Medical Genetics Commons^™

High-Resolution Metabolomics Of Exposure To Tobacco Smoke During Pregnancy And Adverse Birth Outcomes In The Atlanta African American Maternal-Child Cohort, Youran Tan, Dana Boyd Barr, P Barry Ryan, Veronika Fedirko, Jeremy A Sarnat, Audrey J Gaskins, Che-Jung Chang, Ziyin Tang, Carmen J Marsit, Elizabeth J Corwin, Dean P Jones, Anne L Dunlop, Donghai Liang

Faculty, Staff and Student Publications

Exposure to tobacco smoke during pregnancy has been associated with a series of adverse reproductive outcomes; however, the underlying molecular mechanisms are not well-established. We conducted an untargeted metabolome-wide association study to identify the metabolic perturbations and molecular mechanisms underlying the association between cotinine, a widely used biomarker of tobacco exposure, and adverse birth outcomes. We collected early and late pregnancy urine samples for cotinine measurement and serum samples for high-resolution metabolomics (HRM) profiling from 105 pregnant women from the Atlanta African American Maternal-Child cohort (2014-2016). Maternal metabolome perturbations mediating prenatal tobacco smoke exposure and adverse birth outcomes were assessed …

Progression Free Survival Prediction For Head And Neck Cancer Using Deep Learning Based On Clinical And Pet/Ct Imaging Data, Mohamed A Naser, Kareem A Wahid, Abdallah S R Mohamed, Moamen Abobakr Abdelaal, Renjie He, Cem Dede, Lisanne V Van Dijk, Clifton D Fuller

Faculty, Staff and Student Publications

Determining progression-free survival (PFS) for head and neck squamous cell carcinoma (HNSCC) patients is a challenging but pertinent task that could help stratify patients for improved overall outcomes. PET/CT images provide a rich source of anatomical and metabolic data for potential clinical biomarkers that would inform treatment decisions and could help improve PFS. In this study, we participate in the 2021 HECKTOR Challenge to predict PFS in a large dataset of HNSCC PET/CT images using deep learning approaches. We develop a series of deep learning models based on the DenseNet architecture using a negative log-likelihood loss function that utilizes PET/CT …

Combining Tumor Segmentation Masks With Pet/Ct Images And Clinical Data In A Deep Learning Framework For Improved Prognostic Prediction In Head And Neck Squamous Cell Carcinoma, Kareem A Wahid, Renjie He, Cem Dede, Abdallah S R Mohamed, Moamen Abobakr Abdelaal, Lisanne V Van Dijk, Clifton D Fuller, Mohamed A Naser

Faculty, Staff and Student Publications

PET/CT images provide a rich data source for clinical prediction models in head and neck squamous cell carcinoma (HNSCC). Deep learning models often use images in an end-to-end fashion with clinical data or no additional input for predictions. However, in the context of HNSCC, the tumor region of interest may be an informative prior in the generation of improved prediction performance. In this study, we utilize a deep learning framework based on a DenseNet architecture to combine PET images, CT images, primary tumor segmentation masks, and clinical data as separate channels to predict progression-free survival (PFS) in days for HNSCC …

Establishment And Validation Of Pre-Therapy Cervical Vertebrae Muscle Quantification As A Prognostic Marker Of Sarcopenia In Patients With Head And Neck Cancer, Brennan Olson, Jared Edwards, Catherine Degnin, Nicole Santucci, Michelle Buncke, Jeffrey Hu, Yiyi Chen, Clifton D Fuller, Mathew Geltzeiler, Aaron J Grossberg, Daniel Clayburgh

Faculty, Staff and Student Publications

Background: Sarcopenia is prognostic for survival in patients with head and neck cancer (HNC). However, identification of this high-risk feature remains challenging without computed tomography (CT) imaging of the abdomen or thorax. Herein, we establish sarcopenia thresholds at the C3 level and determine if C3 sarcopenia is associated with survival in patients with HNC.

