Screening mammography performance according to breast density: a comparison between radiologists versus standalone intelligence detection.

BACKGROUND: Artificial intelligence (AI) algorithms for the independent assessment of screening mammograms have not been well established in a large screening cohort of Asian women. We compared the performance of screening digital mammography considering breast density, between radiologists and AI standalone detection among Korean women. METHODS: We retrospectively included 89,855 Korean women who underwent their initial screening digital mammography from 2009 to 2020. Breast cancer within 12 months of the screening mammography was the reference standard, according to the National Cancer Registry. Lunit software was used to determine the probability of malignancy scores, with a cutoff of 10% for breast cancer detection. The AI's performance was compared with that of the final Breast Imaging Reporting and Data System category, as recorded by breast radiologists. Breast density was classified into four categories (A-D) based on the radiologist and AI-based assessments. The performance metrics (cancer detection rate [CDR], sensitivity, specificity, positive predictive value [PPV], recall rate, and area under the receiver operating characteristic curve [AUC]) were compared across breast density categories. RESULTS: Mean participant age was 43.5 ± 8.7 years; 143 breast cancer cases were identified within 12 months. The CDRs (1.1/1000 examination) and sensitivity values showed no significant differences between radiologist and AI-based results (69.9% [95% confidence interval [CI], 61.7-77.3] vs. 67.1% [95% CI, 58.8-74.8]). However, the AI algorithm showed better specificity (93.0% [95% CI, 92.9-93.2] vs. 77.6% [95% CI, 61.7-77.9]), PPV (1.5% [95% CI, 1.2-1.9] vs. 0.5% [95% CI, 0.4-0.6]), recall rate (7.1% [95% CI, 6.9-7.2] vs. 22.5% [95% CI, 22.2-22.7]), and AUC values (0.8 [95% CI, 0.76-0.84] vs. 0.74 [95% CI, 0.7-0.78]) (all P < 0.05). Radiologist and AI-based results showed the best performance in the non-dense category; the CDR and sensitivity were higher for radiologists in the heterogeneously dense category (P = 0.059). However, the specificity, PPV, and recall rate consistently favored AI-based results across all categories, including the extremely dense category. CONCLUSIONS: AI-based software showed slightly lower sensitivity, although the difference was not statistically significant. However, it outperformed radiologists in recall rate, specificity, PPV, and AUC, with disparities most prominent in extremely dense breast tissue.

Use of a commercial artificial intelligence-based mammography analysis software for improving breast ultrasound interpretations.

OBJECTIVES: To evaluate the use of a commercial artificial intelligence (AI)-based mammography analysis software for improving the interpretations of breast ultrasound (US)-detected lesions. METHODS: A retrospective analysis was performed on 1109 breasts that underwent both mammography and US-guided breast biopsy. The AI software processed mammograms and provided an AI score ranging from 0 to 100 for each breast, indicating the likelihood of malignancy. The performance of the AI score in differentiating mammograms with benign outcomes from those revealing cancers following US-guided breast biopsy was evaluated. In addition, prediction models for benign outcomes were constructed based on clinical and imaging characteristics with and without AI scores, using logistic regression analysis. RESULTS: The AI software had an area under the receiver operating characteristics curve (AUROC) of 0.79 (95% CI, 0.79-0.82) in differentiating between benign and cancer cases. The prediction models that did not include AI scores (non-AI model), only used AI scores (AI-only model), and included AI scores (integrated model) had AUROCs of 0.79 (95% CI, 0.75-0.83), 0.78 (95% CI, 0.74-0.82), and 0.85 (95% CI, 0.81-0.88) in the development cohort, and 0.75 (95% CI, 0.68-0.81), 0.82 (95% CI, 0.76-0.88), and 0.84 (95% CI, 0.79-0.90) in the validation cohort, respectively. The integrated model outperformed the non-AI model in the development and validation cohorts (p < 0.001 for both). CONCLUSION: The commercial AI-based mammography analysis software could be a valuable adjunct to clinical decision-making for managing US-detected breast lesions. CLINICAL RELEVANCE STATEMENT: The commercial AI-based mammography analysis software could potentially reduce unnecessary biopsies and improve patient outcomes. KEY POINTS: • Breast US has high rates of false-positive interpretations. • A commercial AI-based mammography analysis software could distinguish mammograms having benign outcomes from those revealing cancers after US-guided breast biopsy. • A commercial AI-based mammography analysis software may improve interpretations for breast US-detected lesions.

