Deterministic Lateral Displacement

Cheaper and Easier Cancer Diagnosis

Mean displacement of the cells in the separation direction for the CTC-Ichip device (simulation results)
90% of cancer metastasis is caused by Circulating Tumor Cells (CTCs) which intravasete the bloodstream from a primary tumors. The detection of CTCs in the blood is one of the most potent methods for the early diagnosis of cancer. However, with 1 CTC present for every 1’000’000’000 RBCs, this detection is equivalent to finding a ”needle in a flowing haystack”. This challenge has led to a wealth of activity in developing microfluidic devices for high throughput cell separation. The design of microfluidic devices relies on the quantitative understanding of the behaviour of cells and fluids confined in complex, microscale geometries. However, the acquisition of information for the microrheology of fluids remains a formidable experimental task. Without such information, the unique, and often unexpected, properties of fluids at the microscale can be a hindrance for new designs, while complex prototypes may be impossible to manufacture at an economy of scales. Predictive simulations capable of handling complex systems at sub-micron resolution, are becoming an invaluable companion to experiments to accelerate the design cycle and innovation in microfluidics.

Designing Efficient Devices

We consider the CTC-iChip microfluidics device, which is arguably the most effective device in isolating CTCs and WBCs from hundreds of thousands RBCs by taking advantage of the deterministic lateral displacement effect. We plan to enhance the accuracy of the simulations and to optimize the geometry by employing detailed models with quantified uncertainties. Our efficient software implementation allows us to investigate flow behavior at high RBCs concentrations, which remain prohibitive for other state-of-the-art methods and codes.

65% Peak Performance

We employ uDeviceX, an in-house particle based code for simulating blood flow. We use Dissipative Particle Dynamics for solvent and a viscoelastic membrane model for both CTCs and RBCs. The code is able to reach 65% peak performance in terms of instructions per seconds on 18’688 nodes of the Titan supercomputer.

Data Driven Simulations: UQ+P

The predictive capabilities of DPD models hinge on their parameters that are usually validated from various experiments. However these parameters may vary drastically and are not transferable across experiments. In our lab we estimate the RBC model parameters using classical and hierarchical Bayesian inference techniques. We revisit classical experiments of stretching and tank treading blood cells in shear flows and examine through the Bayesian model uncertainty the capabilities and limitations of the RBC model.

Elastic properties calibration

During the course of their circulation, RBCs undergo severe elastic deformation as they flow through narrow capillaries whose diameter is as small as 3um. For the calibration of the elastic properties of the RBC model, the elongation of the RBC is analyzed under various stretching forces. We consider data available in the literature, from studies of the mechanics of single cells, subjected to a variety of well-controlled stress-states.

Elastic properties calibration

Viscous properties calibration

There exist several experimental datasets describing the motion of a single RBC in a linear shear flow. It is known that the motion of single RBCs strongly depends on the shear rate of the flow. At the high shear rate regime, RBCs undergo a tank-treading motion, where they acquire a stable orientation while their membrane performs a rotation around itself. The data for this regime, consist of the RBC tank-treading frequency (TTF), measured at various shear rates. In our inference, we take into account the variability between the different datasets, aiming towards a proper integration of experimental data for reproducible blood flow simulations.

Viscous properties calibration

People: Dmitry Alexeev, Lucas Amoudruz, Athena Economides, Panagiotis Hadjidoukas, Sergey Litvinov, Georgios Arampatzis

2017

  • A. Economides, L. Amoudruz, S. Litvinov, D. Alexeev, S. Nizzero, P. E. Hadjidoukas, D. Rossinelli, and P. Koumoutsakos, “Towards the virtual rheometer,” in Proceedings of the platform for advanced scientific computing – PASC ’17, 2017.

BibTeX

@inproceedings{economides2017a,
author = {Athena Economides and Lucas Amoudruz and Sergey Litvinov and Dmitry Alexeev and Sara Nizzero and Panagiotis E. Hadjidoukas and Diego Rossinelli and Petros Koumoutsakos},
booktitle = {Proceedings of the Platform for Advanced Scientific Computing - {PASC} {\textquotesingle}17},
doi = {10.1145/3093172.3093226},
publisher = {{ACM} Press},
title = {Towards the Virtual Rheometer},
url = {https://cse-lab.seas.harvard.edu/files/cse-lab/files/economides2017a.pdf},
year = {2017}
}

Abstract

Recent advances in medical research and bio-engineering have led to the development of devices capable of handling fluids and biological matter at the microscale. The operating conditions of medical devices are constrained to ensure that characteristic properties of blood flow, such as mechanical properties and local hemodynamics, are not altered during operation. These properties are a consequence of the red blood cell (RBC) microstructure, which changes dynamically according to the device geometry. The understanding of the mechanics and dynamics that govern the interactions between the RBCs is crucial for the quantitative characterization of blood flow, a stepping stone towards the design of medical devices specialized to the patient, in the context of personalized medicine. This can be achieved by analyzing the microstructural characteristics of the RBCs and study their dynamics. In this work we focus on the quantification of the microstructure of high and low hematocrit blood flows, in wall bounded geometries. We present distributions of the RBCs according to selected deformation criteria and dynamic characteristics, and elaborate on mechanisms that control their collective behavior, focusing on the interplay between cells and shear induced effects.

2015

  • D. Rossinelli, Y. Tang, K. Lykov, D. Alexeev, M. Bernaschi, P. Hadjidoukas, M. Bisson, W. Joubert, C. Conti, G. Karniadakis, M. Fatica, I. Pivkin, and P. Koumoutsakos, “The In-Silico Lab-on-a-Chip: petascale and high-throughput simulations of microfluidics at cell resolution,” in Proceedings of the international conference for high performance computing, networking, storage and analysis – SC ’15, 2015.

BibTeX

@inproceedings{rossinelli2015b,
author = {Diego Rossinelli and Yu-Hang Tang and Kirill Lykov and Dmitry Alexeev and Massimo Bernaschi and Panagiotis Hadjidoukas and Mauro Bisson and Wayne Joubert and Christian Conti and George Karniadakis and Massimiliano Fatica and Igor Pivkin and Petros Koumoutsakos},
booktitle = {Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis – {SC} {\textquotesingle}15},
doi = {10.1145/2807591.2807677},
number = {Article 2},
publisher = {{ACM} Press},
title = {The {In-Silico Lab-on-a-Chip}: Petascale and High-Throughput Simulations of Microfluidics at Cell Resolution},
url = {https://cse-lab.seas.harvard.edu/files/cse-lab/files/rossinelli2015b.pdf},
year = {2015}
}

Abstract

We present simulations of blood and cancer cell separation in complex microfluidic channels with subcellular resolu- tion, demonstrating unprecedented time to solution, per- forming at 65.5% of the available 39.4 PetaInstructions/s in the 18, 688 nodes of the Titan supercomputer. These simulations outperform by one to three orders of magnitude the current state of the art in terms of numbers of simulated cells and computational elements. The com- putational setup emulates the conditions and the geometric complexity of microfluidic experiments and our results re- produce the experimental findings. These simulations pro- vide sub-micron resolution while accessing time scales rele- vant to engineering designs. We demonstrate an improvement of up to 45X over com- peting state-of-the-art solvers, thus establishing the frontiers of simulations by particle based methods. Our simulations redefine the role of computational science for the develop- ment of microfluidics – a technology that is becoming as important to medicine as integrated circuits have been to computers