SPI common mistakes
SPI · Understand the traps and how to avoid them.
When asked which control or parameter determines axial resolution, the candidate answers 'beam width' or 'focal zone,' and when asked about lateral resolution answers 'pulse duration' or 'number of cycles in the pulse.'
Both axial and lateral resolution improve as transmitted frequency rises, so the two blur together under exam pressure and the candidate substitutes whichever factor comes to mind first.
What to do instead
Anchor each resolution to the one physical quantity that governs it: axial resolution = spatial pulse length / 2 = (cycles x wavelength)/2, so it is set by frequency, bandwidth, and damping; lateral resolution = beam width, so it is set by frequency, aperture, and focusing. Use the classic groupings - axial/longitudinal/range/radial/depth (LARRD) versus lateral/azimuthal/transverse/angular (LATA) - and always attach the 'divide by 2' to the axial formula.
The candidate states that turning up overall gain or adjusting time-gain compensation increases the acoustic output delivered to the patient, or recommends lowering gain as the way to reduce patient exposure.
Both gain knobs visibly brighten the image, which makes them feel like 'more power,' and the exam deliberately pairs a display change with the exposure question.
What to do instead
Sort controls into two bins. Receiver-side (overall gain, TGC, compression/dynamic range, gray-scale map) changes only how returning echoes are displayed and has no effect on exposure. Transmitter-side (output power, pulse amplitude, PRF in some modes) changes exposure and moves the thermal and mechanical indices. For an exposure question, the only correct lever is output power / transmit intensity.
In calculation items the candidate solves for pulse repetition period or reflector depth without the round-trip factor of two, or mixes kHz with Hz, MHz with Hz, or milliseconds with microseconds and reports a value off by a factor of 2, 1000, or 10^6.
The exam supplies numbers in whichever units make the arithmetic tidy, and under time pressure the candidate plugs values straight into the formula without a unit-and-geometry check.
What to do instead
Before computing, write the formula with units attached and say the trip out loud: depth d = c x t / 2 because the pulse travels to the reflector and back; maximum PRF for a given depth = c / (2d); Nyquist limit = PRF / 2. Convert everything to meters, seconds, and hertz first, then convert the answer back. A quick sanity check (soft tissue at 1540 m/s means about 13 microseconds per centimeter of round trip) catches most order-of-magnitude errors.
The candidate measures a vessel at a Doppler angle of 70-85 degrees, leaves the angle correction at zero or sets it to the wrong value, and then reports the resulting low velocity shift as the true velocity.
The spectral display looks crisp at a steep angle because the sample volume stays in the vessel, so nothing signals that the cosine term has collapsed toward zero.
What to do instead
Remember the shift is proportional to the cosine of the angle: at 0 degrees the cosine is 1, at 60 degrees it is 0.5, at 90 degrees it is 0 (no detectable shift). Uncorrected or under-corrected measurements systematically underestimate velocity - an uncorrected 60-degree reading is half the true value. Keep the angle at or below 60 degrees, use a steerable or angled approach when the vessel runs perpendicular, and verify that the angle cursor parallels the vessel wall, not the beam direction.
To correct aliasing, the candidate recommends decreasing the PRF, increasing depth of field, or raises transmitted frequency, and also believes continuous-wave Doppler can alias.
Aliasing is remembered only as 'the waveform wraps around,' without linking it to the Nyquist limit and to the inverse relationship between PRF and imaging depth.
What to do instead
State the rule chain: aliasing occurs when the Doppler shift exceeds the Nyquist limit (PRF/2); the maximum PRF is limited by depth because PRF = c/(2d), so increasing depth lowers the available PRF and worsens aliasing; a higher transmitted frequency raises the Doppler shift and also worsens aliasing. Effective fixes are raising the PRF scale, shifting the baseline, lowering transmitted frequency, increasing the Doppler angle toward 60 degrees, using a higher-velocity scale preset, or switching to continuous wave - which has no range resolution but cannot alias because it has no sampling PRF.
The candidate places the Doppler sample volume against the vessel wall or straddling the wall, keeps a large gate in a small vessel, or tries to eliminate aliasing with the wall filter, and then misreads the resulting spectral broadening or low-velocity clutter as true flow.
Wall filter and sample-gate size both change the appearance of the spectrum, so their distinct purposes get merged; the candidate also forgets that flow profile depends on vessel size and branching.
