Artificial Intelligence Applied to Breast MRI for Improved Diagnosis.

Background Recognition of salient MRI morphologic and kinetic features of various malignant tumor subtypes and benign diseases, either visually or with artificial intelligence (AI), allows radiologists to improve diagnoses that may improve patient treatment. Purpose To evaluate whether the diagnostic performance of radiologists in the differentiation of cancer from noncancer at dynamic contrast material-enhanced (DCE) breast MRI is improved when using an AI system compared with conventionally available software. Materials and Methods In a retrospective clinical reader study, images from breast DCE MRI examinations were interpreted by 19 breast imaging radiologists from eight academic and 11 private practices. Readers interpreted each examination twice. In the "first read," they were provided with conventionally available computer-aided evaluation software, including kinetic maps. In the "second read," they were also provided with AI analytics through computer-aided diagnosis software. Reader diagnostic performance was evaluated with receiver operating characteristic (ROC) analysis, with the area under the ROC curve (AUC) as a figure of merit in the task of distinguishing between malignant and benign lesions. The primary study end point was the difference in AUC between the first-read and the second-read conditions. Results One hundred eleven women (mean age, 52 years ± 13 [standard deviation]) were evaluated with a total of 111 breast DCE MRI examinations (54 malignant and 57 nonmalignant lesions). The average AUC of all readers improved from 0.71 to 0.76 (P = .04) when using the AI system. The average sensitivity improved when Breast Imaging Reporting and Data System (BI-RADS) category 3 was used as the cut point (from 90% to 94%; 95% confidence interval [CI] for the change: 0.8%, 7.4%) but not when using BI-RADS category 4a (from 80% to 85%; 95% CI: -0.9%, 11%). The average specificity showed no difference when using either BI-RADS category 4a or category 3 as the cut point (52% and 52% [95% CI: -7.3%, 6.0%], and from 29% to 28% [95% CI: -6.4%, 4.3%], respectively). Conclusion Use of an artificial intelligence system improves radiologists' performance in the task of differentiating benign and malignant MRI breast lesions. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Krupinski in this issue.

Comparison of Abbreviated Breast MRI vs Digital Breast Tomosynthesis for Breast Cancer Detection Among Women With Dense Breasts Undergoing Screening.

Importance: Improved screening methods for women with dense breasts are needed because of their increased risk of breast cancer and of failed early diagnosis by screening mammography. Objective: To compare the screening performance of abbreviated breast magnetic resonance imaging (MRI) and digital breast tomosynthesis (DBT) in women with dense breasts. Design, Setting, and Participants: Cross-sectional study with longitudinal follow-up at 48 academic, community hospital, and private practice sites in the United States and Germany, conducted between December 2016 and November 2017 among average-risk women aged 40 to 75 years with heterogeneously dense or extremely dense breasts undergoing routine screening. Follow-up ascertainment of cancer diagnoses was complete through September 12, 2019. Exposures: All women underwent screening by both DBT and abbreviated breast MRI, performed in randomized order and read independently to avoid interpretation bias. Main Outcomes and Measures: The primary end point was the invasive cancer detection rate. Secondary outcomes included sensitivity, specificity, additional imaging recommendation rate, and positive predictive value (PPV) of biopsy, using invasive cancer and ductal carcinoma in situ (DCIS) to define a positive reference standard. All outcomes are reported at the participant level. Pathology of core or surgical biopsy was the reference standard for cancer detection rate and PPV; interval cancers reported until the next annual screen were included in the reference standard for sensitivity and specificity. Results: Among 1516 enrolled women, 1444 (median age, 54 [range, 40-75] years) completed both examinations and were included in the analysis. The reference standard was positive for invasive cancer with or without DCIS in 17 women and for DCIS alone in another 6. No interval cancers were observed during follow-up. Abbreviated breast MRI detected all 17 women with invasive cancer and 5 of 6 women with DCIS. Digital breast tomosynthesis detected 7 of 17 women with invasive cancer and 2 of 6 women with DCIS. The invasive cancer detection rate was 11.8 (95% CI, 7.4-18.8) per 1000 women for abbreviated breast MRI vs 4.8 (95% CI, 2.4-10.0) per 1000 women for DBT, a difference of 7 (95% CI, 2.2-11.6) per 1000 women (exact McNemar P = .002). For detection of invasive cancer and DCIS, sensitivity was 95.7% (95% CI, 79.0%-99.2%) with abbreviated breast MRI vs 39.1% (95% CI, 22.2%-59.2%) with DBT (P = .001) and specificity was 86.7% (95% CI, 84.8%-88.4%) vs 97.4% (95% CI, 96.5%-98.1%), respectively (P < .001). The additional imaging recommendation rate was 7.5% (95% CI, 6.2%-9.0%) with abbreviated breast MRI vs 10.1% (95% CI, 8.7%-11.8%) with DBT (P = .02) and the PPV was 19.6% (95% CI, 13.2%-28.2%) vs 31.0% (95% CI, 17.0%-49.7%), respectively (P = .15). Conclusions and Relevance: Among women with dense breasts undergoing screening, abbreviated breast MRI, compared with DBT, was associated with a significantly higher rate of invasive breast cancer detection. Further research is needed to better understand the relationship between screening methods and clinical outcome. Trial Registration: ClinicalTrials.gov Identifier: NCT02933489.

