|Approvals/Requirements Satisfied by eRADIMAGING Courses|
|~ ASRT accreditation for ARRT Category A credit (All Courses)||~ MDCB accreditation by the Medical Dosimetrist Certification (Selected Courses)|
|~ ARMRIT accepted (All MRI Courses)||~ CAMRT and Sonography Canada recognize the ASRT approval (All Courses)|
|~ ARDMS accepted (All Courses)||~ Florida approval for all courses 1 credit or more|
|~ NMTCB accepted (All Courses)||~ California CE requirements met for all radiography courses|
The Role of Image Fusion in Medical Dosimetry
Nishele Lenards, MS, CMD, RT(R)(T)
*University of Wisconsin-La Crosse, Medical Dosimetry Program, College of Science and Health -Department of Health Professions, La Crosse, Wisconsin.
Address correspondence to: Nishele Lenards, MS, CMD, RT(R)(T), University of Wisconsin-La Crosse, Medical Dosimetry Program, College of Science and Health -Department of Health Professions, 1725 State Street - 4033 HSC, La Crosse, WI 54601. E-mail: email@example.com.
Disclosure Statement: Ms Lenards reports having no significant financial or advisory relationships with corporate organizations related to this activity.
Medical imaging fusion plays an important role in the diagnosis and treatment of patients with cancer who are receiving radiation therapy. Image fusion and registration are essential in the daily treatment planning duties of a medical dosimetrist. The image fusion and registration process involves combining multimodality images to delineate the anatomical and physiological differences from one dataset to another. This article will discuss imaging modalities, such as computed tomography, magnetic resonance imaging, positron emission tomography, single photon emission computed tomography, and sonography, and examine how significant their roles are in the treatment and management of radiation therapy patients. Image registration processes will be introduced as well as the recommended quality assurance procedures. The DICOM (Digital Imaging and Communications in Medicine) networking requirements will be explained as they apply to image fusion, registration, and storage. DICOM-RT (Radiotherapy) networking will be discussed as it applies to the radiation therapy department. Various case studies are presented to demonstrate the technique and benefit of image fusion in the treatment and management of the radiation therapy patient.
edical imaging is a fundamental tool in radiation therapy.1 For medical dosimetrists, it is part of their daily planning duties. Anatomic images of high quality are required to accurately delineate target volumes and structures for the purpose of radiation treatment planning.2 In the management of patients with cancer, image data are used for diagnosis, staging, treatment planning and delivery, and patient follow-up.1 Treatment planning is referred to as medical dosimetry, which involves measuring and calculating doses in cancer treatment. Medical dosimetrists use their expertise in physics, anatomy, and radiobiology to develop an optimal arrangement of radiation portals to spare normal and radiosensitive tissues while applying a prescribed dose to the targeted disease. Improvements in radiation therapy techniques are attributed in large part to the image fusion capabilities available for medical dosimetry treatment planning. Image fusion allows for more precise tumor localization with respect to size, shape, and location of the target volume, as well as nearby critical anatomic structures. The additional information from the fused images is transcribed to the computed tomography (CT) image dataset as contoured structures which are then used for treatment planning.3 This has a significant effect on the outcome of treatment for patients as well as the associated side effects they may encounter.
Computed tomography is the primary imaging modality for the medical dosimetry treatment planning process. Other imaging modalities such as magnetic resonance imaging (MRI), positron emission tomography (PET), single photon emission computed tomography (SPECT), and sonography are also used to improve patient management in radiation therapy. These modalities alone do not provide all of the geometric and physical information needed for treatment planning.1 Fusion of these imaging modalities provides the information needed in certain circumstances, such as when it is difficult to detect lymph node involvement or boundaries of primary tumors where inflammatory changes and metal artifacts are present. The addition of any one of these image datasets to the primary CT dataset provides advantageous results for the patient, thus demonstrating the role of fusion in medical dosimetry.
