Integrating deep learning with microfluidics for biophysical classification of sickle red blood cells adhered to laminin

Sickle cell disease, a genetic disorder affecting a sizeable global demographic, manifests in sickle red blood cells (sRBCs) with altered shape and biomechanics. sRBCs show heightened adhesive interactions with inflamed endothelium, triggering painful vascular occlusion events. Numerous studies employ microfluidic-assay-based monitoring tools to quantify characteristics of adhered sRBCs from high resolution channel images. The current image analysis workflow relies on detailed morphological characterization and cell counting by a specially trained worker. This is time and labor intensive, and prone to user bias artifacts. Here we establish a morphology based classification scheme to identify two naturally arising sRBC subpopulations-deformable and non-deformable sRBCs-utilizing novel visual markers that link to underlying cell biomechanical properties and hold promise for clinically relevant insights. We then set up a standardized, reproducible, and fully automated image analysis workflow designed to carry out this classification. This relies on a two part deep neural network architecture that works in tandem for segmentation of channel images and classification of adhered cells into subtypes. Network training utilized an extensive data set of images generated by the SCD BioChip, a microfluidic assay which injects clinical whole blood samples into protein-functionalized microchannels, mimicking physiological conditions in the microvasculature. Here we carried out the assay with the sub-endothelial protein laminin. The machine learning approach segmented the resulting channel images with 99.1±0.3% mean IoU on the validation set across 5 k-folds, classified detected sRBCs with 96.0±0.3% mean accuracy on the validation set across 5 k-folds, and matched trained personnel in overall characterization of whole channel images with R2 = 0.992, 0.987 and 0.834 for total, deformable and non-deformable sRBC counts respectively. Average analysis time per channel image was also improved by two orders of magnitude (∼ 2 minutes vs ∼ 2-3 hours) over manual characterization. Finally, the network results show an order of magnitude less variance in counts on repeat trials than humans. This kind of standardization is a prerequisite for the viability of any diagnostic technology, making our system suitable for affordable and high throughput disease monitoring.

Compressed sensing of remotely detected MRI velocimetry in microfluidics

NMR and MRI can yield detailed chemical and dynamic information about flow at microscopic resolutions, but suffer from low signal to noise relative to alternative techniques for flow measurements. In porous media and microfluidic devices, this sensitivity problem is further exacerbated by magnetic susceptibility broadening and low coil filling factor. Fortunately, remote detection can mitigate these issues by physically separating signal detection from the other steps of the experiment. The technique requires, however, that any measured interactions be encoded in indirectly sampled dimensions, leading to experiments of high dimensionality and correspondingly long acquisition times. We have applied compressed sensing, a reconstruction technique used in MRI, to dramatically reduce these experiment times by 8-64x through partial sampling (sub-sampling) of k-space, allowing for the collection of images with significantly higher resolutions in reasonable amounts of time. Here, we demonstrate this reconstruction technique to remotely detected flow measurements in a serpentine mixing chip and in a microfluidic channel harboring a constriction. We find that compressed sensing allows for significantly higher resolution images to be collected in a practical amount of time, thus significantly enhancing the applicability of remote detection to flow imaging.

Beating Poisson encapsulation statistics using close-packed ordering

Loading drops with discrete objects, such as particles and cells, is often necessary when performing chemical and biological assays in microfluidic devices. However, random loading techniques are inefficient, yielding a majority of empty and unusable drops. We use deformable particles that are close packed to insert a controllable number of particles into every drop. This provides a simple, flexible means of efficiently encapsulating a controllable number of particles per drop.