More than 300,000 patients in the United States undergo cardiac surgery each year, and substantial morbidity and mortality are common. Acute kidney injury(AKI) is one of the most frequent post-surgical complications, with estimates ranging from 20-50% depending on the procedure. AKI is strongly correlated with reduced quality of life, progression to renal-replacement therapy (i.e. dialysis),and mortality. The pathophysiology of AKI is complex and poorly understood, but likely includes hemodynamic changes (low cardiac output, blood pressure), mechanical factors (emboli, venous congestion), and other mechanisms (vasoconstriction, drugs, inflammation)—many of which are highly modifiable. One of the key challenges to developing preventative strategies for AKI is our inability to monitor intraoperative kidney function.

This overarching goal of this proposal is to develop an automated platform to continuously forecast the risk for AKI based on real-time intraoperative physiological data. This information would allow clinicians to optimize intraoperative variables on-the-fly to reduce the likelihood of kidney injury.

Radiation therapy is a clinically popular treatment for cancer, and knowledge-based radiation therapy treatment planning is becoming an integral part of clinical practice.However, there is considerable variation in clinical guidance—which identify the feasibility constraints for the plans—of treatment plans among different institutes, hospitals, and oncologists. In this proposed study, we develop a novel methodology that combines machine learning and inverse optimization to infer utility functions and hidden constraints of a black-box decision-making process. We will evaluate the method in the context of personalized radiation therapy treatment planning, building on a large dataset of historical radiation therapy treatment plans that were clinically approved and delivered. Using a data-driven approach, we apply our iterative approach of machine learning and inverse optimization on the radiation therapy plans to find the underlying criteria that the oncologists considered when approving the plans. Once these criteria are known, a forward optimization model can be used to design personalized treatment plans automatically.

Glaucoma is the second leading cause of blindness globally, with approximately 79.6 million people expected to be affected by the disease by 2020.  A small subset of glaucoma patients will experience rapid glaucoma worsening that, if untreated, can quickly lead to complete vision loss. Identifying these rapid progressors early is critical for ophthalmologists to perform interventions that can preserve a patient’s eyesight.

The project team will develop machine learning algorithms that can accurately identify patients at risk for undergoing rapid glaucoma progression, based on multimodal clinical data that is acquired as part of routine glaucoma care at Wilmer Eye Institute. To this end, the team plans to aggregate functional, structural and clinical information – automated visual field tests, optical coherence tomography images, and clinical parameters such as intraocular pressure – to better understand the factors that contribute to rapid progression.

Older adults are highly susceptible to postoperative complications following major surgery. A growing body of research shows that physical conditioning prior to surgery, often called “prehabilitation,” can provide significant benefit to patients, including reduced complications, shorter recoveries, and lower costs. However, tools to measure the effectiveness of prehabilitation are currently lacking.

Anton Dahbura and team are looking too wearable technology for new insights on the benefits of prehabilitation. His team has developed a platform that can collect data from wearable activity monitors and display feedback to patients on their smartphones. By continuously monitoring patient activity with wearable devices, the team hopes to apply analytics to determine the true relationship between pre-surgery activity levels and surgical outcomes.

Recent studies show that hand dexterity and strength recover mostly within the first three months after stroke, and that these two critical components of hand function are supported by separate biological systems. However, in most stroke patients, dexterous hand function does not fully recover with the standard rehabilitation therapy. These findings strongly suggest the need for intense rehabilitation targeting hand dexterity in the early post-stroke stages. Currently, the stroke rehabilitation field is missing effective tools to meet this need. Researchers from the Malone Center and the  Brain, Learning, Animation, and Movement Laboratory (BLAM) are conducting pilot studies on the Hand Articulation Neurotraining Device (HAND), a portable rehabilitation device for the hand, which can be used in various clinical settings, starting within the hospital, immediately after brain injury, and after discharge, in the patient’s home.

The Cardiac Surgery Division at the Johns Hopkins Hospital is a highly complex organization that involves multi-disciplinary services and care teams.  The Cardiac Surgery team is routinely monitoring their quality of care, forming and testing hypotheses as to which parts of the care system need to be more efficient. This project establishes the foundation to investigate and pinpoint inefficiencies in cardiac surgery care by analyzing comprehensive data sets taken from a large cohort of patients. Researchers and clinicians from the Malone Center and Cardiac Surgery team up to validate and investigate four potential inefficiencies in the Cardiac Surgery care system: (1) Prediction of length of stay, (2) Early diagnosis of Heparin Induced Thrombocytopenia (HIT), (3) Measuring deviation from routine daily pathways, and (4) Prediction of Bounce-back patients to ICU. The larger goal of the project is to explore new data-driven approaches to improve the quality and efficiency of care for cardiac surgery patients.

The team published a paper on the project in the European Heart Journal, October 2019:
Finding predictors and causes of cardiac surgery ICU readmission using machine learning and causal inference

The overall goal of the SpineCloud project is to gain understanding and predictive power of the factors underlying spine surgery outcomes – particularly the unacceptably broad variability associated with spinal fusion. Such capability will yield evidence-based insight and decision support to improve patient selection, enable more optimal surgical planning, identify perioperative sentinel factors, and help to guide the most effective post-operative, therapeutic, and rehabilitative process on a patient-specific basis. The SpineCloud project aims to curate a database consisting of patient demographic data, image and specific anatomy, surgical procedures, and pathologies. By developing this data-intensive approach to future spine surgeries, SpineCloud will provide more favorable and consistent outcomes for patients.