Radiology Commons^™

Comparing Ai-Generated And Human-Written Radiology Reports: Evidence From The Literature, Lakenya Webb

Radiologic Sciences

Purpose To evaluate the current evidence comparing artificial intelligence (AI)-generated radiology reports with radiologist-written reports, emphasizing diagnostic accuracy, clinical reasoning, workflow efficiency, and patient safety.

Method A literature review was performed using peer-reviewed articles identified through The University of Texas MD Anderson Cancer Center Research Medical Library. Studies evaluating AI-generated radiology reports were analyzed using AI-assisted thematic synthesis to identify recurring patterns related to report quality, diagnostic performance, and clinical implementation.

Results The literature consistently identified three overarching themes: (a) AI-generated reports demonstrated improvements in structured reporting and factual consistency but exhibited variable diagnostic accuracy; (b) current AI systems had …

Deep Learning Reconstruction In Pediatric Ct Imaging, Cody Wyatt

Radiologic Sciences

Purpose To synthesize current evidence on dose optimization using deep learning reconstruction (DLR) with emphasis on computed tomography (CT), pediatrics, contrast to noise ratio (CNR), signal to noise ratio (SNR), and radiation dosage.

Method The methodology was a systematic literature review, synthesizing peer-reviewed literature on the use of DLR in pediatric CT. Databases searched included Scopus, IEEE Xplore, and Web of Science. Keywords include combinations of deep learning reconstruction, DLR, CT, pediatric, dose, and optimization. Boolean operators were used to refine search results.

Results Three themes emerged: dosage reduction, noise reduction, and improved spatial resolution with the use of DLR …

National Pacs And Electronic Health Records: Considerations, Challenges, And Implications For Radiomics, Darchelle Scott

Radiologic Sciences

Purpose To examine the considerations, challenges, and implications of implementing a national picture archiving and communication system (PACS) integrated with electronic health records (EHRs), with particular emphasis on interoperability, radiomics, and patient-centered data sharing.

Method A systematic literature review was conducted using books, peer-reviewed articles, original research studies, and government publications addressing radiomics, TEFCA, blockchain technology, PACS, and EHR integration. Literature was analyzed to evaluate national health information exchange, radiomics applications, and implementation challenges associated with a connected healthcare infrastructure.

Results Three primary themes emerged: (a) interoperable national PACS and EHR systems have the potential to improve continuity of care …

Ai-Driven Segmentation And Volumetric Response Modeling Of Liver Regions To Radiotherapy, Aashish Chandra Gupta

Dissertations and Theses (Open Access)

In liver-directed radiotherapy (RT), liver regions receiving higher doses typically undergo atrophy while contralateral/adjacent lower-dose regions may exhibit compensatory hypertrophy through regeneration of healthy tissue. Optimizing the RT plan to promote regional hypertrophy while minimizing the risk of developing atrophy has the potential to enhance post-RT liver function and long-term survivorship. However, current clinical practice largely relies on global liver dose-volume metrics during RT-planning, which may obscure favorable dose-response correlation and limit actionable guidance for clinicians. Therefore, we hypothesized that post-RT regional liver response is governed by a combination of region-specific dose-volume and patient clinical features, and that these responses …

Developing A Deep-Learning Model To Segment White Matter Lesions In Cerebral Palsy, Varsha Kumar, Hyukjin Yun, Rose N. Gelineau-Morel, Avner Meoded

Posters

No abstract provided.

Computed Tomography Findings Of Normal Appendix In Pediatric Population, Thanwarat Koonthong, Teerasak Phewplung*

Chulalongkorn Medical Journal

Background: Appendicitis remains the most common indication for emergency abdominal surgery in children. Computed tomography (CT) plays a vital role in the evaluation of pediatric patients with suspected appendicitis. Yet, there is limited of literature describing the normal appendiceal diameter in the pediatric Asian population.

Objectives: To characterize the normal pediatric appendix and to identify the variables that influence its diameter in pediatric patients aged one day to 18 years who underwent abdominal CT.