Clinical outcomes after incomplete cycles of R-CHOP for diffuse large B-cell lymphoma: 10 years' real-world experience in a single institute.

Although the current standard of care for diffuse large B-cell lymphoma (DLBCL) is six cycles of rituximab/cyclophosphamide/doxorubicin/vincristine/prednisolone combination chemotherapy (R-CHOP), a larger than expected number of patients cannot complete planned six cycles for various reasons in the real world. We aimed to evaluate the prognosis of patients with DLBCL after incomplete treatment by analyzing the chemotherapy response and survival according to the cause of discontinuation and the number of cycles. We analyzed a retrospective cohort of patients diagnosed with DLBCL who underwent incomplete cycles of R-CHOP at Seoul National University Hospital and Boramae Medical Center from January 2010 to April 2019. A total of 1183 patients were diagnosed with DLBCL, of which 260 (22%) did not complete six cycles of R-CHOP. The most common cause of discontinuation of chemotherapy was life-threatening infection, and the most common pathogen was Pneumocystis jirovecii. Overall survival (OS) and progression-free survival (PFS) were significantly better in patients who achieved complete response (CR) or partial response (PR) at the first response evaluation. Patients underwent three or more cycles of chemotherapy had a longer OS than those who did not. In patients with limited-stage disease, consolidative radiotherapy showed a significant improvement in OS and PFS. Advanced stage, high comorbidity score, and poor primary response to chemotherapy were poor prognostic factors in patients with unplanned treatment shortening. This study provides real-world outcomes for patients who could not complete the planned six cycles of R-CHOP.

Prognostic roles of leptin-signaling proteins, PD-L1, and tumor-infiltrating lymphocytes in surgically-resected biliary tract cancers.

BACKGROUND: Biliary tract cancers are rare, with a poor patient prognosis. Leptin and programmed death-ligand 1 (PD-L1) influence CD8+ and forkhead box P3 (FOXP3)+ lymphocytes, and thus, cancer cell growth. We aimed to define the prognostic implications of these variables and the clinicopathological features of biliary tract cancers. METHODS: Immunohistochemistry for leptin signaling-related proteins (leptin, leptin receptor, pSTAT3, extracellular-regulated kinase, mammalian target of rapamycin), PD-L1, CD8, and FOXP3 and in situ hybridization for Epstein-Barr virus-encoded small RNAs were performed in 147 cases of surgically-resected biliary tract cancers. RESULTS: Immune cell PD-L1-positivity, tumor size < 3 cm, adjuvant chemotherapy, no recurrence, and early-stage tumors were correlated with better 5-year survival in the tumoral PD-L1(-) and leptin(-) subgroups, and extrahepatic cholangiocarcinoma through multivariate analysis (all p < 0.05). Immune cell PD-L1 and adjuvant chemotherapy lost its prognostic significance in the tumoral PD-L1+ and leptin+ subgroups. CONCLUSIONS: The prognostic implication of the variables may depend upon tumoral protein expression and the anatomical site. Immune cell PD-L1-positivity and the administration of adjuvant chemotherapy may indicate the favorable survival of patients with surgically-resected biliary tract cancers, specifically, in the tumoral PD-L1(-) or tumor leptin(-) subgroups and extrahepatic cholangiocarcinoma. PD-L1- or leptin-targeted therapy combined with conventional chemotherapy may benefit the tumoral PD-L1+ or leptin+ subgroups.

FBI-1 inhibits epithelial-to-mesenchymal transition, migration, and invasion in lung adenocarcinoma A549 cells by downregulating transforming growth factor-ß1 signaling pathway.

