Image Reconstruction Using Deep Learning for Near-Infrared Optical Tomography: Generalization Assessment.

Time is one of the most critical factors in preventing brain lesions due to hypoxic ischemia in preterm infants. Since early detection of low oxygenation is vital and the time window for therapy is narrow, near-infrared optical tomography (NIROT) must be able to process the high-dimensional data provided by today's advanced systems in the shortest possible time. Deep learning approaches are attractive because they can exploit such high information density while reducing inference time. The aim of this study was to evaluate the performance of a hybrid convolutional neural network, designed for NIROT image reconstruction and trained on synthetic data. Generalization capability was assessed using measurements on phantoms of a surface topology more divergent than the range of variation in the geometries of the in-silico data, with unseen, non-spherical inclusion shapes, and with source and detector arrangements different from those used for data generation. Substantial gains in speed, localization accuracy, and high image quality were achieved even under the highly varied measurement conditions.

Time Domain Near-Infrared Optical Tomography Utilizing Full Temporal Data: A Simulation Study.

The analysis of full temporal data in time-domain near-infrared optical tomography (TD NIROT) measurements enables valuable information to be obtained about tissue properties with good temporal and spatial resolution. However, the large amount of data obtained is not easy to handle in the image reconstruction. The goal of the project is to employ full-temporal data from a TD NIROT modality. We improved TD data-based 3D image reconstruction and compared the performance with other methods using frequency domain (FD) and temporal moments. The iterative reconstruction algorithm was evaluated in simulations with both noiseless and noisy in-silico data. In the noiseless cases, a superior image quality was achieved by the reconstruction using full temporal data, especially when dealing with inclusions at 20 mm and deeper in the tissue. When noise similar to measured data was present, the quality of the recovered image from full temporal data was no longer superior to the one obtained from the analysis of FD data and temporal moments. This indicates that denoising methods for TD data should be developed. In conclusion, TD data contain richer information and yield better image quality.

BIAN: A Multilayer Microfluidic-Based Tissue-Mimicking Phantom for Near-Infrared Imaging.

Near-infrared spectroscopy (NIRS) is a non-invasive optical method for monitoring cerebral oxygenation. Changes in regional blood flow and oxygenation due to neurovascular coupling are important biomarkers of neuronal activation. So far, there has been little research on multilayer tissue phantoms with tuneable blood flow, blood volume, and optical properties to simulate local changes in oxygenation at different depths. The aim of this study is to design, fabricate and characterize a complex dynamic phantom based on multilayer microfluidics with controllable blood flow, blood volume, and optical properties for testing NIRS instruments. We developed a phantom prototype with two microfluidic chips embedded at two depths inside a solid silicone phantom to mimic the vessels in the scalp and in the cortex. To simulate the oxygenation and perfusion of tissue, a solution with blood-like optical properties was sent into the microchannels by a pump with a programmable pressure controller. The pressure adjusted the volume of the microfluidic chips representing a distension of blood vessels. The optical changes in the superficial and deep layers were measured by a commercially available frequency domain NIRS instrument. The NIRS successfully detected the changes in light intensity elicited by the changes in the pressure input to the two layers. In conclusion, the microfluidics-based imaging phantom was successfully designed and fabricated and mimics brain functional activity. This technique has great potential for testing other optical devices, e.g., diffuse correlation spectroscopy, pulse oximetry, and optical coherence tomography.

Hybrid Convolutional Neural Network (hCNN) for Image Reconstruction in Near-Infrared Optical Tomography.

Near-infrared optical tomography (NIROT), a promising imaging modality for early detection of oxygenation in the brain of preterm infants, requires data acquisition at the tissue surface and thus an image reconstruction adaptable to cephalometric variations and surface topologies. Widely used model-based reconstruction methods come with the drawback of huge computational cost. Neural networks move this computational load to an offline training phase, allowing much faster reconstruction. Our aim is a data-driven volumetric image reconstruction that generalises well to different surfaces, increases reconstruction speed, localisation accuracy and image quality. We propose a hybrid convolutional neural network (hCNN) based on the well-known V-net architecture to learn inclusion localisation and absorption coefficients of heterogenous arbitrary shapes via a joint cost function. We achieved an average reconstruction time of 30.45 s, a time reduction of 89% and inclusion detection with an average Dice score of 0.61. The CNN is flexible to surface topologies and compares well in quantitative metrics with the traditional model-based (MB) approach and state-of-the-art neuronal networks for NIROT. The proposed hCNN was successfully trained, validated and tested on in-silico data, excels MB methods in localisation accuracy and provides a remarkable increase in reconstruction speed.

Resolution and penetration depth of reflection-mode time-domain near infrared optical tomography using a ToF SPAD camera.

In a turbid medium such as biological tissue, near-infrared optical tomography (NIROT) can image the oxygenation, a highly relevant clinical parameter. To be an efficient diagnostic tool, NIROT has to have high spatial resolution and depth sensitivity, fast acquisition time, and be easy to use. Since many tissues cannot be penetrated by near-infrared light, such tissue needs to be measured in reflection mode, i.e., where light emission and detection components are placed on the same side. Thanks to the recent advance in single-photon avalanche diode (SPAD) array technology, we have developed a compact reflection-mode time-domain (TD) NIROT system with a large number of channels, which is expected to substantially increase the resolution and depth sensitivity of the oxygenation images. The aim was to test this experimentally for our SPAD camera-empowered TD NIROT system. Experiments with one and two inclusions, i.e., optically dense spheres of 5mm radius, immersed in turbid liquid were conducted. The inclusions were placed at depths from 10mm to 30mm and moved across the field-of-view. In the two-inclusion experiment, two identical spheres were placed at a lateral distance of 8mm. We also compared short exposure times of 1s, suitable for dynamic processes, with a long exposure of 100s. Additionally, we imaged complex geometries inside the turbid medium, which represented structural elements of a biological object. The quality of the reconstructed images was quantified by the root mean squared error (RMSE), peak signal-to-noise ratio (PSNR), and dice similarity. The two small spheres were successfully resolved up to a depth of 30mm. We demonstrated robust image reconstruction even at 1s exposure. Furthermore, the complex geometries were also successfully reconstructed. The results demonstrated a groundbreaking level of enhanced performance of the NIROT system based on a SPAD camera.