See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/51393421

A simple dose gradient measurement tool to

complement the conformity index. J Neurosurg

Article in Journal of Neurosurgery · December 2006

Impact Factor: 3.74 · DOI: 10.3171/sup.2006.105.7.194 · Source: PubMed

CITATIONS READS

84 379

2 authors, including:

Ian Paddick
Medical Physics Limited
16 PUBLICATIONS 520 CITATIONS

SEE PROFILE

Available from: Ian Paddick

Retrieved on: 18 April 2016
 J Neurosurg (Suppl 3) 93:219–222, 2000

A simple scoring ratio to index the conformity of

radiosurgical treatment plans

Technical note

IAN PADDICK, M.SC.

Gamma Knife Center, Cromwell Hospital, London, England

 A conformity index is a measure of how well the volume of a radiosurgical dose distribution conforms to the size
and shape of a target volume. Because the success of radiosurgery is related to the extremely conformal irradiation of
the target, an accurate method for describing this parameter is important. Existing conformity ratios and indices used
in radiosurgery are reviewed and criticized. It will be demonstrated that previously proposed measurements of con-
formity can, under certain conditions, give false perfect scores. A new conformity index is derived that gives an objec-
tive score of conformity for a treatment plan and gives no false scores.
An analysis of five different treatment plans is made using both the existing scoring methods and the new confor-
mity index.

KEY WORDS • gamma knife • radiosurgery • conformal planning • conformity index

HE quest for a simple and universal scoring system
T that measures the conformity of radiosurgical treat-
ment has been important but as yet not entirely suc-
cessful. A conformity index should be an objective mea-
sure of how well the distribution of radiation conforms to
the shape of the radiosurgical target. Indeed, radiosurgery
is only possible because of this conformity of the dose.
Without it, fractionation of the dose would be necessary as
a method of reducing lethal damage to normal tissue. The
conformity of a dose plan is crucial to the effectiveness of
the treatment and it should be expressed in a value that is
easy to understand. Not only is a conformity index invalu-
able in decision making when comparing two potential
treatment plans for the same patient, but it is also a useful
tool to compare different methods of radiosurgery such as
the gamma knife, LINAC, and particle accelerator. This
means an ideal conformity index should be universally
applicable, using easy to obtain parameters, usually from
a dose–volume histogram.
Shaw, et al.,5 proposed the PITV ratio as a measure of
the conformity of a treatment plan. However, this ratio has
a fundamental flaw that will be examined. Further scoring
methods are also reviewed and criticized. A new confor-
mity index is proposed, its derivation explained, and a
comparative study, with existing scoring methods, pre-
sented.
Abbreviations used in this paper: PI = prescription isodose;
PIV = PI volume; TV = target volume; TVR = TV ratio.
Materials and Methods
In the Radiation Therapy Oncology Group radiosurgery
guidelines in 1993, Shaw, et al.,5 proposed the PITV ratio
as one such conformity index. This is defined as the vol-
ume of the PIV divided by the TV:
PITV = PIV/TV. (1)
Larson (unpublished data) has suggested the TVR ratio
is a useful tool that is simply the inverse of the ratio above:
TVR = TV/PIV. (2)
These ratios are now routinely used at many radiosurgi-
cal centers. The PITV ratio has been used to compare the
conformity of treatments delivered among gamma knife,
conventional LINAC, and intensity-modulated radiother-
apy.1,4,6 Both ratios have the advantage that they are easy
to calculate; however, as will now be demonstrated, they
both have fundamental flaws.
By convention, a perfect conformity score is unity,
whereas a deviation from perfect conformity gives a high-
er or lower score. With the PITV ratio, a value greater than
one suggests overtreatment in most cases, whereas a value
of less than one suggests undertreatment. However, the
ratio has been criticized because it does not take into
account the location of the PIV relative to the TV.
The PITV ratio assumes that the center of the treatment
volume is always located at the center of the radiosurgical
target and that the TV is the same shape as the target.
Using the PITV ratio, a treatment plan would receive the
same perfect score of unity, whether the prescription iso-

J. Neurosurg. / (Suppl 3) / Volume 93 / December, 2000 219

 I. Paddick

FIG. 2. Diagram demonstrating the volume TVPIV (hatched

area).