Modalities of Image Fusion
Image fusion is the process that matches 2 or more image datasets resulting in a single image dataset.4 This merged image dataset is essential in clinical interpretation of patient disease. CT, MRI, SPECT, PET, and sonography each have their own special advantages for imaging certain types of tumors. A brief overview of these imaging modality characteristics is presented to demonstrate the advantages and limitations of their use in treatment planning.
Computed tomography is the primary dataset used for radiation therapy treatment planning because it provides the electron density map needed for accurate radiation dose calculations. CT offers excellent spatial and anatomic delineation of tumors and their relationship to nearby critical structures.5 CT provides also the best geometric accuracy, and therefore is considered a reference for anatomic landmarks. The spiral or helical CT scanners allow continuous rotation of the X-ray tube as the patient is transferred through the scanner aperature.2 This provides an acquisition of the large number of thin slices needed for the high quality CT images and digitally reconstructed radiographs (DRRs). DRRs are reconstructed images in planes other than that of the original transverse image (Figure 1). The requirements for high quality DRR images include high contrast, high resolution, and small slice thicknesses. Therefore, the advantages of using CT as the primary image dataset for radiation therapy treatment planning are high-quality spatial resolution images, superior cortical bone contrast, and the ability to generate DRRs in any plane. Original treatment planning CT scans are also fused with new treatment planning CT scans during patient treatment when anatomical patient data have changed significantly. This method is also used when the original treatment planning CT scan is fused with a cone-beam CT scan of the patient while under radiation treatment. This assists in daily setup to ensure alignment of treatment volumes which can be affected by patient contour changes (Figure 2).
In many cases, CT is superior in the evaluation of the extent of disease; however, it is limited in soft-tissue contrast, which is needed for differentiating tumors from scar tissue or abnormalities of the central nervous system (CNS) that have poor tissue differentiation on CT images. CT is also less advantageous in treatment planning where artifacts are concerned. The scatter distortions from the scanner or by the patient could lead to inaccuracies in dose calculations for patient treatment.
Magnetic Resonance Imaging
Magnetic resonance imaging plays an important role in treatment planning for several disease sites. MRI is considered superior to CT in soft tissue discrimination. This soft tissue contrast permits better visualization of tumors or tissue abnormalities in the CNS, head and neck area, sarcomas, prostate gland, and lymph nodes (Figure 3). For the CNS tumors, MRI distinguishes tissue fluid differences, such as tumor and edema (Figure 4). These images also demonstrate the heterogeneity of CNS tumors by using the signal intensity between T1 and T2 weighting. The MRI pulse sequences improve image contrast by enhancing or suppressing specific tissues (eg, fat) and conditions (eg, edema).1 MRI is also advantageous because of its ability to directly generate scans in the axial, sagittal, coronal, or oblique planes.2 Functional MRI has the potential to be useful in treatment planning by showing physiologic activity as it happens and may be useful in delineating target volumes and critical structures for advanced conformal treatment planning cases.2
Similar to CT, MRI has limitations and disadvantages in detection, diagnosis, or planning of radiation therapy patients. It takes much longer to scan the patient when using MRI than with CT, and therefore images are susceptible to artifacts from patient movement. MRI is more susceptible to spatial distortions and intensity artifacts, has a lack of signal from cortical bone, and has image intensity values that have no relationship to electron or physical density.1
Emission Computed Tomography (Nuclear Medicine Imaging)
Positron emission tomography and SPECT permit imaging of the flow and accumulation of biologically active tracers.1 Unique to PET and SPECT, the imaging information provided from these procedures is physiological rather than anatomical. These modalities provide important information about tumor metabolism and tissue function. In radiation treatment planning, PET is useful for discriminating between tumor recurrence and radiation necrosis, whereas SPECT is useful for demonstrating regional lung function to determine beam directions.
