Overall Survival Prediction in Stereotactic Radiosurgery Patients with Glioblastoma Via a Deep-Learning Approach

PURPOSE/OBJECTIVE(S)

Accurate and automated early survival prediction is critical for glioblastoma (GBM) patients as their poor prognosis requires timely treatment decision-making. We have developed a deep learning (DL)-based GBM overall survival (OS) prediction model based on a multi-institutional public dataset using only pre-operative basic structural multi-parametric magnetic resonance images (MRIs). The purpose of this study is to evaluate this DL-based OS prediction model with an institutional stereotactic radiosurgery (SRS) clinical trial dataset.

MATERIALS/METHODS

The task of this study is to classify GBM patients into 3 OS classes: long-survivors (>15 months), mid-survivors (between 10 and 15 months) and short-survivors (< 10 months). The proposed OS prediction model is an ensemble of a ResNet-based classifier and a K-NN classifier. The ResNet-based classifier is trained in a Siamese fashion to explore inter-class differences. During testing, training sample features are implemented with a K-NN classifier to ensemble with the ResNet-based classifier. A public dataset from Medical Image Computing and Computer Assisted Intervention (MICCAI) Brain Tumor Segmentation (BraTS) challenge 2020 (235 patients) were used for model establishing and initial validation. Then the validated model was evaluated on 19 GBM patients from an institutional SRS clinical trial. Each data entry consists of pre-operative basic structural multi-parametric MRIs and survival days, as well as patient ages for BraTS data and basic clinical characteristics for institutional data. GBM sub-regions, including contrast-enhancing tumor, peri-tumoral edema, and necrotic/non-enhancing tumor core, were segmented in the multi-parametric MRIs by an in-house DL model for both datasets. The OS prediction model was trained on 90% of the segmented BraTS data and validated on the rest 10%, then further evaluated on the institutional data. The model performance was assessed by prediction accuracy (ACC) and the area under the curve (AUC).

RESULTS

For this 3-class OS classification task, our DL-based prediction model achieved an ACC of 65.22% and an AUC of 0.81 on the BraTS dataset compared with the top-ranked result from the BraTS challenge 2020 (Rank 1st: ACC 61.7%), and an ACC of 52.63% and an AUC of 0.69 on the institutional dataset. Further analysis of the institutional dataset found that the predicted OS class had a statistically significant correlation with treatment volume (p = 0.012) and age (p = 0.006), which matches the analysis that the patients' ground truth OS class is statistical significantly correlated with treatment volume (p = 0.045).

CONCLUSION

Our DL-based OS prediction model for GBM using basic structural multi-parametric pre-operative MRIs has demonstrated promising performance in both public and institutional dataset with minimal manual processing requirements. This OS prediction model can be potentially applied to assist timely clinical decision-making.

Magnetic Susceptibility of Andreev Bound States in Superfluid ^{3}He-B

Nuclear magnetic resonance measurements of the magnetic susceptibility of superfluid ^{3}He imbibed in anisotropic aerogel reveal anomalous behavior at low temperatures. Although the frequency shift clearly identifies a low-temperature phase as the B phase, the magnetic susceptibility does not display the expected decrease associated with the formation of the opposite-spin Cooper pairs. This susceptibility anomaly appears to be the predicted high-field behavior corresponding to the Ising-like magnetic character of surface Andreev bound states within the planar aerogel structures.

Injectable carbon nanotube-functionalized hydrogel as a tool for cardiac tissue engineering

Background/Introduction

Heart failure (HF) is an expensive major public health problem in the United States and around the world (1). The current treatments for HF are aimed at reducing symptoms, slowing disease progression, and reducing mortality and not aimed at repairing heart muscle or restoring function. Furthermore, even with these treatments, approximately half of patients with HF will die within 5 years of diagnosis (2). Cardiac transplantation remains the only definitive treatment for those affected with end-stage HF, but availability of donor hearts remains a major limitation (3).

Purpose

The ability of the adult heart to regenerate cardiomyocytes (CMs) lost after injury is limited, generating interest in developing tissue engineering therapies to avoid progression towards HF. Rigid carbon nanotubes (CNTs) scaffolds have been used to improve CMs viability, proliferation, and maturation (4), but require undesirable invasive surgeries for implantation. To overcome this limitation, we engineered an injectable reverse thermal gel (RTG) functionalized with CNTs (RTG-CNT) that transitions from a liquid-solution to a gel-based matrix shortly after reaching body temperature allowing for a liquid-based delivery rapidly followed by a stable-gel localization (5).