The factor binding inducer of short transcripts-1 (FBI-1) is a POZ-domain Kruppel-like (POK) family of transcription factors and is known as a proto-oncogene or tumor suppressor in various carcinomas. However, the role of FBI-1 on epithelial-to-mesenchymal transition (EMT) and invasiveness in lung cancer remains unknown. Preliminarily, clinical data such as tissue microarray, Kaplan-Meier, and Oncomine were analyzed to confirm the correlation between lung cancer metastasis and FBI-1. To investigate the function of FBI-1 in EMT in lung cancer, EMT was measured in FBI-1-deficient or FBI-1-overexpressing cells. FBI-1 showed decreased expression in tumors metastasized to lymph nodes than in the primary tumor. In addition, it was also associated with improved survival rates of lung cancer patients. FBI-1 knockdown improved E-to-N-cadherin switching, migration, and invasion in A549 cells, similar to the initiation of EMT stimulated by transforming growth factor- ß1 (TGF-ß1). In contrast, overexpression of FBI-1 inhibited the transcription and activation of Smad2, thereby interfering with EMT, despite stimulation by TGF-ß1. These results suggest that FBI-1 plays a negative role in EMT in lung cancer via the TGF-ß1 signaling pathway, implying its use as a new potential therapeutic target and diagnostic indicator for early stage of lung cancer metastasis.

Varying undersampling directions for accelerating multiple acquisition magnetic resonance imaging.

In this study, we propose a new sampling strategy for efficiently accelerating multiple acquisition MRI. The new sampling strategy is to obtain data along different phase-encoding directions across multiple acquisitions. The proposed sampling strategy was evaluated in multicontrast MR imaging (T1, T2, proton density) and multiple phase-cycled (PC) balanced steady-state free precession (bSSFP) imaging by using convolutional neural networks with central and random sampling patterns. In vivo MRI acquisitions as well as a public database were used to test the concept. Based on both visual inspection and quantitative analysis, the proposed sampling strategy showed better performance than sampling along the same phase-encoding direction in both multicontrast MR imaging and multiple PC-bSSFP imaging, regardless of sampling pattern (central, random) or datasets (public, retrospective and prospective in vivo). For the prospective in vivo applications, acceleration was performed by sampling along different phase-encoding directions at the time of acquisition with a conventional rectangular field of view, which demonstrated the advantage of the proposed sampling strategy in the real environment. Preliminary trials on compressed sensing (CS) also demonstrated improvement of CS with the proposed idea. Sampling along different phase-encoding directions across multiple acquisitions is advantageous for accelerating multiacquisition MRI, irrespective of sampling pattern or datasets, with further improvement through transfer learning.

AfSec1 is a signal peptidase and removes signal peptides of 1,3-ß-glucanosyltransferases in Aspergillus fumigatus.

To identify the infection mechanism of Aspergillus fumigatus, which is an opportunistic fungal pathogen, we analyzed the expression profile of the whole genome of A. fumigatus during the infection of murine macrophages. A previously reported RNA-seq data analysis showed that many genes involved in cell wall synthesis were upregulated during the infection process. Interestingly, AfSec1 (3g12840), which encodes a putative signal peptidase, was upregulated dramatically, and its putative target protein Gel1, which encodes a 1,3-ß-glucanosyltransferase, was also upregulated. Instead of the AfSec1 deletion strain, the AfSec1-ΔP strain was constructed, in which the promoter region of AfSec1 was deleted, and AfSec1 expression was not detected in the AfSec1-ΔP strain. The expression of AfSec1 was recovered by the introduction of the promoter region (the AfSec1-ΔP/P strain). The nonprocessed form of Gel1 was identified in the AfSec1-ΔP strain, which lacked the promoter, but mature forms of Gel1 were found in the wild-type and in AfSec1-ΔP/P, which was the promoter complementation strain. In the plate assay, the AfSec1-ΔP strain showed higher sensitivity against caspofungin than the wild-type. However, compared with the wild-type, the deletion strain showed no difference in the sensitivity to other antifungal drugs, such as amphotericin B and voriconazole, which inhibit different targets compared with caspofungin. The AfSec1-ΔP strain exhibited ∼20% lower levels of ß-glucan in the cell wall than the wild-type. Finally, the virulence decreased when the promoter region of AfSec1 was deleted, as observed in the murine infection test and conidia-killing assay using human macrophages and neutrophils. These results suggest that AfSec1 exerts signal peptidase activity on its target Gel1 and has an important role in fungal pathogenesis.

