· web viewimage registration in veterinary radiation oncology: indications, implications and...
TRANSCRIPT
IMAGE REGISTRATION IN VETERINARY RADIATION ONCOLOGY: INDICATIONS,
IMPLICATIONS AND FUTURE ADVANCES
Yang Feng1, 2, Jessica Lawrence2, Kun Cheng1, Dean Montgomery1, Lisa Forrest3, Duncan B.
McLaren1, Stephen McLaughlin4, David J. Argyle2, William H. Nailon1
1The University of Edinburgh, Department of Oncology Physics, Edinburgh Cancer Centre,
Western General Hospital, Edinburgh, UK. 2The University of Edinburgh, Royal (Dick) School
of Veterinary Studies and Roslin Institute, UK. 3The University of Wisconsin-Madison,
Department of Surgical Sciences, 2015 Linden Drive, Madison WI, USA. 4Heriot-Watt
University, School of Engineering and Physical Sciences, Edinburgh, EH14 4AS, UK.
Keywords: Oncology, Image Registration, Computed tomography, Radiation therapy, Magnetic
resonance imaging
Running Head: Image Registration in Radiation Oncology
Funding Sources: This work was generously supported by NHS Lothian, Edinburgh and
Lothians Health Foundation (charity number SC007342), the James Clerk Maxwell Foundation,
the Jamie King Uro-Oncology Endowment Fund, a University of Edinburgh Darwin Award and
a University of Edinburgh Individual Stipend.
Previous Abstracts: Material has not previously been presented.
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
123456789
Abstract
The field of veterinary radiation therapy (RT) has gained substantial momentum in recent
decades with significant advances in conformal treatment planning, image-guided radiation
therapy (IGRT) and intensity-modulated (IMRT) techniques. At the root of these advancements
lie improvements in tumor imaging, image alignment (registration), target volume delineation,
and identification of critical structures. Image registration has been widely used to combine
information from multi-modality images such as computerized tomography (CT), magnetic
resonance imaging (MRI) and positron emission tomography (PET) to improve the accuracy of
radiation delivery and reliably identify tumor-bearing areas. Many different techniques have
been applied in image registration. This review provides an overview of medical image
registration in RT and its applications in veterinary oncology. A summary of the most commonly
used approaches in human and veterinary medicine is presented along with their current use in
IGRT and adaptive radiation therapy (ART). It is important to realize that registration does not
guarantee that target volumes, such as the gross tumor volume (GTV), are correctly identified on
the image being registered, as limitations unique to registration algorithms exist. Research
involving nNovel registration-frameworks for automatic segmentation of tumor volumes is
ongoing and comparative oncology programs offer a unique opportunity to test the efficacy of
proposed algorithms.
2
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
101112131415161718
Introduction
Imaging data from multiple anatomical and functional imaging studies is becoming a routine
component of veterinary patient management for a variety of medical and surgical conditions.
Spanning from initial diagnosis to determining therapeutic options to assessing response or
recrudescence of disease, these data help direct decisions regarding disease management,
efficacy of treatment and patient outcome. While computed tomography (CT) and magnetic
resonance (MR) imaging are readily available in veterinary medicine, novel imaging techniques
such as molecular imaging offer complementary information to aid in disease recognition, extent
and treatment planning. In order to optimally use information gathered from these various
techniques, the data must be easily compared despite differences in image acquisition and
presentation. Image registration is defined as the process of aligning two or more images from
the same or different imaging modalities to allow for data mapping (data fusion) and
interpretation. While registration carries importance across imaging disciplines, the purpose of
this overview review is to provide a broad overview of image registration and data fusion
techniques used in radiation therapy (RT), given the high frequency with which it is utilized and
radiation oncologists’ reliance on registered images. While RT is the focus of the review, the
principles outlined for registration techniques can be broadly applied to non-oncologic
applications.
It is estimated that greater than 60% of human patients with cancer will receive RT and
although there is no similar strong data in dogs and cats, RT has become a common therapeutic
3
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
192021222324252627
modality in veterinary oncology.1 There is a large body of evidence supporting the use of RT for
optimal local tumor control, therefore identification of the tumor volume is of utmost importance
when determining a treatment plan or evaluating changes in tumor volume. In practice, medical
images acquired from different imaging modalities are used to guide the entire RT process from
the initial treatment plan to fractionated radiation delivery through to dose verification. While
radiation oncologists have always been guided by images in some form, the advent of image-
guided radiation therapy (IGRT) has revolutionized how integral images are to modern radiation
oncology. In a broad sense, IGRT may reflect any aspect of RT that utilizes imaging to improve
treatment, such as weekly or twice weekly portal imaging. In a stricter sense, IGRT refers to
contemporaneous functional and structural imaging to improve target delineation, adjust for
target motion and/or uncertainties in patient positioning, and potentially adapt treatment to the
response of the tumor during adaptive radiation therapy (ART).2 Often without recognition of the
process involved, radiationWhile radiation oncologists are highly dependent on image
registration to ensure that IGRT is successful atin targeting tumor and limiting dose to adjacent
normal tissue, the underlying process is complex and often overlooked. A typical IGRT process,
using veterinary images as an example, where image registration is vital in order to merge
information from multi-modality images and therefore provide an accurate guide for radiation
delivery is illustrated in Fig. 1.
Image registration provides a geometric transform that makes it possible to map
information between the images, often with sub-pixel resolution. As the use of multimodality
4
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
282930313233343536
image data in RT increases, medical image registration is essential to combine the information
from each modality. As a result, it has become a very active area of research.3 In RT, Iimages
utilized in RT can be registered to obtain comprehensive information whether theyregardless of
patient positioning, time point with respect to therapy, or type of imaging acquisition. were
obtained on patients in dissimilar positions or immobilization devices, from different points of
view, at various time points, or by different sensors. Registration may also be used to combine
multiple images from the same imaging modality,4 or to combine information from multiple
modalities such as CT, MR imaging, positron emission tomography (PET) and single-photon
emission computed tomography (SPECT).5 For the purpose of this review, athe reference image
is defined as the source image containing the reference information, while athe target image is
defined as the movable image spatially matched to the reference. After registration, information
from the reference image, for examplesuch as contoured structures, can be used on the target
image.
Image registration methods have been widely used both in human medicine6-11 and in
preclinical studies12-18 to improvefor matching between images acquired on different modalities
such as PET and-MR,19 SPECT and-CT,20 and PET and-CT.21 Registration approaches applied to
humans are also suitable for companion animals.22-31 and may prove useful in comparative
oncology, which refers to the study of cancer etiology, biology and treatment in companion
animals that develop spontaneously-occurring cancer.1, 32-34 Comparative oncology approaches
presents a unique opportunity to improve RT for both human and veterinary patient groups given
5
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
373839404142434445
the number of similarities in tumor biology, imaging techniques and therapeutic
managementmodalities. Image registration approachesmethodologies, regardless of the species
imaged, will help facilitate this by finding the optimal geometric transformation between
corresponding image data.35
In general, an image registration algorithm can be divided into four parts: image
conditioning, geometric transformation, similarity function and optimization (Fig. 2). Any step
that alters the original data to make it more suitable for applying image registration is known as
image conditioning. Due to the diversity of the methods used for this, a comprehensive review of
image conditioning falls outside the scope of this manuscript, but may be found elsewhere.5, 36-37
Image Registration
Table 15 provides a comprehensive list of the most commonly used terminology in image
registration.5 A summary of the kKey technical details of the transformation function, similarity
function and optimization process (Fig. 2) are introduced in this section.
