Purpose To evaluate the performance of an artificial intelligence (AI) tool using a deep learning algorithm for detecting hemorrhage, mass effect, or hydrocephalus (HMH) at non-contrast material-enhanced head computed tomographic (CT) examinations and to determine algorithm performance for detection of suspected acute infarct (SAI). Materials and Methods This HIPAA-compliant retrospective study was completed after institutional review board approval. A training and validation dataset of noncontrast-enhanced head CT examinations that comprised 100 examinations of HMH, 22 of SAI, and 124 of noncritical findings was obtained resulting in 2583 representative images. Examinations were processed by using a convolutional neural network (deep learning) using two different window and level configurations (brain window and stroke window). AI algorithm performance was tested on a separate dataset containing 50 examinations with HMH findings, 15 with SAI findings, and 35 with noncritical findings. Results Final algorithm performance for HMH showed 90% (45 of 50) sensitivity (95% confidence interval [CI]: 78%, 97%) and 85% (68 of 80) specificity (95% CI: 76%, 92%), with area under the receiver operating characteristic curve (AUC) of 0.91 with the brain window. For SAI, the best performance was achieved with the stroke window showing 62% (13 of 21) sensitivity (95% CI: 38%, 82%) and 96% (27 of 28) specificity (95% CI: 82%, 100%), with AUC of 0.81. Conclusion AI using deep learning demonstrates promise for detecting critical findings at noncontrast-enhanced head CT. A dedicated algorithm was required to detect SAI. Detection of SAI showed lower sensitivity in comparison to detection of HMH, but showed reasonable performance. Findings support further investigation of the algorithm in a controlled and prospective clinical setting to determine whether it can independently screen noncontrast-enhanced head CT examinations and notify the interpreting radiologist of critical findings. RSNA, 2017 Online supplemental material is available for this article.

Glioblastoma (GBM) is the most common and aggressive histologic subtype of brain cancer with poor outcomes and limited treatment options. Here we report the selective overexpression of the protein arginine methyltransferase PRMT5 as a novel candidate theranostic target in this disease. PRMT5 silences the transcription of regulatory genes by catalyzing symmetric di-methylation of arginine residues on histone tails. PRMT5 overexpression in patient-derived primary tumors and cell lines correlated with cell line growth rate and inversely with overall patient survival. Genetic attenuation of PRMT5 led to cell cycle arrest, apoptosis and loss of cell migratory activity. Cell death was p53-independent but caspase-dependent and enhanced with temozolomide, a chemotherapeutic agent used as a present standard of care. Global gene profiling and chromatin immunoprecipitation identified the tumor suppressor ST7 as a key gene silenced by PRMT5. Diminished ST7 expression was associated with reduced patient survival. PRMT5 attenuation limited PRMT5 recruitment to the ST7 promoter, led to restored expression of ST7 and cell growth inhibition. Lastly, PRMT5 attenuation enhanced GBM cell survival in a mouse xenograft model of aggressive GBM. Together, our findings defined PRMT5 as a candidate prognostic factor and therapeutic target in GBM, offering a preclinical justification for targeting PRMT5-driven oncogenic pathways in this deadly disease.

Brain Metastases (BM) complicate 20-40% of cancer cases. BM lesions can present as punctate (1 mm) foci, requiring high-precision Magnetic Resonance Imaging (MRI) in order to prevent inadequate or delayed BM treatment. However, BM lesion detection remains challenging partly due to their structural similarities to normal structures (e.g., vasculature). We propose a BMdetection framework using a single-sequence gadoliniumenhanced T1-weighted 3D MRI dataset. The framework focuses on detection of smaller (< 15 mm) BM lesions and consists of: (1) candidate-selection stage, using Laplacian of Gaussian approach for highlighting parts of a MRI volume holding higher BM occurrence probabilities, and (2) detection stage that iteratively processes cropped region-of-interest volumes centered by candidates using a custom-built 3D convolutional neural network ("CropNet"). Data is augmented extensively during training via a pipeline consisting of random gamma correction and elastic deformation stages; the framework thereby maintains its invariance for a plausible range of BM shape and intensity representations. This approach is tested using five-fold cross-validation on 217 datasets from 158 patients, with training and testing groups randomized per patient to eliminate learning bias. The BM database included lesions with a mean diameter of ~5.4 mm and a mean volume of ~160 mm 3 . For 90% BM-detection sensitivity, the framework produced on average 9.12 falsepositive BM detections per patient (standard deviation of 3.49); for 85% sensitivity, the average number of falsepositives declined to 5.85. Comparative analysis showed that the framework produces comparable BM-detection accuracy with the state-of-art approaches validated for significantly larger lesions.Index Terms-magnetic resonance imaging, brain metastases, convolutional neural networks, deep learning, scale-space representations, computer-aided detection, medical image analysis.