A radioactive tracer, such as 18flurodeoxyglucose (FDG) for PET and 131I or 99mTc for SPECT, is administered intravenously and then transported intracellularly across membranes. The inconsistent metabolic activity of cancer cells compared to their normal counterpart allows for imaging of disease areas. During the imaging process, the equipment detects the emitted particles from the radiopharmaceutical tracer. The scanner detectors localize and quantify the interactions and then register the amount of metabolic activity.4
Positron emission tomography is based on the physical properties of certain radioactive isotopes known as positron emitters.6 For a PET scan, a small amount of radioactivity is attached to a biologic substance that is similar to what is already found in the body. Organs and tissues process these radioactive agents as part of their normal function. The scanner detects the location of radiation in the body, and then the computer creates pictures based on activity by using colors or brightness to highlight the different levels of function.7 PET images can then be displayed in the axial, sagittal, and coronal planes. For patients with cancer, the PET scan detects active tumors with high sensitivity. The active tumors have a high metabolism and therefore a high demand for glucose, which is visible on the PET scan.6
The use of FDG-PET in radiation therapy has drastically increased with the ability to fuse functional image sets with anatomical image sets such as CT.8 The use of 18FDG in PET imaging aids in defining the target volumes and reveals tumor-bearing tissues that may have been previously excluded from the gross tumor volume (Figure 5).9 However, PET-only images are difficult to use for radiation therapy planning because of the lack of anatomical details necessary. PET-only images lack the simple methodology to fuse PET data with the radiation therapy treatment planning CT. There is also the possibility of false results if the patient's chemical balances are not normal. Another disadvantage is the physiologic uptake by fat, muscle, normal, and lymphoid tissues which may confound interpretation.6 According to Workman and Coleman, a limitation of FDG-PET is the overlap that exists between inflammatory disease and malignancy.6 The inflammatory process has increased uptake which is just as intense as a malignancy. Radiation therapy changes can also have increased FDG uptake that mimics residual disease.6 These limitations are what lead to false-positives in FDG-PET. All of these disadvantages provided the direction for implementation of PET/CT scanners. The PET/CT scanners decrease acquisition time by using CT for attenuation correction of PET.7 With PET/CT, one imaging procedure provides the functional and morphologic data necessary for the treatment planning process.
Single photon emission CT/CT provides information on perfusion distribution for lung parenchyma. The 99mTc radioisotope tracer is injected intravenously into the patient. This is distributed to the pulmonary blood pool and trapped by the precapillary bed of the entire lung tissue. Photons emitted are collected by the gamma camera and used to reconstruct the 3-dimensional (3D) SPECT images. The intensity of the SPECT images is proportional to the precapillary blood density (blood perfusion) in a specific voxel. This information is used to clinically evaluate the blood-gas exchange and thus the function distribution in the lung parenchyma. During the SPECT/CT image acquisition, the CT images are obtained with the patient lying down in the treatment position. Afterwards, SPECT images are acquired with the patient in the same position to ensure spatial registration. A cardiac SPECT (also called myocardial perfusion imaging) is also useful in treatment planning to assess the heart's structure and function.10 A small amount of radioactive tracers is injected into the patient intravenously and detected by the camera to produce images of the heart. Clinical information, such as blood flow through the heart or abnormal function of the heart muscle, is demonstrated on the SPECT images.10
For radiation treatment planning, the radiation will contribute to the change of blood perfusion and lung or heart function. The SPECT image intensity shows the perfusion status of different portions of the lung or heart. The brighter (higher value) voxels represents greater perfusion, thus, better function. These portions of the organ will generally be the preferred area to be preserved (Figure 6). Vice versa, darker portions of the organ reflect poor perfusion and poor function which may be less sensitive to radiation damage. Follow-up SPECT scanning of the heart or lungs can be used to detect damage to either of these organs due to the radiation treatment of patient disease (Figure 7).
