Adaptive soft erasure with edge self-attention for weakly supervised semantic segmentation: Thyroid ultrasound image case study

[S U M M A R Y] Weakly supervised segmentation for medical images ease the reliance of models on pixel-level annotation while advancing the field of computer-aided diagnosis. However, the differences in nodule size in thyroid ultrasound images and the limitations of class activation maps in weakly supervised segmentation methods typically lead to under- and/or over-segmentation problems in real predictions. To alleviate this problem, we propose a weakly supervised segmentation neural network approach. This new method is based on a dual branch soft erase module that expands the foreground response region while constraining the erroneous expansion of the foreground region by the enhancement of background features. The sensitivity of this neural network to the nodule scale size is further enhanced by the scale feature adaptation module, which in turn generates integral and high-quality segmentation masks. In addition, while the nodule area can be significantly expanded through soft erase module and scale feature adaptation module, the activation effect in the nodule edge area is still not satisfactory, so that we further add an edge-based attention mechanism to strengthen the nodule edge segmentation effect. The results of experiments performed on the thyroid ultrasound image dataset showed that our new approach significantly outperformed existing weakly supervised semantic segmentation methods, e.g., 5.9% and 6.3% more accurate than the second-based results in terms of Jaccard and Dice coefficients, respectively.

A balanced social LSTM for PM2.5 concentration prediction based on local spatiotemporal correlation

Reliable prediction for the concentration of PM_2.5 has become a hot topic in pollution prevention. However, the prediction for PM_2.5 concentration remains a challenge, one of the reasons is that current prediction methods do not consider the relevance of PM_2.5 concentration among surrounding areas. In this paper, we propose the assumption that the PM_2.5 concentration has spatial interaction, which includes two parts: 1) The PM_2.5 concentrations observed by adjacent stations usually present relevant trends; 2) Stations with higher PM_2.5 concentration tend to show higher influences on neighboring areas. Based on the spatial interaction assumption, we propose a balanced social long short-term memory (BS-LSTM) neural network for the prediction of PM_2.5 concentration. BS-LSTM is composed of two kernel components: a social-LSTM based prediction model and a new balanced mean squared error (B-MSE) based loss function. On the one hand, to capture the spatiotemporal correlation of the PM_2.5 concentration among adjacent stations, we develop a social-LSTM based model which has advantages in describing the trend information of neighboring locations. On the other hand, considering the unbalanced influence caused by various local pollution levels, we design a new B-MSE loss function to assign different attention to the observation stations. In the experiments, we evaluate the proposed method on two real-world PM_2.5 datasets. The results indicate that BS-LSTM is promising, especially in the case of heavy pollution.

A dose-response meta-analysis to evaluate the relationship between high-density lipoprotein cholesterol and all-cause and cardiovascular disease mortality

PURPOSE

Previous studies have not fully described the relationship between high-density lipoprotein cholesterol (HDL-C) and death risks from all cause and cardiovascular disease (CVD). This study quantitatively evaluates HDL-C-mortality associations.

METHODS

Embase and PubMed databases were searched for relevant articles published up to 1 June 2019. Random-effects models were used to pool relative risks (RRs) and 95% confidence intervals (CIs). We used restricted cubic splines to model the dose-response association.

RESULTS

We identified 32 prospective cohort studies including 369,904 participants and 33,473 total deaths (9426 CVD deaths). Compared to the lowest HDL-C levels, all cause and CVD mortality risks were reduced by 18% (RR 0.82; 95% CI, 0.73-0.93) and 36% (0.64, 0.46-0.89), respectively, for the highest HDL-C levels. All cause and CVD mortality risks were reduced by 15% (0.85, 0.79-0.92) and 23% (0.77, 0.69-0.87), respectively, with each 1 mmol/L increment of HDL-C. We found evidence of nonlinear and negative dose-response associations of HDL-C with all cause and CVD mortality (P_nonlinearity < 0.001), and the lowest death risks from all cause and CVD were observed at approximately 1.34 and 1.55 mmol/L, respectively.

CONCLUSION

HDL-C is inversely associated with all cause and CVD mortality risks under approximately 2.05 and 2.33 mmol/L, respectively. Optimal doses require investigation via clinical practice or high-quality research.