Methods and results

Here we show experimental evidences the RTG-CNT hydrogel, used as a three-dimensional (3D) niche to culture human induced pluripotent stem cells (hiPSC)-CMs, promotes hiPSC-CMs alignment and elongation with increased Cx43 localization and improved contraction function when compared with traditional two-dimensional (2D) fibronectin controls and plain 3D RTG system without CNTs. Moreover, the short-term (4-week) biocompatibility of the RTG-CNT hydrogel was also assessed in a mouse model (intracardial injection). The results confirmed that the RTG-CNT hydrogel is well tolerated by the cardiac tissue.

Conclusion

Our results indicated that the injectable RTG-CNT hydrogel has the potential to be used as a minimally invasive tool for cardiac tissue engineering efforts.

Funding Acknowledgement

Type of funding sources: Other. Main funding source(s): NATIONAL HEART, LUNG, AND BLOOD (NHLBI) INSTITUTE

Dysregulation of immune cell and cytokine signaling correlates with clinical outcomes in myelodysplastic syndrome (MDS)

OBJECTIVES

Myelodysplastic syndromes (MDS) are characterized by ineffective hematopoiesis. Although hypomethylating agents (HMA) have improved survival in higher-risk MDS, most patients eventually succumb to progressive disease. Utilizing samples collected prospectively from three MDS clinical trials, we analyzed genetic and immunological biomarkers and correlated them with clinical outcomes.

METHODS

154 samples were analyzed from 133 de novo MDS patients for T-cell and myeloid cell immunophenotyping and gene expression analysis. Treatments were with HMA or immunomodulatory drug (IMiD) alone or in combination.

RESULTS

We observed differences in immune cell subsets between lower and higher risk IPSS groups with NKT cells, MDSCs, intermediate-proinflammatory and non-classical monocytes being higher in the latter group while naïve CD4+ T-cells were reduced. Intermediate-proinflammatory monocytes were increased in non-responders and those failing to achieve at least a hematological improvement. Pro-inflammatory NKT cells were increased at diagnosis for patients failing to derive clinical benefit after 12 months of treatment. Gene expression analysis of paired bone marrow (BM) colony-forming units (CFUs) from diagnosis and 4 cycles post-treatment confirmed that genes involved in cytokine signaling were downregulated in C4 normal colonies.

CONCLUSIONS

These findings support the central roles of dysregulation in innate immunity and inflammatory signaling in the pathogenesis of MDS which correlated with clinical outcomes post-treatment.

Thyroid surgery during coronavirus-19 pandemic phases I, II and III: lessons learned in China, South Korea, Iran and Italy

OBJECTIVE

We compared demographic and clinic-pathological variables related to the number of surgeries for thyroid conditions or for cancer, morbidity, and fine needle aspiration (FNA) practices among Covid19 pandemic phases I, II, III and the same seasonal periods in 2019.

METHODS

The prospective database of the Division of Thyroid Surgery, China-Japan Union Hospital of Jilin University, Changchun, China was used for this study. Covid19 emergency levels were stratified according to the World Health Organization: phase I (January 25-February 25, 2020), phase II (February 26-March 19), phase III (March 20-April 20).

RESULTS

There were fewer outpatient FNAs and surgeries in 2020 than in 2019. There were no thyroid surgeries during phase I. There were also fewer surgeries for cancer with a significant reduction of advanced stage cancer treatments, mainly stage T1b N1a in phase II and T3bN1b in phase III. Operative times and postoperative stays were significantly shorter during the pandemic compared to our institutional baseline. In phase III, vocal cord paralysis (VCP) increased to 4.3% of our baseline numbers (P = 0.001). There were no cases of Covid19-related complications during the perioperative period. No patients required re-admission to the hospital.

CONCLUSION

The Covid19 outbreak reduced thyroid surgery patient volumes. The decrease of Covid19 emergency plans contributed to unexpected outcomes (reduction of early stage cancer treatment, decreased operative times and hospital stays, increased VCP rate).