Methods: This retrospective cohort study was conducted in consecutive patients with a squamous cell carcinoma of the head and neck with cross-sectional abdominal or neck imaging within 60 days prior to treatment. Measurement of cross-sectional muscle area at L3 and C3 levels was performed from CT …

Deep-Learning-Based Generation Of Synthetic 6-Minute Mri From 2-Minute Mri For Use In Head And Neck Cancer Radiotherapy, Kareem A Wahid, Jiaofeng Xu, Dina El-Habashy, Yomna Khamis, Moamen Abobakr, Brigid Mcdonald, Nicolette O' Connell, Daniel Thill, Sara Ahmed, Christina Setareh Sharafi, Kathryn Preston, Travis C Salzillo, Abdallah S R Mohamed, Renjie He, Nathan Cho, John Christodouleas, Clifton D Fuller, Mohamed A Naser

Faculty, Staff and Student Publications

Background: Quick magnetic resonance imaging (MRI) scans with low contrast-to-noise ratio are typically acquired for daily MRI-guided radiotherapy setup. However, for patients with head and neck (HN) cancer, these images are often insufficient for discriminating target volumes and organs at risk (OARs). In this study, we investigated a deep learning (DL) approach to generate high-quality synthetic images from low-quality images.

Methods: We used 108 unique HN image sets of paired 2-minute T2-weighted scans (2mMRI) and 6-minute T2-weighted scans (6mMRI). 90 image sets (~20,000 slices) were used to train a 2-dimensional generative adversarial DL model that utilized 2mMRI as input and …

Head And Neck Cancer Primary Tumor Auto Segmentation Using Model Ensembling Of Deep Learning In Pet/Ct Images, Mohamed A Naser, Kareem A Wahid, Lisanne V Van Dijk, Renjie He, Moamen Abobakr Abdelaal, Cem Dede, Abdallah S R Mohamed, Clifton D Fuller

Faculty, Staff and Student Publications

Auto-segmentation of primary tumors in oropharyngeal cancer using PET/CT images is an unmet need that has the potential to improve radiation oncology workflows. In this study, we develop a series of deep learning models based on a 3D Residual Unet (ResUnet) architecture that can segment oropharyngeal tumors with high performance as demonstrated through internal and external validation of large-scale datasets (training size = 224 patients, testing size = 101 patients) as part of the 2021 HECKTOR Challenge. Specifically, we leverage ResUNet models with either 256 or 512 bottleneck layer channels that demonstrate internal validation (10-fold cross-validation) mean Dice similarity coefficient …

Use Of Specific Duodenal Dose Constraints During Treatment Planning Reduces Toxicity After Definitive Paraaortic Radiation Therapy For Cervical Cancer, David S Lakomy, Juliana Wu, Bhavana V Chapman, Zhiqian Henry Yu, Belinda Lee, Ann H Klopp, Anuja Jhingran, Patricia J Eifel, Lilie L Lin

Faculty, Staff and Student Publications

Purpose: This study aimed to validate the safety of paraaortic nodal (PAN) radiation therapy (RT) for patients with cervical cancer when the duodenal dose is limited to V55 < 15 cm3 and V60 < 2 cm3.

Methods and materials: A total of 97 patients who were treated with RT for cervical cancer between 2010 and 2018 received at least 56 Gy to grossly involved PANs. Patients were treated with concurrent chemoradiation (n = 88; 91%), with 93% of patients (n = 90) receiving intensity modulated RT to the initial PAN field and 98% (n = 95) receiving intensity modulated RT to a sequential PAN boost. The V55 < 15 cm3 and V60 < 2 cm3 criteria were implemented in 2014. Normal tissues were contoured on computed tomography (CT) simulation data sets, and the duodenum was contoured from the gastric outlet to the duodenojejunal flexure. Sixty-six patients (68%) had a resimulation scan after approximately 20 fractions. Composite duodenal doses were calculated using the initial CT scan for 50 patients (52%) and the resimulation CT scan for 47 patients (48%) depending on the anatomic changes throughout treatment.

Results: …

Clinical Characteristics And Outcomes Of Castleman Disease: A Multicenter Consortium Study Of 428 Patients With 15-Year Follow-Up, Wanying Liu, Qingqing Cai, Tiantian Yu, Paolo Strati, Frederick B Hagemeister, Qiongli Zhai, Mingzhi Zhang, Ling Li, Xiaosheng Fang, Jianyong Li, Ruifang Sun, Shanxiang Zhang, Hanjin Yang, Zhaoming Wang, Wenbian Qian, Noriko Iwaki, Yasuharu Sato, Eric Oksenhendler, Zijun Y Xu-Monette, Ken H Young, Li Yu