Transformation
As it is common for the patient orientation and immobilization to vary between imaging studies,
the first and keyessential component of registration is transformation. This describes the
geometrical shift that is required for the target image to match it to the reference image. Rigid
transformation methods are most commonly used and are usually applied on images that have no
6
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
464748495051525354
distortion, such as CT to CT. Generally, a rigid geometric transformation can be achieved by
translation and rotation and is commonly used to match between bones on medical images, such
as the skull.38-39 An affine transformation, which is an extension of the rigid transformation,
allows for rotations, translations, scaling and shearing.40 Rigid transformations can also be
applied before a non-rigid registration as an image conditioning step.41-42
The assumption that rigid movement of anatomy occurs globally is incorrect in any
number of situations, therefore limiting the widespread use of rigid registrations for sites other
than the head.40 Most non-rigid registration methods are based on deformable models. These
range in complexity and may be veryvary from simple with relatively few parameters toor more
complicated where each point or voxel moves independently.40 There are two directions in
deformable registration: free form deformable registration (often abbreviated FFD) and guided
deformable registration, which are controlled by models based on prior knowledge of the
registered objects or organs. The fundamental difference between them is that the free form
deformable registration allows any deformations by moving the positions of its control grid. The
most commonly used guided deformable registration methods are elastic-based and flow-based.
Elastic-based methods treat organs as elastic solids and define two forces: internal forces that
oppose the deformation and external forces that try to deform the images. The best transform can
be obtained by finding equilibrium between both internal and external forces. Flow-based
methods such as include fluid flow and optical flow . These methods treat the registration
problem as a motion problem and therefore achieves the best match is achieved by meeting a
7
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
555657585960616263
pre-set constraint in a physical model. Fig. 3 shows tThe application of the popular B-spline
based FFD method registering pre- and post-treatment CT images of a dog with a sinonasal
tumor is demonstrated in Fig 3.43 A similar method called thin-plate spline (often abbreviated
TPS) can also generate FFD registration. However, compared to the B-spine method this
approach is limited by the fact that the deformation is applied globally. 44
Similarity Function
The transformation, or alignment, of the image datasets is assessed by measuring the similarity
between them.40 Generally, tThere are two predominant kinds of similarity measurements:
intensity-based and feature-based measurements. The most popular intensity-based similarity
function is based on intensity difference or intensity correlation. Feature-based similarity
functions depend on the feature structure extracted, such as anatomic structures (bones) or
artificial landmarks (fiducials). Most feature-based algorithms use points, lines, or surfaces for
matching.40, 45-46 A minimum of three or four pairs of points are required in order to compute the
rotations and translations for a rigid or affine transformation.40 Unlike point matching, line and
surface matching do not require a one-to-one match between images but rather attempt to
maximize the overlap between equivalent lines and surfaces such as the skull surface, ribs or
pelvis.40, 45-46 Fig. 4 illustrates tThe transformation process and the corresponding similarity
calculation used in the registration of images from a dog with a nasal tumor is illustrated in Fig
4.
8
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
646566676869707172
Optimization
Optimization is the process used to search for a numerical value produced by the similarity
function, often the minimum value, which is indicative of when the best registration, or match,
between the images has been found. The goal of the optimization algorithm is to find a
maximum or minimum value of the similarity measure accepted; it is common for optimal
registration to be accomplished when the similarity measures are defined by their minimal value.
It is an important step in a registration method because the similarity function can often produce
several local minima, which will trap the search and giveproduces sub-optimal results.
Assessment
The obvious motivation for registering images from different studies is to map clinically useful
information from one study onto another, for example in treatment planning prior to radiation
delivery. If data can be successfully fused, radiation oncologists can then map tumor volumes
such as the GTV, clinical target volume (CTV) and planning target volume (PTV) or organs at
risk (OAR) directly onto the CT needed for dose calculations (structure mapping).40 Therefore,
relying on the ability of a bespoke software system to adequately register images is of utmost
importance in order to adhere to current radiation guidelines – namely to appropriately outline
tumor volumes andto limit normal tissue toxicity. In this vain, it is important to realize the
difficulties in accurately validating the performance of a complex registration method. In some
9
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
737475767778798081
cases, a registration algorithm can be assessed using phantom data,47 but this is not commonly
done as performance on phantoms cannot ensure comparable performance on clinical cases. A
nother commonly used approach is to use synthetic images on which manual definition of
corresponding points such as fiducial markers can provides a straightforward assessment of a
registration method.48-49 The Dice coefficient, which measures the overlap between clinical and
registered contours, placesuts a numerical value on the registration accuracy.50-51 Within the
range 0% to 100%, a Dice value of 100% indicates excellent geometric agreement while a Dice
value of 0% represents poor geometric agreement (Fig. 5). The Jaccard index and Tanimoto
coefficient (both vary from 0% to 100%), which can be calculated directly from the Dice
coefficient, are also used to measure the similarity between regions. However, they need prior
knowledge (contours of the same region) on both reference and target images.
Once the registration process has been completed, there are a number of techniques that
can be used to visualize fused data, including the use of overlays, pseudo-coloring, grey scale
coloring, and side-by-side display of anatomic planes.40 This not only allows the radiation
oncologist to evaluate the fused images but can also be used forand to map ping of the calculated
3D dose distributions fromon the CT treatment plan to the coordinate system of another imaging
study, such as MR. Not only is this helpful to define the tumor and normal tissue dose for tissues
that are not optimally imaged with CT (such as canine or feline intracranial tumors), but the
fusion of 3D dose distributions to post-treatment imaging scans to improve detection and
understanding of radiation changes over time (Fig. 5). As veterinary radiation oncology advances
10
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
828384858687888990
into the realm of IGRT and ART with the use of volumetric imaging systems (on-board
kilovoltage CT or megavoltage CT units), the concept of accurate registration becomes
increasingly criticalmore important to provide input into the ART process.
Current State of Image Registration in Veterinary Radiation Oncology
Image registration and data fusion are useful in veterinary radiation oncology with most
modern treatment planning systems now facilitating the use of a secondary datasets for target and
OAR delineation. In addition iIt is now also possible to map dose information to the secondary
dataset following an appropriate registration.
Dose distributions generated for conformal radiation therapy (CRT) have been accurately
predicted with CT-based treatment planning systems and are already widely used in veterinary
oncology.53 CT data includes the Hounsfield units as a linear transformation of radiation beam
attenuation that varies with electron densities of materials as the beam progresses through tissue.