Comprehensive analysis of peripheral blood non-coding RNAs identifies a diagnostic panel for fungal infection after transplantation

The occurrence of fungal infection seriously affects the survival and life quality of transplanted patients. The accurate diagnosis is of particular importance in the early stage of infection. To develop a novel diagnostic method for this kind of patient, we established a post-transplant immunosuppressed mice model with fungus inoculation and collected their peripheral blood at specific time points after infection. After screening by microarray, differentially expressed miRNAs and lncRNAs were selected and homologously analyzed with those of human beings from the gene database. These miRNAs and lncRNAs candidates were validated by qRT-PCR in peripheral blood samples from transplanted patients. We found that, compared with normal transplanted patients, the levels of miR-215 and miR-let-7 c were up-regulated in the plasma of patients with fungal infection (P < 0.01), while levels of miR-154, miR-193a, NR_027669.1, and NR_036506.1 were down-regulated in their peripheral blood mononuclear cells (P < 0.01). Principal component analysis shows that the expression pattern of the above RNAs was different between the two groups. A 6-noncoding-RNA detection panel was established by the support vector machine analysis, whose area under the ROC curve was 0.927. The accuracy, precision, sensitivity, and specificity of this model were 0.928, 0.919, 0.944, and 0.910, respectively. Though our detection panel has excellent diagnostic efficacy, its clinical application value still needs to be further confirmed by multi-center prospective clinical trials.

Effect of temperature on PTEs deportment and ecological risks of the biochars obtained from sewage sludge

Sewage sludge-derived biochars (SSBCs) were obtained at temperatures of 300, 500, and 700 °C to investigate the potentially toxic elements (PTEs) behaviors and assess the environmental acceptability for the possible application in the environment. Results indicated that PTEs exhibited diversely in the distribution of chemical speciation, while all elements tended to be immobilized in biochar matrix and the total amount elevated during the pyrolysis. The risk assessment of biochars implied a low degree of environmental risk for the utilization of SSBCs prepared at high temperatures. In addition, higher pyrolysis temperature alleviated the inhibition on the early seedling growth of Triticum aestivum L., with root elongation more sensitive to the biochar addition. PTEs, especially Cr, contributed much to the phytotoxicity of biochars as revealed by the principle component analysis (PCA) and leaner correlation analysis. Findings from this work illustrated that SSBCs prepared at higher temperatures might be more conductive to a wide range of applications with acceptable environmental risk.

Heterogeneity and evolution of tumour immune microenvironment in metastatic gastroesophageal adenocarcinoma

BACKGROUND

Tumour immune microenvironment heterogeneity is prevalent in numerous cancers and can negatively impact immunotherapy response. Immune heterogeneity and evolution in gastroesophageal adenocarcinoma (GEA) have not been studied in the past.

METHODS

Together with a multi-region sampling of normal, primary and metastatic tissues, we performed whole exome sequencing, TCR sequencing as well as immune cell infiltration estimation through deconvolution of gene expression signals.

RESULTS

We discovered high TCR repertoire and immune cell infiltration heterogeneity among metastatic sites, while they were homogeneous among primary and normal samples. Metastatic sites shared high levels of abundant TCR clonotypes with blood, indicating immune surveillance via blood. Metastatic sites also had low levels of tumour-eliminating immune cells and were undergoing heavy immunomodulation compared to normal and primary tumour tissues. There was co-evolution of neo-antigen and TCR repertoire, but only in patients with late diverging mutational evolution. Co-evolution of TCR repertoire and immune cell infiltration was seen in all except one patient.

CONCLUSIONS

Our findings revealed immune heterogeneity and co-evolution in GEA, which may inform immunotherapy decision-making.

Dependability of Automated Breast Ultrasound (ABUS) in Assessing Breast Imaging Reporting and Data System (BI-RADS) Category and Size of Malignant Breast Lesions Compared with Handheld Ultrasound (HHUS) and Mammography (MG)

Purpose

This study aimed to evaluate the dependability of automated breast ultrasound (ABUS) compared with handheld ultrasound (HHUS) and mammography (MG) on the Breast Imaging Reporting and Data System (BI-RADS) category and size assessment of malignant breast lesions.

Patients and Methods

A total of 344 confirmed malignant lesions were recruited. All participants underwent MG, HHUS, and ABUS examinations. Agreements on the BI-RADS category were evaluated. Lesion size assessed using the three methods was compared with the size of the pathological result as the control. Regarding the four major molecular subtypes, correlation coefficients between size on imaging and pathology were also evaluated.