Materials and Methods: The study was approved by the IRB. We conducted a retrospective study of 358 children (≤18 years of age) who underwent abdominal CT …

Fractured Focus: Understanding Stress In Second-Year Radiologic Science Students Across Alabama, Kelsi Clotfelter

Honors Theses

Higher education is a period of rapid transition and growing responsibility, where stress often becomes a routine part of students’ lives. In health professional programs, the combined demands of intensive coursework and clinical training can intensify emotional strain and negatively affect overall well-being. Radiography students experience uniquely high levels of academic, clinical, and emotional stress due to the simultaneous demands of didactic coursework, competency-based clinical training, and high-stakes professional expectations. A cross-sectional mixed-methods study was employed to evaluate the prevalence, sources, and perceived impact of stress among second-year radiography students across multiple Alabama institutions with the goal of identifying strategies …

Hypovascular Aggressive Angiomyxoma In A Pregnant Adolescent: A Case Report And Review Of Its Multimodal Imaging Appearances, Kyler Shin, Kweku Agyepong, Craig Kym, Mohammad Ghasemi-Rad, David Léon

Faculty, Staff and Students Publications

Aggressive angiomyxoma (AAM) is a rare mesenchymal tumor that primarily arises in the perineum and pelvis of young females. It is sometimes identified incidentally on pelvic examination and obstetric imaging. Although asymptomatic at small sizes, AAM is typically hypervascular and can grow large enough to complicate a patient's obstetric and gynecologic course of treatment. While AAM is a rare diagnosis reported in the literature, there are limited descriptions of its multimodal radiologic appearance, and uncommonly, it can present as a hypovascular mass. We describe a rare case of hypovascular AAM in pregnancy that did not respond favorably to targeted embolization, …

Blinded, Bias-Controlled Multi-Rater Evaluation Of Human-Versus-Ai Brain Metastasis Segmentation Using A Hybrid Foundation-Model Framework, Yiding Han, Enze Zhu, Mhd Hasan Ai Mekdash, Omar Awad, Piyush Pathak, Shixiao Liang, Daniel Allan Hamstra, Xizhe Zhang, Zaid Ali Siddiqui, Baozhou Sun

Faculty, Staff and Students Publications

Background: Accurate segmentation of brain metastases (BM) is essential for diagnosis, stereotactic radiosurgery planning, and longitudinal assessment. However, manual contouring is time-intensive, limiting clinical scalability, and exhibits substantial inter-observer variability. This variability complicates objective assessment of automated segmentation methods and challenges interpretation of model performance.

Purpose: To address these limitations, we developed TUM-SAM, a hybrid foundation-model framework for fully automated BM segmentation, and introduced a bias-controlled, blinded multi-rater evaluation paradigm to determine whether AI-based BM segmentation has reached expert-level performance and whether AI-generated contours are preferred by human experts under unbiased assessment.

Methods: TUM-SAM integrates nnU-Net-based lesion detection with a …

Hiding In Plain Sight: A Case Of Intestinal Tuberculosis Masquerading As Benign Lower Gi Bleeding, Lavender Otom, Priyanka Panwar, Gerald Odong, Fatma Kassim, Patricia Okiro, Christopher Opio, Saleem Abdulkarim

Internal Medicine, East Africa

Background Intestinal tuberculosis (TB) represents the most common form of abdominal TB, often affecting the ileocaecal region. Lower gastrointestinal bleeding is an uncommon presentation and may create significant diagnostic uncertainty, particularly in immunocompromised patients and in settings where inflammatory bowel disease is also encountered.

Case Presentation A woman in her sixties from Africa presented with recurrent diarrhoea and intermittent haematochezia over a 6-month period. Initial endoscopic evaluation was nondiagnostic. Cross-sectional imaging later demonstrated multifocal bowel involvement affecting both the small and large bowel, including ileocaecal and rectoanal disease, with associated abdominopelvic lymphadenopathy. Histopathology from repeat colonoscopy biopsy showed caseating granulomas …

Letter To The Editor: Hemodynamic Role In Hepatic Encephalopathy Pathogenesis Following Transjugular Intrahepatic Portosystemic Shunt Placement, John Halphen, Yash S Dayal, Cameron Laurie, Mohadese Ahmadzade, David Wynne, Mohammad Ghasemi-Rad

Faculty, Staff and Students Publications

Hepatic encephalopathy (HE) is a common complication following transjugular intrahepatic portosystemic shunt (TIPS) placement and is thought to result from reduced hepatic detoxification of neurotoxins. However, Zhuang et al in the World Journal of Radiology demonstrate that this understanding may be incomplete through exploration of the pathophysiology from a hemodynamic perspective. They found significant changes in cerebral blood flow (CBF) and spontaneous brain activity following TIPS. Initially, CBF increased in specific areas within the first month, followed by a return to baseline at three months. These CBF changes coincided with increased spontaneous brain activity that persisted in some areas, suggesting …