Faculty, Staff and Student Publications

Castleman disease (CD) has been reported as a group of poorly understood lymphoproliferative disorders, including unicentric CD (UCD) and idiopathic multicentric CD (iMCD) which are human immunodeficiency virus (HIV) negative and human herpes virus 8 (HHV-8) negative. The clinical and independent prognostic factors of CD remain poorly elucidated. We retrospectively collected the clinical information of 428 patients with HIV and HHV-8 negative CD from 12 large medical centers with 15-year follow-up. We analyzed the clinicopathologic features of 428 patients (248 with UCD and 180 with iMCD) with a median age of 41 years. The histology subtypes were hyaline-vascular (HV) histopathology …

Targetable Vulnerability Of Deregulated Foxm1/Plk1 Signaling Axis In Diffuse Large B Cell Lymphoma, Fang Yu, Hua He, Loretta J Nastoupil, Zijun Y Xu-Monette, Ky Pham, Yong Liang, Guang Chen, Nathan H Fowler, C Cameron Yin, Dongfeng Tan, Yaling Yang, Shimin Hu, Ken H Young, Lan V Pham, M James You

Faculty, Staff and Student Publications

FOXM1 is a transcription factor that controls cell cycle regulation, cell proliferation, and differentiation. Overexpression of FOXM1 has been implicated in various cancer types. However, the activation status and functional significance of FOXM1 in diffuse large B cell lymphoma (DLBCL) have not been well investigated. Using proteomic approaches, we discovered that the protein expression levels of FOXM1 and PLK1 were positively correlated in DLBCL cell lines and primary DLBCL. Expression levels of FOXM1 and PLK1 mRNAs were also significantly higher in DLBCL than in normal human B cells and could predict poor prognosis of DLBCL, particularly in patients with germinal …

Head-Mounted Digital Metamorphopsia Suppression As A Countermeasure For Macular-Related Visual Distortions For Prolonged Spaceflight Missions And Terrestrial Health, Joshua Ong, Nasif Zaman, Ethan Waisberg, Sharif Amit Kamran, Andrew G Lee, Alireza Tavakkoli

Faculty, Staff and Student Publications

During long-duration spaceflight, astronauts are exposed to various risks including spaceflight-associated neuro-ocular syndrome, which serves as a risk to astronaut vision and a potential physiological barrier to future spaceflight. When considering exploration missions that may expose astronauts to longer periods of microgravity, radiation exposure, and natural aging processes during spaceflight, more severe changes to functional vision may occur. The macula plays a critical role in central vision and disruptions to this key area in the eye may compromise functional vision and mission performance. In this article, we describe the development of a countermeasure technique to digitally suppress monocular central visual …

Disruption Of Growth Hormone Receptor Signaling Abrogates Hepatocellular Carcinoma Development, Abedul Haque, Vishal Sahu, Jamie Lynne Lombardo, Lianchun Xiao, Bhawana George, Robert A Wolff, Jeffrey S Morris, Asif Rashid, John J Kopchick, Ahmed O Kaseb, Hesham M Amin

Faculty, Staff and Student Publications

Introduction: Hepatocellular carcinoma (HCC) is the most common type of primary liver cancers. It is an aggressive neoplasm with dismal outcome because most of the patients present with an advanced-stage disease, which precludes curative surgical options. Therefore, these patients require systemic therapies that typically induce small improvements in overall survival. Hence, it is crucial to identify new and promising therapeutic targets for HCC to improve the current outcome. The liver is a key organ in the signaling cascade triggered by the growth hormone receptor (GHR). Previous studies have shown that GHR signaling stimulates the proliferation and regeneration of liver cells …

Editorial: The Role Of The Igf Axis In Tumorigenesis And Cancer Treatment: From Genes To Metabolites, Guo Mengzhe, Xie Shanshan, Zhao Di, Wu Shihua

Faculty, Staff and Student Publications

No abstract provided.

Blockade Of Growth Hormone Receptor Signaling By Using Pegvisomant: A Functional Therapeutic Strategy In Hepatocellular Carcinoma, Ahmed O Kaseb, Abedul Haque, Deeksha Vishwamitra, Manal M Hassan, Lianchun Xiao, Bhawana George, Vishal Sahu, Yehia I Mohamed, Roberto Carmagnani Pestana, Jamie Lynne Lombardo, Rony Avritscher, James C Yao, Robert A Wolff, Asif Rashid, Jeffrey S Morris, Hesham M Amin

