Decentralized digital twins of complex dynamical systems.

In this article, we introduce a decentralized digital twin (DDT) modeling framework and its potential applications in computational science and engineering. The DDT methodology is based on the idea of federated learning, a subfield of machine learning that promotes knowledge exchange without disclosing actual data. Clients can learn an aggregated model cooperatively using this method while maintaining complete client-specific training data. We use a variety of dynamical systems, which are frequently used as prototypes for simulating complex transport processes in spatiotemporal systems, to show the viability of the DDT framework. Our findings suggest that constructing highly accurate decentralized digital twins in complex nonlinear spatiotemporal systems may be made possible by federated machine learning.

A Real-World Study of Safety, Immunogenicity and Efficacy of Bevacizumab in Patients With Solid Malignancies: A Phase IV, Post-Marketing Study in India.

Objective: The aim of this study was to evaluate the post-marketing safety, tolerability, immunogenicity and efficacy of Bevacizumab (manufactured by Hetero Biopharma) in a broader population of patients with solid tumors. Patients And Methods: This phase IV, prospective, multi-centric clinical study was carried out in Indian patients with solid malignancies (metastatic colorectal cancer, non-squamous non-small-cell lung cancer, metastatic renal cell carcinoma) treated with Bevacizumab between April 2018 and July 2019. This study included 203 patients from 16 tertiary care oncology centers across India for safety assessment, of which a subset of 115 patients who have consented were also evaluated for efficacy and immunogenicity. This study was prospectively registered in the Clinical Trial Registry of India (CTRI), and was commenced only after receiving approval from the competent authority (Central Drugs Standard Control Organization, CDSCO). Results: Out of the 203 enrolled patients, 121 (59.6%) patients reported 338 adverse events (AEs) during this study. Of 338 reported AEs, 14 serious adverse events (SAEs) were reported by 13 patients including 6 fatal SAEs, assessed as unrelated to the study medication and 7 non-fatal SAEs, 5 assessed as related, and 3 unrelated to Bevacizumab. Most AEs reported in this study (33.9%) were general disorders and administration site conditions, followed by gastrointestinal disorders (29.1%). The most frequently reported AEs were diarrhea (11.3%), asthenia (10.3%), headache (8.9%), pain (7.4%), vomiting (7.9%), and neutropenia (5.9%). At the end of the study, 2 (1.75%) of 69 patients reported antibodies to Bevacizumab without affecting safety and efficacy. However, at the end of 12 months, no patient had reported antibodies to Bevacizumab. Complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD) were reported in 18.3%, 22.6%, 9.6%, and 8.7% of patients, respectively. The overall response rate (CR + PR) was reported in 40.9% of patients at the end of the study. Disease control rate (DCR), also known as the clinical benefit rate (CBR) was reported in 50.4% of patients. Conclusions: Bevacizumab (Cizumab, Hetero Biopharma) was observed to be safe, well tolerated, lacking immunogenicity, and efficacious in the treatment of solid tumors. The findings of this phase IV study of Bevacizumab, primarily as a combination therapy regimen suggest its suitability and rationality for usage in multiple solid malignancies. Clinical Trial Registry Number: CTRI/2018/4/13371 [Registered on CTRI http://ctri.nic.in/Clinicaltrials/advsearch.php : 19/04/2018]; Trial Registered Prospectively.

Effect of Peritumoral Infiltration of Local Anesthetic Before Surgery on Survival in Early Breast Cancer.