MRI, Clinical Examination, and Mammography for Preoperative Assessment of Residual Disease and Pathologic Complete Response After Neoadjuvant Chemotherapy for Breast Cancer: ACRIN 6657 Trial.

OBJECTIVE: The objective of our study was to determine the accuracy of preoperative measurements for detecting pathologic complete response (CR) and assessing residual disease after neoadjuvant chemotherapy (NACT) in patients with locally advanced breast cancer. SUBJECTS AND METHODS: The American College of Radiology Imaging Network 6657 Trial prospectively enrolled women with ≥ 3 cm invasive breast cancer receiving NACT. Preoperative measurements of residual disease included longest diameter by mammography, MRI, and clinical examination and functional volume on MRI. The accuracy of preoperative measurements for detecting pathologic CR and the association with final pathology size were assessed for all lesions, separately for single masses and nonmass enhancements (NMEs), multiple masses, and lesions without ductal carcinoma in situ (DCIS). RESULTS: In the 138 women with all four preoperative measures, longest diameter by MRI showed the highest accuracy for detecting pathologic CR for all lesions and NME (AUC = 0.76 and 0.84, respectively). There was little difference across preoperative measurements in the accuracy of detecting pathologic CR for single masses (AUC = 0.69-0.72). Longest diameter by MRI and longest diameter by clinical examination showed moderate ability for detecting pathologic CR for multiple masses (AUC = 0.78 and 0.74), and longest diameter by MRI and longest diameter by mammography showed moderate ability for detecting pathologic CR for tumors without DCIS (AUC = 0.74 and 0.71). In subjects with residual disease, longest diameter by MRI exhibited the strongest association with pathology size for all lesions and single masses (r = 0.33 and 0.47). Associations between preoperative measures and pathology results were not significantly influenced by tumor subtype or mammographic density. CONCLUSION: Our results indicate that measurement of longest diameter by MRI is more accurate than by mammography and clinical examination for preoperative assessment of tumor residua after NACT and may improve surgical planning.

Value of breast MRI for patients with a biopsy showing atypical ductal hyperplasia (ADH).

PURPOSE: To evaluate the diagnostic value of dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) for patients with atypical ductal hyperplasia (ADH) in predicting malignant upgrade. MATERIALS AND METHODS: 3T DCE-MRI was performed for 17 patients with ADH (median age 52, range 42-76) proven by stereotactic biopsy (n = 15), and ultrasound-guided biopsy (n = 2) from January 2011 to April 2015. All patients underwent surgical excision after the MRI. Two radiologists prospectively reviewed the MRI to determine the presence or absence of suspicious findings at the site of biopsy, and evaluated the MR features of any lesion present according to the Breast Imaging Reporting and Data System (BI-RADS) lexicon. MRI findings and clinical information were correlated with the final surgical pathology by multivariate analysis. RESULTS: Nine of 17 lesions were upgraded to malignancy. MRI demonstrated suspicious nonmass enhancement (NME) at the site of biopsy in all upgraded patients. The median size was 19.5 mm (range, 9-44 mm). In the eight patients without upgrade, no enhancement (n = 2), linear enhancement along the biopsy track (n = 4), thin rim enhancement around hematoma (n = 1), and a focal NME (n = 1) were seen. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of MRI findings were 100, 87.5, 90, and 100%, respectively. Multivariate analysis revealed that the presence of suspicious enhancement on MRI was the most significant predictor of upgrade to malignancy (P = 0.0006) CONCLUSION: Our study revealed a high NPV of DCE-MRI for patients with ADH in terms of malignant upgrade at subsequent surgery. This suggests that patients with ADH without suspicious enhancement on DCE-MRI might be followed with DCE-MRI rather than undergoing surgical excision. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2017;46:1738-1747.

Fast bilateral breast coverage with high spectral and spatial resolution (HiSS) MRI at 3T.