Imaging And Non-Imaging Approaches For The Diagnosis And Monitoring Of Necrotizing Enterocolitis-What Lies Ahead?, Indrani Bhattacharjee, Catalina Le Cacheux, Eric B. Ortigoza, Jonathan Dillman, Sherwin S. Chan, Alain Cuna

Manuscripts, Articles, Book Chapters and Other Papers

Necrotizing enterocolitis (NEC) remains one of the most serious gastrointestinal emergencies in preterm infants, and imaging plays a central role in diagnosis and clinical management. Historically, evaluation has relied primarily on abdominal radiography, which remains widely available and embedded in established diagnostic frameworks. However, the hallmark radiographic signs of NEC (i.e., pneumatosis intestinalis, portal venous gas, and free air) reflect relatively advanced manifestations of intestinal injury that indicate established mucosal disruption or transmural necrosis. Bowel ultrasound has increasingly complemented radiography by enabling real-time assessment of bowel wall integrity, perfusion, motility, and intra-abdominal fluid, providing physiologic information that may refine clinical …

A Novel Qa Method For Adaptive Aperture Position Verification Of Proton Therapy System, Shuxin Fan, Tiezhi Zhang, Xiandong Zhao, Winter Green, Baozhou Sun, Tianyu Zhao, Yao Hao

Faculty, Staff and Students Publications

Accurate Adaptive Aperture (AA) leaf positioning is essential for some proton therapy systems to ensure precise dose delivery and patient safety. However, conventional quality assurance (QA) methods can be suboptimal or inefficient for routinely assessing the dynamic nature of AA in the Mevion S250i HYPERSCAN system. This study introduces a novel automated QA method to verify AA positions utilizing a scintillation detector. A specialized QA plan consisting of 7 rectangular fields was generated in the RayStation treatment planning system (TPS) and delivered using the Mevion S250i HYPERSCAN proton therapy system. Dose distributions were measured with an IBA Lynx PT scintillation …

Retrospective Analysis Of A Washu-Based Knowledge-Based Planning (Kbp) Model For Breast And Chest Wall Planning In Guatemala: Compliance With American Society For Radiation Oncology (Astro) 2026 Guidelines, Milton E Ixquiac Cabrera, Erick O Montenegro, Matthew Schmidt, Baozhou Sun, James A Kavanaugh, Taoran Li, Angel Velarde, Vicky De Falla, Francisco Reynoso

Faculty, Staff and Students Publications

Background/purpose

Breast cancer is among the most prevalent malignancies treated at the Liga Nacional Contra el Cáncer (LNCC) in Guatemala, representing a significant proportion of annual radiotherapy cases. Access to high-quality, standardized treatment planning in resource-constrained settings remains a critical challenge. This study evaluates the dosimetric performance of knowledge-based planning (KBP) models adapted from Washington University (WashU) in St. Louis for breast and chest wall radiotherapy at LNCC, validated against a retrospective 2025 clinical cohort, and benchmarked against the ASTRO 2026 Practical Radiation Oncology guidelines.

Materials and methods

A retrospective analysis of 84 treatment plans (40 left, 44 right) for …

Mri-Based Deep Learning Radiomics Model For Automated Classification Of Disc Degeneration In The Lumbar Spine, Shiv Patil, Om Gandhi, Mert Karabacak, Matthew Carr, Konstantinos Margetis

Student Papers, Posters & Projects

Disc degeneration in the lumbar spine is a major cause of low back pain (LBP). The accurate grading of disc degeneration on magnetic resonance imaging (MRI) is critical for clinical management and patient selection for spine surgery. This study aims to develop and evaluate machine learning (ML) models that combine features from deep learning (DL) and radiomics for the automated prediction of Pfirrmann grade (PG), a measure of disc degeneration, using multi-parametric lumbar spine MRI. Sagittal T1, T2, and T2 SPACE MRIs of 218 patients with LBP were acquired from the SPIDER dataset. For each intervertebral disc and available sequence, …

A Multi-Agent Large Language Model Framework To Automatically Assess Performance Of A Clinical Ai Triage Tool, Adam Flanders, Yifan Peng, Luciano Prevedello, Robyn Ball, Errol Colak, Prahlad Menon, George Shih, Hui-Ming Lin, Paras Lakhani