PURPOSE: Preventing metastases by using perioperative interventions has not been adequately explored. Local anesthesia blocks voltage-gated sodium channels and thereby prevents activation of prometastatic pathways. We conducted an open-label, multicenter randomized trial to test the impact of presurgical, peritumoral infiltration of local anesthesia on disease-free survival (DFS). METHODS: Women with early breast cancer planned for upfront surgery without prior neoadjuvant treatment were randomly assigned to receive peritumoral injection of 0.5% lidocaine, 7-10 minutes before surgery (local anesthetics [LA] arm) or surgery without lidocaine (no LA arm). Random assignment was stratified by menopausal status, tumor size, and center. Participants received standard postoperative adjuvant treatment. Primary and secondary end points were DFS and overall survival (OS), respectively. RESULTS: Excluding eligibility violations, 1,583 of 1,600 randomly assigned patients were included in this analysis (LA, 796; no LA, 804). At a median follow-up of 68 months, there were 255 DFS events (LA, 109; no LA, 146) and 189 deaths (LA, 79; no LA, 110). In LA and no LA arms, 5-year DFS rates were 86.6% and 82.6% (hazard ratio [HR], 0.74; 95% CI, 0.58 to 0.95; P = .017) and 5-year OS rates were 90.1% and 86.4%, respectively (HR, 0.71; 95% CI, 0.53 to 0.94; P = .019). The impact of LA was similar in subgroups defined by menopausal status, tumor size, nodal metastases, and hormone receptor and human epidermal growth factor receptor 2 status. Using competing risk analyses, in LA and no LA arms, 5-year cumulative incidence rates of locoregional recurrence were 3.4% and 4.5% (HR, 0.68; 95% CI, 0.41 to 1.11), and distant recurrence rates were 8.5% and 11.6%, respectively (HR, 0.73; 95% CI, 0.53 to 0.99). There were no adverse events because of lidocaine injection. CONCLUSION: Peritumoral injection of lidocaine before breast cancer surgery significantly increases DFS and OS. Altering events at the time of surgery can prevent metastases in early breast cancer (CTRI/2014/11/005228).[Media: see text].

Variational multiscale reinforcement learning for discovering reduced order closure models of nonlinear spatiotemporal transport systems.

A central challenge in the computational modeling and simulation of a multitude of science applications is to achieve robust and accurate closures for their coarse-grained representations due to underlying highly nonlinear multiscale interactions. These closure models are common in many nonlinear spatiotemporal systems to account for losses due to reduced order representations, including many transport phenomena in fluids. Previous data-driven closure modeling efforts have mostly focused on supervised learning approaches using high fidelity simulation data. On the other hand, reinforcement learning (RL) is a powerful yet relatively uncharted method in spatiotemporally extended systems. In this study, we put forth a modular dynamic closure modeling and discovery framework to stabilize the Galerkin projection based reduced order models that may arise in many nonlinear spatiotemporal dynamical systems with quadratic nonlinearity. However, a key element in creating a robust RL agent is to introduce a feasible reward function, which can be constituted of any difference metrics between the RL model and high fidelity simulation data. First, we introduce a multi-modal RL to discover mode-dependant closure policies that utilize the high fidelity data in rewarding our RL agent. We then formulate a variational multiscale RL (VMRL) approach to discover closure models without requiring access to the high fidelity data in designing the reward function. Specifically, our chief innovation is to leverage variational multiscale formalism to quantify the difference between modal interactions in Galerkin systems. Our results in simulating the viscous Burgers equation indicate that the proposed VMRL method leads to robust and accurate closure parameterizations, and it may potentially be used to discover scale-aware closure models for complex dynamical systems.

Physics guided neural networks for modelling of non-linear dynamics.

The success of the current wave of artificial intelligence can be partly attributed to deep neural networks, which have proven to be very effective in learning complex patterns from large datasets with minimal human intervention. However, it is difficult to train these models on complex dynamical systems from data alone due to their low data efficiency and sensitivity to hyperparameters and initialisation. This work demonstrates that injection of partially known information at an intermediate layer in a DNN can improve model accuracy, reduce model uncertainty, and yield improved convergence during the training. The value of these physics-guided neural networks has been demonstrated by learning the dynamics of a wide variety of nonlinear dynamical systems represented by five well-known equations in nonlinear systems theory: the Lotka-Volterra, Duffing, Van der Pol, Lorenz, and Henon-Heiles systems.

Parameter Identification of Cardiovascular System Model Used for Left Ventricular Assist Device Algorithms.

