Past Events

Date: 6/16/26

Address: 500 West 120th Street. New York, NY

Presenter: Christine O'Brien (Washington University in St. Louis)

Moderator: Christine Hendon and John Kymissis

Topic: Sensors

Abstract/Description:

"Postpartum hemorrhage (PPH) is the leading cause of maternal mortality globally, with the majority of PPH deaths occurring in low- and middle-income countries. Importantly, PPH has been noted as the most preventable cause of maternal mortality, and the leading factors causing preventable PPH are delays in diagnosis and treatment. This is especially critical in rural and low resource settings that often have low blood stores for transfusion, relying primarily on early pharmacologic treatment. There is an urgent need for a technology that can monitor occult blood loss, which would otherwise go undetected, in real-time to detect PPH in early stages. To this end, our lab has developed a compact, low-cost, wrist-worn wearable device that continuously monitors perfusion via laser speckle flow index (LSFI) to detect early signs of blood loss. The system was designed as a wireless, battery-powered reflectance mode configuration using off-the-shelf components. The device is Bluetooth enabled and can process and transmit LSFI signal to a receiving device for near real-time data visualization. Preliminary data in a swine hemorrhage model demonstrated that the system was sensitive to blood loss volumes of less than 5% with a highly linear response to blood loss volume (correlation coefficient of -0.94). To quantify blood loss, we moved to mining LSFI cardiac waveform features and preliminary data in swine show blood loss volume prediction errors of less than 100 mL over a 1000 mL blood loss range. These promising results motivate further development of hardware, which is being both simplified and expanded to also include a wired version of the device that hosts greater computational power and longer patient monitoring duration. Additionally, generalization of our data processing and analysis pipeline increases the robustness of our feature extraction algorithms and expansion into additional cardiovascular parameters. Finally, device adaptations informed by patient and healthcare provider interviews from a Nigerian hospital setting will be discussed.

Bio: Christine O'Brien is an Assistant Professor of Biomedical Engineering at Washington University in St. Louis with a secondary appointment in the department of Obstetrics & Gynecology. Her research is focused on developing and translating non-invasive optical spectroscopy and imaging tools to solve global problems in maternal-fetal health. She obtained her PhD in Biomedical Engineering at Vanderbilt University and completed postdoctoral training at Washington University School of Medicine in the Department of Radiology. She launched her independent research program with projects focused on the development of novel wearable sensors for the early detection of postpartum hemorrhage and novel strategies for preterm birth detection and investigation. She has a startup company, Armor Medical Inc, that is working to commercialize the postpartum hemorrhage sensor technology that has been developed in her lab."

 

Date: 5/8/26

Address: 500 West 120th Street. New York, NY

Presenter: Dr. Diana Mojahed (LightFinder)

Moderator: Christine Hendon

Topic: Sensors 

Abstract/Description:

High-performance optical sensing has long been trapped on the optics bench - constrained by the size, fragility, and cost of traditional instrumentation. What if you could put that capability on a chip? Just as the Apple Watch and WHOOP brought health monitoring from the clinic to your wrist, integrated photonics is bringing optical sensing from the lab to the field—unlocking a new paradigm for industries ranging from telecommunications and manufacturing to energy and healthcare.

In this talk, I will share the technical journey from Columbia SEAS to founding Lightfinder, an MIT spinout now growing in Kendall Square. The foundation began at Columbia, developing an ultrahigh-speed optical coherence tomography (OCT) system. Applying this system to tissue pathology prompted a longer-term look at where the industry would be in five to ten years. The next era of optical sensing would not be confined to a benchtop—it would be on a chip. A subsequent collaboration on a millimeter-scale Si₃N₄ supercontinuum source demonstrated that integrated photonics could compete with its benchtop counterpart while maintaining, if not exceeding, performance.

While the light source could be miniaturized on-chip, analyzing the light still relied on bulky, traditional optics. Miniaturization of the spectrometer was addressed during postdoctoral research at MIT. The digital Fourier transform (dFT) spectrometer is an ultrabroadband, ultrahigh-resolution spectrometer on a commercial silicon photonic platform, delivering 0.05 nm resolution across 380 nm of bandwidth (1260-1680nm).

Today, this technology underpins Lightfinder's mission to commercialize ultrahigh-performance chip-scale optical sensing systems. The talk will cover the core technology development, the journey of bringing innovation from lab to market, and how the next generation of engineers can join in building the future of optical sensing.