PURPOSE: To develop and assess a full-coverage, sensitivity encoding (SENSE)-accelerated breast high spatial and spectral resolution (HiSS) magnetic resonance imaging (MRI) within clinically reasonable times as a potential nonenhanced MRI protocol for breast density measurement or breast cancer screening. MATERIALS AND METHODS: Sixteen women with biopsy-proven cancer or suspicious lesions, and 13 women who were healthy volunteers or were screened for breast cancer, received 3T breast MRI exams, including SENSE-accelerated HiSS MRI, which was implemented as a submillimeter spatial resolution echo-planar spectroscopic imaging (EPSI) sequence. In postprocessing, fat and water resonance peak height and integral images were generated from EPSI data. The postprocessing software was custom-designed, and new algorithms were developed to enable processing of whole-coverage axial HiSS datasets. Water peak height HiSS images were compared to pre- and postcontrast T1 -weighted images. Fat suppression was quantified as parenchymal-to-suppressed-fat signal ratio in HiSS water peak height and nonenhanced T1 -weighted images, and artifact levels were scored. RESULTS: Approximately a 4-fold decrease in acquisition speed, with a concurrent 2.5-fold decrease in voxel size, was achieved, with low artifact levels, and with spectral signal-to-noise ratio (SNR) of 45:1. Fat suppression was 1.9 times more effective (P < 0.001) in HiSS images than in T1 -weighted images (SPAIR), and HiSS images showed higher SNR in the axilla. HiSS MRI visualized 10 of 13 malignant lesions identified on dynamic contrast-enhanced (DCE)-MRI, and did not require skin removal in postprocessing to generate maximum intensity projection images. CONCLUSION: We demonstrate full-coverage, SENSE-accelerated breast HiSS MRI within clinically reasonable times, as a potential protocol for breast density measurement or breast cancer screening. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2017;46:1341-1348.

MR spectroscopy of breast cancer for assessing early treatment response: Results from the ACRIN 6657 MRS trial.

PURPOSE: To estimate the accuracy of predicting response to neoadjuvant chemotherapy (NACT) in patients with locally advanced breast cancer using MR spectroscopy (MRS) measurements made very early in treatment. MATERIALS AND METHODS: This prospective Health Insurance Portability and Accountability Act (HIPAA)-compliant protocol was approved by the American College of Radiology and local-site institutional review boards. One hundred nineteen women with invasive breast cancer of ≥3 cm undergoing NACT were enrolled between September 2007 and April 2010. MRS measurements of the concentration of choline-containing compounds ([tCho]) were performed before the first chemotherapy regimen (time point 1, TP1) and 20-96 h after the first cycle of treatment (TP2). The change in [tCho] was assessed for its ability to predict pathologic complete response (pCR) and radiologic response using the area under the receiver operating characteristic curve (AUC) and logistic regression models. RESULTS: Of the 119 subjects enrolled, only 29 cases (24%) with eight pCRs provided usable data for the primary analysis. Technical challenges in acquiring quantitative MRS data in a multi-site trial setting limited the capture of usable data. In this limited data set, the decrease in tCho from TP1 to TP2 had poor ability to predict either pCR (AUC = 0.53, 95% confidence interval [CI]: 0.27-0.79) or radiologic response (AUC = 0.51, 95% CI: 0.27-0.75). CONCLUSION: The technical difficulty of acquiring quantitative MRS data in a multi-site clinical trial setting led to a low yield of analyzable data, which was insufficient to accurately measure the ability of early MRS measurements to predict response to NACT. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 2 J. MAGN. RESON. IMAGING 2017;46:290-302.

Neoadjuvant Chemotherapy for Breast Cancer: Functional Tumor Volume by MR Imaging Predicts Recurrence-free Survival-Results from the ACRIN 6657/CALGB 150007 I-SPY 1 TRIAL.