Results

The agreement between ABUS and HHUS on the BI-RADS category was 86.63% (kappa = 0.77), whereas it was 32.22% (kappa = 0.10) between ABUS and MG. Imaging lesion size compared to pathologic lesion size was assessed correctly in 36.92%/52.91% (ABUS), 33.14%/48.84% (HHUS) and 33.44%/43.87% (MG), with the threshold of 3 mm/5 mm, respectively. The correlation coefficient of size of ABUS-Pathology (0.75, Spearman) was statistically higher than that of the MG-Pathology (0.58, Spearman) with P < 0.01, but not different from that of the HHUS-Pathology (0.74, Spearman) with P > 0.05. The correlation coefficient of ABUS-Pathology was statistically higher than that of MG-Pathology in the triple-negative subtype, luminal B subtype, and luminal A subtype (P<0.01).

Conclusion

The agreement between ABUS and HHUS in the BI-RADS category was good, whereas that between ABUS and MG was poor. ABUS and HHUS allowed a more accurate assessment of malignant tumor size compared to MG.

Gastrointestinal Goblet Cell Adenocarcinomas Harbor Distinctive Clinicopathological, Immune, and Genomic Landscape

Goblet cell adenocarcinoma (GCA) is a rare amphicrine tumor and difficult to diagnose. GCA is traditionally found in the appendix, but extra-appendiceal GCA may be underestimated. Intestinal adenocarcinoma with signet ring cell component is also very rare, and some signet ring cell carcinomas are well cohesive, having some similar morphological features to GCAs. It is necessary to differentiate GCA from intestinal adenocarcinomas with cohesive signet ring cell component (IACSRCC). The goal of this study is to find occurrence of extra-appendiceal GCA and characterize the histological, immunohistochemical, transcriptional, and immune landscape of GCA. We collected 12 cases of GCAs and 10 IACSRCCs and reviewed the clinicopathologic characters of these cases. Immunohistochemical stains were performed with synaptophysin, chromogranin A, CD56, somatostatin receptor (SSTR) 2, and Ki-67. Whole transcriptome RNA-sequencing was performed, and data were used to analyze differential gene expression and predict immune cell infiltration levels in GCA and IACSRCC. RNA-sequencing data for colorectal adenocarcinoma were gathered from TCGA data portal. Of the 12 patients with GCA, there were 4 women and 8 men. There were three appendiceal cases and nine extra-appendiceal cases. GCAs were immunohistochemically different from IACSRCC. GCA also had different levels of B-cell and CD8+ T-cell infiltration compared to both colorectal adenocarcinoma and cohesive IACSRCCs. Differential gene expression analysis showed distinct gene expression patterns in GCA compared to colorectal adenocarcinoma, with a number of cancer-related differentially expressed genes, including upregulation of TMEM14A, GOLT1A, DSCC1, and HSD17B8, and downregulation of KCNQ1OT1 and MXRA5. GCA also had several differentially expressed genes compared to IACSRCCs, including upregulation of PRSS21, EPPIN, RPRM, TNFRSF12A, and BZRAP1, and downregulation of HIST1H2BE, TCN1, AC069363.1, RP11-538I12.2, and REG4. In summary, the number of extra-appendiceal GCA was underestimated in Chinese patients. GCA can be seen as a distinct morphological, immunohistochemical, transcriptomic, and immunological entity. The classic low-grade component of GCA and the immunoreactivity for neuroendocrine markers are the key points to diagnosing GCA.

Stress Ulcer Prophylaxis in Critically Ill Adult Patients with Sepsis at Risk of Gastrointestinal Bleeding: A Retrospective Cohort Study

BACKGROUND

The Surviving Sepsis Campaign Guidelines recommend stress ulcer prophylaxis (SUP) for patients with sepsis who have gastrointestinal (GI) bleeding risks, however, the effect of SUP has not been specially studied in these patients.

AIMS

The purpose of this study is to determine the effects of SUP versus no prophylaxis on patient-important outcomes in critically ill adult patients with sepsis who have risk factors for GI bleeding.

METHODS

This retrospective cohort study utilized data from the Medical Information Mart for Intensive Care III database. We compared those who received SUP with proton pump inhibitors or histamine-2 receptor antagonists for ≥3 days with those who received no prophylaxis. Propensity score matching (PSM) was conducted to make comparisons between groups with similar distributions of study variables. The primary outcome was in-hospital mortality.

