Quantifying the potential benefits of early detection for pancreatic cancer through a counterfactual simulation modeling analysis.

The benefits of cancer early detection depend on various factors, including cancer type, screening method performance, stage at diagnosis, and subsequent treatment. Although numerous studies have evaluated the effectiveness of screening interventions for identifying cancer at earlier stages, there is no quantitative analysis that studies the optimal early detection time interval that results in the greatest mortality benefit; such data could serve as a target and benchmark for cancer early detection strategies. In this study, we focus on pancreatic ductal adenocarcinoma (PDAC), a cancer known for its lack of early symptoms. Consequently, it is most often detected at late stages when the 5-year survival rate is only 3%. We developed a PDAC population model that simulates an individual patient's age and stage at diagnosis, while replicating overall US cancer incidence and mortality rates. The model includes "cancer sojourn time," serving as a proxy for the speed of cancer progression, with shorter times indicating rapid progression and longer times indicating slower progression. In our PDAC model, our hypothesis was that earlier cancer detection, potentially through a hypothetical screening intervention in the counterfactual analysis, would yield reduced mortality as compared to a no-screening group. We found that the benefits of early detection, such as increased life-years gained, are greater when the sojourn time is shorter, reaching their maximum when identification is made 4-6 years prior to clinical diagnosis (e.g., when a symptomatic diagnosis is made). However, when early detection occurs even earlier, for example 6-10 years prior to clinical diagnosis, the benefits significantly diminish for shorter sojourn time cancers, and level off for longer sojourn time cancers. Our study clarifies the potential benefits of PDAC early detection that explicitly incorporates individual patient heterogeneity in cancer progression and identifies quantitative benchmarks for future interventions.

Structured deep embedding model to generate composite clinical indices from electronic health records for early detection of pancreatic cancer.

The high-dimensionality, complexity, and irregularity of electronic health records (EHR) data create significant challenges for both simplified and comprehensive health assessments, prohibiting an efficient extraction of actionable insights by clinicians. If we can provide human decision-makers with a simplified set of interpretable composite indices (i.e., combining information about groups of related measures into single representative values), it will facilitate effective clinical decision-making. In this study, we built a structured deep embedding model aimed at reducing the dimensionality of the input variables by grouping related measurements as determined by domain experts (e.g., clinicians). Our results suggest that composite indices representing liver function may consistently be the most important factor in the early detection of pancreatic cancer (PC). We propose our model as a basis for leveraging deep learning toward developing composite indices from EHR for predicting health outcomes, including but not limited to various cancers, with clinically meaningful interpretations.

Machine Learning in Cardiac Surgery: Predicting Mortality and Readmission.

Predicting outcomes in open-heart surgery can be challenging. Unexpected readmissions, long hospital stays, and mortality have economic implications. In this study, we investigated machine learning (ML) performance in data visualization and predicting patient outcomes associated with open-heart surgery. We evaluated 8,947 patients who underwent cardiac surgery from April 2006 to January 2018. Data visualization and classification were performed at cohort-level and patient-level using clustering, correlation matrix, and seven different predictive models for predicting three outcomes ("Discharged," "Died," and "Readmitted") at binary level. Cross-validation was used to train and test each dataset with the application of hyperparameter optimization and data imputation techniques. Machine learning showed promising performance for predicting mortality (AUC 0.83 ± 0.03) and readmission (AUC 0.75 ± 0.035). The cohort-level analysis revealed that ML performance is comparable to the Society of Thoracic Surgeons (STS) risk model even with limited number of samples ( e.g. , less than 3,000 samples for ML versus more than 100,000 samples for the STS risk models). With all cases (8,947 samples, referred as patient-level analysis), ML showed comparable performance to what has been reported for the STS models. However, we acknowledge that it remains unknown at this stage as to how the model might perform outside the institution and does not in any way constitute a comparison of the performance of the internal model with the STS model. Our study demonstrates a systematic application of ML in analyzing and predicting outcomes after open-heart surgery. The predictive utility of ML in cardiac surgery and clinical implications of the results are highlighted.

Deep learning on time series laboratory test results from electronic health records for early detection of pancreatic cancer.