No abstract

Radiology and Enterprise Medical Imaging Extensions (REMIX) is a platform originally designed to both support the medical imaging-driven clinical and clinical research operational needs of Department of Radiology of The Ohio State University Wexner Medical Center. REMIX accommodates the storage and handling of “big imaging data,” as needed for large multi-disciplinary cancer-focused programs. The evolving REMIX platform contains an array of integrated tools/software packages for the following: (1) server and storage management; (2) image reconstruction; (3) digital pathology; (4) de-identification; (5) business intelligence; (6) texture analysis; and (7) artificial intelligence. These capabilities, along with documentation and guidance, explaining how to interact with a commercial system (e.g., PACS, EHR, commercial database) that currently exists in clinical environments, are to be made freely available.

The detection of brain metastases (BM) in their early stages could have a positive impact on the outcome of cancer patients. The authors previously developed a framework for detecting small BM (with diameters of <15 mm) in T1-weighted contrast-enhanced 3D magnetic resonance images (T1c). This study aimed to advance the framework with a noisy-student-based self-training strategy to use a large corpus of unlabeled T1c data. Accordingly, a sensitivity-based noisy-student learning approach was formulated to provide high BM detection sensitivity with a reduced count of false positives. This paper (1) proposes student/teacher convolutional neural network architectures, (2) presents data and model noising mechanisms, and (3) introduces a novel pseudo-labeling strategy factoring in the sensitivity constraint. The evaluation was performed using 217 labeled and 1247 unlabeled exams via two-fold cross-validation. The framework utilizing only the labeled exams produced 9.23 false positives for 90% BM detection sensitivity, whereas the one using the introduced learning strategy led to ~9% reduction in false detections (i.e., 8.44). Significant reductions in false positives (>10%) were also observed in reduced labeled data scenarios (using 50% and 75% of labeled data). The results suggest that the introduced strategy could be utilized in existing medical detection applications with access to unlabeled datasets to elevate their performances.

High grade astrocytomas are aggressive brain tumors that are associated with a dismal prognosis and are incurable with a median survival of less than 15 months despite intensive multimodal therapy. The linability to effectively target grade III and grade IV (glioblastoma multiforme, GBM) astrocytomas highlights the need for novel therapeutic approaches. Recent studies have shown that epigenetic regulation of chromatin plays a central role in the control of cell growth, differentiation, and survival. Chromatin remodeling enzymes like histone deacetylases, DNA methyltransferases and protein arginine methyltransferase 5 (PRMT5) are involved in silencing tumor suppressor gene (TSG) expression and may contribute towards cellular transformation. PRMT5 silences the transcription of key regulatory genes by symmetric di-methylation (S2Me) of arginine (R) residues on histone proteins (H4R3 and H3R8) and works more efficiently when associated with other co-repressor enzymes. Eight patient-derived GBM cell lines and 45 primary GBM tumors showed abundant expression of PRMT5 protein. Confocal microscopy and immunohistochemical staining showed the PRMT5 signal to localize primarily to the nucleus. Normal brain tissue, normal human astrocytes and low/intermediate grade astrocytomas failed to show PRMT5 expression. The degree of PRMT5 over expression inversely correlated with survival of GBM patients (r=-0.57, p=0.0001) and correlated with proliferation of GBM cell lines (r=0.81, p&lt;0.0001). Elevated PRMT5 expression was observed in high grade astrocytomas that spontaneously develop in a pre-clinical mouse model of GBM employing conditional Nf1, TP53 and PTEN haploinsufficiency. Small inhibitory RNAs (siRNA) specific for PRMT5 led to loss of PRMT5 protein expression and S2Me of H4R3, a histone target of PRMT5. PRMT5 inhibition led GBM cell lines to cell cycle arrest, apoptosis, and complete loss of cell migratory activity. Apoptosis occurred independent of caspase and p53 pathways. Furthermore, PRMT5 knockdown led GBM cell lines to become sensitized to the toxic effects of temazolomide, a drug frequently used to manage patients with GBM. We utilized gene microarray analysis of cDNA isolated from siRNA (or control RNA) treated GBM cells to identify potential targets of PRMT5 and identified the TSG ST7 and three chemokine transcripts (RANTES, IP10, CXCL11) to be upregulated with PRMT5 knockdown. We used chromatin immuno precipitation to show that siRNA treatment led to loss of PRMT5 recruitment on ST7 and chemokine gene promoters that coincided with restoration of transcriptional and translational activity leading to marked elevation in protein expression. PRMT5 knock-down led to secretion of all chemokines into growth medium. These findings identify PRMT5 as an independent prognostic factor for GBM and an attractive therapeutic target for high grade astrocytomas. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 1584.