RESULTS

A total of 7,744 patients were included in the analysis, with 1,088 (14.0%) in the non-SUP group and 6,656 (86.0%) in the SUP group. A 1:1 PSM created 866 patients in each cohort. No significant differences were noted between the two groups with regard to in-hospital mortality (22.3%vs.20.4%, p=0.379), GI bleeding (4.7%vs.6.4%, p=0.172), pneumonia(38.9%vs.36.6%,p=0.346), Clostridium difficile infection(6.4% vs. 8.9%, p=0.0.057), or ICU length of stay (LOS) (4.2 d vs. 4.6 d, p=0.394).

CONCLUSIONS

Among critically ill, septic, adult patients at risk for gastrointestinal bleeding, stress ulcer prophylaxis showed no effect on hospital mortality, the rate of GI bleeding, pneumonia, CDI, and ICU LOS. This article is protected by copyright. All rights reserved.

[Functional MRI-based study on neuromechanism of trans-auricular vagus nerve stimulation for treatment-resistant depression]

OBJECTIVE

To explore the neuromechanism of trans-auricular vagus nerve stimulation (taVNS) for treatment-resistant depression(TRD) based on functional brain network.

METHODS

Twenty-eight patients with TRD were recruited from the psychiatric clinic or by the advertisement. The patients were treated by taVNS (5 Hz/20 Hz, 4-8 mA) at the auricular concha for 30 min, twice daily for 8 weeks. The symptom severity was assessed by 17-Item Hamilton Rating Scale for Depression (HAMD-17, ranging from 0 to 54 points, higher score indicates more severe conditions). Resting state fMRI data of the brain were collected to analyze changes of the regional homogeneity (ReHo), amplitude of low frequency fluctuation (ALFF) and resting state functional connectivity (rs-FC) before and after 8 weeks' taVNS by using DPARSF toolkit and the correlation between the rs-FC and clinical scale score was analyzed to assess the related brain mechanisms.

RESULTS

Twenty-four patients finished the clinical study, and 23 patients finished the fMRI tests. After the treatment, the average score of HAMD-17 was significantly decreased (P<0.01), with the reduction rate being 66.95%; the ALFF and ReHo values of the right insula and putamen, the ReHo values of the right caudate nucleus and thalamus, as well as the rs-FC values of the right insula, left superior frontal gyrus and middle frontal gyrus were all significantly decreased (P<0.05). The reduced ReHo value in the right insular lobe was negatively correlated with the HAMD score reduction (P=0.001, r=-0.633). The rs-FC values of the right insula lobe and the left superior frontal gyrus were significantly negatively correlated with the reduced HAMD score(P=0.012, r=-0.512).

CONCLUSION

TaVNS significantly relieves the symptoms of TRD patients, which may be related to its functions in regulating functional changes of the right insular and the left frontal gyrus network, and the limbic area and basal ganglia.

Analysis of single-mode fiber-optic extrinsic Fabry-Perot interferometric sensors with planar metal mirrors

We theoretically study the spectral characteristics and noise performance of wavelength-interrogated fiber-optic sensors based on an extrinsic Fabry-Perot (FP) interferometer (EFPI) formed by thin metal mirrors. We develop a model and use it to analyze the effect of key sensor parameters on the visibility and spectral width of the sensors, including the beam width of the incident light, metal coating film thickness, FP cavity length, and wedge angle of the two mirrors. Through Monte Carlo simulations, we obtain an empirical equation that can be used to estimate the wavelength resolution from the visibility and spectral width, which can be used as a figure-of-merit that is inherent to the sensor and independent on the system noises. The work provides a useful tool for designing, constructing, and interrogating high-resolution fiber-optic EFPI sensors.

Organoid technology for tissue engineering

For centuries, attempts have been continuously made to artificially reconstitute counterparts of in vivo organs from their tissues or cells. Only in the recent decade has organoid technology as a whole technological field systematically emerged and been shown to play important roles in tissue engineering. Based on their self-organizing capacities, stem cells of versatile organs, both harvested and induced, can form 3D structures that are structurally and functionally similar to their in vivo counterparts. These organoid models provide a powerful platform for elucidating the development mechanisms, modeling diseases, and screening drug candidates. In this review, we will summarize the advances of this technology for generating various organoids of tissues from the three germ layers and discuss their drawbacks and prospects for tissue engineering.