Most RT planning software obtains the relative electron density from the relationship between
the linear attenuation coefficients and Hounsfield unit values in order to determine heterogeneity
in the dog or cat’s body. Non-contrast CT images are typically utilized for treatment planning,
given that contrast agents are high-Z radio-opaque materials, which would attenuate a radiation
beam more than normal, resulting in higher than normal electron density. Post-contrast CT data
used for a treatment plan would theoretically give rise to higher monitor unit (MU) values, and
therefore radiation dose, compared to MU values taken from calculating a plan using pre-contrast
11
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
919293949596979899
CT data. As veterinary radiation oncologists often use commercially available human treatment
planning software, automatic registration methods may be built-in to the software, although other
“in house” registration algorithms may be adapted for use.22, 24-29 One of the most common
automatic registrations used in routine treatment planning is the merging of pre-contrast and
post-contrast CT data in order to delineate the GTV, CTV, PTV and OAR. Because regions of
interest are often best identified on post contrast CT images, many commercial software
programs (such as Eclipse™, Varian Medical Systems, Palo Alto, used at the authors’
institutions) automatically permit DICOM-coordinated registration and facilitate contouring on
fused multimodality images. However, there are often limited registration capabilities between
multimodal image acquisitions; our current planning system is capable of a fully automatic,
mutual-information-based rigid registration but results are often unsatisfactory between
modalities (i.e. CTs obtained with different DICOM origins or MRI and CT data acquired with
vastly in different animal positionings). This eliminatesAutomatic registration partially
eliminates a potential source of error as radiation oncologists historically may have estimated
tumor volume on the pre-contrast CT data based on visual examination of alternative imaging
studies. Upon eEvaluationng of the veterinary literature to date, it is often difficult to firmly
understand how tumor volumes were identified prior to treatment; for example in one study
evaluating RT for canine intracranial tumors, the GTV was defined on planning CT scans by the
contrast-enhancing area on CT or MR data.54 It is presumed that contours for the GTV were
drawn on pre-contrast CT images after evaluation of diagnostic scans, however this presumption
12
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
100101102103104105106107108
may be incorrect. In most other veterinary studies, it is unclear how the GTV was derived, as
irradiated volumes have been inconsistently defined.52 The delineation of target volumes is
paramount to effective RT, as geographic miss could occur if the GTV, CTV and PTV are
ineffectually contoured and/or unexpected toxicity could result from a poorly defined OAR.
Proper radiation reporting is also critical to permit reproducibility of clinical studies across
institutions.52 It is also important to note that dose distributions generated by treatment planning
are estimated dose distributions, thereby introducing an additional uncertainty to planning. The
variation in registration methods introduces bias that could lead to erroneous conclusions about
the extent and location of various volumes compounded by variations in dose distribution
algorithms. As veterinary reporting of RT planning improves, information regarding registration
should be included.
In addition to RT planning, image registration has made its way into the routine
assessment of daily patient positioning, enabling better tumor targeting.25-29, 55-57 One of the key
concepts of IMRT, stereotactic radiation therapy (SRT) or radio surgery (SRS) and IGRT
planning is that tight dose constraints can beare placed around the PTV provided the clinician is
confident that the patient’s animal’s positioning will be reproducible during treatment. One of
the most common practices immediately prior to treatment delivery is to generate a pair of
orthogonal digitally reconstructed radiographs (DRRs) from the planning CT and to register
simulated radiographs with MV radiographs acquired by a flat-panel imager (electronic portal
imaging devices or EPIDs) attached to the linear accelerator (Fig. 65).55, -56 More advanced on-
13
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
109110111112113114115116117
board imaging equipment involves cone-beam CT systems or helical CT systems built into the
treatment unit.25-29, 57 Automated image registration at the treatment unit can determine the
rotations and translations required to align the datasets, thus decreasing some of the variability
with daily patient setup and intra-patient shifts.25-29, 40 Challenges still exist with registration of
different modalities with different positioning and therefore fiducials and other landmarks may
be needed to improve alignment (Fig. 76). Dose delivered to patients throughout a treatment
protocol also rely on accurate patient positioning and registration (i.e. daily cone beam CT
registered with treatment-planning CT) in order to sum dose delivered over time; monitoring of
dose deviations from the expected distribution allows for intra-treatment adjustments in order to
better target tumor yet limit normal tissue dose. Little veterinary literature has evaluated the
utility or validity of various registration algorithms other than an early study demonstrating an
in-house computer program that was developed for brain imaging.22 This study highlighted the
widespread potential of image registration by evaluating corresponding images from CT and
MR, CT images before and after surgical treatment, and CT and post-mortem cryosection images
in dogs or cats.22 Currently at our institution, for follow-up imaging, standardized response
criteria are applied when assessing tumor response following treatment.58 Responses for tumors
of the head and neck (intracranial, sino-nasal, cervical) are often assessed by the radiology and
radiation oncology team on serial CT scans, without consideration for image registration over
time, which may not accurately reflect the response of the tumor or the assessment of treatment-
induced toxicity.27 Often tToxicities are often inferred by tissue changes, however easier
14
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
118119120121122123124125126
implementation of image registration in the clinic may provide more objective information
regarding tumor and normal tissue changes following radiation, particularly as varying
fractionated protocols are used across veterinary radiation oncologists.52,5 3,59,60
With the incorporation of functional imaging into veterinary oncology, determination of
structural and, indirectly, pathological changes may be better assessed. Changes such as fibrosis
and tumor recurrence can be difficult to distinguish on CT and MR; functional imaging with PET
may help clarify differences and alter future therapy. Currently in veterinary medicine, 2-deoxy-
2-[18F]-fluoro-deoxyglucose (FDG), 3’-deoxy-3’-[18F]-fluorothymidine (FLT), and copper(II)-
diacetyl-bis (N4-methylthiosemicarbazone) (Cu-ATSM) have been used to assess various
neoplastic conditions in dogs and cats.61 59-70 68 In particular, several studies have evaluated
functional imaging over time in an attempt to correlate pre-treatment scans to post-treatment
scans; ideal comparisons would require registration of image datasets to avoid interpretation
errors.67 5-70 68 One study that assessed pre-radiation therapy FDG, FLT and Cu-ATSM PET/CT
scans to post-treatment FDG PET/CT scans demonstrated the complexity and utility of
registration methods in dogs with sinonasal tumors.70 68 Dogs in this study underwent a radiation-
planning CT scan that served as the reference dataset; bony anatomy from pre-radiation PET/CT
scans was registered to the bony anatomy of the radiation-planning CT; rigid registration using
cross-correlation resulted in affine transformations that were applied to the corresponding PET
images. The translated and rotated PET matrices were subsequently resampled using a spline
filter defined by the reference PET image that was obtained at the same time as the treatment-15
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
127128129130131132133134135
planning CT.70 68 While complex, the registration process ensured that each voxel index of the
pre-treatment PET/CT scans corresponded as closely as possible to the same region in the post-
treatment FDG PET image for each patient, enabling assessment of changes in tracer uptake.70 68
Imaging and registration advances will undoubtedly lead to more work being performed in a
similar fashion and may alter current criteria for response assessment and/or response prediction.