PURPOSE: To evaluate volumetric magnetic resonance (MR) imaging for predicting recurrence-free survival (RFS) after neoadjuvant chemotherapy (NACT) of breast cancer and to consider its predictive performance relative to pathologic complete response (PCR). MATERIALS AND METHODS: This HIPAA-compliant prospective multicenter study was approved by institutional review boards with written informed consent. Women with breast tumors 3 cm or larger scheduled for NACT underwent dynamic contrast-enhanced MR imaging before treatment (examination 1), after one cycle (examination 2), midtherapy (examination 3), and before surgery (examination 4). Functional tumor volume (FTV), computed from MR images by using enhancement thresholds, and change from baseline (ΔFTV) were measured after one cycle and before surgery. Association of RFS with FTV was assessed by Cox regression and compared with association of RFS with PCR and residual cancer burden (RCB), while controlling for age, race, and hormone receptor (HR)/ human epidermal growth factor receptor type 2 (HER2) status. Predictive performance of models was evaluated by C statistics. RESULTS: Female patients (n = 162) with FTV and RFS were included. At univariate analysis, FTV2, FTV4, and ΔFTV4 had significant association with RFS, as did HR/HER2 status and RCB class. PCR approached significance at univariate analysis and was not significant at multivariate analysis. At univariate analysis, FTV2 and RCB class had the strongest predictive performance (C statistic = 0.67; 95% confidence interval [CI]: 0.58, 0.76), greater than for FTV4 (0.64; 95% CI: 0.53, 0.74) and PCR (0.57; 95% CI: 0.39, 0.74). At multivariate analysis, a model with FTV2, ΔFTV2, RCB class, HR/HER2 status, age, and race had the highest C statistic (0.72; 95% CI: 0.60, 0.84). CONCLUSION: Breast tumor FTV measured by MR imaging is a strong predictor of RFS, even in the presence of PCR and RCB class. Models combining MR imaging, histopathology, and breast cancer subtype demonstrated the strongest predictive performance in this study.

Evaluation of Kinetic Entropy of Breast Masses Initially Found on MRI using Whole-lesion Curve Distribution Data: Comparison with the Standard Kinetic Analysis.

OBJECTIVES: To quantify kinetic heterogeneity of breast masses that were initially detected with dynamic contrast-enhanced MRI, using whole-lesion kinetic distribution data obtained from computer-aided evaluation (CAE), and to compare that with standard kinetic curve analysis. METHODS: Clinical MR images from 2006 to 2011 with breast masses initially detected with MRI were evaluated with CAE. The relative frequencies of six kinetic patterns (medium-persistent, medium-plateau, medium-washout, rapid-persistent, rapid-plateau, rapid-washout) within the entire lesion were used to calculate kinetic entropy (KE), a quantitative measure of enhancement pattern heterogeneity. Initial uptake (IU) and signal enhancement ratio (SER) were obtained from the most-suspicious kinetic curve. Mann-Whitney U test and ROC analysis were conducted for differentiation of malignant and benign masses. RESULTS: Forty benign and 37 malignant masses comprised the case set. IU and SER were not significantly different between malignant and benign masses, whereas KE was significantly greater for malignant than benign masses (p = 0.748, p = 0.083, and p < 0.0001, respectively). Areas under ROC curve for IU, SER, and KE were 0.479, 0.615, and 0.662, respectively. CONCLUSION: Quantification of kinetic heterogeneity of whole-lesion time-curve data with KE has the potential to improve differentiation of malignant from benign breast masses on breast MRI. KEY POINTS: • Kinetic heterogeneity can be quantified by computer-aided evaluation of breast MRI • Kinetic entropy was greater in malignant masses than benign masses • Kinetic entropy has the potential to improve differentiation of breast masses.

Classification of breast lesions pre-contrast injection using water resonance lineshape analysis.

Inhomogeneously broadened, non-Lorentzian water resonances have been observed in small image voxels of breast tissue. The non-Lorentzian components of the water resonance are probably produced by bulk magnetic susceptibility shifts caused by dense, deoxygenated tumor blood vessels (the 'blood oxygenation level-dependent' effect), but can also be produced by other characteristics of local anatomy and physiology, including calcifications and interfaces between different types of tissue. Here, we tested the hypothesis that the detection of non-Lorentzian components of the water resonance with high spectral and spatial resolution (HiSS) MRI allows the classification of breast lesions without the need to inject contrast agent. Eighteen malignant lesions and nine benign lesions were imaged with HiSS MRI at 1.5 T. A new algorithm was developed to detect non-Lorentzian (or off-peak) components of the water resonance. After a Lorentzian fit had been subtracted from the data, the largest peak in the residual spectrum in each voxel was identified as the major off-peak component of the water resonance. The difference in frequency between these off-peak components and the main water peaks, and their amplitudes, were measured in malignant lesions, benign lesions and breast fibroglandular tissue. Off-peak component frequencies were significantly different between malignant and benign lesions (p < 0.001). Receiver operating characteristic (ROC) analysis was used to assess the diagnostic performance of HiSS off-peak component analysis compared with dynamic contrast-enhanced (DCE) MRI parameters. The areas under the ROC curves for the 'DCE rapid uptake fraction', 'DCE washout fraction', 'off-peak component amplitude' and 'off-peak component frequency' were 0.75, 0.83, 0.50 and 0.86, respectively. These results suggest that water resonance lineshape analysis performs well in the classification of breast lesions without contrast injection and could improve the diagnostic accuracy of clinical breast MR examinations. In addition, this approach may provide an alternative to DCE MRI in women who are at risk for adverse reactions to contrast media.

