lesions. METHODS: We retrospectively reviewed consecutive 3-T MRI performed from February to October 2017, during which ultrafast DCE-MRI was performed as part of a hybrid clinical protocol with conventional DCE-MRI. In total, 301 female patients with 369 biopsy-proven breast lesions were included. Ultrafast DCE-MRI was acquired continuously over approximately 60 s (temporal resolution, 2.7-7.1 s/phase) starting simultaneously with the start of contrast injection. Four ultrafast DCE-MRI-derived kinetic parameters (maximum slope [MS], contrast enhancement ratio [CER], bolus arrival time [BAT], and initial area under gadolinium contrast agent concentration [IAUGC]) and one conventional DCE-MRI-derived kinetic parameter (signal enhancement ratio [SER]) were calculated for each lesion. Wilcoxon rank sum test or Fisher's exact test was performed to compare kinetic parameters, volume, diameter, age, and BI-RADS morphological descriptors between subcentimeter carcinomas and benign lesions. Univariate/multivariate logistic regression analyses were performed to determine predictive parameters for subcentimeter carcinomas. RESULTS: In total, 125 lesions (26 carcinomas and 99 benign lesions) were identified as BI-RADS 4-5 subcentimeter lesions. Subcentimeter carcinomas demonstrated significantly larger MS and SER and shorter BAT than benign lesions (p = 0.0117, 0.0046, and 0.0102, respectively). MS, BAT, and age were determined as significantly predictive for subcentimeter carcinoma (p = 0.0208, 0.0023, and < 0.0001, respectively). CONCLUSIONS: Ultrafast DCE-MRI-derived kinetic parameters may be useful in differentiating subcentimeter BI-RADS 4 and 5 carcinomas from benign lesions. KEY POINTS: • Ultrafast DCE-MRI can generate kinetic parameters, effectively differentiating breast carcinomas from benign lesions. • Subcentimeter carcinomas demonstrated significantly larger maximum slope and shorter bolus arrival time than benign lesions. • Maximum slope and bolus arrival time contribute to better management of suspicious subcentimeter breast lesions.