Proposed Advances in Veterinary Image Registration
In general, RT is delivered to patients over the course of several weeks in both human and
veterinary oncology. While the GTV, CTV and PTV are currently considered standard targets in
radiation oncology, the PTV is created to make sure that dose plans are robust to
uncertainties.71 69 The GTV and CTV are oncological-anatomical concepts, which are heavily
dependent on the experience and judgment of the radiation oncologist and radiologist. The PTV,
on the other hand, is a geometrical scheme used for treatment planning in order to ensure that the
prescribed dose targets the CTV.71 69 Alternatively, one might view the PTV as decreasing the risk
that patient positioning, setup uncertainties, and OAR or target position uncertainties will lead to
a severe under-dosage of the CTV. As the impact of setup uncertainties decreases with IGRT, the
CTV to PTV margins may be reduced, thus decreasing normal tissue toxicity. Taking this a step
further, if diagnostic imaging, interpretation, and registration with the initial treatment plan are
performed frequently while a patient is on therapy, the treatment plan can be “adapted” to
optimize tumor control and minimize normal tissue damage.
16
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
136137138139140141142143144
ART is defined as changing a radiation plan during RT to account for anatomy or biology
changes of the patients.72 0 Studies have shown clear advantages and gains in clinical application
of adaptive radiation therapy. However, clinical implementation is limited as ART is time-
consuming and places an additional workload on clinicians to quickly contour follow-up images
manually. Adaptive planning could be used at different fractions (off-line) and at different time
points within one fraction (on-line).73 1 Identifying changes in the GTV between images acquired
for the purpose of RT planning and images acquired for the purpose of assessing response to RT
is challenging. This is mainly because of: 1. uncertainties of soft tissues and internal organ
motion;74 2 2. the response of the GTV to radiation or other treatments is difficult to predict,
which may cause anatomical changes;75 3 3. distortions or contrast variability between different
modalities, which makes it difficult to register the information between multi-modality images.76 4
To overcome these difficulties, image guided adaptive radiation therapy (IGART) is the most
reliable method for a comprehensive ART approach.77 5 Research into known, documented
uncertainties in soft tissue and organ motions in prostate,78 6 head and neck79 7 and lung80 78 is well
developed and directly impact ART approaches. Furthermore, registration methods that combine
information and realize deformable mapping between structural and functional images are
essential to make individualized IGART more practicalrapid and feasible and enable
individualized plans.81 79
Preliminary Experience With A Novel Image Registration Framework
17
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
145146147148149150151152153
To develop an efficient and practical IGART approach, an automatic and robust method for the
registration of images acquired at different time points with different modalities is required. In
addition, an accurate segmentation method for non-rigidly tracking the GTV on follow-up
images is also required. Our research group has developed an automatic registration-
segmentation framework, which has the potential to achieve these aims. In this framework the
planning contour of the GTV from the pre-treatment reference image is first registered to the
post-treatment target image both rigidly and non-rigidly. A segmentation method, which uses the
registered data and planning contour as an initial condition, is applied to find the new contour on
the target image. Four Dice coefficients (D1: between the original reference and target contour;
D2: between the rigidly registered reference contour and target contour; D3: between the non-
rigidly registered reference contour and target contour; D4: between the final segmented result
and target contour) are used to assess the step-by-step improvement in this framework (Fig. 7).
Rigid registration is deemed sufficient when there is limited distortion between the target
and reference images, while in the presence of distortion, a non-rigid transform is required. The
proposed automatic registration-segmentation framework has non-rigid capability and to date,
has been applied to images obtained from five dogs (P1-P5) with sino-nasal tumors treated with
helical Tomotherapy.27 Details of data sets used to test this framework are presented in Table 2.
CT imaging was performed using a single detector row unit (GE HiLight Advantage, GE
Medical Systems, Milwaukee, WI, USA) with dogs in sternal recumbency while MR was
performed at 1.0T (GE Signa Advantage) with dogs in dorsal recumbency. All CT images were
18
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
154155156157158159160161162
acquired with 512x512 pixels and a pixel size of 0.39x0.39 mm. All MR images were acquired
with 256x256 pixels and a pixel size of 1.6x1.6 mm. The number of CT and MR image slices
containing the tumor volume varied between dogs and data sets. A board-certified radiation
oncologist involved in the canine trial retrospectively evaluated all images and contoured GTV,
CTV, PTV and OAR for evaluation (JL).
The performance of the framework for automatic registration and segmentation of target
volumes was initially tested on pre-treatment and 6-week post-treatment images from case P5,
where the images were from the same modality (CT) (Fig. 8). The registration results are shown
in blue, target GTV contours are shown in yellow and the automatically segmented GTV contour
produced by the image analysis algorithm are illustrated by a red contour. The Dice coefficient
between the clinical and automatically generated contour for this case was 96.73% indicating
good agreement. The framework was next tested on multi-modality data from case P5
comprising of pre-treatment MR and CT acquired 6 weeks after treatment (Fig. 9). In this case,
the reference contour from the MR image shown in blue was pre-processed by scaling and
rotating prior to matching as different pixel sizes and orientations existed between CT and MR.
With a Dice coefficient of 93.7% between the target contour (yellow) and the automatically
segmented GTV contour (red), once again there is good agreement. The overall performance, as
indicated by the Dice coefficient, in the first 4 dogs was found to be > 80% (Fig. 10), however
additional investigation and analysis is required to assess the performance of the approach on
estimating the position of the GTV for different tumor histologies and over a longer time course.
19
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
163164165166167168169170171
The preliminary results from the proposed registration-segmentation framework
demonstrate that by using prior information from pre-treatment images, it is possible to
automatically segment the GTV on post-radiation therapy images with acceptable clinical
accuracy (mean Dice, 84.44%). The approach also has the potential to be used as part of an ART
strategy for personalized RT that is tailored to time-of-treatment variations of the GTV as it has
the advantage of automatically defining topology changes (Fig. 11). Further work is required to
assess the performance on a larger data set, with different targets (CTV) and using other imaging
modalities such as PET, however, this framework shows promise as a tool to define the GTV
over time and has the potential to be used for ART.
Conclusion
In its simplest context, image registration simply refers to the comparison of one dataset
to another. The complexity behind making quantitative use of this imaging data has been
reviewed here, illustrating that it is necessary to determine the transformations needed to relate
coordinates of one dataset to another. The impact of image registration on RT from treatment-
planning to image-guided delivery to post-treatment follow-up is enormous and has already
impacted veterinary patients. Novel registration frameworks may be ideally tested on companion
animals undergoing RT given similarities between human and veterinary radiation oncology.
Ultimately optimized image registration and data fusion techniques should aid in targeting tumor
20
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
172173174175176177178179180
volumes and organs-at-risk, improve response assessment and increase the availability and utility
of an efficient rapid ART approach.