Utility of preoperative ultrasound for predicting pN2 or higher stage axillary lymph node involvement in patients with newly diagnosed breast cancer.

OBJECTIVE: The objective of our study was to report the positive predictive value (PPV) of ultrasound of the axilla to predict pN2 or higher disease in breast cancer patients. MATERIALS AND METHODS: A retrospective study of 559 patients with newly diagnosed invasive breast cancer from 2005 through 2009 was performed. All patients underwent ipsilateral axillary ultrasound for staging purposes. Ultrasound findings were considered suspicious for metastasis if cortical thickening or nonhilar blood flow to the cortex was present. Suspicious lymph nodes were classified on the basis of their features as high, intermediate, or low suspicion. The standard of truth was confirmed pathologically. RESULTS: Either pN2 or pN3 disease was found in 50 of 181 (28%) patients with positive findings on an ultrasound study and 10 of 378 (3%) patients with a negative ultrasound study (p < 0.01). When two or more lymph nodes of high suspicion or a total of three or more lymph nodes of any combination of high suspicion and intermediate suspicion were detected, patients were likely to have pN2 or pN3 disease (PPV, 82%). Either pN2 or pN3 disease was found in two of 122 (2%) patients whose primary cancers were up to 10 mm and 58 of 437 (13%) patients whose primary cancers were larger than 10 mm (p < 0.001). Ultrasound of the patient with tumors larger than 10 mm showing at least two highly suspicious nodes had a PPV of 87% for predicting pN2 or higher disease. CONCLUSION: Ultrasound was useful for predicting pN2 or higher axillary disease in breast cancer patients. Preoperative ultrasound assessment for staging of axillary lymph nodes might help avoid underestimation at sentinel lymph node biopsy.

Decision making for breast lesions initially detected at contrast-enhanced breast MRI.

OBJECTIVE: The purpose of this study was to assess the clinical significance of breast lesions initially detected at contrast-enhanced breast MRI and to consider how to manage those lesions in accordance with the imaging findings and the indication for MRI. MATERIALS AND METHODS: A retrospective study of 4260 consecutive breast MRI examinations was performed to identify MRI-detected enhancing lesions. In 4260 studies, 554 MRI-detected lesions were found in 417 patients, and 134 (24%) of the lesions were malignant. Pathologic confirmation was obtained for 319 (58%) lesions. Results of the subsequent imaging workup, biopsy, surgery, and imaging follow-up were reviewed. RESULTS: The median size of the lesions was 89 mm (malignant, 15.45 mm; benign, 7.48 mm). Irregular shape, irregular or spiculated margins, and heterogeneous or rim enhancement were seen significantly more often in malignant mass lesions (p < 0.001). Malignant lesions were more likely to exhibit rapid enhancement (p < 0.001). Benign lesions were more likely to have persistent kinetics (p < 0.001). There was a statistically significant difference (p < 0.001) between malignant (58/87, 67%) and benign lesions (128/287, 45%) with respect to sonographic detection at second-look ultrasound examinations. Malignant lesions were most often detected in patients with metastatic axillary lymph nodes with an unknown primary tumor (8/8, 100%), followed by patients with positive or close margins in recent breast cancer surgery (45/76, 59%), and patients with newly diagnosed breast cancer (44/115, 38%). CONCLUSION: Management of MRI-detected lesions should be based on both MRI findings and the patient's indication for MRI.

Breast cancer detected on an incident (second or subsequent) round of screening MRI: MRI features of false-negative cases.

OBJECTIVE: The purpose of this article is to evaluate the nature of breast cancers detected in the incident round of screening MRI to determine MRI features of early breast cancer. MATERIALS AND METHODS: From 2003 to 2012, there were 16 incident breast cancers in 15 patients on screening MRI, including nine cancers that were retrospectively identifiable on the prior MRI (false-negative [FN] cancers at prior screening examination). We evaluated the BI-RADS features of these incident cancers in previous and current MRI scans. RESULTS: Of 16 incident cancers, there were 11 mass lesions (69%), three foci (19%), and two nonmasslike enhancement lesions (13%). Of the nine FN cancers (five foci, two masses, and two nonmasslike enhancement lesions), all showed increases in size on the current examination (median, 80% increase); four lesions showed rapid uptake kinetics on prior examinations, and five lesions showed a change in kinetic pattern from slow to rapid uptake. Among the five foci, one focus was isolated and four foci were in a background of other foci, where two foci could be distinguished for their higher signal intensity. CONCLUSION: On screening MRI, any lesion that increases in size, has rapid uptake kinetics or a change in kinetic pattern, or is an isolated focus or focus showing more enhancement than other foci should be viewed with a high degree of suspicion, and a biopsy should be considered.