Bio: Dr. Diana Mojahed is the CEO of Lightfinder, an MIT spinout pioneering integrated photonics to deliver compact, high-performance spectroscopy and imaging across industries. Mojahed earned her Ph.D. in biomedical engineering from Columbia University and completed postdoctoral training in MIT’s Department of Materials Science & Engineering. At MIT, she was named a Kavanaugh Fellow and received an award at the MIT $100K Entrepreneurship Competition finals. Mojahed is a 2025 Activate Fellow and serves as an ambassador for Optica, the scientific society for photonics.

Date: 1/30/26

Address: Casa Italiana, Columbia University, 1161 Amsterdam Ave, New York, NY 10027

Abstract/Description:

Hosted by Columbia Engineering, this symposium was held to showcase advancements at the intersection of research, engineering, and medicine with a focus on breakthroughs in diagnostic and therapeutic solutions that enhance gynecologic, pregnancy, labor, and pelvic floor heatth; bringing together researchers, industry leaders, and collaborators for a fulll-day event featuring panels, several keynote addresses, and a student poster session.

Date: 12/17/25

Address: 500 West 120th Street. New York, NY

Presenter: Jessica Ramella-Roman, Ph.D. Professor (FIU)

Moderator: Christine Hendon and Kristin Myers

Topic: Sensors 

Abstract/Description:

Cervical remodeling is a critical biomechanical process that facilitates parturition, yet the structural dynamics underlying this phenomenon remain poorly understood. This study employs polarization-sensitive imaging modalities, including polarization-resolved Second Harmonic Generation (p-SHG) microscopy, to quantitatively analyze the reorganization of collagen fibers within the cervical extracellular matrix throughout murine pregnancy. By integrating polarization-sensitive detection, these imaging approaches enable enhanced contrast and specificity for assessing the anisotropic properties of collagen, providing sub-micron resolution and precise mapping of collagen fiber orientation and polarization signatures across the entire cervix.

Early gestational stages reveal a highly organized, circumferential arrangement of collagen fibers concentrated around the cervical os. As pregnancy advances, significant disruption in collagen alignment is observed, accompanied by a progressive loss of structural order. Polarization analysis further elucidates changes in collagen birefringence and molecular organization, offering insights beyond conventional intensity-based imaging. Statistical analyses corroborate these findings, demonstrating a strong correlation between the degradation of collagen organization and the reduction in mechanical stiffness necessary for cervical dilation during childbirth. This study provides a comprehensive characterization of collagen remodeling through polarization-sensitive optical methods and highlights their implications for understanding the biomechanics of both normal and pathological labor.

Date: 4/18/25

Presenter: Christine O'Brien (Washington University in St. Louis)

Moderator: Christine Hendon
 

 

Date: 3/28/25

Presenter: Matthew Bersi (Washington University in St. Louis)

Moderator: Sara Roccabianca

Abstract/Description:

The uterus is a dynamic organ that experiences large amounts of tissue remodeling to support fetal growth and gestation throughout pregnancy. Following parturition, the uterus undergoes a programmed phase of reverse remodeling that serves to contract the tissue back to its pre-pregnancy size and shape. This process, known as postpartum involution, is unique to the female reproductive system and is important for decreasing the likelihood of birth-induced complications, such as postpartum hemorrhage. Involution also returns the tissues of the pelvic floor back to a state that can accommodate subsequent pregnancies. While involution is an essential aspect of pregnancy and delivery, the mechanobiology that drives reverse remodeling during postpartum involution is not well understood. My group’s research is focused on soft tissue biomechanics and mechanobiology with a specific interest in immune-mediated mechanisms of extracellular matrix remodeling. In this talk, I will discuss recent and ongoing work focused on characterizing the relationship between biomechanics, inflammation, and uterine remodeling throughout postpartum involution in the mouse. I will also discuss our initial efforts to investigate murine uterine wound healing with a specific focus on Cesarean delivery and scar integrity in subsequent pregnancies. Using techniques ranging from small animal surgical models to tissue mechanical testing to cellular and molecular biology, we seek to identify the mechanobiological and immunological factors that contribute to postpartum involution. Insights into such processes can serve as a first step toward biologically-focused strategies to improve healing and limit postpartum complications and may also offer unique insights into the reversal of extracellular matrix remodeling that can be translated to physiological systems throughout the body.

