Mammography is the current standard for breast cancer screening. This study aimed to develop an artificial intelligence (AI) algorithm for diagnosis of breast cancer in mammography, and explore ...whether it could benefit radiologists by improving accuracy of diagnosis.
In this retrospective study, an AI algorithm was developed and validated with 170 230 mammography examinations collected from five institutions in South Korea, the USA, and the UK, including 36 468 cancer positive confirmed by biopsy, 59 544 benign confirmed by biopsy (8827 mammograms) or follow-up imaging (50 717 mammograms), and 74 218 normal. For the multicentre, observer-blinded, reader study, 320 mammograms (160 cancer positive, 64 benign, 96 normal) were independently obtained from two institutions. 14 radiologists participated as readers and assessed each mammogram in terms of likelihood of malignancy (LOM), location of malignancy, and necessity to recall the patient, first without and then with assistance of the AI algorithm. The performance of AI and radiologists was evaluated in terms of LOM-based area under the receiver operating characteristic curve (AUROC) and recall-based sensitivity and specificity.
The AI standalone performance was AUROC 0·959 (95% CI 0·952–0·966) overall, and 0·970 (0·963–0·978) in the South Korea dataset, 0·953 (0·938–0·968) in the USA dataset, and 0·938 (0·918–0·958) in the UK dataset. In the reader study, the performance level of AI was 0·940 (0·915–0·965), significantly higher than that of the radiologists without AI assistance (0·810, 95% CI 0·770–0·850; p<0·0001). With the assistance of AI, radiologists' performance was improved to 0·881 (0·850–0·911; p<0·0001). AI was more sensitive to detect cancers with mass (53 90% vs 46 78% of 59 cancers detected; p=0·044) or distortion or asymmetry (18 90% vs ten 50% of 20 cancers detected; p=0·023) than radiologists. AI was better in detection of T1 cancers (73 91% vs 59 74% of 80; p=0·0039) or node-negative cancers (104 87% vs 88 74% of 119; p=0·0025) than radiologists.
The AI algorithm developed with large-scale mammography data showed better diagnostic performance in breast cancer detection compared with radiologists. The significant improvement in radiologists' performance when aided by AI supports application of AI to mammograms as a diagnostic support tool.
Lunit.
Background
Nipple-sparing mastectomy (NSM) has become increasingly prevalent for patients with ductal carcinoma in situ (DCIS) requiring mastectomy. However, few data regarding recurrence outcomes ...after NSM are available for this patient population. This study evaluated the locoregional recurrence (LRR) rate for patients with pure DCIS who underwent NSM followed by immediate breast reconstruction without adjuvant radiotherapy and investigated potential risk factors for LRR and/or nipple-areola complex recurrence (NR).
Methods
A retrospective chart review was performed for 199 consecutive patients with pure DCIS who underwent NSM and immediate breast reconstruction between March 2003 and December 2015. Risk factors for LRR and NR were analyzed using univariate (Chi square test) and multivariate (Cox model) methods.
Results
The median follow-up duration after surgery was 97 months (range, 39–186 months). At 10 years, the LRR rate was 4.5%, and the NR rate was 3%. The univariate analysis showed that high nuclear grade, negative receptor status, positive human epidermal growth factor receptor 2 (HER2) status, and negative hormone receptor/positive HER2 subtype were associated with increased risk for NR. The multivariate analysis demonstrated that negative progesterone receptor status was an independent risk factor for LRR. However, margin status and tumor-to-nipple distance (TND) were not associated with increased risk for either LRR or NR.
Conclusions
The study findings suggest that NSM can be a feasible surgical option even for DCIS with a TND of 1 cm or less if the retroareolar resection margin is negative for malignancy. Determining the molecular subtype of DCIS might be helpful in identifying patients at high risk for recurrence.
The main concern associated with nipple-sparing mastectomy (NSM) is the risk of local breast cancer recurrence at the retained nipple-areola complex (NAC) consequent to occult nipple involvement. ...Long-term follow-up data regarding the oncologic safety of modern therapeutic NSM in terms of cancer recurrence at the NAC and survival are limited.