Faculty, Staff and Student Publications

Hepatocellular carcinoma (HCC) is an aggressive neoplasm with poor clinical outcome because most patients present at an advanced stage, at which point curative surgical options, such as tumor excision or liver transplantation, are not feasible. Therefore, the majority of HCC patients require systemic therapy. Nonetheless, the currently approved systemic therapies have limited effects, particularly in patients with advanced and resistant disease. Hence, there is a critical need to identify new molecular targets and effective systemic therapies to improve HCC outcome. The liver is a major target of the growth hormone receptor (GHR) signaling, and accumulating evidence suggests that GHR signaling …

Harmonization Of Multi-Scanner In Vivo Magnetic Resonance Spectroscopy: Enigma Consortium Task Group Considerations, Ashley D Harris, Houshang Amiri, Mariana Bento, Ronald Cohen, Christopher R K Ching, Christina Cudalbu, Emily L Dennis, Arne Doose, Stefan Ehrlich, Ivan I Kirov, Ralf Mekle, Georg Oeltzschner, Eric Porges, Roberto Souza, Friederike I Tam, Brian Taylor, Paul M Thompson, Yann Quidé, Elisabeth A Wilde, John Williamson, Alexander P Lin, Brenda Bartnik-Olson

Faculty, Staff and Student Publications

Magnetic resonance spectroscopy is a powerful, non-invasive, quantitative imaging technique that allows for the measurement of brain metabolites that has demonstrated utility in diagnosing and characterizing a broad range of neurological diseases. Its impact, however, has been limited due to small sample sizes and methodological variability in addition to intrinsic limitations of the method itself such as its sensitivity to motion. The lack of standardization from a data acquisition and data processing perspective makes it difficult to pool multiple studies and/or conduct multisite studies that are necessary for supporting clinically relevant findings. Based on the experience of the ENIGMA MRS …

Elastic Shape Analysis Of Brain Structures For Predictive Modeling Of Ptsd, Yuexuan Wu, Suprateek Kundu, Jennifer S Stevens, Negar Fani, Anuj Srivastava

Faculty, Staff and Student Publications

It is well-known that morphological features in the brain undergo changes due to traumatic events and associated disorders such as post-traumatic stress disorder (PTSD). However, existing approaches typically offer group-level comparisons, and there are limited predictive approaches for modeling behavioral outcomes based on brain shape features that can account for heterogeneity in PTSD, which is of paramount interest. We propose a comprehensive shape analysis framework representing brain sub-structures, such as the hippocampus, amygdala, and putamen, as parameterized surfaces and quantifying their shape differences using an elastic shape metric. Under this metric, we compute shape summaries (mean, covariance, PCA) of brain …

Urinary T Cells Are Detected In Patients With Immune Checkpoint Inhibitor-Associated Immune Nephritis That Are Clonotypically Identical To Kidney T Cell Infiltrates, Shailbala Singh, Leticia C Clemente, Edwin R Parra, Amanda Tchakarov, Chao Yang, Yisheng Li, James P Long, Cassian Yee, Jamie S Lin

Faculty, Staff and Student Publications

Acute kidney injury (AKI) occurs in ~20% of patients receiving immune checkpoint inhibitor (ICI) therapy; however, only 2-5% will develop ICI-mediated immune nephritis. Conventional tests are nonspecific in diagnosing disease pathology and invasive procedures (i.e. kidney biopsy) may not be feasible. In other autoimmune renal diseases, urinary immune cells correlated with the pathology or were predictive of disease activity. Corresponding evidence and analysis are absent for ICI-mediated immune nephritis. We report the first investigation analyzing immune cell profiles of matched kidney biopsies and urine of patients with ICI-AKI. We demonstrated the presence of urinary T cells in patients with immune …

Doxorubicin-Induced Cardiotoxicity Is Mediated By Neutrophils Through Release Of Neutrophil Elastase, Anchit Bhagat, Pradeep Shrestha, Prince Jeyabal, Zhanglong Peng, Stephanie S Watowich, Eugenie S Kleinerman

Faculty, Staff and Student Publications

The mechanisms by which Doxorubicin (Dox) causes acute and late cardiotoxicity are not completely understood. One understudied area is the innate immune response, and in particular the role of neutrophils in Dox-induced cardiotoxicity. Here, using echocardiography, flow cytometry and immunofluorescence staining, we demonstrated increased infiltration of neutrophils that correlated with decreased heart function, disruption of vascular structures and increased collagen deposition in the heart after Dox treatment. Depleting neutrophils protected the heart from Dox-induced cardiotoxicity and changes in vascular structure. Furthermore, our data using neutrophil elastase (NE) knock-out mice and the NE inhibitor AZD9668 suggest that neutrophils cause this damage …