We identified the novel signal peptidase AfSec1 from A. fumigatus. AfSec1 which removes the signal peptide of 1,3-ß-glucanosyltransferases, is a virulence factor and is a potential target for a new antifungal drug.

Outliers in SAR and QSAR: 4. effects of allosteric protein-ligand interactions on the classical quantitative structure-activity relationships.

Effects of allosteric interactions on the classical structure-activity relationship (SAR) and quantitative SAR (QSAR) have been investigated. Apprehending the outliers in SAR and QSAR studies can improve the quality, predictability, and use of QSAR in designing unknown compounds in drug discovery research. We explored allosteric protein-ligand interactions as a possible source of outliers in SAR/QSAR. We used glycogen phosphorylase as an example of a protein that has an allosteric site. Examination of the ligand-bound x-ray crystal structures of glycogen phosphorylase revealed that many inhibitors bound at more than one binding site. The results of QSAR analyses of the inhibitors included a QSAR that recognized an outlier bound at a distinctive allosteric binding site. The case provided an example of constructive use of QSAR identifying outliers with alternative binding modes. Other allosteric QSARs that captured our attention were the inverted parabola/bilinear QSARs. The x-ray crystal structures and the QSAR analyses indicated that the inverted parabola QSARs could be associated with the conformational changes in the allosteric interactions. Our results showed that the normal parabola, as well as the inverted parabola QSARs, can describe the allosteric interactions. Examination of the ligand-bound X-ray crystal structures of glycogen phosphorylase revealed that many inhibitors bound at more than one binding site. The results of QSAR analyses of the inhibitors included a QSAR that recognized an outlier bound at a distinctive allosteric binding site.

Outliers in SAR and QSAR: 3. Importance of considering the role of water molecules in protein-ligand interactions and quantitative structure-activity relationship studies.

It is frequently mentioned that QSARs have not generally lived up to expectations, especially in cases where high predictability is expected yet failed to deliver satisfactory results. Even though outliers can provide an increased opportunity in drug discovery research, outliers in SAR and QSAR can contort predictions and affect the accuracy if proper attention is not given. The percentages of outliers in QSARs have not changed appreciably over the last decade. In our previous studies, we suggested two possible sources of outliers in SAR and QSAR. In this paper, we suggest an additional possible source of outliers in QSAR. We presented several literature examples that show one or more water molecules that play a critical role in protein-ligand binding interactions as observed in their crystal structures. These examples illustrate that failing to account for the effects of water molecules in protein-ligand interactions could mislead interpretation and possibly yield outliers in SAR and QSAR. Examples include cases where QSAR, considering the role of water molecules in protein-ligand crystal structures, provided deeper insight into the understanding and interpretation of the developed QSAR.

Efficacy of tetracyclines and fluoroquinolones for the treatment of macrolide-refractory Mycoplasma pneumoniae pneumonia in children: a systematic review and meta-analysis.