Axillary staging of breast cancer: what the radiologist should know.

Identifying the presence of axillary node and internal mammary node metastases in patients with invasive breast cancer is critical for determining prognosis and for deciding on appropriate treatment. Sentinel lymph node biopsy (SLNB) is the definitive method to exclude axillary metastases. Patients with positive SLNB results generally undergo axillary lymph node dissection (ALND). The benefit of preoperative identification of axillary metastases is that it allows the surgeon to proceed directly to ALND and to avoid an unnecessary SLNB and the need for a second surgical procedure involving the axillary nodes. Knowledge of the important anatomic landmarks of the axilla is important in finding and accurately reporting suspicious lymph nodes. The pathologic features of nodal metastases illuminate the imaging appearances of these nodes, as depicted with all modalities. Ultrasonography (US) is the primary imaging modality for evaluating axillary nodes. Morphologic criteria, such as cortical thickening, hilar effacement, and nonhilar cortical blood flow, are more important than size criteria in the identification of metastases. US-guided lymph node sampling, especially with core biopsy, is invaluable in confirming the presence of a metastasis in a suspicious node. Core biopsy has been shown to be equal in safety to fine needle aspiration and has a significantly lower false-negative rate. Magnetic resonance imaging is also useful, with the added benefit of providing a global view of both axillae. Computed tomography and radionuclide imaging play a lesser role in imaging the axilla. Preoperative image-based identification and sampling of abnormal lymph nodes that have a high positive predictive value for metastases is an extremely important component in the management of patients with invasive breast cancer.

Locally advanced breast cancer: MR imaging for prediction of response to neoadjuvant chemotherapy--results from ACRIN 6657/I-SPY TRIAL.

PURPOSE: To compare magnetic resonance (MR) imaging findings and clinical assessment for prediction of pathologic response to neoadjuvant chemotherapy (NACT) in patients with stage II or III breast cancer. MATERIALS AND METHODS: The HIPAA-compliant protocol and the informed consent process were approved by the American College of Radiology Institutional Review Board and local-site institutional review boards. Women with invasive breast cancer of 3 cm or greater undergoing NACT with an anthracycline-based regimen, with or without a taxane, were enrolled between May 2002 and March 2006. MR imaging was performed before NACT (first examination), after one cycle of anthracyline-based treatment (second examination), between the anthracycline-based regimen and taxane (third examination), and after all chemotherapy and prior to surgery (fourth examination). MR imaging assessment included measurements of tumor longest diameter and volume and peak signal enhancement ratio. Clinical size was also recorded at each time point. Change in clinical and MR imaging predictor variables were compared for the ability to predict pathologic complete response (pCR) and residual cancer burden (RCB). Univariate and multivariate random-effects logistic regression models were used to characterize the ability of tumor response measurements to predict pathologic outcome, with area under the receiver operating characteristic curve (AUC) used as a summary statistic. RESULTS: Data in 216 women (age range, 26-68 years) with two or more imaging time points were analyzed. For prediction of both pCR and RCB, MR imaging size measurements were superior to clinical examination at all time points, with tumor volume change showing the greatest relative benefit at the second MR imaging examination. AUC differences between MR imaging volume and clinical size predictors at the early, mid-, and posttreatment time points, respectively, were 0.14, 0.09, and 0.02 for prediction of pCR and 0.09, 0.07, and 0.05 for prediction of RCB. In multivariate analysis, the AUC for predicting pCR at the second imaging examination increased from 0.70 for volume alone to 0.73 when all four predictor variables were used. Additional predictive value was gained with adjustments for age and race. CONCLUSION: MR imaging findings are a stronger predictor of pathologic response to NACT than clinical assessment, with the greatest advantage observed with the use of volumetric measurement of tumor response early in treatment.

T(2)∗ relaxation times of intraductal murine mammary cancer, invasive mammary cancer, and normal mammary gland.