Donor Selection For Kir Alloreactivity Is Associated With Superior Survival In Haploidentical Transplant With Ptcy, Jun Zou, Piyanuch Kongtim, Samer A Srour, Uri Greenbaum, Johannes Schetelig, Falk Heidenreich, Henning Baldauf, Brandt Moore, Supawee Saengboon, Yudith Carmazzi, Gabriela Rondon, Qing Ma, Katayoun Rezvani, Elizabeth J Shpall, Richard E Champlin, Stefan O Ciurea, Kai Cao

Faculty, Staff and Student Publications

With the continuous increase in the use of haploidentical donors for transplantation, the selection of donors becomes increasingly important. Haploidentical donors have been selected primarily based on clinical characteristics, while the effects of killer cell immunoglobulin-like receptors (KIRs) on outcomes of haploidentical-hematopoietic stem cell transplantation (haplo-HSCT) with post-transplant cyclophosphamide (PTCy) remain inconclusive. The present study aimed to thoroughly evaluate the effect of KIRs and binding ligands assessed by various models, in addition to other patient/donor variables, on clinical outcomes in haplo-HSCT. In a cohort of 354 patients undergoing their first haplo-HSCT, we found that a higher Count Functional inhibitory KIR …

Distal-Less Homeobox Genes Dlx5/6 Regulate Müllerian Duct Regression, Rachel D Mullen, Brice Bellessort, Giovanni Levi, Richard R Behringer

Faculty, Staff and Student Publications

Dlx5 and Dlx6 encode distal-less homeodomain transcription factors that are present in the genome as a linked pair at a single locus. Dlx5 and Dlx6 have redundant roles in craniofacial, skeletal, and uterine development. Previously, we performed a transcriptome comparison for anti-Müllerian hormone (AMH)-induced genes expressed in the Müllerian duct mesenchyme of male and female mouse embryos. In that study, we found that Dlx5 transcripts were nearly seven-fold higher in males compared to females and Dlx6 transcripts were found only in males, suggesting they may be AMH-induced genes. Therefore, we investigated the role of Dlx5 and Dlx6 during AMH-induced Müllerian …

Clinical Implementation And Initial Experience With A 15 Tesla Mr-Linac For Mr-Guided Radiation Therapy For Gynecologic Cancer: An R-Ideal Stage 1 And 2a First In Humans Feasibility Study Of New Technology Implementation, David S Lakomy, Jinzhong Yang, Sastry Vedam, Jihong Wang, Belinda Lee, Angela Sobremonte, Pamela Castillo, Neil Hughes, Mustefa Mohammedsaid, Anuja Jhingran, Ann H Klopp, Seungtaek Choi, C David Fuller, Lilie L Lin

Faculty, Staff and Student Publications

PURPOSE: Magnetic resonance imaging-guided linear accelerator systems (MR-linacs) can facilitate the daily adaptation of radiation therapy plans. Here, we report our early clinical experience using a MR-linac for adaptive radiation therapy of gynecologic malignancies.

METHODS AND MATERIALS: Treatments were planned with an Elekta Monaco v5.4.01 and delivered by a 1.5 Tesla Elekta Unity MR-linac. The system offers a choice of daily adaptation based on either position (ATP) or shape (ATS) of the tumor and surrounding normal structures. The ATS approach has the option of manually editing the contours of tumors and surrounding normal structures before the plan is adapted. Here, …

A Novel Group Of Genes That Cause Endocrine Resistance In Breast Cancer Identified By Dynamic Gene Expression Analysis, Arvand Asghari, Katherine Wall, Michael Gill, Natascha Del Vecchio, Farnaz Allahbakhsh, Jacky Wu, Nan Deng, W Jim Zheng, Hulin Wu, Michihisa Umetani, Vahed Maroufy

Faculty, Staff and Student Publications

Breast cancer (BC) is the most common type of cancer diagnosed in women. Among female cancer deaths, BC is the second leading cause of death worldwide. For estrogen receptor-positive (ER-positive) breast cancers, endocrine therapy is an effective therapeutic approach. However, in many cases, an ER-positive tumor becomes unresponsive to endocrine therapy, and tumor regrowth occurs after treatment. While some genetic mutations contribute to resistance in some patients, the underlying causes of resistance to endocrine therapy are mostly undetermined. In this study, we utilized a recently developed statistical approach to investigate the dynamic behavior of gene expression during the development of …