Modified phase-generated carrier demodulation of fiber-optic interferometric ultrasound sensors

We propose and demonstrate a modified phase-generated carrier (PGC) demodulation scheme optimized for detection of ultrasound using interferometric sensors with sinusoidal fringes. The sensor used in demonstration is made from a pair of weak fiber Bragg-gratings at the ends of a coiled fiber that form a low-finesse Fabry-Perot interferometer. The phase of the laser source is modulated using an electro-optic phase modulator to generate the carrier signal and obtain 2 quadrature (the sine and cosine) terms at the first and the second order carrier frequencies. The signal of interest (ultrasound) has much higher frequency than the environmental perturbation but a very small amplitude that causes only small phase shift. Using small-signal approximation, for each of the 2 quadrature terms, we separate the contributions from the environmental perturbations (quasi-DC component) and from the ultrasound (AC component). The AC components that contain the information of the ultrasound signal are then further amplified with a large gain. The signal of interest is constructed by simple algebraic operations on the 2 quasi-DC components and the 2 amplified AC components involving multiplying and summing. This work provides a simple and robust demodulation method with potentially high sensitivity for fiber-optic interferometric ultrasound sensors.

[Fast Separation Method of Continuous Recording Data in Dynamic Electrocardiogram Database]

The square waves and pulses with obvious porter characteristics are inserted into each different adjacent existing dynamic electrocardiogram data to solve the time-consuming problem in the currently used manual input verification method. The standard database files are converted into analog siginals, then output automatically to the ECG acquisition device according to the sequence of the database files in one time. The data recorded in the acquisition equipment is separated according to the interval data protocol, so as to achieve the purpose of rapid digital examination of dynamic electrocardiogram.

Relationship of expression of heat shock transcription factor 1 with sensitivity to radiotherapy and chemotherapy in esophageal squamous cell carcinoma

BACKGROUND

The morbidity and mortality of esophageal cancer are extremely high all over the world, and the treatment effect is not good. As the pathogenesis of esophageal cancer is not yet fully understood, this is not conducive to the study of specific therapeutic drugs for esophageal cancer. Heat shock transcription fact 1 (HSF1) is closely related to the occurrence and development of a variety of malignant tumors. Is HSF1 also closely related to the occurrence and development of esophageal squamous cell carcinoma (ESCC)? Will HSF1 become a biological target for the treatment of ESCC? Different patients with advanced ESCC have different sensitivity to radiotherapy and chemotherapy. Studies have shown that in hepatocellular carcinoma, HSF1 can weaken the toxic effect of radiotherapy and chemotherapy on tumors and reduce the curative effect. Does HSF1 affect the sensitivity of ESCC to radiotherapy and chemotherapy?

AIM

To investigate the expression of HSF1 in ESCC and its effect on the sensitivity of ESCC to radiotherapy and chemotherapy.

METHODS

Ninety-two patients were divided into four groups: 20 stage Ⅰ/Ⅱ ESCC patients undergoing surgical resection, 18 stage Ⅲ/Ⅱ ESCC patients undergoing surgical resection, and 44 stage Ⅲ/Ⅱ ESCC patients undergoing radiotherapy and chemotherapy. Among the 44 stage Ⅲ/Ⅱ ESCC patients undergoing radiotherapy and chemotherapy, 16 had low HSF1 expression and 28 had high expression. Ten cases of esophageal dysplasia. Ten esophagitis tissues were used as a control group. The expression of HSF1 in each group was detected by immunohistochemistry. The changes of non-tumorous lesion size, tumor diameter, and CEA value were compared between the HSF1 low expression group and high expression group before and after radiotherapy and chemotherapy to assess the sensitivity of patients to radiotherapy and chemotherapy. Factors that might affect the 3-year survival of ESCC patients were identified, and the 3-year overall survival rate of ESCC patients was calculated.

RESULTS

HSF1 was highly expressed in each ESCC group, but lowly expressed in esophagitis group and esophageal dysplasia group, and there was a significant difference in the expression of HSF1 between each ESCC group and esophagitis group and esophageal dysplasia group (P = 0.001). HSF1 expression was not significantly associated with age, gender, tumor location , tumor size, degree of differentiation, T stage, N stage, or M stage (P > 0.05). In the HSF1 low expression group, the non-tumor lesion was more significantly relieved, the tumor diameter was more significantly reduced, and the CEA value was more significantly decreased after radiotherapy and chemotherapy compared with those in the HSF1 high expression group (P < 0.05). In the ESCC surgical resection group, the 3-year survival period was significantly related to age (P = 0.019), HSF1 expression (P = 0.028), T stage (P = 0.007), and N stage (P = 0.016), but not related to gender, tumor location, tumor diameter, or degree of differentiation (P > 0.05). In stage Ⅲ/Ⅱ ESCC patients undergoing radiotherapy and chemotherapy, the HSF1 low expression group had a significantly higher 3-year overall survival rate than the HSF1 high expression group (P = 0.016). The 1-, 2-, and 3-year survival rates of the HSF1 low expression group were significantly higher than those of the HSF1 high expression group (P < 0.05). The HSF1 low expression group had a significantly higher 3-year overall survival rate than the HSF1 high expression group (P = 0.03).