Identification of Pistacia-associated flexivirus 1, a putative mycovirus of the family Gammaflexiviridae, in the mastic tree (Pistacia lentiscus) transcriptome

In this study, we identified the genome sequence of the novel virus Pistacia-associated flexivirus 1 (PAFV1), a putative member of the mycovirus family Gammaflexiviridae (the order Tymovirales), via analysis of a transcriptome dataset for the mastic tree (Pistacia lentiscus, the family Anacardiaceae). PAFV1 was predicted to have three open reading frames (ORFs): ORF1, encoding a replicase (REP) with RNA-dependent RNA polymerase activity; ORF2, a movement protein (MP); and ORF3, a hypothetical protein. The PAFV1 REP sequence showed high similarity to those of three known members of the family Gammaflexiviridae i.e., Entoleuca gammaflexivirus 1 (EnFV1), Entoleuca gammaflexivirus 2 (EnFV2), and Botrytis virus F (BVF). A genome contig of the fungus Monosporascus cannonballus also contained a sequence of an endogenous virus similar to that of PAFV1. Sequence comparison and phylogenetic analysis indicated that PAFV1, EnFV1, and the endogenous virus of M. cannonballus formed a distinct subgroup (apart from EnFV2 and BVF), and may be the founding members of a novel genus in the family Gammaflexiviridae. Notably, MP sequences of PAFV1/EnFV1 showed similarity to the MP sequences of the mycovirus group called tobamo-like mycoviruses (an unassigned taxon), implying that genomic recombination occurred between members of the family Gammaflexiviridae and tobamo-like mycoviruses. Since PAFV1 is phylogenetically related to mycoviruses, PAFV1 may also be a mycovirus that infected a fungus associated with the mastic tree sample, which is evidenced by the presence of fungal ribosomal RNA sequences in the mastic tree transcriptome. Thus, the PAFV1 genome sequence may be useful in elucidating the genome evolution of Gammaflexiviridae and tobamo-like mycoviruses. Keywords: Pistacia-associated flexivirus 1; Gammaflexiviridae; mycovirus, mastic tree.

Clinical features of subungual melanoma according to the extent of Hutchinson's nail sign: a retrospective single-centre study

BACKGROUND

Hutchinson's nail sign (HS) is among the diagnostic criteria for subungual melanoma (SUM). However, there is minimal evidence supporting the overall clinical significance of HS in SUM.

OBJECTIVES

To identify clinicopathological features of SUM according to the extent of HS.

METHODS

Retrospective cohort study was performed with consecutive SUM patients at a single centre from January 2006 to December 2017. The extent of HS was defined by the number of affected nail folds (range 0-4). Comparison groups were organized as follows: patients with HS (affecting ≥1 nail folds) vs. without HS; patients with HS affecting ≥2 nail folds vs. HS affecting <2 nail folds; patients with HS affecting ≥3 nail folds vs. HS affecting <3 nail folds. Clinicopathological characteristics of SUM were compared between the groups.

RESULTS

Sixty-one SUM patients were included. Forty-six (75.4%) exhibited HS; 22 (47.8%) on a toe and 24 (52.2%) on a finger. In multivariate analysis, nail destruction [hazard ratio (HR), 10.00; 95% confidence interval (CI), 2.61-38.30; P = 0.001] was significantly associated with the presence of HS and amputation was significantly associated with HS affecting ≥2 nail folds (HR, 4.75; 95% CI, 1.36-16.61; P = 0.015). High T stage (HR, 1.85; 95% CI, 1.20-2.85; P = 0.005, Fig. 2) was significantly associated with HS appearing in ≥3 nail folds.

CONCLUSION

Besides its value of detecting SUM, HS provides useful clinical information. The number of nail folds exhibiting HS could be a useful clinical clue for planning therapeutic strategies for SUM.

TEK is a novel prognostic marker for clear cell renal cell carcinoma

OBJECTIVE

Clear cell renal cell carcinoma (ccRCC) is the most common type of kidney cancer. However, effective therapeutics for ccRCC are lacking. Novel biomarkers could provide critical information when determining prognoses for patients with ccRCC. In this study, we sought to determine if the expression of receptor tyrosine kinase (TEK) could be a potential novel prognostic biomarker for ccRCC. TEK, originally identified as an endothelial cell-specific receptor, plays an important role in the modulation of vasculogenesis and remodeling. Altered TEK expression has been observed in tumor tissues (e.g., oral squamous cell carcinomas, leukemia) and breast, gastric and thyroid cancers. However, the role of TEK in ccRCC remains unknown.