Acknowledgements:
We are grateful for the generous support from NHS Lothian, Edinburgh and Lothians Health
Foundation (charity number SC007342), the James Clerk Maxwell Foundation, the Jamie King
Uro-Oncology Endowment Fund, a University of Edinburgh Darwin Award and a University of
Edinburgh Individual Stipend.
References
1. Paoloni M, Khanna C. Translation of new cancer treatments from pet dogs to humans.
Nat Rev Cancer. 2008;8:147-56.
2. Mell LK, Song WY, Pawlicki T, Mundt AJ. Image-guided radiation therapy. In: Halperin EC,
Wazer DE, Perez CA, Brady LW (eds): Principles and Practice of Radiation Oncology, 6th
ed. Philadelphia: Lippincott Williams and Wilkins, 2013;246-76.
3. Pluim JPM, Fitzpatrick JM. Image registration. IEEE Trans Med Imaging.
2003;22(11):1341–3.
4. Gooding MJ, Rajpoot K, Mitchell S, Chamberlain P, Kennedy SH, Noble JA. Investigation
into the fusion of multiple 4-D fetal echocardiography images to improve image quality.
Ultrasound Med Biol. 2010;36(6):957-66.
5. Maintz JBA, Viergever MA. A survey of medical image registration. Med Image Anal. 21
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
181182183184185186187188189
1998;2(1):1-36.
6. Freeborough PA, Fox NC. Modeling brain deformations in Alzheimer disease by fluid
registration of serial 3D MR images. J Comput Assist Tomogr. 1998;22(5):838–43.
7. Collins DL, Evans AC. ANIMAL: validation and applications of non-linear registration-
based segmentation. Intern J Pattern Recognit Artif Intell. 1997;11(8):1271–94.
8. Lavely WC, Scarfone C, Cevikalp H, Li R, Byrne DW, Cmelak AJ, et al. Phantom validation
of co-registration of PET and CT for image-guided radiotherapy. Med Phys.
2004;31(4):1083–92.
9. Maurer CR, Fitzpatrick JM, Wang MY, Galloway RL, Maciunas RJ, Allen GS. Registration of
head volume images using implantable fiducial markers. IEEE Trans Med Imaging.
1997;16(4):447–62.
10. Grosland NM, Bafna R, Magnotta VA. Automated hexahedral meshing of anatomic
structures using deformable registration. Comput Methods Biomech Biomed Engin.
2009;12(1):35-43.
11. Maksimov D, Hesser J, Brockmann C, Jochum S, Dietz T, Schnitzer A. Graph-Matching
based CTA. IEEE Trans Med Imaging. 2009;28(12):1940–1954.
12. Martin B, Julien M, Jouke D, Tobias DH, Axel WW, Ivo Q, et al. Atlas-based whole-body
segmentation of mice from low-contrast Micro-CT data. Med Image Anal.
2010;14(5):723-37.
22
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
190191192193194195196197198
13. Anand AJ, Abhijit JC, Changqing L, Joyita D, Simon RC, David WS, et al. DigiWarp: a
method for deformable mouse atlas warping to surface topographic data. Phys Med Biol.
2010;55(20):6197–214.
14. Johnson CA, Seidel J, Carson RE, Gandler WR, Sofer A, Green MV. Evaluation of 3D
reconstruction algorithms for a small animal PET camera. IEEE Trans Nucl Sci.
1997;44(3):1303-8.
15. Green MV, Seidel J, Vaquero JJ, Jagoda E, Lee I, Eckelman WC. High resolution PET, SPECT
and projection imaging in small animals. Comput Med Imaging Graph. 2001;25(2):79-
86.
16. Wildeman M, Baiker M, Reinders MJT, Lowik CWGM, Reiber JHC, Lelieveldt BPF. 2D/3D
registration of micro-CT data to multi-view photographs based on a 3D distance map.
Proc IEEE Int Symp Biomed Imaging. In: Proceedings of IEEE International Symposium
on Biomedical Imaging; 2009; Boston, Massachusetts, USA.
17. Zheng X, Xishi H, Xiaobo Z, Youxian S, Vasilis N, Stephen TCW. Registration of 3-D CT and
2-D flat images of mouse via affine transformation. IEEE Trans Inf Technol Biomed.
2008;12(5):569-78.
18. Yang CC, Wu TH, Lin MH, Huang YH, Guo WY, Chen CL, et al. Multimodality imaging
combination in small animal via point-based registration. Nucl Instrum Methods Phys
Res A. 2006;569(2):240-4.Johnson HJ, McCormick M. ITK software guide. 2013.
23
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
199200201202203204205206207
19. Hayakawa N, Uemura K, Ishiwata K, Shimada Y, Ogi N, Nagaoka T. A PET-MRI
registration technique for PET studies of the rat brain. Nucl Med Biol. 2000;27(2):121-5.
20. Sakdinawat AE, Iwata K, Hwang AB, Tang HR, Wong KH, Hasegawa BH. Development of
external fiducial markers for image registration in small animal SPECT/CT. IEEE Nucl
Sci Symp Conf Rec. 2002;2:842-5.
21. Woo SK, Kim KM, Lee T, Kim JY, Jung JH, Woo KS, et al. Experimental condition and
registration method for the tumor detection of lung metastasis small animal PET and CT
whole body images. IEEE Nucl Sci Symp Conf Rec. 2007;5:3372-5.
22. Yubei S, Koji U, Babak AA, Tsukasa N, Kiichi I, Hinako T, et al. Application of PET-MRI
registration techniques to cat brain imaging. J Neurosci Methods. 2000;101(1):1-7.
23. Bottcher P, Maierl J, Hecht S, Matis U, Leibich HG. Automatic image registration of three-
dimensional images of the head of cats and dogs by use of maximization of mutual
information. Am J Vet Res. 2004;65:1680-7.
24. Shimada Y, Uemura K, Ardenkani BA, Nagaoka T, Ishiwata K, Toyama H, et al. Application
of PET-MRI registration techniques to cat brain imaging. J Neurosci Methods.
2000;101:1-7.
25. Nolan MW, Kogan L, Griffin LR, Custis Jt, Harmon JF, Biller BJ, et al. Intensity-modulated
and image-guided radiation therapy for treatment of genitourinary carcinomas in dogs.
J Vet Intern Med. 2012;26:987-95.
24
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
208209210211212213214215216
26. Harmon J, Yoshikawa H, Custis J, LaRue S. Evlauation of canine prostate intrafractional
motion using serial cone beam computed tomography imaging. Vet Radiol Ultrasound.
2013;54:93-8.
27. Murphy S, Gutierrez A, Lawrence J, Bjorling D, Mackie T, Forrest L. Laparoscopically
implanted tissue expander radiotherapy in canine transitional cell carcinoma. Vet Radiol
Ultrasound. 2008;49:400-5.
28. Lawrence JA, Forrest LJ, Turek MM, Miller PE, Mackie TR, Jaradat HA, et al. Proof of
principle of ocular sparing in dogs with sinonasal tumors treated with intensity-
modulated radiation therapy. Vet Radiol Ultrasound. 2012;51:561-70.