CONCLUSION

HSF1 is highly expressed in ESCC and the higher the HSF1 expression, the worse the prognosis of patients. HSF1 expression is not related to patients' clinical characteristics. In stage Ⅲ/Ⅳ ESCC patients receiving radiotherapy and chemotherapy, the higher the expression of HSF1, the worse the sensitivity of patients to radiotherapy and chemotherapy. In ESCC patients undergoing surgical resection and stage Ⅲ/Ⅳ ESCC patients receiving radiotherapy and chemotherapy, the 3-year overall survival rate is higher in the HSF1 low expression group than in the HSF1 high expression group.

Genomic landscape and evolution of arm aneuploidy in lung adenocarcinoma

For lung adenocarcinoma, arm aneuploidy landscape among primary and metastatic sites, and among different driver and frequently mutated gene groups have not been previously studied. We collected the largest cohort of LUAD patients (n=3533) to date and analyzed the profiles of chromosome arm aneuploidy (CAA), and its association with different metastatic sites and mutated gene groups. Our results showed distant metastasis (bone, brain, liver) were characterized by high CAA burden and biased towards arm losses compared to regional metastasis (pleura, chest) and primary tumors. Moreover, EGFR, MET, PIK3CA, PKHD1 and RB1 mutant groups were found to have high CAA burden, while those with BRAF, ERBB2 and KRAS mutations belonged to the low CAA burden group. Comparing EGFR L858R and EGFR 19del mutants, distinct CAA co-occurrences were observed. Network-based stratification with population based genomic evolution analysis revealed two distinct subtypes of LUAD with different CAA signatures and unique CAA order of acquisition. In summary, our study presented a comprehensive characterization of arm aneuploidy landscape and evolutionary trajectories in lung adenocarcinoma, which could provide basis for both biological and clinical investigations in the future.

Radiomic Phenotypes Distinguish Atypical Teratoid/Rhabdoid Tumors from Medulloblastoma

BACKGROUND AND PURPOSE

Atypical teratoid/rhabdoid tumors and medulloblastomas have similar imaging and histologic features but distinctly different outcomes. We hypothesized that they could be distinguished by MR imaging-based radiomic phenotypes.

MATERIALS AND METHODS

We retrospectively assembled T2-weighted and gadolinium-enhanced T1-weighted images of 48 posterior fossa atypical teratoid/rhabdoid tumors and 96 match-paired medulloblastomas from 7 institutions. Using a holdout test set, we measured the performance of 6 candidate classifier models using 6 imaging features derived by sparse regression of 900 T2WI and 900 T1WI Imaging Biomarker Standardization Initiative-based radiomics features.

RESULTS

From the originally extracted 1800 total Imaging Biomarker Standardization Initiative-based features, sparse regression consistently reduced the feature set to 1 from T1WI and 5 from T2WI. Among classifier models, logistic regression performed with the highest AUC of 0.86, with sensitivity, specificity, accuracy, and F1 scores of 0.80, 0.82, 0.81, and 0.85, respectively. The top 3 important Imaging Biomarker Standardization Initiative features, by decreasing order of relative contribution, included voxel intensity at the 90th percentile, inverse difference moment normalized, and kurtosis-all from T2WI.

CONCLUSIONS

Six quantitative signatures of image intensity, texture, and morphology distinguish atypical teratoid/rhabdoid tumors from medulloblastomas with high prediction performance across different machine learning strategies. Use of this technique for preoperative diagnosis of atypical teratoid/rhabdoid tumors could significantly inform therapeutic strategies and patient care discussions.

Adherence to antihypertensive medications for secondary prevention of cardiovascular disease events: a dose-response meta-analysis

OBJECTIVES

The aim of the study was to explore the association between adherence to antihypertensive medications (AHMs) and the risk of recurrence of cardiovascular disease (CVD) events in patients with a history of CVD events from cohort studies.

STUDY DESIGN

This is a dose-response meta-analysis.

METHODS

PubMed and Embase databases were searched up to March 4, 2021, to identify English-language reports of cohort studies that assessed the association of AHM adherence with risk of recurrence of CVD events. Pooled relative risks (RRs) and 95% confidence intervals (CIs) were estimated by using a fixed- or random-effects model. Restricted cubic splines were used to evaluate the possible linear or non-linear association.