Advancement of implanted left ventricular assist device (LVAD) technology includes modern sensing and control methods to enable online diagnostics and monitoring of patients using on-board sensors. These methods often rely on a cardiovascular system (CVS) model, the parameters of which must be identified for the specific patient. Some of these, such as the systemic vascular resistance (SVR), can be estimated online while others must be identified separately. This paper describes a three-staged approach for designing a parameter identification algorithm (PIA) for this problem. The approach is demonstrated using a two-element Windkessel model of the systemic circulation (SC) with a time-varying elastance for the left ventricle (LV). A parameter identifiability stage is followed by identification using an unscented Kalman filter (UKF), which uses measurements of LV pressure (Plv), aortic pressure (Pao), aortic flow (Qa), and known input measurement of LVAD flowrate (Qvad). Both simulation and experimental data from animal experiments were used to evaluate the presented methods. By bounding the initial guess for left ventricular volume, the identified CVS model is able to reproduce signals of Plv, Pao, and Qa within a normalized root mean squared error (nRMSE) of 5.1%, 19%, and 11%, respectively, during simulations. Experimentally, the identified model is able to estimate SVR with an accuracy of 3.4% compared with values from invasive measurements. Diagnostics and physiological control algorithms on-board modern LVADs could use CVS models other than those shown here, and the presented approach is easily adaptable to them. The methods also demonstrate how to test the robustness and accuracy of the identification algorithm.

Role of ado-trastuzumab emtansine (T-DM1) in HER2 positive salivary gland tumour with brain metastasis.

Malignant neoplasms of salivary gland neoplasms are rare and often involve the parotid gland. The primary treatment of these malignancies is surgery with or without adjuvant therapy. Chemotherapy or systemic therapy is indicated in recurrent or metastatic disease where surgery or radiotherapy is not possible. Salivary gland carcinomas, which are human epidermal growth factor receptor 2 (HER2) positive, show an aggressive behaviour with a poor prognosis. Targeting the HER2 pathway with drugs designed to block this pathway is an interesting novel therapy to treat salivary gland carcinomas. We report a case of a patient with HER 2-overexpressing parotid gland adenocarcinoma with brain metastasis, who was managed with ado-trastuzumab emtansine (T-DM1): a monoclonal antibody-cytotoxic drug conjugate that combines trastuzumab with the microtubule inhibitor, emtansine. The patient showed excellent response to the therapy. This case highlights the role of systemic chemotherapy with T-DM1 in HER2 positive salivary gland tumours that could be considered a part of the treatment regimen.

Multi-fidelity information fusion with concatenated neural networks.

Recently, computational modeling has shifted towards the use of statistical inference, deep learning, and other data-driven modeling frameworks. Although this shift in modeling holds promise in many applications like design optimization and real-time control by lowering the computational burden, training deep learning models needs a huge amount of data. This big data is not always available for scientific problems and leads to poorly generalizable data-driven models. This gap can be furnished by leveraging information from physics-based models. Exploiting prior knowledge about the problem at hand, this study puts forth a physics-guided machine learning (PGML) approach to build more tailored, effective, and efficient surrogate models. For our analysis, without losing its generalizability and modularity, we focus on the development of predictive models for laminar and turbulent boundary layer flows. In particular, we combine the self-similarity solution and power-law velocity profile (low-fidelity models) with the noisy data obtained either from experiments or computational fluid dynamics simulations (high-fidelity models) through a concatenated neural network. We illustrate how the knowledge from these simplified models results in reducing uncertainties associated with deep learning models applied to boundary layer flow prediction problems. The proposed multi-fidelity information fusion framework produces physically consistent models that attempt to achieve better generalization than data-driven models obtained purely based on data. While we demonstrate our framework for a problem relevant to fluid mechanics, its workflow and principles can be adopted for many scientific problems where empirical, analytical, or simplified models are prevalent. In line with grand demands in novel PGML principles, this work builds a bridge between extensive physics-based theories and data-driven modeling paradigms and paves the way for using hybrid physics and machine learning modeling approaches for next-generation digital twin technologies.