Sonography is a useful imaging tool for delineating surface contours and localizing internal structures such as the prostate gland.11 Sonography delineates tissues that differ only slightly. Sonography can also be used to image blood flow, for example, with fusion of Doppler ultrasound with a magnetic resonance angiography procedure. The sonography transducer is used to generate a mechanical disturbance (pressure or sound wave) that moves through the tissue. The waves are created by the piezoelectric crystals within the transducer. As the sound wave passes through the body, it encounters a variety of tissue interfaces and reflects the energy. The reflected energy depends on the tissue density and speed of ultrasound through tissue.11
For treatment planning and delivery, real-time sonography is used for localization of the prostate gland in the lower pelvis before radiation therapy is administered. Several 2-dimensional (2D) images are acquired and then rendered into a 3D matrix covering the volume of interest. This information is used to generate an in-room coordinate system each day to determine absolute position of the prostate. It is also used for radioactive brachytherapy prostate seed implantation. Real-time sonography can also be used for localization of a breast tumor bed for radiation therapy boost treatments. Breast tumors are at high risk for recurrence in the region close to the tumor bed. If there are inaccuracies in patient positioning, there will be an underdose of radiation which leads to recurrence. Distinct from CT, breast sonography can differentiate solid from fluid-filled structures with high specificity.12 For 2D sonography, the limitation is the lack of spatial orientation in 3 dimensions and the challenge is to achieve accurate co-registration of sonography and CT. With the introduction of the high resolution 3D sonography, these challenges do not exist. The 3D sonography allows images to be referenced to the room coordinates and to that of the CT isocenter.11
The clear advantages of sonography image fusion are that images are produced in real-time, the apparatus is relatively small, and it does not involve ionizing radiation. With the most recent advancements in 3D sonography, there will most likely be an increase in the number of image fusion cases for the use in radiation treatment planning. Stereotactic sonography is one of the first image guidance systems established in the last few years and is fast, simple, and cost-effective for soft-tissue positioning.13 One major disadvantage of sonography for fusion studies is that it cannot image all regions of the body.
Each image modality discussed has its drawbacks but the goal is to overlap the strengths of each one.1 Overall, the information gained from various imaging modalities in a complementary manner includes:
- CT delineation of bone tissue and for the accurate computation of radiation dose;
- MRI-defined anatomic structures of soft tissue with high accuracy;
- PET and SPECT provide tumor extension and biologic information; and
- Sonography images acquired real-time and used for delineation of soft tissue interfaces that differ slightly.14
The term image registration means a process of correlating different image sets to identify corresponding structures or regions.2 The process involves comparison of images from one study to another and then fuses them into one dataset. Pelizzari describes the process as establishing a coordinate transformation between the inherent coordinate spaces of 2 image studies such that homologous points in the 2 spaces are mapped accurately onto each other.15 He also noted that image registration is not limited to medical applications because it is also used for satellite imaging, astronomy, robot vision, fingerprint analysis, and "smart" weapon guidance.15 Various techniques of image registration include point-to-point fitting, interactively superimposing images into 2 datasets, and surface topography matching.2 During the image registration process, the pixels within an image are identified and classified based on specific properties. Pixels are identified on the basis of their appearance, such as density or texture, as well as belonging to a group or class, such as a particular organ.9 In radiation therapy treatment planning, these groups can be identified as organs or target volumes. More specifically, the image registration uses a manual or automatic slice-by-slice delineation of anatomic regions of interest such as external contours, target volumes, critical structures, or anatomic landmarks.2 Depending on the case, simple line drawings of contours or projections are sufficient whereas more complex cases require 3D style display based on geometric models of segmented objects or 3D geometric definition of the volume of a structure.14
Image segmentation is the process by which pixels within an image are identified and classified based on specific properties.9 As discussed previously, these groups can be organs or target volumes (Figure 8). In radiation therapy treatment planning, image segmentation is a very tedious and labor intensive process, but the most important step. Khan states that although there is an automatic delineation based on structures, the target volume delineation requires clinical judgment.2 Accurate information about the shapes and locations of anatomic structures are necessary for the qualitative and quantitative evaluation of treatment plans such as the dose distribution displays, calculation of dose volume histograms for selected regions of interest, and predicted values of biologic indices.9 The anatomic structure volumes can also be displayed on DRRs and then used to correlate the planning geometry with the treatment geometry when compared to portal images.
