healthmedicinet

A total of 377 CBCT scans were reviewed in 40 post-prostatectomy patients (median
number of CBCTs per patient was 9, range of 8 to 11). Forty-five planning CT scans
were reviewed.

PTV size

The 0.5 cm uniform expansion had the smallest PTV volume (median?=?222.3 cc), followed
by smaller anisotropic (238.9 cc), NSCC (331.5 cc), 1?+?0.5 cm posterior (337.7 cc),
van Herk (354.1 cc), and 1 cm uniform (361.7 cc).

Volume of bladder and rectum inside the PTV

The volume of bladder and rectum inside each PTV is shown in Fig. 1. The NSCC PTV had only 1 % more bladder in the PTV than the 1 cm uniform expansions
despite having a 0.5 cm larger PTV expansion anteriorly. The larger expansions of
the van Herk PTV covered much of the rectum. The NSCC PTV had 5 % less rectum in the
volume than the 1 cm uniform PTV although the upper prostate bed margin is the same,
due to the decreased posterior expansion in the lower prostate bed.

Fig. 1. Rates of geographic miss with volume of bladder and rectum inside the PTV using bony
anatomy matching. The percentage of all bony anatomy matched images that showed a
geographic miss for each PTV compared to the percentage of the bladder and rectum
located inside the PTV

Rate of geographic miss using different matching techniques and PTVs

The rate of geographic miss using three different matching techniques and five different
PTV expansions (as previously described, excluding the van Herk margin) is shown in
Fig. 2. The van Herk margin was excluded because this PTV covered a prohibitive amount of
the rectum, so could not be implemented clinically. With all matching types, rates
of geographic miss increased as the PTV expansions became smaller.

Fig. 2. Geographic miss rates using different matching techniques and PTV expansions. The
percentage of images displaying a geographic miss for three matching techniques (bone,
superior soft tissue, and average soft tissue) for different PTV expansions

BA matching resulted in the highest geographic miss rate for all PTVs, followed by
SST matching then AST matching (see Fig. 2). Changing from a BA to an AST match decreased potential geographic miss by half
to two thirds, depending on the PTV expansion. Changing soft tissue matching from
the SST to the AST technique resulted in an approximate 50 % decrease in geographic
miss for all PTVs except the 1 cm uniform expansion. The AST matching technique reduced
geographic miss rates to below 10 % for all PTV expansions tested and to less than
6 % for four of the PTV expansions. Using the anisotropic NSCC PTV expansion with
the AST matching technique would result in a 3.2 % geographic miss rate compared to
9.3 % with BA matching. The smaller anisotropic PTV expansions resulted in a slightly
higher rate of geographic miss of 5.6 % (a 2.4 % increase) than the NSCC PTV expansion
when using the AST match.

Location of soft tissue geographic misses using BA or AST matching

The AST match reduced most lateral geographic misses compared with BA matching for
all PTV expansions (see Fig. 3), with the majority of geographic misses, although reduced in number, in the central
area of the PTV, especially in the upper prostate bed. AST matching resulted in a
smaller increase in anterior geographic misses in the lower prostate bed than with
BA matching, but the overall number of geographic misses was reduced.

Fig. 3. Area of prostate bed where soft tissue geographic miss occurred. Schematic diagrams
of the areas of prostate bed where soft tissue geographic miss occurred for the 0.5 cm
uniform (a f), 1 cm uniform (b g), 1?+?0.5 cm posterior (c h), NSCC (d i), and smaller anisotropic (e j) PTV expansions when bony anatomy matching (left column) or averaged soft tissue matching (right column) was used. Numbers refer to total percentage of misses in each area over 377 images