To assess the incidence, risk factors, treatment, and long-term outcomes associated with cancer recurrence at the NAC in a large series of patients with invasive breast cancer who underwent NSM and immediate breast reconstruction.
In this retrospective cohort study at a single institution (Asan Medical Center) in Seoul, Republic of Korea, 962 breasts from 944 patients who underwent NSM and immediate breast reconstruction for invasive breast cancer were analyzed between March 3, 2003, and December 31, 2015. Patients who underwent neoadjuvant systemic therapy or palliative surgery were excluded. Data analysis was performed from June 4, 2018, to August 31, 2018.
Univariate and multivariate Cox proportional hazards regression models were used to analyze the association between clinicopathologic variables and cancer recurrence at the NAC. To evaluate the association between cancer recurrence at the NAC and prognosis, distant metastasis-free survival, and overall survival were estimated using the Kaplan-Meier method and compared using the log-rank test.
Among the 944 study patients (median age at diagnosis, 43 years range, 23-67 years) during a median follow-up of 85 months (range, 14-185 months), 39 cases (4.1%) of cancer recurrence at the NAC were identified as the first event after NSM. The 5-year cumulative incidence of cancer recurrence at the NAC was 3.5% (n = 34). In multivariate analyses, multifocality or multicentricity (hazard ratio HR, 3.309; 95% CI, 1.501-7.294; P = .003), negative hormone receptor or ERBB2 (formerly HER2 or HER2/neu)-positive subtype (HR, 3.051; 95% CI, 1.194-7.796; P = .02), high histologic grade (HR, 2.641; 95% CI, 1.132-6.160; P = .03), and extensive intraductal component (HR, 3.338; 95% CI, 1.262-8.824; P = .02) were independently associated with cancer recurrence at the NAC after NSM. All 39 recurrent cases involved wide local excision. Patients with and without cancer recurrence at the NAC as the first event did not differ significantly with regard to distant metastasis-free survival (P = .95) or overall survival (P = .21). The 10-year distant metastasis-free survival rates were 89.3% among patients with cancer recurrence at the NAC and 94.3% among patients without recurrence. The 10-year overall survival rates were 100% among patients with cancer recurrence at the NAC and 94.5% among those without recurrence.
Patients had a low incidence of cancer recurrence at the NAC after NSM and immediate breast reconstruction in this study. The findings suggest that multifocal or multicentric disease, hormone receptor-negative/ERBB2-positive subtype, high histologic grade, and positive extensive intraductal component should be considered before determining the NSM procedure.
Mammography is the primary screening method for breast cancers. However, the sensitivity of mammographic screening is lower for dense breasts, which are an independent risk factor for breast cancers. ...Automated breast ultrasound (ABUS) is used as an adjunct to mammography for screening breast cancers in asymptomatic women with dense breasts. It is an effective screening modality with diagnostic accuracy comparable to that of handheld ultrasound (HHUS). Radiologists should be familiar with the unique display mode, imaging features, and artifacts in ABUS, which differ from those in HHUS. The purpose of this study was to provide a comprehensive review of the clinical significance of dense breasts and ABUS screening, describe the unique features of ABUS, and introduce the method of use and interpretation of ABUS.
Injury to axons in the peripheral nervous system induces rapid and local regenerative responses to form a new growth cone, and to generate a retrogradely transporting injury signal. The evidence for ...essential roles of intra-axonal protein synthesis during regeneration is now compelling. MicroRNA (miRNA) has recently been recognized as a prominent player in post-transcriptional regulation of axonal protein synthesis. Here, we directly contrast temporal changes of miRNA levels in the sciatic nerve following injury, as compared to those in an uninjured nerve using deep sequencing. Small RNAs (<200 nucleotides in length) were fractionated from the proximal nerve stumps to improve the representation of differential miRNA levels. Of 141 axoplasmic miRNAs annotated, 63 rat miRNAs showed significantly differential levels at five time points following injury, compared to an uninjured nerve. The differential changes in miRNA levels responding to injury were processed for hierarchical clustering analyses, and used to predict target mRNAs by Targetscan and miRanda. By overlapping these predicted targets with 2,924 axonally localizing transcripts previously reported, the overlapping set of 214 transcripts was further analyzed by the Gene Ontology enrichment and Ingenuity Pathway Analyses. These results suggest the possibility that the potential targets for these miRNAs play key roles in numerous neurological functions involved in ER stress response, cytoskeleton dynamics, vesicle formation, and neuro-degeneration and-regeneration. Finally, our results suggest that miRNAs could play a direct role in regenerative response and may be manipulated to promote regenerative ability of injured nerves.
