Units#
dcmri
uses a standardized system of units for all input and output
quantities. The system is internally consistent: as long as data and
parameters are entered in these units, all return values will automatically
also be in the same units. Hence it is not needed to track units through
intermediate computations, which greatly reduces the risk of unit conversion
errors in complex analysis pipelines.
Table of units#
The following table shows for each quantity the dcmri
unit, and wehere
relevant also other units that are commonly found in the literature on DC-MRI:
Quantity |
dcmri unit |
common alternative units |
---|---|---|
Time |
sec |
min |
Length |
cm |
|
Area |
cm2 |
|
Volume |
cm3 |
mL |
Angle |
deg |
rad |
Fluid volume |
mL |
|
Weight |
kg |
g |
Amount of indicator |
mmol |
mol |
Concentration |
M |
mmol/L, mM |
Tissue concentration |
mmol/cm3 |
mmol/L, mM |
Indicator flux |
mmol/sec |
|
MRI relaxation time |
sec |
msec |
MRI relaxation rate |
1/sec |
1/msec |
Magnetization |
A/cm |
A/m |
Contrast agent relaxivity |
1/sec*M |
1/mM/sec |
Perfusion |
mL/sec/cm3 |
mL/100mL/min, 1/sec, mL/sec/mL |
Permeability-surface area |
mL/sec/cm3 |
1/sec, 1/min |
(Fluid) volume fraction |
mL/cm3 |
mL/100mL, mL/100g, %, dimensionless |
Fluid flow |
mL/sec |
Rationale#
We have chosen to deviate from the system of SI units in order to align more closely to the natural units of most of the quantities, and historical conventions in the field. However we have also chosen to deviate from conventions in some places in order the arrive at an internally consistent set of units.
The use of a single systematic system of units minimises the risk of errors due to incorrect conversions between units in code, which are hard to track especially when different functions from different modules are assembled.
It does mean that some of the quantities are not in their natural units -
for instance tissue concentrations are typically in the order of 0.001 M and
are therefore commonly expressed in units of mM. Equally, contrast agent
relaxivities are conventionally expressed in units of 1/mM*sec. In dcmri
,
concentrations and relaxivities will nevertheless be required and returned in
units of M and 1/M*sec, respectively.
It is the responsability of the user to convert any output in more
conventional units for publication or presentation purposes. Equally, it
is the user’s responsability to convert any input data into dcmri
units
before applying any of the functions in the package.
Units of volume#
Fluid volumes of mL and volumes of cm3 are of course numerically
interchangeable. dcmri
nevertheless distinguishes explicitly between
(tissue) volumes (cm3) or voxel volumes (cm3) and fluid volumes (mL) to avoid
confusion in the interpretation of important quantities such as perfusion or
concentration.
In the dcmri
system of units, perfusion is expressed in mL/sec/cm3. This
directly reflects the physical interpretation as volume of blood (mL)
delivered per unit of time (sec) to a unit of tissue (cm3). If tissue
volumes were to be expressed in units of mL instead of cm3, the units of
perfusion would be mL/sec/mL. This looks odd and is often reduced to units
of 1/sec, cancelling out the mL’s. In this form the units suggest that
perfusion is a rate, which is physically confusing.
The units of mL/sec/cm3 also align more directly to the historical units of perfusion as mL/sec/g. They reflect the original practice of first measuring whole-organ blood flow (mL/min), excising and weighing the organ, and dividing out the weight to arrive at organ perfusion. In an imaging context normalising to weight is awkward as voxel volume is known, but voxel weight is not. Converting to historical units of mL/min/g would require inserting of literature values for tissue density in all models, and keeping track of these throughout computations. This is cumbersome and unnecessary, since conversion to mL/sec/g can always be done at the end of computations on the final output.
The use of cm3 for tissue volumes also ensures that important distinctions
such as that between concentration and tissue concentration, are explicit.
The term concentration in dcmri
documentation is consistently used to
designate amounts of indicator (mmol) relative to fluid volumes (mL), whereas
the term tissue concentration refers to indicator amounts (mmol) per unit
tissue volume (cm3). They are different physical quantities and in an imaging
context the distinction is important.
Units in plots#
While dcmri
rigorously adheres to standard units for all arguments and return
values in functions - exceptions are made in plots or other presentations of
data and results. Here dcmri
will choose the units that are most natural and
intuitive in conveying the key messages. For instance concentrations in plots
will typically be shown in units of mM.