Image Registration Techniques
The process of image registration involves 2 main tasks: (1) data registration; and (2) structure mapping. Data registration refers to parameter estimation of coordinate points between 2 studies. Structure mapping uses the data registration transformation to map structures or regions of interest from one study to another or to directly combine the grayscale data from the 2 studies.1
Dataset registration can be done with numerous techniques. Khan reported point-to-point, superimposition, and surface topography.2 These geometric structures are used to compute the extracted data from the datasets or native grayscale.1 An example of a geometric feature is the use of manually placed fiducials and stereotactic frames. The parameters used to model coordinate transformation between 2 datasets depend on the modality and clinical site but it is only necessary to account for different patient orientation at the time of imaging.1 For rigid anatomy, 3 rotations (φx, φy, φz) and 3 translations (tx, ty, tz) are required. For non-rigid anatomy, a more complicated spatially variant transformation is required involving a larger number of degrees of freedom.1
The most popular algorithms for dataset registration used in the clinical setting include surface based registration and image based registration. Surface based registration uses the surface of 1 or more anatomic structures and extracts them from the image data in order to calculate and minimize the mismatch between datasets. Typically, these are easily extracted using automated techniques and minor hand editing.1 The surfaces from one study set are assigned as a binary volume or polygon surface and the surfaces from another study set are assigned as a set of point samples from the surface. The metric is the degree of mismatch between the 2 datasets and is computed as the sum or average of the distances between the points and the surfaces.1 For image-based registration, the grayscale data are used to compute the measure of mismatch or similarities between the 2 datasets. The metric used for this technique involves measuring similarities between the grayscale distributions of the 2 datasets.1 Kessler and Kelvin state that the joint and individual probability densities are computed from the histogram of the grayscale pairs from the datasets at each iteration of the coordinate transformation.1
Structure mapping is another method of image registration which involves mapping the outlines of anatomic structures or treatment volumes from one imaging study to the other. Another approach called "image mapping" or "image fusion" involves transforming and reformatting image data from one study to match the orientation and scale of another study which allows for simultaneous visualization of grayscale information from corresponding anatomic planes.1 Structure mapping produces outlines of structures defined from one study on the images of another. A 3D reconstruction of 2D outlined image structures is performed. This reconstruction is then mapped to the other study by computed transformation. These transformed surface model image outlines are used as input for the treatment planning process. For image mapping or fusion, there is a simultaneous display of images of corresponding planes from 2 studies. This involves resampling between studies to match scales and orientations from one study to the other. There is flexibility in how the relationship of data between the 2 studies is displayed for the user. For example, planes can be identified in a side-by-side fashion or a cursor can be used to track where movement over certain areas reveals the image of the secondary dataset. The information can be displayed using different coloring, grayscale variations, and so on.
Image fusion and registration provide the tools for advanced treatment planning in the management of patients receiving radiation. However, these tools are not available without the network application layer named DICOM (Digital Imaging and Communications in Medicine) and its RT (Radiotherapy) extension. DICOM formats are the standard for diagnostic file transfers and DICOM-RT formats are the standard for radiation oncology files.9 Khan states that our current infrastructure is burdened by the inefficient flow of data and lack of industry standards to define objects and workflow.9 The DICOM and DICOM-RT are the industry standards that improve the connectivity and operations of medical imaging data. DICOM-RT consists of 5 objects: (1) RT dose; (2) RT structure set; (3) RT image; (4) RT plan; and (5) RT treatment record.
Table 1 provides a description of the DICOM-RT objects and their functions. With the advent of DICOM, it became possible to store digital image data in a common repository called picture archiving and communications system (PACS). PACS not only provides archiving, but also provides digital image processing such as window/level, magnification/shrinking, and convolution contouring. Just as DICOM-RT is an extension of DICOM, RT-PACS is an extension of PACS and is used in radiation therapy for improving clinical workflow and operational efficiencies.