RESULTS

We included nine cohort studies (54,349 patients) in the present meta-analysis. The pooled RR of CVD events was 0.66 (95% CI, 0.54-0.78) for the highest versus lowest AHM adherence category. We did not find any evidence of non-linearity association between AHM adherence and risk of CVD events (P_{non-linearity} = 0.534); for patients with a history of CVD events, the risk of CVD events was reduced by 9% for each 20% increase in AHM adherence (RR, 0.91; 95% CI, 0.85-0.97). The results of sensitivity analysis and subgroup analysis were virtually unchanged.

CONCLUSIONS

The high level of adherence to AHM is an effective strategy for preventing recurrence of CVD events. Patients with a history of CVD events should adhere to AHM.

On Image Segmentation Based on Local Entropy Fitting Under Nonconvex Regularization Term Constraints

The energy functional of the CV and LBF model is single, which makes the curve to get into the local minimum easily during the evolution process, and results inaccurate segmentation of the images with nonuniform grayscale and nonsmooth edges. The proposed algorithm, which is based on local entropy fitting under the constraint of nonconvex regularization term, is used to deal with such problems. In this algorithm, global information and local entropy are fitted to avoid segmentation falling into local optimum, and nonconvex regularization term is imported for constraint to protect edge smoothing. First, global information is used to evolve the approximate contour curve of the target segmentation. Then, a local energy functional with local entropy information is constructed to avoid the segmentation process from falling into a local minimum, and to precisely segment the image. Finally, nonconvex regularization terms are used in the energy functional to protect the smoothness of edge information during image segmentation process. The experimental results clearly indicate that the new algorithm can effectively resist noise, precisely segment images with nonuniform grayscale, and achieve the global optimal.

Abstract 2210: Chromosome arm aneuploidy landscape in lung adenocarcinoma

Background: Lung adenocarcinoma (LUAD) is the most common subtype of non-small cell lung cancer. Overall aneuploidy frequency in LUAD have been previously documented. We aimed to assess whether specific chromosome arm aneuploidy (CAA) patterns can be found in primary tumor and different metastatic sites, and across patients with different driver mutations. We also investigated CAA and driver mutation co-occurrence patterns in LUAD, as well as probable order of CAA acquisition.

Methods: We performed targeted panel sequencing (425 cancer-related genes) on 3533 tissue samples from LUAD patients. CAA status (arm gain or loss) was derived from segmental copy number analysis using Sequenza, accounting for purity and ploidy. 4 types of CAA analysis were conducted: total CAA burden, gain-loss pattern, co-occurrence/exclusivity, and population level CAA progression assessed by Tronco on network-based stratified (NBS) subtypes.

Results: Use of panel sequencing data for analysis of CAA status was validated against WES results, and achieved high concordance (Kappa: 0.75, Pearson: 0.85). Primary tumor and different metastatic sites had dramatic differences in CAA burdens (p=3.4e-06). Liver and brain had more arm losses than gains compared to primary (q=0.047 and 0.049 respectively). Different driver mutation groups also had significant difference in CAA burden (p&lt;2.2e-16). BRAF group had more arm gains than losses; whereas PIK3CA had more arm losses than gains. TP53 mutation status had a moderating effect on both CAA burden and gain-loss pattern across metastatic sites and driver mutation groups. EGFR 19del and L858R mutations showed significant differences in CAA co-occurrence/exclusion. EGFR 19del tend to co-occur with 4q-, 5p-, 5q-, 8q+, 9p+, 16q+, 17p+, 17q-, 22q-, and was in exclusion with 14q-. EGFR L858R co-occurred with 4q+, but was in exclusion with 1p-, 8p+ and 16p-. MET significantly co-occurred with 1q-, and PIK3CA with 2p+. Clustering of LUAD patients using NBS gave us 2 tight clusters with high silhouette score (0.86). These two LUAD subtypes showed distinct orders of CAA acquisition. Both clustered subtypes began with acquisition of 7p gain. Subtype 1 is dominated by initial arm gains, followed by few arm losses, seen in previously published literature; whereas subtype 2 began with a simultaneous acquisition of 19p loss, with further additions of many arm losses, before eventual accumulation of arm gains. Proportion of these 2 subtypes varied across primary and metastatic sites, and driver mutant groups.

