Next Generation Frequency Domain Near Infrared Spectroscopy

NIRS is an imaging technique that uses red and near-infrared light in the first near infrared biological optical window (~600-1000nm, NIR-I) to safely and non-invasively measure the optical properties of centimeter thick tissue. In this wavelength range the primary absorbers in tissue are oxygenated hemoglobin, deoxygenated hemoglobin, lipid, and water. By using this optical window, the information content provided with the detected light necessarily contains biologically relevant functional information about the chromophores of interest unlike standard imaging modalities which typically provide structural information alone. Using this functional information, NIRS is able to determine benign from malignant tumors, measure functional brain activity, monitor the effects of exercise on the body as well as many other applications.

Despite these exciting and promising applications, NIRS faces many limitations which need to be overcome in order to become a widely utilized clinical imaging technology. fdNIRS imaging equipment is typically a large highly expensive cart-based system that must be stored and wheeled around the hospital when needed for use. The

technology also tends to be quite complicated to use for the average medical professional requiring a number of steps for both calibration and imaging. Typical images require nearly one hour to complete and even more time to process the acquired data which must be done by a scientist familiar with the process to provide the clinician a usable image. Therefore, using fdNIRS technology is not only a challenge for medical staff but is also time-consuming, placing more strain on already time constrained personnel. Finally, when the clinician receives this information, they are forced to ask themselves how it compares to imaging modalities with which they are intimately familiar such as CT, MRI, and ultrasound. Although NIRS provides physiological information (hemoglobin, water, and lipid concentration among others) not available by these other imaging methods, the imaging depth, resolution, and field of view tend to be significantly reduced.

To address some of these issues an approach for performing frequency domain (fd) NIRS utilizing a near-infrared tunable vertical cavity surface emitting laser (VCSEL) that enables high spectral resolution optical sensing in a miniature format is presented. At room temperature, the laser is tunable across 14nm from 769-782nm covering a portion of the NIRI window and has a peak output power sufficient for fdNIRS. The tunable VCSEL was used to recover optical properties (leading to chromophores) which were within 13% of a reference system. The results indicate that tunable VCSELs are an attractive choice to enable high spectral resolution optical sensing in a wearable format.

The imaging depth and sensitivity of fdNIRS is almost always constrained by photodetector sensitivity. We demonstrated that by using highly sensitive silicon photomultiplier (SiPM) detectors, these limitations may be overcome. Though SiPMs are typically employed in photon-counting or pulsed time-domain measurement applications, they provide increased signal to noise ratio (SNR) and deeper tissue sensitivity, as compared to avalanche photodiodes (APDs) in fdNIRS. A basic investigation of an fdNIRS system utilizing SiPM detectors, and thus enabling extended source-detector separations (sds) and increased depth penetration is presented. The SiPM was found to have 10-30dB greater SNR than a comparably sized APD while detecting 1.5-2 orders of magnitude lower light levels. The SiPM and APD recover optical property values of tissue simulating phantoms within 13% agreement and are stable with 1% coefficient of variation over one hour. Finally, the SiPM is used to accurately recover optical properties at s/d separations up to 48mm (12mm increase compared to reference) in phantoms mimicking human breast tissue, enabling increased depth sensitivity.

The question still remains, how fundamental SiPM characteristics affect the detection of fdNIRS signals. In order to address this issue, simulations are developed which are capable of evaluating and optimizing SiPM parameters (microcell size, microcell number, and recharge time constant) utilizing a Monte Carlo approach to SiPM response from modulated fdNIRS signals. The effects of these parameters are examined in relation to bandwidth and signal to noise ratio, thus providing a methodology for choosing the best SiPM for this unconventional application, and further improving fdNIRS performance.

The details of integrating silicon photomultiplier detectors in fdNIRS imaging systems to provide performance improvements are further investigated to include the operation of SiPMs and how they differ from other fdNIRS photodetectors. It is shown that SiPMs offer similar sensitivity to PMTs while having a lower cost, size, and operating voltage. Experimental data shows drastically increased signal to noise ratio performance as compared to an avalanche photodiode, up to 25dB on human breast, head, and muscle tissue. Finally, we extend the dynamic range of the SiPM through a nonlinear calibration technique which reduced absorption error by a mean 16 percentage points.

Finally, limitations in speed and accessibility are addressed with a systems-based approach. Currently no fdNIRS systems are completely handheld and only one is capable of displaying chromophore data in real time, which could allow for real-time studies of tissue hemodynamics, spatial chromophore concentrations, greater ease of use by clinicians, and more generally, a simple platform for quantitative tissue spectroscopy. Evaluation was performed on a handheld real-time capable fdNIRS system based upon an all-digital FPGA coupled hardware approach. Quantitative optical scattering and absorption measurements are found to be within 10% agreement as compared to a reference system. It is concluded that handheld high-speed quantitative tissue chromophore assessments are possible with this system-on-a-chip fdNIRS approach, which will enable real-time handheld monitoring of rapid physiological changes.

This work concludes with a summary of the progress that has been made as well as the work that is yet to be done when considering this text within the scope of fdNIRS. Distinct areas where further improvements and research are required to fully enable a handheld fdNIRS systems utilizing the advantages provided by tunable VCSELs and SiPMs is provided. Lastly the authors vision for the future of fdNIRS and a clinically viable fdNIRS device is presented.