BACKGROUND: Mycoplasma pneumoniae is a common pathogen that causes community-acquired pneumonia in school-age children. Macrolides are considered a first-line treatment for M. pneumoniae infection in children, but macrolide-refractory M. pneumoniae (MRMP) strains have become more common. In this study, we assessed the efficacy of tetracyclines and fluoroquinolones in MRMP treatment in children through a systematic review and meta-analysis. METHODS: Two reviewers individually searched 10 electronic databases (Medline/Pubmed, Embase, the Cochrane Library, and core Korean, Chinese, and Japanese journals) for papers published from January 1, 1990 to March 8, 2018. The following data for each treatment group were extracted from the selected studies: intervention (tetracyclines and fluoroquinolones/comparator), patient characteristics (age and sex), and outcomes (fever duration, hospital stay length, treatment success rate, and defervescence rates 24, 48, and 72 h after starting treatment). RESULTS: Eight studies involving 537 participants were included. Fever duration and hospital stay length were shorter in the tetracycline group than in the macrolide group (weighted mean difference [WMD] = - 1.45, 95% confidence interval [CI]: - 2.55 to - 0.36, P = 0.009; and WMD = - 3.33, 95% CI: - 4.32 to - 2.35, P < 0.00001, respectively). The therapeutic efficacy was significantly higher in the tetracycline group than in the macrolide group (odds ratio [OR]: 8.80, 95% CI: 3.12-24.82). With regard to defervescence rate, patients in the tetracycline group showed significant improvement compared to those in the macrolide group (defervescence rate after 24 h, OR: 5.34, 95% CI: 1.81-15.75; after 48 h, OR 18.37, 95% CI: 8.87-38.03; and after 72 h, OR: 40.77, 95% CI: 6.15-270.12). There were no differences in fever improvement within 24 h in patients in the fluoroquinolone group compared to those in the macrolide group (OR: 1.11, 95% CI: 0.25-5.00), although the defervescence rate was higher after 48 h in the fluoroquinolone group (OR: 2.78, 95% CI: 1.41-5.51). CONCLUSION: Tetracyclines may shorten fever duration and hospital stay length in patients with MRMP infection. Fluoroquinolones may achieve defervescence within 48 h in patients with MRMP infection. However, these results should be carefully interpreted as only a small number of studies were included, and they were heterogeneous.

Immunogenicity and Safety of a Newly Developed Tetanus-Diphtheria Toxoid (Td) in Healthy Korean Adolescents: a Multi-center, Randomized, Double-blind, Active-Controlled Phase 3 Trial.

BACKGROUND: Although the combination tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis vaccine (Tdap) is recommended at adolescence in developed countries, the tetanus and diphtheria toxoid vaccine (Td), which is less costly, is recommended instead in some parts of the world. A new Td, BR-TD-1001, was developed by a Korean manufacturer for distribution to endemic regions and for use in the initial step of novel Tdap development. METHODS: This phase 3, randomized, double-blind, multi-center trial, conducted in Korea, aimed to evaluate the immunogenicity and safety of BR-TD-1001. Healthy children aged 10 to 12 years were randomized 1:1 to receive either BR-TD-1001 or the control Td (Td-pur, GlaxoSmithKline). Antibodies were measured using enzyme-linked immunosorbent assay. RESULTS: A total of 218 subjects (BR-TD-1001, n = 108; control, n = 110) were enrolled and included in the safety analysis. Vaccine-mediated antibody responses were similar in both groups. We confirmed the non-inferiority of BR-TD-1001 against the control, Td; 100% of both groups achieved seroprotection against diphtheria and tetanus. Furthermore, there was no significant difference between groups in the proportion of participants who demonstrated boost responses against diphtheria and tetanus toxoids. The incidence of solicited local and systemic adverse events (AEs), unsolicited AEs, and serious AEs did not differ significantly between groups. CONCLUSION: The BR-TD-1001 satisfied the immunological non-inferiority criterion against diphtheria and tetanus, with a clinically acceptable safety profile. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT04618939.

Secure and Efficient High Throughput Medium Access Control for Vehicular Ad-Hoc Network.

The evolution of the internet has led to the growth of smart application requirements on the go in the vehicular ad hoc network (VANET). VANET enables vehicles to communicate smartly among themselves wirelessly. Increasing usage of wireless technology induces many security vulnerabilities. Therefore, effective security and authentication mechanism is needed to prevent an intruder. However, authentication may breach user privacy such as location or identity. Cryptography-based approach aids in preserving the privacy of the user. However, the existing security models incur communication and key management overhead since they are designed considering a third-party server. To overcome the research issue, this work presents an efficient security model namely secure performance enriched channel allocation (S-PECA) by using commutative RSA. This work further presents the commutative property of the proposed security scheme. Experiments conducted to evaluate the performance of the proposed S-PECA over state-of-the-art models show significant improvement. The outcome shows that S-PECA minimizes collision and maximizes system throughput considering different radio propagation environments.