In order for DICOM-RT to be successful, the network connection and protocols must be properly configured. The systems that export and import the image datasets must be compatible for DICOM-RT and the image fusion process. Finally, the coordination between the imaging department and radiation therapy department is essential for the success of transferring image datasets. Table 2 demonstrates the general flow of digital imaging data when using DICOM and DICOM-RT.
Quality Assurance of Image Registration
Image registration in treatment planning and delivery requires verification of the results.16 There is commercial software available for the registration of rigid objects. The American Association of Physicists in Medicine (AAPM) Task Group 53 Report outlined general commissioning and routine procedural quality assurance (QA) checks to be used for treatment planning.17 They also stated that multimodality image registration is a complex area and required further development and its own task group.17 A recent task group (TG 132) was created by the AAPM to review the techniques for image registration, to identify issues related to clinical implementation, to determine the best methods to assess accuracy, and to outline issues related to acceptance and QA. The main concern is the demand for objective metrics of image registration quality within radiation therapy. Phantom testing determines algorithm variations and confirms metrics such as linearity calibration limits.16 However, phantoms do not completely capture factors that corrupt image registration algorithms, such as variations in slice thickness, resolution, distortion, noise, and patient movement. Setup variations and irregularity of patients are limiting factors that can be reduced by cross-comparison of redundant structures. This type of qualitative assessment includes image overlay, side-by-side comparison, split screens, and checkerboard displays.16 In summary, we can confirm performance of imaging devices, registration software, and network storage and retrieval using phantoms, but there is still a strong need for visual checks to ensure consistency between patient cases. This requires logical, sequential, and reproducible processes to eliminate the possibility of human error.
Image Fusion Case Studies
Image fusion has proven to be a valuable tool in radiation therapy due to its ability to provide additional diagnostic information that was not previously available for the treatment planning process. The power of image fusion lies in the ability to better visualize the anatomical and physiologic structures that are not available on the individual image datasets. This image fusion process provides more accuracy in delineating tumor volumes and decreasing the potential of local disease recurrence for radiation therapy patients. In the following series of case studies, the benefits of image fusion for the treatment planning of radiation therapy patients will be presented.
Case Study #1: Squamous Cell Carcinoma of the Head and Neck
A 41-year-old male presented with dysphagia. A CT scan of the neck revealed diffuse esophageal wall thickening and a possible esophageal mass measuring 2.5 x 1.8 x 5 cm. He also had a left hypopharynx mass measuring 1.9 x 1.6 cm with additional asymmetry in the hypopharyngeal lumen. There were enlarged lymph nodes in the right neck with the largest measuring 4.1 x 1.9 cm. There were also enlarged lymph nodes in the left neck with the largest measuring 2.1 x 0.8 cm.
The patient underwent an upper endoscopy which revealed a 5-cm area of thickness. This area was biopsied and revealed moderately differentiated squamous cell carcinoma. He also underwent a panendoscopy with several biopsies. The bilateral aryepiglottic folds were positive for poorly differentiated squamous cell carcinoma. The physician recommended a PET/CT scan to complete his staging workup as well as to determine the local extent of disease. Furthermore, the PET scan would help clarify whether there was involvement of the hypopharynx as well as rule out distant metastases.
The PET/CT scan revealed a large soft tissue mass engulfing the esophagus measuring 6 x 2.5 x 3.5 cm and displacing the glottis anteriorly. The right vocal cord was thickened and hypermetabolic. The aryepiglottic folds were thickened and hypermetabolic. There was right posterior triangle lymphadenopathy producing a nodal mass measuring 4.2 cm obliquely, 2.2 cm in thickness, and 5.3 cm in length. Left posterior triangle lymphadenopathy revealed involvement of level 2, level 3, and right supraclavicular lymph nodes which measured 1 to 2 cm in size range. No metastases in the chest, pelvis or skeleton were noted.
