CSCaroline Elizabeth Serafini

Caroline Elizabeth Serafini

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Caroline Elizabeth Serafini has a total of 57 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Automated linac downtime tracking across 4 radiation oncology clinics within a single hospital system utilizing statistical process control

    by Caroline E. Serafini, Matthew Thomas, Xiaofeng Yang, Pretesh Patel, Justin Roper, Maxine Washington, Anees Dhabaan, Hania Al‐Hallaq

    Background: Linear accelerator (linac) downtime disrupts clinical workflow and may delay patient treatment; however, downtime events are frequently undocumented or recorded in informal reporting systems. Data‐driven approaches using routinely collected operational metrics may improve downtime surveillance and provide insight into system‐level adaptation. Purpose: This study aims to evaluate if statistical process control (SPC) methods can be applied to routine treatment and scheduling data from an oncology information system (OIS) to identify linac downtime events and characterize operational response across multiple radiation oncology clinics within a single hospital system. Methods: Daily machine‐level operational data from 10 linacs across 4 clinics were analyzed over 29 months (January 2024–May 2026). Machine metrics included daily patient volume, cancellations, and operating hours. For each machine, rolling 45‐day baseline means and standard deviations were computed, and SPC rules (± 3σ control limits and > 10 cancellations / day) were applied to detect downtime days. SPC‐flagged events were compared with email‐based downtime event logs to calculate sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Downtime events were categorized by cause, and intra‐hospital patient redistribution was assessed by identifying concurrent > 4σ volume increases on alternate machines. Results: Over the 29‐month study period, 138 downtime events were documented in downtime event logs. SPC identified 23 additional events exhibiting operational signatures consistent with downtime. Using reported events as the reference standard, SPC achieved sensitivity 0.70, specificity 0.99, PPV 0.79, and NPV 0.99. Downtime frequently resulted in compensatory increases in patient volume on alternate machines, with 41 transfer pairs identified across three hospitals. When the reported downtime cause was specified, machine mechanical causes accounted for the largest proportion (48%) of downtime events. Conclusions: SPC applied to routinely collected operational data provides a scalable framework for automated linac downtime detection. Integration of automated monitoring with structured reporting may improve downtime surveillance and inform proactive operational planning in radiation oncology

  • Qualitative and quantitative assessment of ex vivo human brain tumors using quantitative oblique back-illumination microscopy (qOBM)

    by Srinidhi Bharadwaj, Paloma Casteleiro Costa, Caroline Serafini, Brienna Heinsz, Alice Hsu, Nischita Kaza, Zhe Guang, Zhenmin Li, Jeffrey J. Olson, Kimberly Hoang, et al.

    Accurate intraoperative diagnosis of brain tumors is critical for improving patient outcomes, yet current histopathology is time-consuming and performed outside the operating room. We investigate the use of quantitative oblique back-illumination microscopy (qOBM) for the identification and classification of tumors in brain tumor core samples ex vivo . We studied samples from 28 patients across four tumor types and identified diagnostic features in qOBM images corroborated by H&E. Using samples from 19 glioma patients, a leave-one-out classifier distinguishes high- from low-grade glioma cases with 86.6% accuracy (95% CI: 85.9-87.4%), demonstrating qOBM’s potential for real-time intraoperative diagnosis in future studies.


  • Longitudinal, label-free, high-resolution imaging of glioblastoma spheroid response to therapy: a translational tool for preclinical evaluation of chemotherapy, radiation, and immunotherapy

    by Caroline E. Serafini, Amin Davarzani, Dan Cappabianca, Zhenmin Li, Daniela Baez Collazos, Deniz Mamaghani, Leidong Mao, Krishanu Saha, Lohitash Karumbaiah, Francisco E. Robles