Survival of breast cancer patients has improved, and treatment-related changes regarding metabolic profile deterioration after neoadjuvant systemic treatment (NST) become important issues in cancer ...survivors. We sought to compare metabolic profile changes and the neutrophil-to-lymphocyte ratio (NLR) between patients undergoing neoadjuvant chemotherapy (NCT) and neoadjuvant endocrine therapy (NET) 3 years after the treatment. In a prospective, randomized, phase III trial which compared 24 weeks of NCT with adriamycin and cyclophosphamide followed by docetaxel and NET with goserelin and tamoxifen (NEST), 123 patients in the Asan Medical Center were retrospectively reviewed to evaluate metabolic changes, such as body mass index (BMI), blood pressure (BP), total cholesterol (TC), fasting glucose, and the NLR. The mean age of patients was 42 years. The changes in BMI, serum glucose, and TC during NST and after 3 years were significantly different between NCT and NET. The proportion of overweight + obese group and the mean BMI were significantly increased during NCT (26.6% to 37.5%, 22.84 kg/m
to 23.87 kg/m
, p < 0.05), and these attributes found to have normalized at the 3-year follow-up. In the NET group, BMI changes were not observed (p > 0.05, all). There were no differences in changes over time among in the Hypertension group during NCT and NET (p = 0.96). The mean value of serum TC and fasting glucose significantly increased (< 0.05, both) during NCT and decreased 3 years after NCT (p < 0.05); however, no significant changes were observed in the NET group. The NLR was increased from 1.83 to 3.18 after NCT (p < 0.05) and decreased from 1.98 to 1.43 (p < 0.05) after NET. Compared with minimal metabolic effect of NET, NCT worsens metabolic profiles, which were recovered over 3 years. The NLR was increased after NCT but decreased after NET.
To evaluate the oncologic outcomes and risk factors for locoregional recurrence (LRR) and nipple-areola complex recurrence (NR) in a large series of breast cancer patients who underwent ...nipple-sparing mastectomy (NSM) and immediate reconstruction after neoadjuvant chemotherapy (NACT).
The use of NSM and immediate reconstruction in breast cancer patients receiving NACT is increasing. However, the oncologic safety of this approach is unclear.
A total of 310 breast cancer patients (319 breasts) who underwent NACT and NSM between February 2010 and November 2016 were retrospectively analyzed. Clinical and pathologic factors associated with increased risks of LRR and NR were analyzed using univariate (Chi-square or Fisher exact test) and multivariate (Cox proportional hazard regression model) analyses.
During a mean follow-up of 63 ± 22 months, 38 cases had LRR as the first event, including 6 cases of NR as the first event. The 5-year cumulative LRR and NR rates were 11.0% and 1.9%, respectively. In univariate analysis, clinical T stage, pathologic nodal status, histologic grade, lymphovascular invasion, and post-NACT Ki67 status were associated with increased LRR risk, and post-NACT Ki67 status was the only significant risk factor for NR. In multivariate analysis, post-NACT Ki67 ≥10% (hazard ratio, 4.245; 95% confidence interval, 1.865-9.663; P = 0.001) was an independent risk factor for LRR.
NSM and immediate reconstruction seem to be oncologically safe with acceptable LRR and NR rates for appropriately selected breast cancer patients treated with NACT. Post-NACT Ki67 ≥10% was associated with increased risk of LRR or NR, and therefore, necessitates cautious follow-up.
