Exploring optically-derived uptake functions in the context of dynamic contrast-enhanced MRI arterial input function measurement

Introduction Arterial input function (AIF) measurement provides a challenging problem in the context of dynamic contrast-enhanced MRI (DCE-MRI). Whilst some form of AIF is commonly incorporated into models of perfusion and permeability, direct measurement requires high spatiotemporal resolution to capture intrinsic variability; this is especially difficult if it is combined with simultaneous tissue measurements. This is most acute during the rapid contrast agent (CA) fluctuation of the first pass peak phase. It has been noted [1] that tracers with high levels of absorption in the visible range can interfere with accurate measurement of pSaO2 (blood oxygen saturation as measured by a pulse oximeter). Work in this area prompts consideration of the feasibility of acquiring high-temporal resolution 1D measurements of an optical tracer in an artery contemporaneously with high-spatial resolution DCE-MRI mapping of tissue uptake of a CA [2,3]. As an initial investigation, this study tested whether methylene blue (MB) could be measured in nude mice using equipment designed to locate arterial signals in a manner facilitating incorporation into a DCE-MRI study. Materials and Methods Figure 1 shows the sensor design based on the LM358 operational amplifier and the DrDAQ (Pico Technology, Cambridge, UK) data logger. A 250 mW 660 nm IFE96 LED, two 3 mm × 9 m fibre optic cables and an IFD92 phototransistor (LasIRvis, Cornwall, UK) formed a transmittance source/sensor combination operating as a pulse photoplethysmograph (PPG) and an absolute light signal intensity monitor (SIM). This was designed to allow an arterial signal to be located using the same sensor as that used during the uptake measurements. The PPG utilised a dual stage band pass filter with frequency corners at 1.5 and 16 Hz, each stage providing up to ×100 amplification. The SIM consisted of a dual stage amplifier in which the first stage provided variable amplification such that the output voltage was directly proportional to the current flowing through the phototransistor, and the second stage provided a low pass filter with a frequency corner at 16 Hz and no gain. These initial proof of concept experiments were conducted as shown in Figure 2, with monitoring of subject breathing rate and temperature and in accordance with UK Home Office regulations. Data were recorded on 8 nude mice (21.2 to 25.2 g) into which 20 μl of 0.75% or 1% MB were injected into the tail vein in a carrier of 120 μl of 100 i.u. heparinised saline under administration of 1.5-2% isoflurane and air. Prior to injection, the PPG signal was observed to locate the light source and sensor appropriately, which were secured on the left leg by a rigid cuff. The sensor was then switched to SIM mode, a 200 s baseline signal was measured (not shown) and the MB injected. An additional control experiment, injecting heparinised saline alone, was conducted to verify that signal changes were due to MB injection. Results and Discussion Figure 3 demonstrates it is feasible to measure uptake curves in vivo. (a) shows a PPG trace with breathing and heart rate components wherein large amplitudes indicate a strong arterial signal component; (b) and (c) show curves demonstrating, respectively, an initial signal fluctuation followed by further signal rise, and an initial fluctuation with minimal further signal rise. Curves were measured in terms of signal voltage, V, and converted to phototransistor current I where V = -g.I for gain g. An initial signal fluctuation, at the injection time, was observed in four test subjects. Large signal rise after injection with no initial fluctuation was observed in one subject. The initial fluctuations occurred at the injection time and may be associated with the first pass peak of the AIF. Signal responses did not appear to be related to injection concentration and the control experiment did not show any initial fluctuation or clear signal change. Subsequent signal rises may be associated with gradual leakage of the MB into the extracellular, extravascular space. As the arterial signal is partial volumed with surrounding tissue, it is plausible that such subsequent signal rise may