The multi-modal and unstructured nature of observational data in Electronic Health Records (EHR) is currently a significant obstacle for the application of machine learning towards risk stratification. In this study, we develop a deep learning framework for incorporating longitudinal clinical data from EHR to infer risk for pancreatic cancer (PC). This framework includes a novel training protocol, which enforces an emphasis on early detection by applying an independent Poisson-random mask on proximal-time measurements for each variable. Data fusion for irregular multivariate time-series features is enabled by a "grouped" neural network (GrpNN) architecture, which uses representation learning to generate a dimensionally reduced vector for each measurement set before making a final prediction. These models were evaluated using EHR data from Columbia University Irving Medical Center-New York Presbyterian Hospital. Our framework demonstrated better performance on early detection (AUROC 0.671, CI 95% 0.667 - 0.675, p < 0.001) at 12 months prior to diagnosis compared to a logistic regression, xgboost, and a feedforward neural network baseline. We demonstrate that our masking strategy results greater improvements at distal times prior to diagnosis, and that our GrpNN model improves generalizability by reducing overfitting relative to the feedforward baseline. The results were consistent across reported race. Our proposed algorithm is potentially generalizable to other diseases including but not limited to cancer where early detection can improve survival.

Immunogenetics of gastrointestinal cancers: A systematic review and retrospective survey of inborn errors of immunity in humans.

BACKGROUND AND AIM: Humans with inborn errors of immunity (IEI), or primary immunodeficiencies, may be associated with a potential risk factor for early-onset gastrointestinal (GI) cancer. METHODS: We systematically reviewed all cases with clinical diagnoses of both an IEI and a GI cancer in three databases (MEDLINE, SCOPUS, and EMBASE). In total, 76 publications satisfying our inclusion criteria were identified, and data for 149 cases were analyzed. We also searched our institutional cancer registry for such cases. RESULTS: We identified 149 patients with both an IEI and a GI cancer, 95 presented gastric cancer, 13 small bowel cancer, 35 colorectal cancer, and 6 had an unspecified cancer or cancer at another site. Gastric and colon adenocarcinomas were the most common. For both gastric and colorectal cancers, age at onset was significantly earlier in patients with IEIs than in the general population, based on the SEER database. Common variable immunodeficiency (CVID) was the most common IEI associated with gastrointestinal cancer. About 12% of patients had molecular genetic diagnoses, the three most frequently implicated genes being ATM, CARMIL2, and CTLA4. Impaired humoral immunity and Epstein-Barr virus (EBV) infection were frequently reported as factors potentially underlying early-onset GI cancers in patients with IEIs. We identified one patient with CVID and early-onset gastric adenocarcinoma, recurrent diarrhea, and gastrointestinal CMV infection from a retrospective survey. CONCLUSION: Patients with IEIs should be considered at risk of early-onset GI cancers and should therefore undergo cancer screening at an earlier age.

Cost-effectiveness of neoadjuvant FOLFIRINOX versus gemcitabine plus nab-paclitaxel in borderline resectable/locally advanced pancreatic cancer patients.

BACKGROUND: The 2020 National Comprehensive Cancer Network guidelines recommend neoadjuvant FOLFIRINOX or neoadjuvant gemcitabine plus nab-paclitaxel (G-nP) for borderline resectable/locally advanced pancreatic ductal adenocarcinoma (BR/LA PDAC). AIM: The purpose of our study was to compare treatment outcomes, toxicity profiles, costs, and quality-of-life measures between these two treatments to further inform clinical decision-making. METHODS AND RESULTS: We developed a decision-analytic mathematical model to compare the total cost and health outcomes of neoadjuvant FOLFIRINOX against G-nP over 12 years. The model inputs were estimated using clinical trial data and published literature. The primary endpoint was incremental cost-effectiveness ratios (ICERs) with a willingness-to-pay threshold of $100 000 per quality-adjusted-life-year (QALY). Secondary endpoints included overall (OS) and progression-free survival (PFS), total cost of care, QALYs, PDAC resection rate, and monthly treatment-related adverse events (TRAE) costs (USD). FOLFIRINOX was the cost-effective strategy, with an ICER of $60856.47 per QALY when compared to G-nP. G-nP had an ICER of $44639.71 per QALY when compared to natural history. For clinical outcomes, more patients underwent an "R0" resection with FOLFIRINOX compared to G-nP (84.9 vs. 81.0%), but FOLFIRINOX had higher TRAE costs than G-nP ($10905.19 vs. $4894.11). A one-way sensitivity analysis found that the ICER of FOLFIRINOX exceeded the threshold when TRAE costs were higher or PDAC recurrence rates were lower. CONCLUSION: Our modeling analysis suggests that FOLFIRNOX is the cost-effective treatment compared to G-nP for BR/LA PDAC despite having a higher cost of total care due to TRAE costs. Trial data with sufficient follow-up are needed to confirm our findings.