Clinical significance of radiation dose-volume parameters and functional status on the patient-reported quality of life changes after thoracic radiotherapy for lung cancer: a prospective study.

PURPOSE: To date, limited data exist about the relationship between radiation dose-volume parameters and patient-reported quality of life (QOL) after thoracic radiotherapy (RT) for lung cancer. We conducted this prospective study to investigate which clinico-dosimetric factors have an impact on functional declines and symptom developments after thoracic RT for lung cancer. MATERIALS AND METHODS: The study included 44 patients who had underwent thoracic three-dimensional conformal RT at our institution from 2016 to 2017. The health-related QOL was assessed using the EORTC QLQ-C30 and QLQ-LC13 questionnaires before RT (preRT), at the end of RT (endRT), and 3, 6, and 12 months after the completion of RT. RT dose-volume parameters of adjacent normal organs such as the lung, heart, and esophagus were retrieved and used for regression analysis. RESULTS: Thoracic RT induced a temporary deterioration of many of the functional statuses and symptoms, but most of those improved and recovered to baseline levels 3 months after RT. However, the role function (RF) decline persisted until 6 months after RT (p < 0.05). Dysphagia showed the most noticeable change at the endRT (p < 0.001). In the multiple regression analysis, the absolute volume of body received at least 50 Gy (p = 0.021) and a preRT RF score (p = 0.001) was significantly associated with the endRT RF scores. Dysphagia at the endRT was significantly associated with the V40 of the esophagus (p = 0.047), preRT emotional function (p = 0.029), and receipt of concurrent chemotherapy (p = 0.022). CONCLUSIONS: Both the dosimetric parameters and preRT functional status have an impact on the weak aspect of patient-reported QOL, which may cause poor treatment compliance during and after thoracic RT. For patients with a low preRT QOL score or those having large tumor which may result in higher dose volumes, careful RT planning could prevent the deterioration of QOL after RT.

Artificial neural network for Slice Encoding for Metal Artifact Correction (SEMAC) MRI.

PURPOSE: To develop new artificial neural networks (ANNs) to accelerate slice encoding for metal artifact correction (SEMAC) MRI. METHODS: Eight titanium phantoms and 77 patients after brain tumor surgery involving metallic neuro-plating instruments were scanned using SEMAC at a 3T Skyra scanner. For the phantoms, proton-density, T1-, and T2-weighted images were acquired for developing both multilayer perceptron (MLP) and convolutional neural network (CNN). For the patients, T2-weighted images were acquired for developing CNN. All networks were trained with the SEMAC factor 4 or 6 as input and the factor 12 as label, yielding an acceleration factor of 3 or 2. Performance of the CNN model was compared against parallel imaging and compressed sensing on the phantom datasets. Two extra T1-weighted in vivo sets were acquired to investigate generalizability of the models to different contrasts. RESULTS: Both multilayer perceptron and CNN provided artifact-suppressed images better than the input images and comparable to the label images visually and quantitatively, a trend observable regardless of input SEMAC factor and image type (P < .01). CNN suppressed the artifacts better than multilayer perceptron, parallel imaging, and compressed sensing (P < .01). Tests on the patient datasets demonstrated clear metal artifact suppression visually and quantitatively (P < .01). Tests on T1 datasets also demonstrated clear visual metal artifact suppression. CONCLUSION: Our study introduced a new effective way of artificial neural networks to accelerate SEMAC MRI while maintaining the comparable quality of metal artifact suppression. Application on the preliminary patient datasets proved the feasibility in clinical usage, which warrants further investigation.

Deep Learning for Chest Radiograph Diagnosis in the Emergency Department.