Enhancement in the efficiency of a TiO2 dye-sensitized solar cell (DSSC) has been demonstrated by introducing ferromagnetic perovskite BiFeO3 and controlling the magnetic field, which induces ...two-dimensional material-like properties in the bulk of the TiO2-BiFeO3 DSSC (a 3-dimensional material). The effect of the concentration of BiFeO3 as well as the magnetization direction on the performance of the TiO2-BiFeO3 DSSC has been investigated. After magnetization, it was confirmed that the current density, efficiency, and open circuit voltage of the TiO2-BiFeO3 DSSC were increased. The observed phenomena have been explained in terms of the Hall effect which is responsible for the reduction of the degree of freedom of the electron movement resulting in the two-dimensional material-like properties in the bulk of the TiO2-BiFeO3 DSSC.
•The information from MRI and DBT was better correlated with pathology than that from MG and ABUS for predicting residual tumor size after NAC.•For predicting pCR, MRI and DBT outperform MG and ...ABUS.•Among molecular subtypes, luminal B tumors showed very good correlation with DBT.
To compare the accuracy of mammography (MG), digital breast tomosynthesis (DBT), automated breast ultrasound (ABUS) and magnetic resonance imaging (MRI) for the assessment of residual tumor extent in breast cancer after neoadjuvant chemotherapy (NAC).
Fifty-one stage II-III breast cancer undergoing NAC were enrolled from March 2015 to December 2016. The longest diameter of residual tumor measured with MG, DBT, ABUS and MRI was compared with the pathologic tumor size. Statistical analysis was performed using intraclass correlation coefficients (ICC) and marginal homogeneity test. Receiver operating characteristics (ROC) analysis was used to evaluate the diagnostic performance for predicting pathologic complete response (pCR).
MRI size correlated well with pathology (ICC = 0.83), significantly better than MG, DBT and ABUS size (ICC = 0.56, ICC = 0.63 and ICC = 0.55, respectively). The discrepancy between MRI and pathology was statistical different from that of MG and ABUS (p = 0.0231 and 0.0039, respectively), but not different from that of DBT (p = 0.5727). For predicting pCR, MRI and DBT had a better performance compared to MG and US (area under the ROC curve: 0.92, 0.84, 0.72, 0.75, respectively; p = 0.3749 for DBT, p = 0.0972 for MG and p = 0.0596 for ABUS, when MRI being reference).
MRI and DBT allow more accurate assessment of tumor size compared to pathology compared with MG and ABUS. MRI and DBT outperform MG and ABUS in the prediction of pathologic complete response.
The purpose of this study was to prospectively evaluate interobserver agreement on lesion detection and characterization in the review of automated ultrasound images of the breast by five ...radiologists.
From August to October 2009, bilateral whole-breast ultrasound examinations were performed with an automated technique and with a handheld device for 55 women consecutively scheduled to undergo diagnostic ultrasound. Three-dimensional volume data from automated ultrasound were reviewed by five radiologists, who were unaware of the results of ultrasound with a handheld device and mammography and of the clinical information. If a lesion was detected with automated ultrasound, clock-face position, distance from the nipple, largest diameter, and BI-RADS final assessment category were evaluated. If the lesion was a mass, shape, orientation, margin, echogenicity, and posterior feature were analyzed. Intraclass correlation coefficients and kappa statistics were used for statistical analysis.
At least two observers identified 145 lesions with automated ultrasound. Among 725 possible detections, 587 (81%) detections were made. Individual investigators detected between 74% (107/145) and 88% (127/145) of the lesions. The rate of detection of lesions larger than 1.2 cm was 92%. Most lesions detected only with handheld ultrasound (11/12, 92%) or automated ultrasound (34/36, 94%) were cysts or probably benign masses. All intraclass correlation coefficients for lesion location and size exceeded 0.75, indicating high reliability. Substantial agreement was found for mass shape (κ = 0.71), orientation (κ = 0.72), margin (κ = 0.61), and BI-RADS final assessment category (κ = 0.63).
Detection of lesions larger than 1.2 cm in greatest diameter was reliable. High reliability was obtained for reporting lesion size and location. Substantial agreement was obtained for description of key feature and final assessment category.