This ratio can never yield a score greater than one (indi-
FIG. 1. Cross-section of four different three-dimensional treat- cating an overtreatment) as TVPIV can never be larger than
ment plans using the PITV ratio. The area shaded in gray is the
radiosurgical target. The dashed line is the prescription isodose. TV. The ratio only measures any undertreatment that
a–d: All have the same PITV ratio of 1.00. occurs, expressing it as the proportion of TV covered by
the PI. For a robust conformity index, we also need to
measure the proportion of tissue outside the target that is
being needlessly treated. This overtreatment ratio can be
described as:
dose surface traced the exact periphery of the TV or
whether the PIV missed the TV altogether. This is a seri- overtreatment ratio = TVPIV/PIV. (5)
ous flaw in the equation that could lead to a less suitable However, if we were to use this ratio alone as a measure
treatment plan being chosen if it were used to compare of conformity, it would be inappropriate as it can easily be
two potential plans. Figure 1 shows four different treat-
ment plans, all of which have the same PITV ratio. No one
could argue that all four plans have equal conformity.
The volume of the target covered by the PI (TVPIV) is an
essential parameter that needs to be included in any ratio
to give a true measure of conformity (Fig. 2). This can be
mathematically expressed as TV  PIV (the symbol 
denoting the volume of overlap between two structures).
By modifying the PITV ratio of Shaw, et al.,5 and substi-
tuting PIV with TVPIV the ratio becomes:
ratio = TVPIV/TV. (3)
Knoos, et al.,2 proposed exactly this ratio and modified
it to the inverse of Eq. 3.3 They applied it to conventional
radiotherapy and defined it as:
radiation conformity index = VPIV/VI, (4)
where VPIV is the TV and VI is the volume of VPIV covered
by the PI.
Although the modified PITV ratio will no longer give
false perfect scores for a radiation distribution distant
from the target, it has a new flaw: as long as the PIV cov-
ers the TV completely then TVPIV = TV and the ratio yields
a perfect score of unity. This is easily satisfied by a large
PI covering not only the target but also much of the sur-
rounding normal tissue (Fig. 3). Even whole-brain radia- FIG. 3. A nonconformal dose plan that satisfies the criteria for
tion would have a perfect conformity score. TVPIV/TV = 1.00.

220 J. Neurosurg. (Suppl 3) / Volume 93 / December, 2000

Scoring to index radiation conformity
FIG. 4. A nonconformal dose plan that satisfies the criteria for
TVPIV/PIV = 1.00.
FIG. 5. Comparison of the PITV ratio and the new conformity index for various treatment plans.
J. Neurosurg. (Suppl 3) / Volume 93 / December, 2000
satisfied by a small PI in the center of the target (Fig. 4).
We need to balance the proportion of the TV covered
(TVPIV/TV) with the proportion of the PIV inside the tar-
get (TVPIV/PIV). Combining these two ratios gives us:
conformity index = TVPIV/TV  TVPIV/PIV = TVPIV2/
(TV  PIV). (6)
The numerator (top) of the equation is squared as this
gives two volumes both in the numerator and denomina-
tor (bottom). This makes the equation volume indepen-
dent and gives a score of unity only for a perfectly con-
formal plan.
The lower the score is, the less conformity the plan has.
The index assigns undertreatment as being equally as bad
as overtreatment. An overtreatment of 10% gives a score
of 0.9, and a 10% undertreatment also gives a score of 0.9.
Converting the conformity index to a percentage would
make the value more intuitive. Thus a percentage confor-
mity index of 100% would mean 100% conformity, 0%
would mean a complete miss. A conformity index of 50%
would mean a 50% coverage of the target, or only 50% of
the PIV is inside the target, or more likely, a combination
221