Date: 3/13/25

Presenter: Sara Roccabianca (Washington University in St. Louis)

Moderator: Kristin Myers

Abstract/Description:

Biological sex has an impact on various aspects of soft tissue biomechanics, physiology, and function, in both health and disease. These differences can influence disease progression and treatment, thereby affecting patients' health and well-being. However, the ways in which these biological characteristics affect soft tissue biomechanics remain largely unknown. Dr. Roccabianca’s group was the first to identify significant effects of endogenous and exogenous sex hormones on the three-dimensional mechanical behavior of the murine urinary bladder as well as on the biomechanical behavior of peripheral resistance arteries. This work is pioneering research in this field, reinforcing the combined efforts of the scientific community to emphasize the importance of including sex as a biological variable in all areas of biomechanical research.

Date: 3/6/25

Presenter: Shalmali Joshi (Columbia)

Moderator: Christine Hendon

Topic: Multimodal AI Integration

Abstract/Description:

Multimodal AI models are advancing many application areas, especially scalable image generation. However, most multimodal models assume access to paired data. However, this assumption typically does not hold in clinical data, which consists of high-dimensional sparse samples. In this talk, we propose a general framework to reason about the generalizability of multimodal machine learning models in healthcare. We propose a scalable partial information decomposition method to evaluate when combining modalities can augment predictive AI performance and discuss methods for improving evaluation of multimodal AI performance in health and biomedicine. We will demonstrate initial results in developing diagnostic models for structural heart disease from ECG and chest radiography. 

Date: 2/27/25

Presenter: Tal Korem (Columbia)

Moderator: Kristin Myers

Abstract/Description:

The vaginal microbiome has been implicated in a variety of adverse pregnancy outcomes, such as preterm birth and preterm premature rupture of membranes, but our understanding of its role in pregnancy is still incomplete. I will share findings from our studies of this important ecosystem, in which we used untargeted metabolomics to discover strong associations between vaginal xenobiotics and prematurity; performed high-resolution genetic analyses to find changes in microbial sub-species diversity that are related to the risk of prematurity; and combined metagenomic sequencing with host immune-factor profiling to detect a strong association between the early-pregnancy vaginal molecular profile and development of severe preeclampsia. Overall, our studies demonstrate the potential of high-resolution multi-modal molecular profiling and analyses to illuminate new aspects of host-microbiome interactions during pregnancy, with implications for new diagnostic modalities.

Date: 2/13/25

Presenter: Noémie Elhadad (Columbia)

Moderator: Kristin Myers

Topic: Multimodal AI Integration

Date: 2/7/25

Presenter: Adrian Buganza Tepole (Columbia)

Moderator: Kristin Myers

Topic: Data-informed, multi-scale Digital Twins

Abstract/Description:

This talk focuses on our recent efforts to develop mechanistic models and leverage machine learning (ML) and data-driven approaches to increase our fundamental understanding of soft tissue, its unique ability to adapt to mechanical cues, and to create personalized treatment strategies based on digital twin frameworks. We explore two complex phenomena, breast wound healing and skin growth in tissue expansion, which are central to the successful reconstruction of the breast following breast cancer treatment. Breast cancer is the most common cancer in women and affects approximately one in eight women over their lifetime. The standard-of-care surgical methods for breast cancer treatment are lumpectomy (or breast-conversing surgery) and mastectomy (or total breast removal). Lumpectomy leads to wound healing which can cause significant breast deformity, while mastectomy requires breast reconstruction, often based on tissue expansion. For the first part of the talk, focused on wound healing, the emphasis is on new numerical methods and analysis tools to couple changes in tissue mechanics to dynamical regulatory systems of cell and cell-cell interactions. The second part of the talk tackles skin growth in tissue expansion, a technique that grows new skin in response to sustained supra-physiological loading. We have created computational models that combine mechanics and mechanobiology to describe the deformation and growth of expanded skin. Together with experiments on a porcine model, and leveraging ML tools such as multi-fidelity Gaussian processes, we have performed Bayesian inference to learn mechanistically how skin grows in response to stretch. These models, together with the experimental data they are built on, are key for the improvement of reconstructive procedures in breast cancer survivors, a substantial burden to the US healthcare system and globally.

Date: 1/31/25

Presenter: Sam Sia (Columbia)

Moderator: Christine Hendon

Topic: Sensors

Abstract/Description:

Dr. Sia will discuss two technologies for women's health. First, a rapid 15-min test in OB/GYN clinics that can diagnose vaginitis, STI's, and UTI's, allowing for the leading indications for a women to visit an OB/GYN clinic to receive results and timely treatment within a single visit. Second, a wearable blood pressure monitor that can help women manage pregnancy complications, menopause, and other conditions leading to hypertension, recognizing that women account for ~52% of deaths from high blood pressure. 

Date: 1/24/25

Presenter: John Kymissis (Columbia)

Moderator: Christine Hendon

Topic: Sensors