Hybrid Mock Circulatory Loop Simulation of Extreme Cardiac Events.

OBJECTIVE: This paper presents preliminary methods of incorporating the pathological conditions of cardiac arrhythmias and valvular stenosis in hybrid mock circulation loop (hMCL) operation for the enhanced verification and validation of mechanical circulatory support devices such as VADs. METHODS: The MGH/MF Waveform datasets from PhysioNet database (including both nominal and clinically diagnosed arrhythmic ECG measurements) as well as cardiovascular system model updates are used to recreate arrhythmic events and valvular stenosis in vitro. RESULTS: Preliminary results show the hMCL can recreate each tested cardiac event within 2% and 4% mean error for reference pressure tracking in the aortic and left ventricular pressure chambers, respectively. Further, frequency spectrum analysis comparisons using the magnitude-squared coherence analysis shows close alignment between measured arrhythmic and hMCL realized pressure frequency content. CONCLUSION: The generation of cardiac arrhythmias and valvular stenosis around a VAD via both model and acute measurement based methods was achieved. SIGNIFICANCE: Pathological conditions such as cardiac arrhythmias and valvular stenosis are limited in documentation despite the large percentage of patients who experience these events. This paper provides a means to begin incorporating these events into hardware-in-the-loop mock circulatory systems for next generation VAD validation and verification.

Kolhapur Cancer Centre Aiming for Perfection at Grass Root Level, Spreading Care to Those in Need!

There is a big need for more comprehensive cancer centres in tier 3 cities in India. These small cities are the most accessible to the majority of the rural population of India. Most of these patients are economically compromised and thus cannot manage treatment options in metropolitan cities. Kolhapur Cancer Centre was started with the philosophy of "paying back to society" and serving these needy patients in rural India.

Minimally Invasive Esophagectomy in Semi-Prone Position (Pawar Technique): Technical Aspects and Outcome in 224 Patients.

Background and Objectives There are two patient positions described for minimally invasive esophagectomy (MIE) for esophageal cancer, viz., left lateral and prone positions. To retain the benefits and overcome the disadvantages of these positions, a semi-prone position was developed by us. Our objective was to analyze the feasibility of performing MIE in this position. Materials and Methods A retrospective review of patients who underwent MIE at our center from January 2007 to December 2017 was done. A semi-prone position is a left lateral position with an anterior inclination of 45 degrees. Intraoperative parameters including conversion rate, immediate postoperative outcomes, and long-term oncological outcomes were analyzed. Statistical Analysis Statistical Package for the Social Sciences version 19 (IBM SPSS, IBM Corp., Armonk, New York, United States) was utilized for analysis. Survival analysis was done using Kaplan-Meier graph. Quantitative data were described as mean or median with standard deviation, and qualitative data were described as frequency distribution tables. Results Consecutive 224 patients with good performance status were included. After excluding those who required conversion (14 [6.6%]), 210 patients were further analyzed. Median age was 60 years (range: 27-80 years). Neoadjuvant treatment recipients were 160 (76%) patients. Most common presentation was squamous cell carcinoma (146 [70%]) of lower third esophagus (140 [67%]) of stage III (126 [60%]). Median blood loss for thoracoscopic dissection and for total operation was 101.5 mL (range: 30-180 mL) and 286 mL (range: 93-480 mL), respectively. Median operative time for thoracoscopic dissection alone was 67 minutes (range: 34-98 minutes) and for entire procedure was 215 minutes (range: 162-268 minutes). There was no intraoperative mortality. Median 16 lymph nodes were dissected (range: 5-32). Postoperative complication rate and mortality was 50% and 3.3%, respectively. Disease-free interval was 18 months (range: 3-108 months) and overall survival was 22 months (range: 6-108 months). Conclusion MIE with mediastinal lymphadenectomy in a semi-prone position is feasible, convenient, oncologically safe, which can combine the benefits of the two conventional approaches. Further prospective and comparative studies are required to support our findings.