Combined evaluation of the PTV expansions and image guidance matching techniques

When evaluating PTV expansions, both the rates of geographic miss and the volume of
critical structures treated were considered (Figs. 1 and 2). The van Herk PTV expansion had the smallest rate of geographic miss using BA matching
(4.2 %). However it included the largest amount of bladder (28.0 %) and rectum (36.0 %),
which would make it difficult to implement clinically, so it was eliminated from the
study prior to assessing soft tissue matching. The 0.5 cm uniform PTV expansion treated
the least bladder (17.1 %) and rectum (10.2 %) but had the highest geographic miss
rate (BA?=?28.4 %, SST?=?19.9 %, AST?=?9.8 %). The 1 cm uniform PTV expansion treated
more rectum (25.0 %) than the NSCC PTV (20.0 %), but only decreased geographic miss
by 1.3 % compared to the NSCC PTV. The 1?+?0.5 cm posterior PTV expansion covered
a similar amount of bladder (25.7 %) to the NSCC and 1 cm uniform PTV expansions with
less rectum (15.0 %), but it had nearly double the rate of geographic miss of the
1 cm uniform and NSCC PTV. The NSCC PTV also had a smaller median volume (331.5 cc)
than the 1 cm uniform (361.7 cc) and 1?+?0.5 cm posterior (337.7 cc) expansions. With
a geographic miss rate of 9.3 % with BA matching, this PTV expansion might be optimal,
but geographic miss rate improved to 3.2 % with AST matching. The smaller anisotropic
PTV expansion was similar to the 0.5 cm uniform PTV expansion for BA matching geographic
miss and bladder and rectal volumes, but the geographic miss rate decreased to 5.6 %
with AST matching. This expansion has the benefits of decreasing normal tissue and
critical structure dose compared with the NSCC volume, with only a small increase
in geographic miss.

Discussion

Intensity modulated radiotherapy (IMRT) treatment techniques are desirable in the
post-prostatectomy setting due to the ability to produce highly conformal plans with
rapid dose fall off 10], enabling decreased dose to be delivered to surrounding critical structures such
as the bladder and rectum. They are, however, susceptible to movement and geographic
miss. Current published PTV guidelines assume uniform prostate bed movement, but recent
findings of non-uniform prostate bed movement 3] support an argument for anisotropic PTV margins. The good soft tissue definition
of CBCT imaging 11] could enable a change in image verification processes to a daily soft tissue matching
approach which would further increase treatment accuracy and enable geographic misses
to be identified and corrected before treatment.

Our study supports the theory that anisotropic expansions are the optimal PTV expansion
in the PP-IMRT setting when BA, SST, or AST matched interfraction imaging is used.
The size of the anisotropic PTV expansion will depend on the imaging matching technique
applied.

If BA matching is used, the NSCC PTV expansion would be the best to use. It was the
third smallest in volume (median?=?331.5 cc), but treated a similar amount of bladder
to the 1 cm uniform and 1?+?0.5 cm posterior PTV expansions, with the fourth smallest
amount of rectum overall, and a BA match geographic miss rate of 9.3 % of images.
With BA matching, the 1 cm uniform PTV covers more rectum but only decreases geographic
miss by 1.3 %, and the 1?+?0.5 cm posterior PTV covers less rectum but nearly doubles
the geographic miss rate.

When AST matching is used, the NSCC PTV expansion produces the lowest geographic miss
rate of 3.2 %. Moreover, application of the AST technique could lead to a reduction
in PTV expansions, such as the smaller anisotropic PTV expansion, which resulted in
a 2.4 % increase in potential geographic miss of 5.6 % compared to 3.2 % with the
NSCC PTV expansion. Using the smaller anisotropic margin also has the added benefit
of reduced critical structure volumes inside the PTV compared to the NSCC PTV, with
bladder reducing from 26.4 % to 19.3 % and rectum reducing from 20.0 % to 16.6 %.
All matching techniques enable detection of geographic miss before treatment delivery
if CBCT scans are taken. The averaged nature of the AST match also gives more flexibility
to facilitate inclusion of all the soft tissue inside the PTV for treatment. The decreased
dose to the surrounding critical structures achieved by the use of the smaller anisotropic
PTV expansions could also enable dose escalation. Additionally, there is potential
for a more adaptive radiotherapy approach with tailored margins for specific patients,
although this would require further investigations into the predictive nature of geographic
miss.

The bladder and rectum volumes inside each PTV were used to assess if planning critical
structure limitations could be met. The planning protocol used in our department is:
PTV (D95%?=?100 % of Target Dose), rectum (V6500cGy??17 %, V6000cGy??20 %, V4000cGy??40 %),
and bladder (V6500cGy??25 %, V4000cGy??50 %). The NSCC PTV expansion is used clinically
in our department and these planning objectives can be met in the majority of cases.
Although no dosimetric analysis was carried out, it is clear that the volumes of rectum
and bladder inside the 1 cm uniform and van Herk PTV expansions would make it difficult
to meet these dose constraints.