Impact of Preoperative Lymphopenia on Survival Following Left Ventricular Assist Device Placement.

Lymphopenia has been implicated in poor outcomes in the heart failure population. However, the prognostic implication of lymphopenia in left ventricular assist device (LVAD) patients is unknown. We examine the impact of lymphopenia on all-cause mortality in this population over a 24-month period post-implantation. A total of 170 patients between June 2011 and July 2018 receiving permanent durable LVAD at a single center formed the study population. Criteria for lymphopenia on admission, defined as an absolute lymphocyte count (ALC) <1500 cells/µl, was met in 99 patients. A total of 11 patients were excluded: two with ALC >4800/µl and nine with incomplete data. Survival across groups was compared with a Kaplan-Meier plot and log-rank statistics. The Cox proportional hazard model was used to examine the association between lymphopenia and 24-month all-cause mortality. In the lymphopenia group, mean ALC was 909.6 ± 331.9 versus 2073.6 ± 501.1 in the non-lymphopenic group. Twenty-four-month all-cause mortality was significantly higher in the lymphopenia group (p = 0.009). The lymphopenic patients had worse unadjusted (hazard ratio [HR] = 2.14, confidence interval [CI] = 1.19-3.82; p = 0.01) and adjusted survival (HR = 2.07, CI = 1.13-3.79; p = 0.02). Further clinical investigations are required to assess the utility of continued clinical monitoring of ALC levels beyond LVAD placement.

Biologically Inspired, Open, Helicoid Impeller Design for Mechanical Circulatory Assist.

Rotating impeller actuated by electromagnet has been a key technological innovation which surpassed earlier limitations of pulsatile pumps. Current impeller design, however, is alien to the functional unit of the human circulatory system and remains a potential cause of adverse prothrombotic events such as hemolysis or pump thrombosis by forcing blood cells to pass over a narrow space available within the rapidly alternating blades attached along its central hub, creating fundamentally a nonphysiologic flow, especially for miniaturized percutaneous blood pumps. Here, we present a biologically inspired, open, helicoid (BiO-H) impeller design for a circulatory assist device that has a fundamentally different footprint from the conventional Archimedean screw-based impeller designs by implementing new design features inspired by an avian right atrioventricular valve. Design parameters including an inner diameter, helix height, overall height, helix revolutions/pitch, blade length, blade thickness, introductory blade angle, number of blades, and blade shape were optimized for maximum output volumetric flow rate through the parametric analysis in computational fluid dynamics simulation. BiO-H shows an improved flow path with 2.25-fold less cross-sectional area loss than the conventional impeller designs. BiO-H with a diameter of 15 mm resulted in a maximum flow rate of 25 L/min at 15,000 revolutions per minute in simulation and showed further improved pressure-flow relationship in benchtop experiments. The design shows promise in increasing flow and could serve as a new impeller design for future blood pumps.

Lab-on-a-CD Platform for Generating Multicellular Three-dimensional Spheroids.

A three-dimensional spheroid cell culture can obtain more useful results in cell experiments because it can better simulate cell microenvironments of the living body than two-dimensional cell culture. In this study, we fabricated an electrical motor-driven lab-on-a-CD (compact disc) platform, called a centrifugal microfluidic-based spheroid (CMS) culture system, to create three-dimensional (3D) cell spheroids implementing high centrifugal force. This device can vary rotation speeds to generate gravity conditions from 1 x g to 521 x g. The CMS system is 6 cm in diameter, has one hundred 400 µm microwells, and is made by molding with polydimethylsiloxane in a polycarbonate mold premade by a computer numerical control machine. A barrier wall at the channel entrance of the CMS system uses centrifugal force to spread cells evenly inside the chip. At the end of the channel, there is a slide region that allows the cells to enter the microwells. As a demonstration, spheroids were generated by monoculture and coculture of human adipose-derived stem cells and human lung fibroblasts under high gravity conditions using the system. The CMS system used a simple operation scheme to produce coculture spheroids of various structures of concentric, Janus, and sandwich. The CMS system will be useful in cell biology and tissue engineering studies that require spheroids and organoid culture of single or multiple cell types.