    Glioblastoma (GBM) is an aggressive brain tumor with limited treatment options and poor patient survival, underscoring the need for novel, to our knowledge, therapeutic strategies and improved preclinical models. Patient-derived tumor spheroids (PDTSs) offer a physiologically relevant in vitro platform for evaluating treatments such as chimeric antigen receptor (CAR) T cell therapy, chemotherapy, and radiation. However, significant challenges remain in monitoring the complex three-dimensional (3D) microenvironment of the GBM PDTSs. Current imaging techniques used for this purpose are primarily endpoint analyses which lack critical real-time, non-invasive capabilities that ultimately preclude longitudinal and continuous monitoring. In this study, we introduce quantitative oblique back-illumination microscopy (qOBM) as a label-free and non-invasive imaging approach for longitudinal and continuous, high-resolution monitoring of GBM PDTSs during treatment. qOBM enables real-time visualization of cellular processes, including apoptosis, cell migration, and T cell-mediated cytotoxicity by leveraging tomographic refractive index-based quantitative imaging. We construct a compact qOBM system that fits within common incubators and apply it to study the effects of radiation, chemotherapy, and immunotherapy on three patient-derived GBM cell lines, extracting both static and dynamic image features over a 72 h treatment period. Additionally, we develop machine learning models to predict spheroid viability and cytotoxicity, demonstrating the potential of qOBM to enhance treatment evaluation. Our findings establish qOBM as a powerful tool for longitudinal and continuous spheroid monitoring, offering a non-destructive, high-resolution alternative to conventional endpoint assays and improving the evaluation of preclinical treatments for GBM.


  • Open Access

    Longitudinal and continuous label-free monitoring of glioblastoma patient-derived tumor spheroid treatment response using quantitative oblique back illumination microscopy

    by Caroline Serafini, Amin Davarzani, Dan Cappabianca, Zhenmin Li, Daniela Baez Collazos, Deniz Mamaghani, Leidong Mao, Krishanu Saha, Lohitash Karumbaiah, Francisco Robles

    Glioblastoma (GBM) is an aggressive brain tumor with limited treatment options and poor patient survival, underscoring the need for novel therapeutic strategies and improved preclinical models. Patient-derived tumor spheroids (PDTSs) offer a physiologically relevant in vitro platform for evaluating treatments such as chimeric antigen receptor (CAR) T cell therapy, chemotherapy, and radiation. However, significant challenges remain in monitoring the complex three-dimensional (3D) microenvironment of the GBM PDTSs. Current imaging techniques used for this purpose are primarily endpoint analyses which lack critical real-time, non-invasive capabilities that ultimately preclude longitudinal and continuous monitoring. In this study, we introduce quantitative Oblique Back-illumination Microscopy (qOBM) as a novel, label-free, and non-invasive imaging approach for longitudinal and continuous, high-resolution monitoring of GBM PDTSs during treatment. qOBM enables real-time visualization of cellular processes, including apoptosis, cell migration, and T cell-mediated cytotoxicity by leveraging tomographic refractive index-based quantitative imaging. We construct a compact qOBM system that fits within common incubators and apply it to study the effects of radiation, chemotherapy, and immunotherapy on three patient-derived GBM cell lines, extracting both static and dynamic image features over a 72-hour treatment period. Additionally, we develop machine learning models to predict spheroid viability and cytotoxicity, demonstrating the potential of qOBM to enhance treatment evaluation. Our findings establish qOBM as a powerful tool for longitudinal and continuous spheroid monitoring, offering a non-destructive, high-resolution alternative to conventional endpoint assays and improving the evaluation of preclinical treatments for GBM.


  • Open Access

    Label-Free In-Line Characterization of Immune Cell Culture using Quantitative Phase Imaging

    by Caroline E. Serafini, Viswanath Gorti, Paloma Casteleiro Costa, Aaron D. Silva Trenkle, Bharat Kanwar, Bryan Wang, Brian Wicker, Linda E. Kippner, Isaac LeCompte, Rui Qi Chen, et al.

    Cell therapies, including T cell immunotherapies, offer promising treatments for previously untreatable diseases, but their widespread use is hindered by challenges in monitoring therapeutic cells during culture -- impacting consistency, potency, and cost. This work demonstrates the use of quantitative phase imaging (QPI), specifically a compact, non-interferometric form called quantitative oblique back illumination microscopy (qOBM), for non-destructive, label-free, in-line assessment of T cell cultures. qOBM enables near real-time feedback on culture growth, contamination, and cell status (viability and activation), comparable to flow cytometry. We further apply this method to characterize genetically modified CAR T cells and explore its potential for advanced T cell phenotyping. Analysis of data from over 50 independent donors shows strong correlation between qOBM metrics and traditional destructive at-line assays. Overall, qOBM provides a powerful tool for continuous, in-line monitoring of therapeutic cell cultures, which can be transformative for improving reproducibility, reducing costs, and advancing the development of cell-based therapies.