Two-thirds of local recurrences after surgery occur at the anastomosis (of these 60 %
were posterior, 20 % anterior, and 15 % lateral), 17 % in the retrovesical space,
10 % in the bladder neck and 10 % elsewhere 6]. Our data shows that, for all matching techniques, the most common areas of potential
geographic miss coincide with these areas of highest risk of recurrence, i.e. posteriorly
in the lower prostate bed and in the central upper prostate bed. Soft tissue matching
does not change the area where geographic miss occurs, but it does result in reduced
potential misses. It is also possible that the soft tissue matching process would
result in the Radiation Therapists noticing a geographic miss prior to treatment delivery
so intervention such as bladder or rectum emptying could occur. Daily CBCT soft tissue
matching and/or the addition of implanted markers 12] could therefore reduce the occurrence of geographic miss, allowing smaller PTV volumes
to be used.

It should be noted that these patients followed a bladder and rectal filling protocol.
Patients were simulated with a comfortably full bladder and empty rectum, with an
enema administered if the rectal diameter was greater than 3.5 cm. Enemas were not
given during treatment but the same bladder and rectal preparation was used. If the
PTV expansions and IGRT protocols recommended here were to be adopted the same bladder
and rectal preparation would be required.

This study had some limitations. Although 377 CBCT images were reviewed, they only
belong to 40 patients. Reviewing a larger number of patients could produce more robust
results and give more information on the factors predictive of geographic miss.

The time it would take to acquire daily CBCT scans and to complete the soft tissue
matching has not been assessed here. CBCT acquisition takes longer than standard orthogonal
imaging acquisition and there could be a difference in the time it takes to complete
the different matching techniques. This might increase the time required to treat
each patient, at least during the initial implementation phase. Radiation Therapist
training would also be necessary, but they have been shown to be adept after using
a locally produced image atlas 13], 14].

CBCT imaging delivers more radiation dose to the patient than daily kV/kV imaging.
It is estimated that a pelvic mode CBCT scan delivers 17.7 mGy compared with 3.54 mGy
for kV/kV imaging 15]. This increased imaging dose needs to be considered before deciding to implement
a daily CBCT imaging policy, but the higher accuracy within the context of the treatment
dose probably outweighs this disadvantage. Departments also should ensure that the
image quality of their CBCT scans is adequate for soft tissue matching. Structure
identification on CBCT scans was assessed at NSCC and it was found that with training,
the anterior rectal wall, posterior bladder wall and prostate bed surgical changes
were well recognised on the majority of scans 14].

PTV expansion reduction should be done with caution and all errors that the PTV accounts
for should be assessed before doing so. We used CBCT scans to assess motion, which
only accounts for interfraction motion, and the effect of intrafraction motion has
not been considered, although this is likely to occur during PP-IMRT. Klayton et al.
12] conducted an intrafraction motion study using radio-frequency transponders and found
that bladder and rectum variability caused deformation of the prostate bed. The 5 mm
tracking limit was exceeded for at least 30 s in 11 % of all fractions, and 15 % of
all treatments were interrupted for repositioning. Intrafraction motion should therefore
be considered before reducing PTV expansions.

Another limitation of this study is that only geographic miss was assessed and not
the encroachment of the PTV coverage onto surrounding critical structures. Geographic
miss could be avoided in some CBCT scans but it resulted in an unacceptable amount
of the rectum potentially being treated (see Fig. 4). Training of the Radiation Therapists is essential for online soft tissue matching.
When completing an online soft tissue match, Radiation Therapists should assess both
the coverage of the prostate bed and whether the coverage of the critical structures
is too great. If the critical structures are included too much in the PTV, intervention
can occur before treatment is delivered, such as filling or emptying the bladder,
emptying the rectum, or rescanning and replanning.

Fig. 4. Increased rectum inside PTV. Geographic miss could be avoided using soft tissue matching
in a number of instances. A planning CT (left) and CBCT scan (right) are shown in this example. In this case averaged soft tissue matching would have
allowed a geographic miss to be corrected using the 1 cm uniform PTV expansion (black line), but a large amount of rectum would have been treated. The white line is the CTV.
Critical structure avoidance therefore needs to be reviewed when matching online