595FN2 Introduction: Mantle cell lymphoma (MCL) is an incurable B-cell non-Hodgkin lymphoma characterized by aberrant genetic (t(11;14)(q13;q32)) and epigenetic (DNA hypermethylation) dysregulation. Chromatin remodeling complexes and associated co-repressors such as histone deacetylases (HDAC), DNA methyltransferases (DNMT) and protein arginine methyltransferase 5 (PRMT5), are involved in silencing tumor suppressor and regulatory gene expression and may contribute to B-cell transformation. PRMT5 silences the transcription of key regulatory genes by symmetric di-methylation (S2Me) of arginine (R) residues on histone proteins (H4R3 and H3R8). We have previously identified PRMT5 over expression to be relevant to MCL pathogenesis and shown it to work concertedly with HDAC2, methyl-CpG binding domain protein 2 (MBD2) and DNMT3a to silence genes with anti-cancer and immune modulatory activities. siRNA-mediated knockdown of PRMT5 in MCL cell lines leads to growth arrest and apoptosis, thus, we explored methods to inhibit PRMT5 activity as a novel experimental therapeutic strategy for this disease. Methods and Results: A rational design of small molecule compounds to inhibit PRMT5 activity led us to construct an in silico model of the human PRMT5 catalytic domain based on available homologous crystal structures from Protein Data Bank (MODELLER9v1 software). We screened a library of 10,000 compounds and eight small molecules were identified for biological investigation based on binding energy in the PRMT5 catalytic site. Enzyme inhibition assays using purified PRMT1 (type I PRMT) and PRMT5 (type II PRMT) showed that two compounds (BLL1 and BLL3) were capable of selectively inhibiting PRMT5 and not PRMT1 activity (p<0.0001). PRMT methylation assays were also performed with SWI/SNF complexes containing PRMT5, PRMT7 (type II) or PRMT4 (type I) and both BLL1 and BLL3 demonstrated selective PRMT5 inhibition. Both drugs interfered with maintenance of S2Me-H4R3 and S2Me-H3R8 in MCL cell lines by western blot and confocal microscopy. Dose titration experiments with BLL1 (10uM - 100uM) showed a dose-dependent response of inhibition of cellular proliferation, induction of cell cycle arrest, and promotion of caspase-independent cell death in 7 MCL cell lines. BLL1 treatment of MCL cells resulted in down modulation of cyclin D1 and Mcl1, critical molecules involved in the pathogenesis of MCL. The loss of cyclin D1 and Mcl1 expression occurred as early as 1 hour after treatment with BLL1 (50uM). PRMT5 associates with the co-repressors HDAC2, MBD2 and DNMT3a on target gene promoters, thus we next evaluated the effect of BLL1 on transcriptional repression of known target anti cancer genes. The association with other co-repressors provided rationale for examining PRMT5 inhibition alone and in combination with agents that target epigenetic processes. Combination treatment of MCL cells with subtoxic doses of BLL1 (25uM), hypomethylating agent (5-azacitidine, 500nM) and HDAC inhibitor (TSA 75nM) showed synergistic induction of cell death and loss of S2Me-H4R3. Analysis of the ST7 tumor suppressor, a target repressed by PRMT5, showed mRNA levels to increase 5–7-fold following treatment with BLL1. Preclinical in vivo studies have shown favorable toxicity and pharmacokinetic profiles for both BLL1 and BLL3. In vivo evaluation of BLL1 in a preclinical, xenograft model of human MCL are currently in progress. Primary tumors of 46 patients with MCL (common, blastoid or pleomorphic histology) demonstrated abundant PRMT5 expression in both cytoplasmic and nuclear compartments (87% PRMT5 pos). Conclusions: We have successfully developed a new class of drug to selectively target PRMT5 enzymatic activity. PRMT5 over expression is linked with post translational modification of both histone and non histone proteins that contribute to key oncogenic pathways in MCL. Inhibition of type II PRMT enzymes reverses transcriptional repression of anti cancer genes and restores important regulatory cellular checkpoints of cell growth and survival. We are currently developing drugs with improved selectivity and potency. The anti tumor activity of this novel class of drug and PRMT5 expression profiles seen in MCL primary tumor specimens, supports further exploration of targeting this pathway in hematologic malignancies. Disclosures: No relevant conflicts of interest to declare.