FIG. 6. Dose–volume histogram and volume analysis tools are
used on GammaPlan to obtain the conformity index’s constituent
parameters.
of the two. Certainly the phrase “90% conformity” is eas-
ier to appreciate than a figure that deviates from a theoret-
ical perfect score of 1.00.
Results
Analyses of five different treatment plans were made,
demonstrating the advantages and disadvantages of differ-
ent scoring methods (Fig. 5). The two existing indices
give false perfect scores on at least one of the nonconfor-
mal plans. Unlike existing indices, the new conformity
index does not produce any false perfect scores and gives
an objective assessment of a treatment plan’s conformity.
Discussion
The proposed conformity index is the only equation that
gives a reasonable assessment of a plan’s conformity. All
other indices gave false scores in certain circumstances. In
addition, the index can be presented as a percentage of
conformity.
Although the index cannot tell the user whether the lack
of conformity is due to under- or overtreatment, the two
ratios that are multiplied to obtain the index tell us all we
need to know. These are basically an undertreatment ratio
(TVPIV/TV) and an overtreatment ratio (TVPIV/PIV). Ex-
ample 2 in Fig. 5 is a good description of this. The con-
formity index of 0.60 will not tell us if we have 60% tar-
get coverage or if 60% of the PIV is covering the target or
a combination of the two. However, if we substitute the
numbers into the under- and overtreatment ratios, we get
the figures 0.60 and 1.00, respectively. This tells us that
the lack of conformity is due to undertreatment and so an
enlargement of the PIV would produce a more conformal
plan.
222
I. Paddick
Because the index is split into its two constituent parts,
however, the comparison of two different plans is com-
plex. A single value for each plan is necessary, hence the
index being a product of the two ratios.
For gamma knife users planning with GammaPlan, Fig.
6 shows how the parameters that make up the index are
presented in the planning system. This takes less than a
minute to calculate and thus can easily be implemented
into the routine planning process.
Conclusions
Previously proposed conformity indices have been
shown to have major flaws. The new proposed conformi-
ty index is simple to understand and easy to calculate as
long as dose–volume histograms of the treatment are
available. Although the proposed scoring system does not
take into account the type of normal tissue irradiated (for
example, optic chiasm compared with cerebrospinal fluid)
or the dose gradients around the treatment volume, it is
nevertheless a valuable tool when an objective compari-
son between two treatment plans is required, measuring
the conformity and selectivity of the radiation dose.
Acknowledgments
I thank Dr. C. Lindquist for his many helpful discussions and D.
Shucksmith for her assistance in preparing this manuscript.
References
1. Khoo V, Oldham M, Adams E, et al: Comparison of intensity-
modulated tomotherapy with stereotactically guided conformal
radiotherapy for brain tumors. Int J Radiat Oncol Biol Phys
45:415–425, 1999
2. Knoos T, Kristensen I, Nilsson P: Results from clinical practice
of 3D conformal radiotherapy at Lund University Hospital, in
Leavitt DD, Starkshall G (eds): Proceedings of the Twelfth
International Conference on the use of Computers in Ra-
diation Therapy. Madison, WI: Medical Physics Publishing,
1997
3. Knoos T, Kristensen I, Nilsson P: Volumetric and dosimetric
evaluation of radiation treatment plans: radiation conformity
index. Int J Radiat Oncol Biol Phys 42:1169–1176, 1998
4. Ma L, Xia P, Verhey L, et al: A dosimetric comparison of fan-
beam intensity modulated radiotherapy with gamma knife ste-
reotactic radiosurgery for treating intermediate intracranial le-
sions. Int J Radiat Oncol Biol Phys 45:1325–1333, 1999
5. Shaw E, Kline R, Gillin M, et al: Radiation Therapy Oncology
Group: radiosurgery quality assurance guidelines. Int J Radiat
Oncol Biol Phys 27:1231–1239, 1993
6. Shaw E, Scott C, Souhami L, et al: Single dose radiosurgical
treatment of recurrent previously irradiated primary brain tu-
mors and brain metastases: final report of RTOG protocol 90-
05. Int J Radiat Oncol Biol Phys 47:291–298, 2000
Address reprint requests to: Ian Paddick, M.Sc., Gamma Knife
Center, Cromwell Hospital, Cromwell Road, London SW5 OHS,
England. email: ian.paddick@dial.pipex.com.
J. Neurosurg. (Suppl 3) / Volume 93 / December, 2000