Leaders in Pharmaceutical Business Intelligence (LPBI) Group

Cardiovascular Imaging

Author: Justin D. Pearlman, MD, PhD, FACC

WORK IN PROGRESS

Multiple physical means are used for cardiovascular imaging: ultrasound, electromagnetics, radioactive agents, xrays and optical.

Ultrasound vibrations pass waves through tissue by exciting an array of piezo-electric crystals (e.g., lead zirconate) with alternating current. The speed of sound wave propagation through tissue is ~1540 meters/second, and the frequency of vibration used for imaging ranges 1-20 Megahertz (million cycles per second). Higher frequencies can report finer details but they attenuate faster (signal drop out with depth).  Sound waves are reflected back by tissue interfaces that represent a change in sound transmission impedance.

In its simplest mode, called A-MODE, the amplitude of return signal is displayed as a function of time. The A-MODE signal for a beam of sound passing through the center of an orange immersed in water will return an A-MODE signal peak for the time of the round trip to the first contact with the orange, another for transition from orange peal to the inner pulp, and another pair for the pulp to peel and peel to exodus at the far side of the orange. Immersion in water solves the problem of 99% attenuation (stoppage) of sound propagation at an air-fluid interface (very high impedance). Thus the distance between the first two peaks reports peel thickness, and the inner distance between the paired peaks from peel report the inner diameter of the orange.

The next more advanced mode of ultrasound imaging, called B-MODE, converts the amplitude of returned sonic energy to a bright dot, so that the display of data from a linear probing of tissue by sound wave reflections is reported on a screen as a line with bright spots for the crossing of each tissue interface. Multiple B-MODE displays can be combined at different positions corresponding to the angle and position of the sonic probe to construct composite mode of display, called C-MODE. Aimed at a pregnant abdomen, the result shows contours of fetal head and limbs, which produced the first dynamic fetal two-dimensional images. The excitement generated a mechanical mimic of a sweeping series of angulations, called SECTOR SCANNING. In sector scanning, the sonic probe pathway sweeps back and forth like a windshield wiper, to produce a triangular 2D image called a sector.

With the invention of phased arrays of sonic crystals, a combination of small sources staggered in time produces a composite sonic beam that is electronically steered by adjusting the timing of activation, so that sector scanning can be accomplished electronically without mechanical moving parts. A linear array is now the preferred means to perform sector scanning. Furthermore, with a 2D phased array, the sonic beam can be swept both vertically and horizontally, to build 3D images (4D if you include the time dimension). Current clinical ultrasound uses phased arrays, sector scanning, and optional 3D views that are acquired within one heartbeat (one cardiac cycle).

Johannes Doppler described a shift of phase due to motion of a wave sensor: if the detector is moving towards a source of sound, it will pick up more peaks per second the faster it approaches. The difference between the number of peaks per second (frequency) with the sensor not moving, and the higher frequency observed with the sensor approaching the source at a velocity V,  is called frequency shift (Doppler shift). The frequency shift is proportionally to the rate of approach (component of velocity towards the source of sound waves).  With red blood cells moving through heart valves. Doppler shift reports their velocity towards the sound wave sensor (the ultrasound transducer). The kinetic energy is easily derived (KE=½ m v²), and the pressure drops across a valve or a vessel narrowing (stenosis), which is substantially from the change in kinetic energy, is thus estimated by 1/2 m V² = 4 V² for blood cells reporting mmHg pressure change.

Color Doppler display of ultrasound modifies gray images by applying a red-blue scale to image data according to the frequency shift, with red indicating velocity towards the transducer, and blue away. Thus blood movement can be visualized within the heart, as color contained in the gray walls.

A simple graphic plot of computed velocity versus time for a linear beam of ultrasound is called “continuous wave”. The maximum height on such a plot reports the maximum pressure gradient observed along the beam. Thus continuous wave Doppler can find the peak gradient across a narrowed (stenotic) aortic valve to help determine the severity of narrowness. Seriously small aortic valve area with an otherwise normal heart, meriting valve replacement surgery even at age 90, generally produce a peak gradient well above 50 mmHg.

It is tempting to correct the amplitude according to the cosine of the angle of the beam to the direction of blood flow, and many textbooks and ultrasound machine manufacturers manuals have incorrectly recommended that. The error is that the red blood cells reporting velocities are individual, so the sonic image is a bit like an image of a swarm of bees. The individual elements have different directions of motion, so changing angles to see less of some sees somewhat more of others, making the fall-off with angulation less than the cosine law predicts (the cosine rule overcompensates).

The Doppler interrogation may be “pulsed” (briefly “on” then “off” in a repeated cycle separated by a time gap). Then the return signal shift will represent the velocity corresponding to the timed round-trip transit, or a multiple thereof. Thus, the velocity sampling can be limited to a specified sampling location depth (and multiples thereof) rather than summarize the peak of all velocities encountered along the entire beam transit. By setting the depth to the visualized depth of the aortic valve, velocity can be sampled there, then the depth can be reduced to sample the velocities in the outflow track just before the aortic valve, for comparison. The change in estimated kinetic energy estimates the pressure drop due to the valve stenosis (narrowness).

Estimation of the peak velocity of blood leakage through an austensibly closed tricuspid valve reports the pressure drop across the tricuspid valve, i.e., pressure drop from the right ventricle (RV) to the right atrium (RA) during peak right ventricular contraction is estimated from the peak velocity of tricuspid valve regurgitation. If the RA pressure is assumed to be 10 mmHg and the 4 V² reports a pressure stop of 50 mmHg, then the systolic pressure in the RV is estimated to be 50 mmHg. In systole (contraction time), the pulmonary valve (PV) is open, so if that is a normal passage, then there will be no significant step across the PV and the pulmonary artery peak systolic pressure will be well represented by the RV peak systolic pressure. High pulmonary pressure values impair blood delivery from right to left, and if not treated, can lead to death within 2 years. A silent killer more easily missed than arterial pressure, because the pulmonary pressure is not accessible by a blood pressure cuff.