Effects Of Tamoxifen Inducible Mercremer On Gene Expression In Cardiac Myocytes In Mice, Leila Rouhi, Siyang Fan, Sirisha M Cheedipudi, Melis Olcum, Hyun-Hwan Jeong, Zhongming Zhao, Priyatansh Gurha, Ali J Marian

Faculty, Staff and Student Publications

The Cre-LoxP technology, including the tamoxifen (TAM) inducible MerCreMer (MCM), is increasingly used to delineate gene function, understand the disease mechanisms, and test therapeutic interventions. We set to determine the effects of TAM-MCM on cardiac myocyte transcriptome. Expression of the MCM was induced specifically in cardiac myocytes upon injection of TAM to myosin heavy chain 6-MCM (Myh6-Mcm) mice for 5 consecutive days. Cardiac function, myocardial histology, and gene expression (RNA-sequencing) were analyzed 2 weeks after TAM injection. A total of 346 protein coding genes (168 up- and 178 down-regulated) were differentially expressed. Transcript levels of 85 genes, analyzed …

An Ensemble-Based Deep Convolutional Neural Network For Computer-Aided Polyps Identification From Colonoscopy, Pallabi Sharma, Bunil Kumar Balabantaray, Kangkana Bora, Saurav Mallik, Kunio Kasugai, Zhongming Zhao

Faculty, Staff and Student Publications

Colorectal cancer (CRC) is the third leading cause of cancer death globally. Early detection and removal of precancerous polyps can significantly reduce the chance of CRC patient death. Currently, the polyp detection rate mainly depends on the skill and expertise of gastroenterologists. Over time, unidentified polyps can develop into cancer. Machine learning has recently emerged as a powerful method in assisting clinical diagnosis. Several classification models have been proposed to identify polyps, but their performance has not been comparable to an expert endoscopist yet. Here, we propose a multiple classifier consultation strategy to create an effective and powerful classifier for …

A Deep Learning-Based Framework For Supporting Clinical Diagnosis Of Glioblastoma Subtypes, Sana Munquad, Tapas Si, Saurav Mallik, Asim Bikas Das, Zhongming Zhao

Faculty, Staff and Student Publications

Understanding molecular features that facilitate aggressive phenotypes in glioblastoma multiforme (GBM) remains a major clinical challenge. Accurate diagnosis of GBM subtypes, namely classical, proneural, and mesenchymal, and identification of specific molecular features are crucial for clinicians for systematic treatment. We develop a biologically interpretable and highly efficient deep learning framework based on a convolutional neural network for subtype identification. The classifiers were generated from high-throughput data of different molecular levels, i.e., transcriptome and methylome. Furthermore, an integrated subsystem of transcriptome and methylome data was also used to build the biologically relevant model. Our results show that deep learning model outperforms …

Dimensionality Reduction And Louvain Agglomerative Hierarchical Clustering For Cluster-Specified Frequent Biomarker Discovery In Single-Cell Sequencing Data, Soumita Seth, Saurav Mallik, Tapas Bhadra, Zhongming Zhao

Faculty, Staff and Student Publications

The major interest domains of single-cell RNA sequential analysis are identification of existing and novel types of cells, depiction of cells, cell fate prediction, classification of several types of tumor, and investigation of heterogeneity in different cells. Single-cell clustering plays an important role to solve the aforementioned questions of interest. Cluster identification in high dimensional single-cell sequencing data faces some challenges due to its nature. Dimensionality reduction models can solve the problem. Here, we introduce a potential cluster specified frequent biomarkers discovery framework using dimensionality reduction and hierarchical agglomerative clustering Louvain for single-cell RNA sequencing data analysis. First, we pre-filtered …

Explanation-Driven Deep Learning Model For Prediction Of Brain Tumour Status Using Mri Image Data, Loveleen Gaur, Mohan Bhandari, Tanvi Razdan, Saurav Mallik, Zhongming Zhao