Mathematical Blueprint of a Mitral Valve.

Mathematical modeling tries to simplify understanding and proposes a fundamental mechanism that governs the motion and function of a complex biological system such as a mitral valve (MV) motion which represents a dynamic interplay between papillary muscle (PM) position in the context of left ventricular (LV) shape dynamics. Current therapeutic strategies to intervene on the MV may not have exploited these relationships due to lack of understanding of the interactions. We present a MV 3D mathematical model characterized by LV shape dynamics to understand fundamental working principles of ventriculo-papillary-mitral complex. A complex 3D functional unit of MV apparatus was mathematically modeled based on a principle of dynamics. The model comprises of primary components including the annulus, anterior leaflet, posterior leaflet, chordae tendineae, anterior and posterior PM, and LV wall based on normal anatomical reference values from published series. Simulations based on Carpentier's classification of MV disease were created as well as based on LV shape dynamics and presented graphically. Autodesk Inventor (Autodesk Inc., San Rafael, CA) and Matlab (Mathworks, Natick, MA) were used for modeling and analysis. A stepwise analysis and mathematical models of the annulus, leaflets, chords, PMs, and LV were obtained by combining finite element analysis and computerized model creations. The model was then applied to Carpentier's functional classification. PM positions extrapolated based on different LV deformation in normal and mitral regurgitation (MR) model resulted in a different degree of MV leaflet coaptation with regurgitation (presented numerically and graphically). Abnormal MV coaptation was amended by manipulating PM positions independent with LV size or shape deformation, demonstrating that PM positioning maneuver may improve leaflet coaptation. LV dilation combined with increased interpapillary muscle distance turned out to intensify the level of leaflet prolapse, creating even greater regurgitation volume. Our mathematical model may provide a clue to complex interactions in play within a mitral, papillary, and LV complex. The model offers a possibility of manipulating various variables to obtain the desired outcome.

Wave Intensity Analysis of Right Ventricular Function during Pulsed Operation of Rotary Left Ventricular Assist Devices.

Changing the speed of left ventricular assist devices (LVADs) cyclically may be useful to restore aortic pulsatility; however, the effects of this pulsation on right ventricular (RV) function are unknown. This study investigates the effects of direct ventricular interaction by quantifying the amount of wave energy created by RV contraction when axial and centrifugal LVADs are used to assist the left ventricle. In 4 anesthetized pigs, pressure and flow were measured in the main pulmonary artery and wave intensity analysis was used to identify and quantify the energy of waves created by the RV. The axial pump depressed the intensity of waves created by RV contraction compared with the centrifugal pump. In both pump designs, there were only minor and variable differences between the continuous and pulsed operation on RV function. The axial pump causes the RV to contract with less energy compared with a centrifugal design. Diminishing the ability of the RV to produce less energy translates to less pressure and flow produced, which may lead to LVAD-induced RV failure. The effects of pulsed LVAD operation on the RV appear to be minimal during acute observation of healthy hearts. Further study is necessary to uncover the effects of other modes of speed modulation with healthy and unhealthy hearts to determine if pulsed operation will benefit patients by reducing LVAD complications.

Quantification of Pulsed Operation of Rotary Left Ventricular Assist Devices with Wave Intensity Analysis.

The current generation of left ventricular assist devices (LVADs) provides continuous flow and has the capacity to reduce aortic pulsatility, which may be related to a range of complications associated with these devices. Pulsed LVAD operation using speed modulation presents a mechanism to restore aortic pulsatility and potentially mitigate complications. We sought to investigate the interaction of axial and centrifugal LVADs with the LV and quantify the effects of continuous and pulsed LVAD operations on LV generated wave patterns under different physiologic conditions using wave intensity analysis (WIA) method. The axial LVAD created greater wave intensity associated with LV relaxation. In both LVADs, there were only minor and variable differences between the continuous and pulsed operations. The response to physiologic stress was preserved with LVAD implantation as wave intensity increased marginally with volume loading and significantly with infusion of norepinephrine. Our findings and a new approach of investigating aortic wave patterns based on WIA are expected to provide useful clinical insights to determine the ideal operation of LVADs.