What to do instead
Keep the sample volume at the center of the lumen and no wider than the vessel, because in laminar (parabolic) flow the center holds the highest velocities and the steepest, cleanest spectral edge; sample near the wall and you record the slow, broadened edge. Know that the wall filter removes high-amplitude, low-frequency signals from moving walls and tissue - it does not fix aliasing and it can hide genuinely slow flow. Distinguish parabolic laminar flow (small, straight vessels) from the flatter plug profile of large vessels and from the disturbed, broadened flow seen distal to a stenosis or at a bifurcation.
For quality-assurance items the candidate answers with generic pass/fail numbers or a fixed schedule instead of the baseline principle, and cannot match a phantom target type (dead zone, vertical or horizontal distance accuracy, axial or lateral resolution, uniformity, maximum depth of visualization) to the parameter it actually measures.
Quality assurance reads as administrative housekeeping, so it gets skimmed; the phantom targets all look like the same pins and cylinders in a description.
What to do instead
Learn the frame: the acceptance test on installation establishes the baseline, every later test is a comparison against that baseline rather than against a universal number, and testing repeats after transducer change, repair, or service. Then pair each phantom feature with its purpose: the near-field group of targets tests the dead zone, the measured spacings test vertical and horizontal distance accuracy, closely spaced target pairs test axial versus lateral resolution, the anechoic cylindrical targets test uniformity and beam/slice-thickness behavior, and the deepest visible target establishes maximum depth of penetration. Set gain and TGC reproducibly before measuring so the comparison is meaningful.
The candidate assumes one passing check covers the whole system, or that quality assurance is limited to image appearance, so they omit the individual transducer-by-transducer requirement, mechanical and electrical inspection, and documentation of results.
Quality assurance is pictured as 'does the picture look okay today,' which excludes the equipment and record-keeping components the exam tests.
What to do instead
Treat quality assurance as a documented program per transducer and per imaging preset, repeated on a schedule and after any repair, and covering equipment condition (cables, lens and face integrity, element dropouts on the uniformity image), electrical and mechanical safety, image quality metrics against baseline, and retained records that allow trend detection rather than a single yes/no verdict.
The candidate treats the thermal index and mechanical index as interchangeable, says the mechanical index measures temperature rise, or claims the mechanical index increases as transmitted frequency increases at the same pressure.
Both indices appear as small on-screen numbers that the operator is told to watch, so their distinct physical meaning and their opposite direction of dependence are easy to blur.
What to do instead
Separate them by mechanism and by formula. The mechanical index estimates cavitation risk and equals peak rarefactional (negative) pressure in MPa divided by the square root of frequency in MHz - because frequency is in the denominator, a higher frequency lowers the mechanical index for the same pressure. The thermal index estimates temperature rise and is the ratio of acoustic power to the power needed to raise tissue 1 degree Celsius, reported as soft-tissue, bone, or cranial variants. Know the regulatory ceilings commonly cited (mechanical index at or below 1.9, general thermal index at or below 6, ophthalmic thermal index at or below 1) and that both indices are reduced by lowering output power - not by adjusting receiver gain.
The candidate labels an artifact from its appearance alone - calling a comet tail 'shadowing,' calling ring-down 'reverberation from two strong reflectors,' calling acoustic enhancement 'through-transmission shadowing,' or failing to recognize that a propagation-speed error shifts a structure axially rather than laterally.
Artifacts are memorized as a list of names and pictures instead of as consequences of specific physical mechanisms, so near-misses become coin flips.
What to do instead
Group artifacts by mechanism and reason from the mechanism. Attenuation-based: shadowing behind a strongly attenuating or reflecting structure, and enhancement behind a weakly attenuating fluid. Reflection/reverberation-based: evenly spaced reverberation between two strong reflectors, comet tail from a small highly reflective structure, ring-down from resonant gas, and mirror image from a large curved reflector. Refraction-based: edge shadowing and duplicate images at curved interfaces. Assumption-violation based: propagation-speed error displacing a reflector along the beam axis, range ambiguity with high PRF, slice-thickness or beam-width partial volume, and grating/side-lobe clutter appearing off-axis. Also keep the intensity descriptors straight for safety questions: spatial peak temporal average and spatial average temporal average describe different exposures, and spatial peak pulse average is the highest of the set.