BackgroundThe performance of a deep learning (DL) algorithm should be validated in actual clinical situations, before its clinical implementation.PurposeTo evaluate the performance of a DL algorithm for identifying chest radiographs with clinically relevant abnormalities in the emergency department (ED) setting.Materials and MethodsThis single-center retrospective study included consecutive patients who visited the ED and underwent initial chest radiography between January 1 and March 31, 2017. Chest radiographs were analyzed with a commercially available DL algorithm. The performance of the algorithm was evaluated by determining the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity at predefined operating cutoffs (high-sensitivity and high-specificity cutoffs). The sensitivities and specificities of the algorithm were compared with those of the on-call radiology residents who interpreted the chest radiographs in the actual practice by using McNemar tests. If there were discordant findings between the algorithm and resident, the residents reinterpreted the chest radiographs by using the algorithm's output.ResultsA total of 1135 patients (mean age, 53 years ± 18; 582 men) were evaluated. In the identification of abnormal chest radiographs, the algorithm showed an AUC of 0.95 (95% confidence interval [CI]: 0.93, 0.96), a sensitivity of 88.7% (227 of 256 radiographs; 95% CI: 84.1%, 92.3%), and a specificity of 69.6% (612 of 879 radiographs; 95% CI: 66.5%, 72.7%) at the high-sensitivity cutoff and a sensitivity of 81.6% (209 of 256 radiographs; 95% CI: 76.3%, 86.2%) and specificity of 90.3% (794 of 879 radiographs; 95% CI: 88.2%, 92.2%) at the high-specificity cutoff. Radiology residents showed lower sensitivity (65.6% [168 of 256 radiographs; 95% CI: 59.5%, 71.4%], P < .001) and higher specificity (98.1% [862 of 879 radiographs; 95% CI: 96.9%, 98.9%], P < .001) compared with the algorithm. After reinterpretation of chest radiographs with use of the algorithm's outputs, the sensitivity of the residents improved (73.4% [188 of 256 radiographs; 95% CI: 68.0%, 78.8%], P = .003), whereas specificity was reduced (94.3% [829 of 879 radiographs; 95% CI: 92.8%, 95.8%], P < .001).ConclusionA deep learning algorithm used with emergency department chest radiographs showed diagnostic performance for identifying clinically relevant abnormalities and helped improve the sensitivity of radiology residents' evaluation.Published under a CC BY 4.0 license.Online supplemental material is available for this article.See also the editorial by Munera and Infante in this issue.

Comparison of quality of life and health behaviors in survivors of acute leukemia and the general population.

We aimed to compare the health-related quality of life and health behaviors of acute leukemia (AL) survivors with that of the general population from two cohorts. AL survivors (n = 149) completed a set of questionnaires to evaluate quality of life, mental status, and health behaviors. AL survivors had more physical and mental difficulties (problems with usual activities, 15% vs. 5%, p < 0.001; anxiety or depression, 24% vs. 9%, p < 0.001; pain, 35% vs. 20%, p = 0.002) and more financial difficulties (p < 0.001) than the general population. Survivors who received stem cell transplantation (SCT) had significantly worse problems with role functioning, fatigue, pain, dyspnea, and insomnia, and had higher depression scores than chemotherapy group (p = 0.024). In terms of health behaviors, AL survivors had lower rates of smoking and drinking and higher influenza vaccination rates than the general population. However, only 17% of survivors had been recommended to receive screening for other cancers from health-care providers, and 67% thought their risk for other cancers was equal or lower than that of the general population. Cancer screening rates were even lower in the SCT group than in the chemotherapy group (p = 0.041). Our study indicates that clinicians should establish more appropriate survivorship care plans.

Evaluation of Waning Immunity at 6 Months after Both Trivalent and Quadrivalent Influenza Vaccination in Korean Children Aged 6-35 Months.