Faculty, Staff and Student Publications

Cancer research has seen explosive development exploring deep learning (DL) techniques for analysing magnetic resonance imaging (MRI) images for predicting brain tumours. We have observed a substantial gap in explanation, interpretability, and high accuracy for DL models. Consequently, we propose an explanation-driven DL model by utilising a convolutional neural network (CNN), local interpretable model-agnostic explanation (LIME), and Shapley additive explanation (SHAP) for the prediction of discrete subtypes of brain tumours (meningioma, glioma, and pituitary) using an MRI image dataset. Unlike previous models, our model used a dual-input CNN approach to prevail over the classification challenge with images of inferior quality …

Breast Cancer Detection: Shallow Convolutional Neural Network Against Deep Convolutional Neural Networks Based Approach, Himanish Shekhar Das, Akalpita Das, Anupal Neog, Saurav Mallik, Kangkana Bora, Zhongming Zhao

Faculty, Staff and Student Publications

Introduction: Of all the cancers that afflict women, breast cancer (BC) has the second-highest mortality rate, and it is also believed to be the primary cause of the high death rate. Breast cancer is the most common cancer that affects women globally. There are two types of breast tumors: benign (less harmful and unlikely to become breast cancer) and malignant (which are very dangerous and might result in aberrant cells that could result in cancer).

Methods: To find breast abnormalities like masses and micro-calcifications, competent and educated radiologists often examine mammographic images. This study focuses on computer-aided diagnosis to help …

Re-Thinking Therapeutic Development For Cns Metastatic Disease, Chantal Saberian, Michael A Davies

Faculty, Staff and Student Publications

There has been unprecedented progress in the development of systemic therapies for patients with metastatic melanoma over the last decade. There is now tremendous potential and momentum to further and markedly reduce the impact of this disease. However, developing more effective treatments for metastases to the CNS remains a critical challenge for patients with melanoma. Melanoma patients with active CNS metastases have largely been excluded from both early-phase and registration trials for all currently approved targeted and immune therapies for this disease. While this exclusion has generally been justified in clinical research due to concerns about poor prognosis, lack of …

Deep Learning Auto-Segmentation Of Cervical Skeletal Muscle For Sarcopenia Analysis In Patients With Head And Neck Cancer, Mohamed A Naser, Kareem A Wahid, Aaron J Grossberg, Brennan Olson, Rishab Jain, Dina El-Habashy, Cem Dede, Vivian Salama, Moamen Abobakr, Abdallah S R Mohamed, Renjie He, Joel Jaskari, Jaakko Sahlsten, Kimmo Kaski, Clifton D Fuller

Faculty, Staff and Student Publications

Background/purpose: Sarcopenia is a prognostic factor in patients with head and neck cancer (HNC). Sarcopenia can be determined using the skeletal muscle index (SMI) calculated from cervical neck skeletal muscle (SM) segmentations. However, SM segmentation requires manual input, which is time-consuming and variable. Therefore, we developed a fully-automated approach to segment cervical vertebra SM.

Materials/methods: 390 HNC patients with contrast-enhanced CT scans were utilized (300-training, 90-testing). Ground-truth single-slice SM segmentations at the C3 vertebra were manually generated. A multi-stage deep learning pipeline was developed, where a 3D ResUNet auto-segmented the C3 section (33 mm window), the middle slice of the …

Evaluation Of Deep Learning-Based Multiparametric Mri Oropharyngeal Primary Tumor Auto-Segmentation And Investigation Of Input Channel Effects: Results From A Prospective Imaging Registry, Kareem A Wahid, Sara Ahmed, Renjie He, Lisanne V Van Dijk, Jonas Teuwen, Brigid A Mcdonald, Vivian Salama, Abdallah S R Mohamed, Travis Salzillo, Cem Dede, Nicolette Taku, Stephen Y Lai, Clifton D Fuller, Mohamed A Naser

Faculty, Staff and Student Publications

Background/purpose: Oropharyngeal cancer (OPC) primary gross tumor volume (GTVp) segmentation is crucial for radiotherapy. Multiparametric MRI (mpMRI) is increasingly used for OPC adaptive radiotherapy but relies on manual segmentation. Therefore, we constructed mpMRI deep learning (DL) OPC GTVp auto-segmentation models and determined the impact of input channels on segmentation performance.

Materials/methods: GTVp ground truth segmentations were manually generated for 30 OPC patients from a clinical trial. We evaluated five mpMRI input channels (T2, T1, ADC, Ktrans, Ve). 3D Residual U-net models were developed and assessed using leave-one-out cross-validation. A baseline T2 model was compared to mpMRI models (T2 + T1, …