PURPOSE: This study investigates the feasibility of T(2)∗ to be a diagnostic indicator of early breast cancer in a mouse model. T(2)∗ is sensitive to susceptibility effects due to local inhomogeneity of the magnetic field, e.g., caused by hemosiderin or deoxyhemoglobin. In these mouse models, unlike in patients, the characteristics of single mammary ducts containing pure intraductal cancer can be evaluated. METHODS: The C3(1)SV40Tag mouse model of breast cancer (n = 11) and normal FVB∕N mice (n = 6) were used to measure T(2)∗ of normal mammary gland tissue, intraepithelial neoplasia, invasive cancers, mammary lymph nodes, and muscle. MRI experiments were performed on a 9.4T animal scanner. High resolution (117 microns) axial 2D multislice gradient echo images with fat suppression were acquired first to identify inguinal mammary gland. Then a multislice multigradient echo pulse sequence with and without fat suppression were performed over the inguinal mammary gland. The modulus of a complex double exponential decay detected by the multigradient echo sequence was used to fit the absolute proton free induction decay averaged over a region of interest to determine the T(2)∗ of water and fat signals. RESULTS: The measured T(2)∗ values of tumor and muscle are similar (∼15 ms), and almost twice that of lymph nodes (∼8 ms). There was a statistically significant difference (p < 0.03) between T(2)∗ in normal mammary tissue (13.7 ± 2.9 ms) and intraductal cancers (11 ± 2.0 ms) when a fat suppression pulse was applied. CONCLUSIONS: These are the first reported T(2)∗ measurements from single mammary ducts. The results demonstrated that T(2)∗ measurements may have utility for identifying early pre-invasive cancers in mouse models. This may inspire similar research for patients using T(2)∗ for diagnostic imaging of early breast cancer.

Comparing post-operative human breast specimen radiograph and MRI in lesion margin and volume assessment.

The purpose of this research is to evaluate the potential for identifying malignant breast lesions and their margins on large specimen MRI, in comparison to specimen radiography and clinical dynamic contrast enhanced MRI (DCE-MRI). Breast specimens were imaged with an MR scanner immediately after surgery, with an IRB-approved protocol and with the patients' informed consent. Specimen sizes were at least 5 cm in diameter and approximately 1 to 4 cm thick. Coronal and axial gradient echo MR images without fat suppression were acquired over the whole specimens using a 9.4T animal scanner. Findings on specimen MRI were compared with findings on specimen radiograph, and their volumes were compared with measurements obtained from clinical DCE-MRI. The results showed that invasive ductal carcinoma (IDC) lesions were easily identified using MRI and the margins were clearly distinguishable from nearby tissue. However, ductal carcinoma in situ (DCIS) lesions were not clearly discernible and were diffused with poorly defined margins on MRI. Calcifications associated with DCIS were visualized in all specimens on specimen radiograph. There is a strong correlation between the maximum diameter of lesions as measured by radiograph and MRI (r = 0.93), as well as the maximum diameter measured by pathology and radiograph/MRI (r>0.75). The volumes of IDC measured on specimen MRI were slightly smaller than those measured on DCE-MRI. Imaging of excised human breast lumpectomy specimens with high magnetic field MRI provides promising results for improvements in lesion identification and margin localization for IDC. However, there are technical challenges in visualization of DCIS lesions. Improvements in specimen imaging are important, as they will provide additional information to standard radiographic analysis.

Evaluation of clinical breast MR imaging performed with prototype computer-aided diagnosis breast MR imaging workstation: reader study.

PURPOSE: To evaluate a computer-aided diagnosis (CADx) system for dynamic contrast material-enhanced magnetic resonance (MR) imaging and compare it with a currently used clinical method of interpreting breast MR image findings that includes the use of commercially available automated software for kinetic image data processing and visualization. MATERIALS AND METHODS: In this HIPAA-compliant, institutional review board-approved study, a training set of 121 breast lesions (77 malignant, 44 benign) was used to train the CADx system. After practicing with 10 training cases, six breast imaging radiologists assessed the likelihood of malignancy and the need for biopsy with a separate test set of 60 lesions (30 malignant, 30 benign). Their performances in differentiating between benign and malignant breast lesions both without (conventional lesion viewing, output from commercially available breast MR imaging analysis software) and with the aid of the CADx workstation (with classification yielding an estimation of the probability of malignancy for each lesion) were evaluated with receiver operating characteristic analysis. RESULTS: When CADx was used, the average performance of the radiologists was significantly improved, as indicated by increases in mean area under the receiver operating characteristic curve (from 0.80 to 0.84, P = .007), mean sensitivity (from 83% to 88%, P = .001), and average number of biopsy recommendations for malignant cases (1.7 more biopsies for malignant lesions with use of CADx, P = .032). Although the mean specificity improved (from 50% to 53%), the improvement was not significant (P = .2). CONCLUSION: Use of the CADx system improved the radiologists' performance in differentiating between malignant and benign MR imaging-depicted breast lesions.

Combined use of T2-weighted MRI and T1-weighted dynamic contrast-enhanced MRI in the automated analysis of breast lesions.