  • Experimental assessment of the optical transfer function for quantitative oblique back illumination microscopy (qOBM)

    by Zhenmin Li, Paloma Casteleiro Costa, Caroline Serafini, Srinidhi Bharadwaj, Zhe Guang, Francisco E. Robles

    Quantitative oblique back illumination microscopy (qOBM) is a recently developed imaging technique that enables 3D quantitative phase imaging (QPI) and refractive index (RI) tomography of thick scattering samples. To quantify the phase and RI information with qOBM, the optical transfer function (OTF) of the system must be known or estimated, which requires knowledge of the angular distribution of light at an imaging plane inside a highly scattering medium. To date, this information has been estimated using a Monte Carlo photon transport method which relies on documented tissue scattering properties. While this numerical approach has shown high-fidelity quantitative results, it is limited by its dependence on published scattering parameters and simulated conditions. Here we propose a novel approach that allows experimental measurement of the angular distribution of the multiple-scattered light at the imaging plane inside a highly scattering medium. Experimental results using samples with known and unknown scattering properties are presented, including excised brain tissue, in-vivo skin, and formalin-fixed and paraffin-embedded (FFPE) tissues. Results further support qOBM’s quantitative fidelity across different tissue types, and show how directly measuring the angular distribution of light can widen qOBM’s utility to more complex samples with unknown or highly variable scattering properties.


  • Quantifying UV-induced photodamage for longitudinal live-cell imaging applications of deep-UV microscopy

    by Viswanath Gorti, Kaitlyn McCubbins, Daniel Houston, Aaron D. Silva Trenkle, Abigail Holberton, Caroline E. Serafini, Levi Wood, Gabriel Kwong, Francisco E. Robles

    Deep-UV microscopy enables high-resolution, label-free molecular imaging by leveraging biomolecular absorption properties in the UV spectrum. Recent advances in UV-imaging hardware have renewed interest in this technique for quantitative live cell imaging applications. However, UV-induced photodamage remains a concern for longitudinal dynamic imaging studies. Here, we quantify UV phototoxicity with several cell types at notable UV wavelengths. We find that the fluence required for cell death via UV phototoxicity with continuous UV exposure varies with cell type and wavelength from ∼0.5µJ/µm2 to 2µJ/µm2, but is independent of typical illumination power/radiant flux of UV microscopy (e.g., 0.1-20 nW/µm2). We also show results from fractionation studies that reveal cell repair following UV exposure, which increases the tolerance to UV radiation by a factor of 2 or more, depending on the fractionation paradigm. Results further show that UV tolerance exceeds ANSI guidelines for maximum permissible exposure. Finally, we calculate imaging limits for a typical application of UV microscopy, such as hematology analysis. Together, this work provides UV fluence thresholds that can serve as guidelines for nondestructive, longitudinal, and dynamic deep-UV microscopy experiments.


  • GAN-based quantitative oblique back-illumination microscopy enables computationally efficient epi-mode refractive index tomography

    by Zhenmin Li, Paloma Casteleiro Costa, Zhe Guang, Caroline Elizabeth Serafini, Francisco E. Robles, Caroline Filan

    Quantitative oblique back-illumination microscopy (qOBM) is a novel imaging technology that enables epi-mode 3D quantitative phase imaging and refractive index (RI) tomography of thick scattering samples. The technology uses four oblique back illumination images captured at the same focal plane and a fast 2D deconvolution reconstruction algorithm to reconstruct 2D phase cross-sections of thick samples. Alternatively, a through-focus z-stack of oblique back illumination images can be used to recover 3D RI tomograms with improved RI quantitative fidelity at the cost of a more computationally expensive reconstruction algorithm. Here, we report on a generative adversarial network (GAN) assisted approach to reconstruct 3D RI tomograms with qOBM that achieves high fidelity and greatly reduces processing time. The proposed approach achieves high-fidelity 3D RI tomography using differential phase contrast images from three adjacent z-planes. A ∼9-fold improvement in volumetric reconstruction time is achieved. We further show that this technique provides high SNR RI tomograms with high quantitative fidelity, reduces motion artifacts, and generalizes to different tissue types. This work can lead to real-time, high-fidelity RI tomographic imaging for in-vivo pre-clinical and clinical applications.