Preoperative Chemoradiation in Locally-Advanced Resectable Carcinoma of the Esophagus in a Single Rural Cancer Hospital in Western India.

Background The current standard of care for the treatment of surgically resectable carcinoma of the esophagus is preoperative chemoradiation followed by surgery. There is strong evidence that this trimodality approach improves survival as compared with surgery alone. Objective The objective of this study is to determine the feasibility of this approach in a rural cancer institute in western India. Materials and Methods The data of all the 157 consecutively treated patients with locally-advanced carcinoma of the esophagus from March 2013 to March 2017 who were started on preoperative chemoradiation were analyzed retrospectively. Results Of the 157 patients who were started on preoperative chemoradiation, 68 patients underwent surgery. There are various practical reasons for not undergoing the definitive surgery, with the important being the socioeconomic support to the patients during the course of treatment. Conclusion This study gave us insight into the strategic selection of patients for the trimodality approach as well as the need for continuous socioeconomic support throughout the treatment course.

Retrospective analysis of patients with carcinoma cervix in a rural/semi-urban setting in Western India.

OBJECTIVES: To compare the presentation of cervical cancer and the treatment modalities received by the patients at a semi-urban/rural area of Western India with that of published literature from urban centers. MATERIALS AND METHODS: We conducted a retrospective analysis of patients with cervical cancer who presented at a semi-urban/rural cancer center between 2010 and 2013. A total of 141 patients with the median age of 51 years (25-81) were studied. The demographic and clinical variables included age, annual family income, profession, comorbidities, baseline hemoglobin, prior screening, clinical stage, treatment administered, and complications. The pathological variables included tumor type and grade. RESULTS: In our study, all patients presented with vaginal bleeding. Majority of the patients (51 patients, 37.7%) had Stage 3B disease. Since majority presented at later stages (Stage 3B), chemotherapy-radiotherapy was the most common treatment modality used in our population. On histopathology, 127 patients (90%) had squamous cell carcinoma while 14 patients (10%) had adenocarcinoma. In 96 patients (68%), the tumor grade was not known while it was a high, intermediate, and low grade in 6 (4%), 18 (13%), and 21 (15%) patients, respectively. The follow-up data of our study were not adequate; hence, the long-term survival results could not be presented. CONCLUSION: Patients in rural India setting present at later stages which could be improved by creating awareness, improving their personal hygiene, and adequate screening.

Evaluation of ER, PR and HER-2 receptor expression in breast cancer patients presenting to a semi urban cancer centre in Western India.

BACKGROUND: Hormone receptor expression has been reported to be low in breast cancer patients from developing countries. The pattern of receptor expression from urban and rural areas is not well studied. MATERIALS AND METHODS: This is a retrospective analysis of 206 consecutive breast cancer patients presenting to a semi urban cancer centre from 2009-2010. The demographic and clinical variables included age, residential area (rural, semi urban, or urban), menopausal status, and clinical stage. The pathological variables included tumor type, the presence of ductal carcinoma in situ, lymphovascular invasion, and expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) receptors by immunohistochemical (IHC) analysis. RESULTS: The majority of patients were postmenopausal with the median age of 50 years. Invasive ductal carcinoma was the most common subtype (94%). The ER status was available in 101 (49.3%), PR in 99 (48.0%), and HER2 in 82 (39.8%) cases. In patients in whom this data were available, ER was positive in 44.6%, PR in 40.4%, and HER2 in 34.2%. Out of the 82 patients in whom data on all three receptors were available, 34.1% patients had triple negative tumors. Analysis of our data showed a trend toward increasing ER and PR expression with age but this was not statistically significant. The average age of menopause was between 40-50 years of age. CONCLUSION: This report is an important documentation of the pathological characteristics in a predominantly rural/semi urban population of Indian breast cancer patients. Further studies from other centers with a similar background are required to confirm these results.