Monitoring of hemodynamic changes induced in the healthy breast through inspired gas stimuli with MR-guided diffuse optical imaging

Abstract

Purpose: The modulation of tissue hemodynamics has important clinical value in medicine for both tumor diagnosis and therapy. As an oncological tool, increasing tissue oxygenation via modulation of inspired gas has been proposed as a method to improve cancer therapy and determine radiation sensitivity. As a radiological tool, inducing changes in tissue total hemoglobin may provide a means to detect and characterize malignant tumors by providing information about tissue vascular function. The ability to change and measure tissue hemoglobin and oxygenation concentrations in the healthy breast during administration of three different types of modulated gas stimuli (oxygen/ carbogen, air/carbogen, and air/oxygen) was investigated.

Methods: Subjects breathed combinations of gases which were modulated in time. MR-guided diffuse optical tomography measured total hemoglobin and oxygen saturation in the breast every 30 s during the 16 min breathing stimulus. Metrics of maximum correlation and phase lag were calculated by cross correlating the measured hemodynamics with the stimulus. These results were compared to an air/air control to determine the hemodynamic changes compared to the baseline physiology.

Results: This study demonstrated that a gas stimulus consisting of alternating oxygen/carbogen induced the largest and most robust hemodynamic response in healthy breast parenchyma relative to the changes that occurred during the breathing of room air. This stimulus caused increases in total hemoglobin and oxygen saturation during the carbogen phase of gas inhalation, and decreases during the oxygen phase. These findings are consistent with the theory that oxygen acts as a vasoconstrictor, while carbogen acts as a vasodilator. However, difficulties in inducing a consistent change in tissue hemoglobin and oxygenation were observed because of variability in intersubject physiology, especially during the air/oxygen or air/carbogen modulated breathing protocols.

Conclusions: MR-guided diffuse optical imaging is a unique tool that can measure tissue hemodynamics in the breast during modulated breathing. This technique may have utility in determining the therapeutic potential of pretreatment tissue oxygenation or in investigating vascular function. Future gas modulation studies in the breast should use a combination of oxygen and carbogen as the functional stimulus. Additionally, control measures of subject physiology during air breathing are critical for robust measurements.

Figure 1

(a) Photograph of the optical fiber holder incorporated into a biopsy attachment for an eight-channel breast coil (USA Instruments, Aurora, OH). Slots in the optical fiber holder allowed for vertical positioning of the optical fibers, while the biopsy attachment allowed positioning in the medical/lateral direction. These degrees of freedom in the fiber interface enabled the optical fibers to maintain contact with the breast. Arrows indicate the fiber holder and the vertical adjustment. (b) Photograph of a healthy subject with fibers attached to the MR coil. The arrow shows the fibers attached to the fiber holder.

Figure 2

Gas timing diagram for the respiratory stimuli. Each stimulus alternated between two gases; the specifics of each stimulus is indicated in Table 1.

Figure 3

(a) Photograph of the gas breathing circuit placed on the MR table. (b) Schematic indicating that the carbogen and oxygen gases were controlled with a custom computer-controlled valve which was synchronized via software to maintain the desired modulation rate. (c) Diagram of circuit operation during gas stimulus. During oxygen or carbogen breathing, pressure from the inlet closes the valve to room air, preventing room air from entering the circuit. All valves were one-way. A fan in the MR bore was turned on to prevent rebreathing of expired gases. (d) Diagram of circuit operation during air stimulus. During air stimulus, the lack of pressure from the inlet opened the room air valve, allowing room air to be drawn into the circuit. A small tube was inserted near the mouthpiece to channel a small amount of expired air to an oxycapnometer to monitor exhaled gas concentrations.

Figure 4

Instrumentation related fluctuations in hemoglobin concentration recordings in a tissue-simulating phantom. Standard deviation in fitted deoxyhemoglobin is 0.19% of its amplitude while standard deviation in fitted HbO is 0.22% of its amplitude.

Figure 5

Temporal and frequency response of HbT for oxygen/carbogen gas stimulus compared to an air control in a typical subject where SNR>1. In (a), white bars indicate delivery of the O₂ stimulus and gray bars symbolize carbogen administration, whereas in (c), both bars indicate the breathing of room air. Total hemoglobin oscillations during gas stimulus show a significantly larger peak at 0.0042 Hz, the stimulus frequency, than the air control [compare (b) vs (d)], indicating the tissue response to the stimulus was larger than the physiological noise. Oscillations in CO₂ measured with the oxycapnograph during the oxygen/carbogen stimulus shown in (e) demonstrate subject compliance during the experimental protocol.

Figure 6

Same as Fig. 5 for oxygen saturation.

Figure 7

Time lag comparison of Sat vs HbT for the three gas stimuli (in units of π). The stimuli with carbogen demonstrate vasodilation, a response that allows highly oxygenated blood to enter the breast tissue, thus washing out deoxygenated hemoglobin and increasing oxygen saturation. Oxygen saturation and total hemoglobin during air/oxygen stimulus is not dominated by changes in vascular tone. Instead, these changes may be influenced by changes in pO₂. See Table 2 for summary.

Figure 8

SNR in HbT concentration and Sat for each gas stimulus in fibroglandular tissue. ( ^*) denotes statistical significance. p-values and number of subjects in each group are given in Table 3.

Figure 9

Average SNR for HbT and Sat for the four subjects who received all three gas stimuli. Error bars indicate the standard deviation in SNR for each gas combination.