The CT scan and the PET/CT scan images were fused together to delineate the tumor volumes and positive lymphadenopathy. Figure 9 demonstrates the CT scan (column 1), the PET/CT scan (column 2), and the fused PET/CT image datasets (column 3).18
Case Study #2: Squamous Cell Carcinoma of the Anal Canal
A 61-year-old female presented with numerous external and internal hemorrhoids and a palpable tumor located approximately 1 cm from the anal verge. The tumor extended approximately 3 to 4 cm in the craniocaudal dimension and 2 cm left to right.
A CT scan of the abdomen and pelvis demonstrated an anal canal mass and some concern for a presacral region node. It was noted that if the node was positive, the staging would change and her 5-year survival would decrease from a range of 70% to 80% to a range of 40% to 50%.
A PET/CT scan was ordered to evaluate disease to the chest as well as the suspicious presacral lymph node. The scan revealed a marked FDG activity in the enlarged presacral lymph node. No other evidence of regional or distant spread was noted.
A CT scan was fused with a PET/CT scan for delineation of the presacral lymph node involvement. Figure 10 demonstrates the fused image datasets and Figure 11 displays the treatment planning volumes and isodose distributions.19
Case Study #3: Squamous Cell Carcinoma of the Right Oropharynx
A 43-year-old female presented with throat pain, dysphagia, and trismus. She also developed nasal and right ear fullness. The patient was seen by an otolaryngologist who ordered a CT scan of the neck. This CT scan revealed swollen mucosa of the oral and nasopharynx, right greater than left. Soft palate mucosa swelling was noted with right Eustachian tube obstruction. The patient had a 1.2- x 2.2-cm right mandibular lymph node and a 1.2- x 1.3-cm right jugulodigastric lymph node. Also noted were several shotty deep cervical lymph nodes bilaterally.
The patient was evaluated in Radiation Oncology where a 3-cm palpable right submandibular lymph node was present as well as a 2-cm fixed lymph node in the right posterior submandibular space. There was no contralateral palpable lymphadenopathy. A PET/CT and MRI were ordered for further staging and improved visualization of the tumor.
The PET/CT scan demonstrated a hypermetabolic right posterolateral nasopharyngeal mass measuring 2.7 x 2.3 x 3 cm. The mass extended from the skull base down into the right palatine tonsil fossa. Distortion of the nasopharynx was observed with no obstruction. The right nasopharyngeal mass appeared to extend to the skull base but did not erode it. The mass also extended to the right pterygoid plates without erosion. Hypermetabolic lymphadenopathy demonstrated a right submandibular nodal metastasis, right level 2 nodes, and left level 5 nodes consistent with metastasis. The PET/CT scan also noted 2 mildly hypermetabolic small fluffy alveolar infiltrates in the right middle lobe of the lung. The CT appearance was worrisome for infiltrative tumor or inflammation.
Figure 12 demonstrates the PET/CT fusion which displays the positive disease (left) and the treatment planning CT with the PET fusion information used for delineation of tumor volumes (right). The isodose distribution is also displayed.20
Case Study # 4: Small-Cell Lung Cancer
A 64-year-old female presented in the emergency department with chest pain and chronic cough. She had a history of 30 pack-year, 2 packs/day smoking. A contrast CT scan was ordered which revealed a left upper lobe mass and positive hilar lymph nodes. There was no evidence of metastases. The patient was recommended to medical oncology for chemotherapy and to radiation oncology for concurrent stereotactic body radiotherapy.
A treatment planning noncontrast CT scan was performed. This would be used to avoid an error in water equivalent depth for range, which could be caused by the contrast media if the patient was treated with protons. The contrast CT is fused with planning CT in order to localize the vasculature within the mediastinum. This would help to delineate the tumor volumes as well as the organs that are at risk for radiation damage.
A 4-dimensional (4D) CT was also performed with all phases of breathing fused together to determine the internal margin affected by breathing. This was used to determine the internal target volume that accounts for motion of target during treatment delivery.