Electrical power to run ventricular assist devices using the Free-range Resonant Electrical Energy Delivery system.

BACKGROUND: Models of power delivery within an intact organism have been limited to ionizing radiation and, to some extent, sound and magnetic waves for diagnostic purposes. Traditional electrical power delivery within the intact human body relies on implanted batteries that limit the amount and duration of delivered power. The efficiency of current battery technology limits the substantial demands required, such as continuous operation of an implantable artificial heart pump within a human body. METHODS: The fully implantable, miniaturized, Free-range Resonant Electrical Energy Delivery (FREE-D) system, compatible with any type of ventricular assist device (VAD), has been tested in a swine model (HVAD) for up to 3 hours. Key features of the system, the use of high-quality factor (Q) resonators together with an automatic tuning scheme, were tested over an extended operating range. Temperature changes of implanted components were measured to address safety and regulatory concerns of the FREE-D system in terms of specific absorption rate (SAR). RESULTS: Dynamic power delivery using the adaptive tuning technique kept the system operating at maximum efficiency, dramatically increasing the wireless power transfer within a 1-meter diameter. Temperature rise in the FREE-D system never exceeded the maximum allowable temperature deviation of 2°C (but remained below body temperature) for an implanted device within the trunk of the body at 10 cm (25% efficiency) and 50 cm (20% efficiency), with no failure episodes. CONCLUSIONS: The large operating range of FREE-D system extends the use of VAD for nearly all patients without being affected by the depth of the implanted pump. Our in-vivo results with the FREE-D system may offer a new perspective on quality of life for patients supported by implanted device.

Design and Development of a Miniaturized Percutaneously Deployable Wireless Left Ventricular Assist Device: Early Prototypes and Feasibility Testing.

The current left ventricular assist devices (LVADs) are limited by a highly invasive implantation procedure in a severely unstable group of advanced heart failure patients. Additionally, the current transcutaneous power drive line acts as a nidus for infection resulting in significant morbidity and mortality. In an effort to decrease this invasiveness and eliminate drive line complications, we have conceived a wireless miniaturized percutaneous LVAD, capable of being delivered endovascularly with a tether-free operation. The system obviates the need for a transcutaneous fluid purge line required in existing temporary devices by utilizing an incorporated magnetically coupled impeller for a complete seal. The objective of this article was to demonstrate early development and proof-of-concept feasibility testing to serve as the groundwork for future formalized device development. Five early prototypes were designed and constructed to iteratively minimize the pump size and improve fluid dynamic performance. Various magnetic coupling configurations were tested. Using SolidWorks and ANSYS software for modeling and simulation, several geometric parameters were varied. HQ curves were constructed from preliminary in vitro testing to characterize the pump performance. Bench top tests showed no-slip magnetic coupling of the impeller to the driveshaft up to the current limit of the motor. The pump power requirements were tested in vitro and were within the appropriate range for powering via a wireless energy transfer system. Our results demonstrate the proof-of-concept feasibility of a novel endovascular cardiac assist device with the potential to eventually offer patients an untethered, minimally invasive support.

Hypergravity-induced multicellular spheroid generation with different morphological patterns precisely controlled on a centrifugal microfluidic platform.

In living tissue, cells exist in three-dimensional (3D) microenvironments with intricate cell-cell interactions. To model these cellular environments, numerous techniques for generating cell spheroids have been proposed and improved. However, previously reported methods still have limitations in uniformity, reproducibility, scalability, throughput, etc. Here, we present a centrifugal microfluidic-based spheroid (CMS) formation method for generating both co-culture and mono-culture 3D spheroids in a highly controlled manner. We designed circularly arrayed microwells to allow the even distribution of cells introduced at the center of a rotating platform and to provide identical hypergravity conditions at each well by the centrifugal forces generated. Compared with conventional well plate-based spheroid formation, the CMS formation method significantly promotes sphericity and consistency in both size and shape with high production yields. In addition to mono-culture spheroids, we successfully generated co-culture spheroids in concentric, Janus, and sandwich shapes using human adipose-derived stem cells and human lung fibroblasts, demonstrating the versatility of our CMS formation method. We believe that our new method for generating 3D spheroids will become one of the essential technologies in the field of 3D cell culture. We also expect that we are providing an innovative means to assess cellular responses, including cell motility under different hypergravity conditions.