Conclusions: There are significant differences in arm aneuploidy patterns across metastatic sites and mutation groups. Overall, LUAD patients exhibit a unique pattern of co-occurrence and exclusivity between CAA and driver mutations, as well as two subtypes with distinct CAA evolution trajectories. These CAA patterns may hold biological meaning and may be useful predictors of clinical outcomes and warrants further investigation.

Citation Format: Hua Bao, Ming Han, Yi Shen, Xue Wu, Yang Shao. Chromosome arm aneuploidy landscape in lung adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2210.

Proton pump inhibitors versus histamine-2 receptor blockers for stress ulcer prophylaxis in patients with sepsis: a retrospective cohort study

OBJECTIVE

We aimed to compare the efficacy and risks of proton pump inhibitor (PPI) versus histamine-2 receptor blocker (H2B) use for stress ulcer prophylaxis (SUP) in critically ill patients with sepsis and risk factors for gastrointestinal bleeding (GIB).

METHODS

In this retrospective cohort study, we used the Medical Information Mart for Intensive Care III Clinical Database to identify critically ill adult patients with sepsis who had at least one risk factor for GIB and received either an H2B or PPI for ≥48 hours. Propensity score matching (PSM) was conducted to balance baseline characteristics. The primary outcome was in-hospital mortality.

RESULTS

After 1:1 PSM, 1056 patients were included in the H2B and PPI groups. The PPI group had higher in-hospital mortality (23.8% vs. 17.5%), GIB (8.9% vs. 1.6%), and pneumonia (49.6% vs. 41.6%) rates than the H2B group. After adjusting for risk factors of GIB and pneumonia, PPI use was associated with a 1.28-times increased risk of in-hospital mortality, 5.89-times increased risk of GIB, and 1.32-times increased risk of pneumonia.

CONCLUSIONS

Among critically ill adult patients with sepsis at risk for GIB, SUP with PPIs was associated with higher in-hospital mortality and higher risk of GIB and pneumonia than H2Bs.

Single-center experience of organ transplant practice during the COVID-19 epidemic

In order to safely carry out organ donation transplants during the outbreak of coronavirus disease 2019 (COVID-19), we have formulated strict procedures in place for organ donation and transplantation. We retrospectively analyzed our transplantation work from January 20 to May 5, 2020, to discuss whether organ transplantation can be carried out safely during the epidemic period. From January 20 to May 5, 43 cases of donation were carried out in our hospital, and the utilization rate of liver, kidney, heart, lung, and pancreas donations was more than 90%. Forty-one cases of liver transplantation and 84 cases of kidney transplantation were performed. No graft loss or recipient death occurred within one month after kidney transplantation, and one patient (2.4%) died after liver transplantation. There was no significant difference in the length of hospital stay compared with that during the same period in the previous three years. More importantly, COVID-19 infection did not occur among healthcare providers, donors, patients, or their accompanying families in our center. Under the premise of correct protection, it is safe and feasible to carry out organ transplantation during the epidemic period. Our experience during the outbreak might provide a clinical reference for countries facing COVID-19 worldwide.

Risk factors associated with postoperative respiratory failure in tuberculous empyema patients

Our objective was to identify independent risk factors for predicting which patients in the Chinese population would likely develop respiratory failure.A descriptive analysis was conducted of demographic and clinical data of patients with tuberculous empyema (TE) admitted to the Beijing Chest Hospital, Capital Medical University between January 2001 and January 2020. Risk factors associated with postsurgical respiratory failure in TE patients were identified based on results of analyses based on univariable and multivariable logistic regression models.A total of 139 TE patients who underwent surgical treatment in the Beijing Chest Hospital, Capital Medical University from January 2001 to January 2020 were enrolled in this study. Cases included 109 male and 30 female patients, with an overall mean age (range 17-73) of 39.3 years. Of 139 TE patients, 26 (18.7%) experienced respiratory failure after surgery. Among significant risk factors for postsurgical respiratory failure, intraoperative blood loss volume greater than 1000 mL had the highest odds ratio value of 6.452. In addition, a pathologic preoperative pulmonary function test result showing a high partial pressure of carbon dioxide level was an independent risk factor for respiratory failure. Moreover, the presence of tuberculosis lesions in the contralateral lung was another significant risk factor for respiratory failure, as determined using multivariate analysis.Respiratory failure is a predominant complication experienced by TE patients undergoing surgery. High intraoperative blood loss, high preoperative high partial pressure of carbon dioxide level, and tuberculosis lesion(s) in the contralateral lung of TE patients were associated with increased risk of postoperative respiratory failure.