BACKGROUND: The titer of influenza vaccine-induced antibodies declines over time, and younger children have lower immunogenicity and shorter duration of immunity. This study aimed to compare persistence of antibody at 6 months after influenza vaccination according to influenza virus strains, vaccine type, antigen dose, and primed status in children aged 6 to 35 months. METHODS: A total 124 healthy children aged 6 to 35 months were enrolled from September to December 2016 at 10 hospitals in Korea and randomly assigned to either a full dose of quadrivalent influenza vaccine or a half dose of trivalent influenza vaccine with Victoria B strain group. Hemagglutination inhibition antibody titers (that measure the seroprotection rates) were assessed for the recommended influenza strains at 6 months post vaccination. RESULTS: The seroprotection rates at 6 months for strains A (H1N1), A (H3N2), B/Yamagata, and B/Victoria were 88.7%, 97.4%, 36.6%, and 27.6%, respectively. The seroprotection rates for A (H1N1), A (H3N2) and B (Victoria) were 91.4%, 98.7% and 27.5% in a full dose of quadrivalent vaccine vs. 83.7%, 94.6% and 27.9% in a half dose trivalent vaccine, respectively. The seroprotection rate for the B (Yamagata) strain was 23.8% in the quadrivalent group and 14.0% in the trivalent group. CONCLUSION: Persistence of antibodies at 6 months was more favorable against the influenza A strains than against the B strains. Persistence of antibodies to additional B strain at 6 months was superior in the quadrivalent vaccine group. The immunity of primed children with different B strains was not superior to that of the unprimed group with another B strain.

Improving Arterial Spin Labeling by Using Deep Learning.

Purpose To develop a deep learning algorithm that generates arterial spin labeling (ASL) perfusion images with higher accuracy and robustness by using a smaller number of subtraction images. Materials and Methods For ASL image generation from pair-wise subtraction, we used a convolutional neural network (CNN) as a deep learning algorithm. The ground truth perfusion images were generated by averaging six or seven pairwise subtraction images acquired with (a) conventional pseudocontinuous arterial spin labeling from seven healthy subjects or (b) Hadamard-encoded pseudocontinuous ASL from 114 patients with various diseases. CNNs were trained to generate perfusion images from a smaller number (two or three) of subtraction images and evaluated by means of cross-validation. CNNs from the patient data sets were also tested on 26 separate stroke data sets. CNNs were compared with the conventional averaging method in terms of mean square error and radiologic score by using a paired t test and/or Wilcoxon signed-rank test. Results Mean square errors were approximately 40% lower than those of the conventional averaging method for the cross-validation with the healthy subjects and patients and the separate test with the patients who had experienced a stroke (P < .001). Region-of-interest analysis in stroke regions showed that cerebral blood flow maps from CNN (mean ± standard deviation, 19.7 mL per 100 g/min ± 9.7) had smaller mean square errors than those determined with the conventional averaging method (43.2 ± 29.8) (P < .001). Radiologic scoring demonstrated that CNNs suppressed noise and motion and/or segmentation artifacts better than the conventional averaging method did (P < .001). Conclusion CNNs provided superior perfusion image quality and more accurate perfusion measurement compared with those of the conventional averaging method for generation of ASL images from pair-wise subtraction images. © RSNA, 2017.

The Derived Neutrophil-to-Lymphocyte Ratio Is an Independent Prognostic Factor in Transplantation Ineligible Patients with Multiple Myeloma.

BACKGROUND: The neutrophil-to-lymphocyte ratio (NLR) is an independent prognostic marker in solid and hematological cancers. While the derived NLR (dNLR) was shown to be non-inferior to the NLR in large cohorts of patients with different cancer types, it has not been validated as a prognostic marker for multiple myeloma (MM) to date. METHODS: Between May 22, 2011 and May 29, 2014, 176 patients with MM from 38 centers who were ineligible for autologous stem cell transplantation were analyzed. The dNLR was calculated using complete blood count differential data. The optimal dNLR cut-off value according to receiver operating characteristic analysis of overall survival (OS) was 1.51. All patients were treated with melphalan and prednisone combined with bortezomib. RESULTS: The complete response rate was lower in the high dNLR group compared to the low dNLR group (7 vs. 26.1%, respectively; p = 0.0148); the corresponding 2-year OS rates were 72.2 and 84.7%, respectively (p = 0.0354). A high dNLR was an independent poor prognostic factor for OS (hazard ratio 2.217, 95% CI 1.015-4.842; p = 0.0458). CONCLUSION: The dNLR is a readily available and cheaply obtained parameter in clinical studies, and shows considerable potential as a new prognostic marker for transplantation-ineligible patients with MM.