A multiparametric computer-aided diagnosis scheme that combines information from T1-weighted dynamic contrast-enhanced (DCE)-MRI and T2-weighted MRI was investigated using a database of 110 malignant and 86 benign breast lesions. Automatic lesion segmentation was performed, and three categories of lesion features (geometric, T1-weighted DCE, and T2-weighted) were automatically extracted. Stepwise feature selection was performed considering only geometric features, only T1-weighted DCE features, only T2-weighted features, and all features. Features were merged with Bayesian artificial neural networks, and diagnostic performance was evaluated by ROC analysis. With leave-one-lesion-out cross-validation, an area under the ROC curve value of 0.77±0.03 was achieved with T2-weighted-only features, indicating high diagnostic value of information in T2-weighted images. Area under the ROC curve values of 0.79±0.03 and 0.80 ± 0.03 were obtained for geometric-only features and T1-weighted DCE-only features, respectively. When all features were considered, an area under the ROC curve value of 0.85±0.03 was achieved. We observed P values of 0.006, 0.023, and 0.0014 between the geometric-only, T1-weighted DCE-only, and T2-weighted-only features and all features conditions, respectively. When ranked, the P values satisfied the Holm-Bonferroni multiple-comparison test; thus, the improvement of multiparametric computer-aided diagnosis was statistically significant. A computer-aided diagnosis scheme that combines information from T1-weighted DCE and T2-weighted MRI may be advantageous over conventional T1-weighted DCE-MRI computer-aided diagnosis.

In vivo MRI of early stage mammary cancers and the normal mouse mammary gland.

Since the advent of screening mammography, approximately one-quarter of newly diagnosed breast cancers are at the earliest preinvasive stage of ductal carcinoma in situ (DCIS). Concomitant with this improvement in early detection has been a growing clinical concern that distinguishing aggressive from indolent DCIS is necessary to optimize patient management. Genetically engineered mouse models offer an appealing experimental framework in which to investigate factors that influence and predict progression of preinvasive neoplasias. Because of the small size of early stage carcinomas in mice, high-resolution imaging techniques are required to effectively observe longitudinal progression. The purpose of the present study was to evaluate the feasibility of MRI for assessment of in situ mammary neoplasias and early invasive mammary cancers that stochastically arise in mammary glands of C3(1) SV40 Tag transgenic mice. Additionally, images of normal mammary glands from wild-type FVB/N mice were acquired and compared with those from transgenic mice. Sixteen mice underwent MR examinations employing axial two-dimensional multi-slice gradient recalled echo scans (TR/TE =∼1000/5.5 ms) with fat suppression in a two-step process targeting both the upper and lower mammary glands. MRI successfully detected in situ and early invasive neoplasias in transgenic mice with high sensitivity and specificity. The average signal-to-noise ratio (SNR) of in situ lesions on fat-suppressed high-resolution T(1) -weighted images was 22.9, which was lower than that of invasive tumors, lymph nodes and muscle (average SNR of 29.5-34.9, p < 0.0001) but significantly higher than that of normal mammary tissue (average SNR = 5.5, p < 0.0001). Evaluation of wild-type mammary glands revealed no cancerous or benign lesions, and comparable image contrast characteristics (average SNR = 5.2) as compared with normal tissue areas of transgenic mice. This present study demonstrates that MRI is an excellent candidate for performing longitudinal assessment of early stage mammary cancer disease progression and response to therapy in the transgenic model system.

The diverse pathology and kinetics of mass, nonmass, and focus enhancement on MR imaging of the breast.

PURPOSE: To compare the pathology and kinetic characteristics of breast lesions with focus-, mass-, and nonmass-like enhancement. MATERIALS AND METHODS: A total of 852 MRI detected breast lesions in 697 patients were selected for an IRB approved review. Patients underwent dynamic contrast enhanced MRI using one pre- and three to six postcontrast T(1)-weighted images. The "type" of enhancement was classified as mass, nonmass, or focus, and kinetic curves quantified by the initial enhancement percentage (E(1)), time to peak enhancement (T(peak)), and signal enhancement ratio (SER). These kinetic parameters were compared between malignant and benign lesions within each morphologic type. RESULTS: A total of 552 lesions were classified as mass (396 malignant, 156 benign), 261 as nonmass (212 malignant, 49 benign), and 39 as focus (9 malignant, 30 benign). The most common pathology of malignant/benign lesions by morphology: for mass, invasive ductal carcinoma/fibroadenoma; for nonmass, ductal carcinoma in situ (DCIS)/fibrocystic change(FCC); for focus, DCIS/FCC. Benign mass lesions exhibited significantly lower E(1), longer T(peak), and lower SER compared with malignant mass lesions (P < 0.0001). Benign nonmass lesions exhibited only a lower SER compared with malignant nonmass lesions (P < 0.01). CONCLUSION: By considering the diverse pathology and kinetic characteristics of different lesion morphologies, diagnostic accuracy may be improved.