Department of Radiology Faculty Papers

Radiology reports can be used as a surrogate for performance of clinical AI tools. Radiology reports were analyzed by an ensemble of eight open-source LLM models and a internal version of GPT-4o using a single multi-shot prompt that assessed for presence of ICH. Performance of the open-source models, consensus of models and GPT-4o were compared to human report review. Three ideal consensus LLM ensembles were tested for rating the performance of the triage tool. The capability of each LLM varied. The highest AUC performance was achieved with llama3.3:70b and GPT-4o. Using MCC the ideal combination of LLMs were: Full-9 Ensemble, …

Ghastly Growths And Bone Chilling Cases Pediatric Tumor Diagnosis Demystified, Alexander Aly Restum, Arif Musa

Medical Student Research Symposium

Background:

Pediatric skeletal tumors range from benign to malignant and require multimodal imaging for accurate diagnosis. Recognizing key features enables early detection, guides management, identifies underlying syndromes, and improves outcomes. This presentation provides an educational overview of pediatric tumor evaluation.

Methods:
We conducted a focused review of peer-reviewed literature and representative pediatric imaging studies to outline the radiologic characteristics of common benign and malignant bone tumors in children. Core imaging patterns, diagnostic challenges, and modality-specific advantages were synthesized to create a practical interpretive framework for radiologists.

Results:
The reviewed literature and imaging demonstrated consistent defining features across pediatric bone tumors. …

Imaging Of The Centromedian Thalamic Nucleus Using 3d Fast Gray Matter Acquisition T1 Inversion Recovery Sequence In 3t Mri In Epilepsy, Tunchanok Paprad, Thanyachat Yutthakasaemsan, Supada Prakkamakul

Chulalongkorn Medical Journal

Background:

The centromedian thalamic nucleus (CM) is an emerging deep brain stimulation (DBS) target for neurological disorders, including drug-resistant epilepsy. Traditional DBS localization relies on predefined stereotactic coordinates, which may be imprecise. Novel MRI sequences enhance gray-white matter contrast, potentially allowing direct CM visualization, improving accuracy in DBS planning and targeting.

Objective:

This study evaluates the 3D Fast Gray Matter Acquisition T1 Inversion Recovery (FGATIR) MRI sequence for the visualization of the CM in epilepsy patients.

Methods:

This prospective IRB-approved descriptive study included adult epilepsy patients who underwent 3D FGATIR MRI sequence at 3 Tesla. Two neuroradiologists independently assessed CM …

X-Ray Versus Ct For Femoral Landmark Identification: A Systematic Review And Meta-Analysis, Akash B. Patel, Marzieh Saeedi-Hosseiny, Mohammad Abedin-Nasab

Rowan-Virtua Research Day

Background Accurate femoral bone landmark identification is essential for orthopedic surgical planning, deformity correction, and fracture risk assessment. Radiography offers accessibility and low cost, while CT provides superior spatial resolution and three-dimensional visualization. Despite technological advances, no quantitative synthesis has compared these modalities specifically for femoral landmark identification.

Hypothesis CT provides superior measurement accuracy and interobserver reliability compared to radiography for femoral bone landmark identification.

Methods Following PRISMA 2020 guidelines, Embase, Scopus, and PubMed were searched from 2015-2025. Studies comparing radiography and CT for femoral landmark identification with quantitative outcomes were included. Dual-reviewer screening of 604 records yielded 27 studies …

Impact Of Intracranial Pressure Monitoring On 30-Day Mortality In Severe Traumatic Brain Injury: A Retrospective Cohort Study In Kenya, George Ongas, Idris Chikophe, Jacqueline Mavuti, David Olunya, Mubashir Qureshi, Beverly Cheserem, Edwin Mogere

General Surgery, East Africa

OBJECTIVE The objective of this study was to evaluate 15-year trends in epidemiology, baseline severity, management, and outcomes of severe traumatic brain injury (sTBI) at a Kenyan tertiary center, and to determine the impact of invasive intracranial pressure (ICP) monitoring on 30-day mortality in a low- and middle-income country (LMIC) context.

METHODS A retrospective quasi-experimental cohort analysis of consecutive sTBI admissions (postresuscitation Glasgow Coma Scale [GCS] score ≤ 8, age ≥ 13 years) from January 1, 2010, to December 31, 2024, was conducted. This time period was divided into two eras: era 1 (2010–2014), premonitoring, and era 2 (2015–2024), routine …