29. Kubicek LN, Seo S, Chappell RJ, Jeraj R, Forrest LJ. Helical tomotherapy setup variations
in canine nasal tumor patients immobilized with a bite block. Vet Radiol Ultrasound.
2012;53:474-81.
30. Forrest LJ, Mackie TR, Ruchala K, Turek M, Kapatoes J, Jaradat H, et al. The utility of
megavoltage computed tomography images from a helical tomotherapy system for
setup verification purposes. Int J Radiat Oncol Biol Phys. 2004;60:1639-44.
31. Kaczka DW, Cao K, Christensen GE, Bates JHT, Simon BA. Analysis of regional mechanics
in canine lung injury used forced oscillations and 3D image registration. Ann Biomed
Eng. 2011;39:1112-24.
32. Vail DM, MacEwen EG. Spontaneously occurring tumors of companion animals as
25
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
217218219220221222223224225
models for human cancer. 2000;18:781-92.
33. Khanna C, Lindblad-Toh K, Vail D, London C, Bergman P, Barber L, Breen M, et al. The
dog as a cancer model. Nat Biotechnol. 2006;24:1065-6.
34. Paoloni M, Webb C, Mazcko C, Cherba D, Hendricks W, Lana S, et al. Prospective
molecular profiling of canine cancers provides a clinically relevant comparative model
of evaluating personalized medicine (PMed) trials. PLoS ONE 2014;9:e90028.
doi:10.1371/journal.pone.0090028
35. Modersitzki J. Numerical Methods for Image Registration. Oxford University Press.
2004.
36. Murphy K, van Ginneken B, Klein S, Staring M, de Hoop BJ, Viergever MA, et al. Semi-
automatic construction of reference standards for evaluation of image registration. Med
Image Anal. 2011;15:71-84.
37. Markelj P, Tomazevic D, Likar B, Pernus F. A review of 3D/2D registration methods for
image-guided interventions. Med Image Anal. 2012;16:642-61.
38. Livyatan H, Yaniv Z, Joskowicz L. Gradient-based 2-D/3-D rigid registration of
fluoroscopic X-ray to CT. IEEE Trans Med Imaging. 2003;22(11):1395–406.
39. Andreetto M, Cortelazzo GM, Lucchese L. Frequency domain registration of computer
tomography data. In: John A, Gabriel T, editors. Proceedings of the 2nd International
Symposium on 3D Data Processing, Visualization and Transmission; 2004 Sep 6-9;
26
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
226227228229230231232233234
Thessaloniki, Greece.
40. Kessler ML. Image registration and data fusion in radiation therapy. Br J Radiol.
2006;79:S99-S108.
41. Mattes D, Haynor DR, Vesselle H, Lewellen TK, Eubank W. PET-CT image registration in
the chest using free-form deformations. IEEE Trans Med Imaging. 2003;22(1):120-8.
42. Hellier P, Barillot C. A hierarchical parametric algorithm for deformable multimodal
image registration. Comput Methods Programs Biomed. 2004;75(2):107-15.
43. Rueckert D, Sonoda LI, Hayes C, Hill DLG, Leach MO, Hawkes DJ. Non-rigid registration
using free-form deformations: application to breast MR images. IEEE Trans Med
Imaging. 1999;18(8):712–21.
44. Meyer CR, Boes JL, Kim B, Bland PH, Zasadny KR, Kison PV, et al. Demonstration of
accuracy and clinical versatility of mutual information for automatic multimodality
image fusion using affine and thin-plate spline warped geometric deformations. Med
Image Anal. 1997;1(3):195–206.
45. Kessler ML, Pitluck S, Petti PL, Castro JR. Integration of multimodality imaging data for
radiotherapy treatment planning. Int J Radiat Oncol Biol Phys. 1991;21:1653-67.
46. Pelizzari CA, Chen GT, Spelbring DR, Weichselbaum RR. Accurate three-dimensional
registration of CT, PET, and/or MR images of the brain. J Comput Assist Tomogr.
1989;13:20-26.
27
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
235236237238239240241242243
47. Studholme C, Constable RT, Duncan JS. Accurate alignment of functional EPI data to
anatomical MRI using a physics-based distortion model. IEEE Trans Med Imaging.
2000;19(11):1115–27.
48. Schnabel JA, Tanner C, Castellano-Smith AD, Degenhard A, Martin OL, Hose DR, et al.
Validation of nonrigid image registration using finite element methods: application to
breast MR images. IEEE Trans Med Imaging. 2003;22(2):238–47.
49. Maes F, Collignon A, Vandermeulen D, Marchal G, Suetens P. Multimodality medical
image registration by maximization of mutual information. IEEE Trans Med Imaging.
1997;16(2):187–98.
50. Alterovitz R, Goldberg K, Pouliot J, Hsu IC, Kim Y, Noworolski SM, et al. Registration of
MR prostate images with biomechanical modeling and nonlinear parameter estimation.
Med Phys. 2006;33(2):446-54.
51. Vercauteren T, Pennec X, Perchant A, Ayache N. Nonparametric diffeomorphic image
registration with the demons algorithm. In: Ayache N, Ourselin S, Maeder A, editors.
Proceedings of the 10th International Conference on Medical Image Computing and
Computer-Assisted Intervention; 2007; Brisbane, Australia.
52. Keyerleber MA, McEntee MC, Farrelly J, Podgorsak M. Completeness of reporting of
radiation therapy planning, dose, and delivery in veterinary radiation oncology
manuscripts from 2005-2010. Vet Radiol Ultrasound. 2012;53:221-30.
28
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
244245246247248249250251252
53. LaRue SM, Gordon IK. Radiation Therapy. In: Withrow SJ, Vail DM, Page RL (eds): Small
Animal Clinical Oncology, 5th ed. St Louis: Elsevier Inc; 2013:180-97.
54. Rohrer Bley C, Sumova A, Roos M, Kaser-Hotz B. Irradiation of brain tumors in dogs with
neurologic disease. J Vet Intern Med. 2005;19:849-54.
55. Kent MS, Gordon IK, Benavides I, Primas P, Young J. Assessment of the accuracy and
precision of a patient immobilization device for radiation therapy in canine head and
neck tumors. Vet Radiol Ultrasound. 2009;50:550-4.
56. McEntee MC. Portal radiography in veterinary radiation oncology: options and
considerations. Vet Radiol Ultrasound. 2009;49:S57-S61.
57. Nieset JR, Harmon JF, LaRue SM. Use of cone-beam computed tomography to
characterize daily urinary bladder variations during fractionated radiotherapy for
canine bladder cancer. Vet Radiol Ultrasound. 2011;52:580-8.
58. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New
response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1).
Eur J Cancer. 2009;45:228-47.
59. Farrelly J, McEntee MC. A survey of veterinary radiation facilities in 2010. Vet Radiol
Ultrasound 2014;55:638-643.
60. McEntee MC. A survey of veterinary radiation facilities in the United States during 2001.
Vet Radiol Ultrasound 2004;45:476-479.