PATIENTS AND METHODS

Differential TEK expression between non-metastatic (stage M0) and metastatic (stage M1) ccRCC patient cohorts was determined from The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC). Furthermore, TEK expression was assessed as a prognostic factor using the time-dependent area under the curve (AUC) of Uno's C-index, the AUC value of the receiver operating characteristics (ROC) at 5 years, Kaplan-Meier survival curves and multivariate analyses.

RESULTS

A Kaplan-Meier curve analysis revealed that the downregulation of TEK expression was associated with a poor prognosis for patients with ccRCC with good discrimination (p<0.0001 and p=0.0044 for the TGCA and ICGC cohorts, respectively). Analyses of C-indices and receiver operating characteristic AUC values further support this discriminative ability. Moreover, multivariate analyses showed the prognostic significance of TEK expression levels (p<0.001).

CONCLUSIONS

Although additional clinical investigations will be needed, our results suggest that TEK is a potential biomarker for ccRCC.

A Tabletop Persistent-Mode, Liquid Helium-Free 1.5-T MgB2 "Finger" MRI Magnet: Construction and Operation of a Prototype Magnet

This paper presents results of construction and operation of a persistent-mode, liquid-helium-free, small-scale prototype magnet for the development of a tabletop 1.5-T "finger" MRI system for osteoporosis screening. The prototype magnet, composed of 2 MgB2 coils, one superconducting joint, and a persistent-current switch (PCS) built from a portion of one coil, was wound with a one continuous ~80-m long unreacted and monofilament MgB2 wire and then reacted. The test magnet was charged successfully and generated the estimated target field of 1.75 T at 5 K with the proposed PCS operation. During initial persistent-mode, the field was slightly decayed due to the index dissipation of the joint; thereafter it sustained the persistent field of 1.7 T for 35 h. The test results validated the joint resistance of < 1.2 × 10-11 as well as the proposed approach involving the PCS coil circuit model.

A novel Waikavirus (the family Secoviridae) genome sequence identified in rapeseed (Brassica napus)

The genome sequence of a novel species of the genus Waikavirus (the family Secoviridae), which we named Brassica napus RNA virus 1 (BnRV1), was identified in a rapeseed (Brassica napus) transcriptome dataset. The BnRV1 genome was 12,293 nucleotides long followed by a poly(A) tail. Two open reading frames (ORFs), called ORF1 and ORFX, were predicted. The larger ORF, ORF1, encodes a polyprotein of 3,471 amino acids and the smaller ORF, ORFX, overlaps ORF1 and encodes an 87 aa long protein of unknown function. The BnRV1 ORF1 polyprotein was predicted to undergo proteolytic processing to yield seven mature proteins, including an RNA-dependent RNA polymerase and three distinct coat proteins. The ORF1 and ORFX proteins share sequence similarities with the respective proteins of viruses in the genus Waikavirus, including the bellflower vein chlorosis virus, rice tungro spherical virus, and maize chlorotic dwarf virus. A phylogenetic tree inferred from a conserved segment of the polyproteins of several Secoviridae viruses confirmed that BnRV1 is a novel species of the genus Waikavirus. The BnRV1 genome sequence identified in this study may be useful for the study of waikavirus biology and waikavirus-derived diseases. Keywords: Brassica napus RNA virus 1; Waikavirus; Secoviridae; rapeseed.

Blastacervulus metrosideri sp. nov. leaf spot on Metrosideros excelsa in New Zealand

A leaf-spotting fungal pathogen common on Metrosideros excelsa in New Zealand is described here as Blastacervulus metrosideri sp. nov. It has previously been identified in the New Zealand literature as Leptomelanconium sp. and as Staninwardia breviuscula. The choice of genus for this new species is supported by a phylogeny based on ITS and LSU sequences. It is phylogenetically close to several morphologically similar Eucalyptus leaf spotting pathogens.