Computational simulation of passive leg-raising effects on hemodynamics during cardiopulmonary resuscitation.

BACKGROUND AND OBJECTIVE: The passive leg-raising (PLR) maneuver has been used for patients with circulatory failure to improve hemodynamic responsiveness by increasing cardiac output, which should also be beneficial and may exert synergetic effects during cardiopulmonary resuscitation (CPR). However, the impact of the PLR maneuver on CPR remains unclear due to difficulties in monitoring cardiac output in real-time during CPR and a lack of clinical evidence. METHODS: We developed a computational model that couples hemodynamic behavior during standard CPR and the PLR maneuver, and simulated the model by applying different angles of leg raising from 0° to 90° and compression rates from 80/min to 160/min. RESULTS: The simulation results showed that the PLR maneuver during CPR significantly improves cardiac output (CO), systemic perfusion pressure (SPP) and coronary perfusion pressure (CPP) by ∼40-65% particularly under the recommended range of compression rates between 100/min and 120/min with 45° of leg raise, compared to standard CPR. However, such effects start to wane with further leg lifts, indicating the existence of an optimal angle of leg raise for each person to achieve the best hemodynamic responses. CONCLUSIONS: We developed a CPR-PLR model and demonstrated the effects of PLR on hemodynamics by investigating changes in CO, SPP, and CPP under different compression rates and angles of leg raising. Our computational model will facilitate study of PLR effects during CPR and the development of an advanced model combined with circulatory disorders, which will be a valuable asset for further studies.

Investigation of Key Circuit Constituents Affecting Drug Sequestration During Extracorporeal Membrane Oxygenation Treatment.

We quantified the influence of the elements of the extracorporeal oxygenation (ECMO) circuit on drug sequestration by focusing on the interactions between materials and drugs. Tubing of three different brands (Tygon/Maquet/Terumo) and oxygenators of two different brands (Maquet/Terumo) were used. Drugs included dexmedetomidine, meropenem, and heparin, which were dissolved in deionized water. Tubing was cut into approximately 7 cm sections and allowed drug solutions enclosed inside by clamping both ends. The oxygenator housing, gas membrane, and heat exchanger were dissected into approximately 1 g pieces and submerged into drug solutions. The experimental samples were then immersed in a water bath at 37°C for 1, 6, 12, and 24 h. After 24 h, the dexmedetomidine concentration was significantly reduced in all three types of tubing (<30.1%), the oxygenator heat exchanger from Maquet Inc. (41.8%), and the gas exchanger from Terumo Inc. (8.6%), while no significant losses were found for meropenem and heparin compared with the control group. The heparin concentration within the Maquet gas exchanger, on the contrary, increased significantly compared with the control group at 1 and 12 h (p < 0.05). Our in vitro study reveals that material selection is a vital part of ECMO development.

Manikin-integrated digital measuring system for assessment of infant cardiopulmonary resuscitation techniques.

The process of cardiopulmonary resuscitation (CPR) involves various components that must be followed to deliver high quality of CPR. While the components commonly apply to CPR for all ages from infant to adult, there are several different suggestions for infant CPR such as two-thumb CPR and two-finger CPR. However, the comprehensive evaluation based on all these components has been difficult in the absence of proper evaluation tool. Here, we developed a new manikin-integrated, digital measuring system that objectively estimates overall performance of infant CPR by evaluating individual CPR components one by one including different hand placements. The system collects and analyzes data to present estimations in digital scores according to a new evaluation index constructed based on the previously verified one. The feasibility of the system was validated through simulations with beginners and experts in first aid, resulting in statistically significant differences between the two groups with the indication of specific weaknesses for each group which may provide a basis for creating customized CPR training strategy in compliance with the personal level. We believe that the system would become a valuable assessment tool not only for infant CPR but also for the CPR technique, in general, by reflecting every component in the evaluation.