29
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
253254255256257258259260261
61. Randall E, Loeber S, Kraft S. Physiologic variants, benign processes, and artifacts from
106 canine and feline FDG-PET/computed tomography scans. Vet Radiol Ultrasound.
2014;55:213-26.
62.[60.] Hansen AE, McEvoy F, Engelholm SA, Law I, Kristensen AT. FDG PET/CT imaging in
canine cancer patients. Vet Radiol Ultrasound. 2011;52:201-6.
63.[61.] Leblanc AK, Jakoby BW, Townsend DW, Daniel GB. 18FDG-PET imaging in canine
lymphoma and cutaneous mast cell tumor. Vet Radiol Ultrasound. 2009;50:215-23.
64.[62.] Yoshikawa H, Randall EK, Kraft SL, LaRue SM. Comparison between 2-18F-fluoro-2-
deoxy-D-glucose positron emission tomography and contrast-enhanced computed
tomography for measuring gross tumor volume in cats with oral squamous cell
carcinoma. Vet Radiol Ultrasound. 2013;54:307-13.
65.[63.] Bradshaw TJ, Bowen SR, Jallow N, Forrest LJ, Jaraj R. Heterogeneity in intratumor
correlations of 18F-FDG, 18-FLT, and 61Cu-ATSM PET in canine sinonasal tumors. J Nucl
Med. 2013;54:1917-31.
66.[64.] LeBlanc AK, Peremans K. PET and SPECT imaging in veterinary medicine. Semin
Nucl Med. 2014;44:47-56.
67.[65.] LeBlanc AK, Miller AN, Galyon GD, Moyers TD, Long MJ, Stuckey AC, et al.
Preliminary evaluation of serial (18)FDG-PET/CT to assess response to toceranib
phosphate therapy in canine cancer. Vet Radiol Ultrasound. 2012;53:348-57.
30
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
262263264265266267268269270
68.[66.] Lawrence J, Vanderhoek M, Barbee D, Jeraj R, Tumas DB, Vail DM. Use of 3’-deoxy-
3’-[18F]fluorothymidine PET/CT for evaluating response to cytotoxic chemotherapy in
dogs with non-Hodgkin’s lymphoma. Vet Radiol Ultrasound. 2009;50:660-8.
69.[67.] Bradshaw TJ, Yip S, Jallow N, Forrest LJ, Jeraj R. Spatiotemporal stability of Cu-
ATSM and FLT positron emission tomography distributions during radiation therapy. Int
J Radiat Oncol Biol Phys. 2014; doi:10.1016/j.ijrobp.2014.02.016.
70.[68.] Bowen SR, Chappell RJ, Bentzen SM, Deveau MA, Forrest LJ, Jeraj R. Spatially
resolved regression analysis of pre-treatment FDG, FLT and Cu-ATSM PET from post-
treatment PDG PET: an exploratory study. Radiother Oncol. 2012;105:41-8.
71.[69.] ICRU Report 62. Prescribing, recording, and reporting photon beam therapy
(supplement to ICRU Report 50). Bethesda, USA: International Commission on
Radiation Unites and Measurements; 1999.
72.[70.] Yan D, Vicini F, Wong J, Martinez A. Adaptive radiation therapy. Phys Med Biol.
1997;42(1):123-32.
73.[71.] Li T, Zhu X, Thongphiew D, Lee WR, Vujaskovic Z, Wu Q, et al. On-line adaptive
radiation therapy: feasibility and clinical study. J Oncol. 2010;
doi:10.1155/2010/407236.
74.[72.] Birkner M, Yan D, Alber M, Liang J, Nüsslin F. Adapting inverse planning to patient
and organ geometrical variation: algorithm and implementation. 2003;30(10):2822-31.
31
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
271272273274275276277278279
75.[73.] Wu Q, Ivaldi G, Liang J, Lockman D, Yan D, Martinez A. Geometric and dosimetric
evaluations of an online image-guidance strategy for 3D-CRT of prostate cancer. Int J
Radiat Oncol Biol Phys. 2006;64(5):1596-609.
76.[74.] Jans HS, Syme AM, Rathee S, Fallone BG. 3D interfractional patient position
verification using 2D-3D registration of orthogonal images. Med Phys.
2006;33(5):1420-39.
77.[75.] Court LE, Tishler RB, Petit J, Cormack R, Chin L. Automatic online adaptive
radiation therapy techniques for targets with significant shape change: a feasibility
study. Phys Med Biol. 2006;51(10):2493-501.
78.[76.] Thongphiew D, Wu QJ, Lee WR, Chankong V, Yoo S, McMahon R, et al. Comparison of
online IGRT techniques for prostate IMRT treatment: adaptive vs repositioning
correction. Med Phys. 2009;36(5):1651-62.
79.[77.] Peroni M, Ciardo D, Spadea MF, Riboldi M, Comi S, Alterio D, et al. Automatic
segmentation and online virtualCT in head-and-neck adaptive radiation therapy. Int J
Radiat Oncol Biol Phys. 2012;84(3):e427-33. doi:10.1016/j.ijrobp.2012.04.003.
80.[78.] Sonke JJ, Belderbos J. Adaptive radiotherapy for lung cancer. Semin Radiat Oncol.
2010;20(2):94-106. doi:10.1016/j.semradonc.2009.11.003.
81.[79.] Jobse M, Davelaar J, Hendriks E, Kattevilder R, Reiber H, Stoel B. A new algorithm
for the registration of portal images to planning images in the verification of
32
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
280281282283284285286287288
radiotherapy, as validated in prostate treatments. Med Phys. 2003;30(9):2274-81.
33
629
630
289290291292293294295296297
TablesTable 1: The Classification Of Registration Method By Maintz And Viergever In 1998.5
Criteria Classification
I. Dimensionality A. Spatial dimensions only: 2D-2D; 2D-3D; 3D-3D
B. Time series (more than two images), with spatial dimensions:
2D-2D; 2D-3D; 3D-3D
II. Nature of Registration Basis A. Extrinsic: 1. Invasive (stereotactic frame, fiducials); 2. Non-
invasive (mould-frame, dental adaptor)
B. Intrinsic: 1. Landmark based; 2. Segmentation-based; 3.
Voxel-property-based
C. Non-image based
III. Nature of Transformation A. Rigid; B. Affine; C. Projective; D. Curved
IV. Domain of Transformation A. Local; B. Global
V. Interaction A. Interactive; B. Semi-automatic; C. Automatic
VI. Optimization Procedure A. Parameters computed; B. Parameters searched for
VII. Modalities Involved A. Mono-modal; B. Multimodal; C. Modality to model; D.
Patient to modality
VIII. Subject A. Intrasubject; B. Intersubject; C. Atlas
IX. Object A. Head; B. Thorax; C. Abdomen; D. Pelvis and peritoneum; E.
Limbs; F. Spine and vertebrae
34
631632
633
298299300301302303304305306
Figure Captions
Fig. 1. A typical flow chart of image-guided radiation therapy (IGRT), highlighting the need for
registration, using a canine nasal tumor as an example. Diagnostic images are obtained and
registered to the reference image for treatment-planning. Each fraction of radiation is
administered with the aid of an electronic portal imaging device (EPID) or cone-beam computed
tomography (CBCT) images that are registered to the treatment-planning reference image.
Fig. 2. A typical flow chart of image registration demonstrating the steps – transformation
function, similarity function, and optimization – that are required to obtain a registration result.
Fig. 3. An example of a B-spline control grid on the reference and target image of a dog with a
nasal carcinoma. This demonstrates a 20x20 control grid on pre-treatment and post-treatment
computed tomography (CT) (512x512) images (manually warped).
Fig. 4. The variability in the similarity function (SSD) is shown and demonstrated from canine
nasal tumor images when the transformation function (rotation) is applied on the target image
and makes it rotate around the reference image from -90° to +90°. The rotation degree was
defined as the angle between the directions of the reference (black) and the target (white) images.
35
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
307308309310311312313314315
As a result, the image in A illustrates an approximately -70° rotation while B illustrates an
approximately 30° rotation.
Fig.5. Illustration of a multi-step registration framework in a dog with a nasal tumor. Images
show registration of a pre-treatment computed tomography (CT) study with the post-treatment
CT 6 weeks after definitive intensity-modulated radiation therapy. The series demonstrates the
target image with the pre-treatment gross tumor volume (GTV) contour (blue), target GTV
contour (yellow) and an automatically generated contour using a multistep process (red). A
represents D1, registration between the original reference and the target contour; B represents
D2, registration between the rigid registered reference contour and target contour; C represents
D3, registration between non-rigid registered reference contour and target contour; D represents
D4, the final segmented result and target contour. The Dice coefficients increase following each
step in the registration framework as shown in Fig. 6: D1 = 42.65 %, D2 = 85.57%, D3 =
90.81%, D4 = 96.73%.
Fig. 65. Portal images obtained from a dog with a sinonasal tumor undergoing definitive intent
radiation therapy prior to 6 MV photon therapy with a linear accelerator (Varian 2100 CD;
Varian Medical Systems, Palo Alto, USA); orthogonal images were obtained and analyzed
immediately prior to treatment delivery. An on-board flat panel imager permits acquisition of the
dog’s position prior to treatment (image in the lower right), which can be compared with the
36
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
316317318319320321322323324
digitally reconstructed radiograph (DRR) obtained from the planning computed tomography
(CT) (image in the upper right). Images were evaluated using commercially available treatment
planning software (Eclipse v.11.0, Varian Medical Systems). The larger image on the left
demonstrates registration of the electronic portal image and the DRR. (A) Images were obtained
with the linear accelerator gantry at 0 degrees; the larger image on the left demonstrates
evaluation of the registered images using a “moving window” technique that allows the window
to move the DRR over the dog’s daily portal image. The orange graticule grid represents the
dog’s position at acquisition of the portal image while the green graticule grid represents shifts
needed in the dog’s position to match the ideal planning position. (B) Images were obtained
with the linear accelerator gantry at 90 degrees; the larger image on the left demonstrates
evaluation of the images using a “split window” technique, in which the central axis can be
moved to reveal more of the DRR image or the electronic portal imaging device (EPID) image as
needed. Manual alignment may be performed as was done here in which the daily image is
moved to match the DRR; alternatively semi-automated or point matching registration can be
performed.
Fig 76. A series of rigidly registered images acquired on a human patient treated for
prostate cancer highlighting the difficulty in interpreting critical structures in different
imaging modalities. T1- (A) and T2- (B) weighted MR images acquired on a Siemens
Symphony 1.5T scanner (Siemens, Munich, Germany) were registered to a radiotherapy 37
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
325326327328329330331332333
planning CT (C) acquired on a Philips Brilliance Big Bore CT scanner fitted with a flat
couch. The hyper-intense gold markers implanted within the prostate gland and used for
positional verification are clearly visible. The corresponding pre-treatment CBCT image
(D) acquired on a Varian Clinac and used for positional verification. Alignment is complete
and treatment begins when the hyper-intense markers are aligned on CT and CBCT. Soft
tissue matching is difficult because of the atrophy induced in the prostate by the neo-
adjuvant hormones administered between MR and CT acquisition.
Fig. 7 The proposed flow chart of the novel registration-segmentation framework applied to
serial, multi-modality images of dogs with nasal tumors. Rigid registration, non-rigid registration
(optional) and level set segmentation are built into the framework. Four Dice (D) coefficients are
used to assess the improvement in the image registration in a step-by-step fashion; D1: between
original reference and target contour; D2: between rigid registered reference contour and target
contour; D3: between non-rigid registered reference contour and target contour; D4: between
final segmented result and target contour.
Fig.8. Illustration of improved registration using a novel framework; images show registration of
a pre-treatment computed tomography (CT) study with the post-treatment CT 6 weeks after
definitive intensity-modulated radiation therapy in a dog with a nasal tumor. The series
demonstrates the target image with the pre-treatment gross tumor volume (GTV) contour (blue),
38
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
334335336337338339340341342
target GTV contour (yellow) and the automatically generated contour (shown in red). A
represents D1, registration between the original reference and the target contour; B represents
D2, registration between the rigid registered reference contour and target contour; C represents
D3, registration between non-rigid registered reference contour and target contour; D represents
D4, the final segmented result and target contour. The Dice coefficients increased following
each step in the registration framework as shown in Fig. 6: D1 = 42.65 %, D2 = 85.57%, D3 =
90.81%, D4 = 96.73%.
Fig. 9. Illustration of a novel registration framework in a dog with a nasal tumor; images show
registration of the pre-treatment magnetic resonance (MR) images with the post-treatment
computed tomography (CT) images 6 weeks after definitive intensity-modulated radiation
therapy for a dog with a nasal tumor. The target image is shown with pre-treatment gross tumor
volume (GTV) contour (blue), target GTV contour (yellow) and the automatically generated
contour (red). A represents D1, registration between the original reference and the target contour;
B represents D2, registration between the rigid registered reference contour and target contour; C
represents D3, registration between non-rigid registered reference contour and target contour; D
represents D4, the final segmented result and target contour. The Dice coefficients at each step in
the framework as shown in Fig. 6 were: D1 = 79.76 %, D2 = 86.78%, D3 = 90.07%, D4 =
93.7%.
39
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
343344345346347348349350351
Fig.10. The Dice coefficient calculated by comparing the (true) clinical gross tumor volume
(GTV) contour obtained from the post-treatment images to the automatically generated contour
on all images acquired on 4 dogs with nasal tumors. The overall Dice was 84.44%.
Fig. 11. Automatic topology changes determined in the gross tumor volume (GTV) based on
level set method determined in 2 dogs with nasal tumors. The image in A represents dog 2 (P2)
while the image in B represents dog 3 (P3). The target images are shown in each image with the
pre-treatment GTV contour (blue), target GTV contour (yellow) and the automatically
generated contour (red).
40
733
734
735
737
738
739
740
741
742
743
352353354355356357358359360