Figure 13 displays a noncontrast CT (top) to contrast CT (bottom) fusion. The patient also had a 4D CT fused with this image dataset to demonstrate contour changes due to the breathing motion.21
Case Study #5: Supratentorial Primitive Neuroectodermal Tumor
A 3-year-old male presented to a local physician with left eye intorsion and neck stiffness.
A CT and MRI was ordered and revealed hydrocephalus with a cystic and solid tumor displacing normal brain in the anterior and middle cranial fossa on the left greater than the right side. The solid portion measured 4 x 3.4 cm while the cystic portion measured 6.6 x 4.5 x 4.7 cm. The patient underwent a subtotal resection.
A postoperative MRI revealed a large cystic tumor with questionable residual disease versus postoperative changes. An MRI of the spine was negative for metastatic disease. Cerebrospinal fluid cytology was negative for metastatic disease.
A brain MRI was ordered approximately 5 to 6 weeks later and revealed recurrent residual tumor along the base of the resection cavity with malignant cysts extending out to the frontal horn of the lateral ventricle. Due to these findings, the patient underwent a gross total resection via a left frontotemporal craniotomy. The next day, another MRI of the brain was ordered and revealed residual disease versus postoperative abnormality measuring 9 x 8 x 6 mm adjacent to the left optic nerve and chiasm. Six days later a repeat MRI was performed and again revealed ill-defined tissue measuring 8 x 5 mm at the left gyrus rectus, most consistent with benign postoperative change.
The patient received 2 cycles of chemotherapy, and a follow-up MRI revealed evolving postoperative changes and decreased volume of the tumor bed. The patient continued with 2 more cycles of chemotherapy and started a radiation oncology workup at that time.
The patient had another MRI, as part of the radiation oncology workup, which revealed a small focus of T2 flair hyperintensity over the superficial anterior inferior left frontal lobe. Convexity of the left leptomeninges without post-contrast enhancement was seen. The final impression was a supracellular and anterior cranial fossa supratentorial primitive neuroectodermal tumor. The patient was treated with proton therapy to the tumor bed plus a small margin. The protons are indicated to reduce radiation dose to the normal tissue structures, specifically the optic apparatus, left optic nerve, and optic chiasm as well as the left temporal lobe and hippocampi.
Figure 14 displays a treatment planning CT scan (left) which is fused with an MRI scan (right). The image fusion is beneficial for delineating the pre- and post-surgical tumor volumes to be used for radiation therapy treatment.22
In treatment planning, multimodality image correlation offers advantages in terms of tumor delineation, discrimination between necrosis and recurrent disease, and evaluation of treatment effect. Therefore, image fusion has become an essential tool for clinical use. Combining morphologic (CT, MRI, and sonography) and functional (PET and SPECT) data improves the possibilities of interpreting 3D data for medical dosimetry treatment planning.13 DICOM networking provides the path for the multimodality images to travel from one department to another. This has provided numerous possibilities for medical dosimetry treatment planning and the patients they care for. The complexity of image fusion and the registration process requires a solid and consistent QA program to ensure the accuracy and safety of radiation treatments to the patient. Although complex, the most important message is that the image fusion process is advantageous for the patient with cancer and their long-term survival.
1. Kessler ML, Kelvin L. Image fusion for conformal radiation therapy. Am Assoc Phys Med. 2001;1-12. Available at: http://www.aapm.org/meetings/2001AM/pdf/7213-95766.pdf. Accessed December 4, 2009.
2. Khan FM. The Physics of Radiation Therapy. 4th ed. Baltimore, MD: Lippincott Williams and Wilkins; 2010:415-416.
3. Chao KSC. Practical Essentials of Intensity Modulated Radiation Therapy. 2nd ed. Baltimore, MD: Lippincott Williams and Wilkins; 2005:5.
4. Saw CB, Chen H, Beatty RE, Wagner H. Multimodality image fusion and planning and dose delivery for radiation therapy. Med Dosim. 2008;33:149-155.
5. Heron DE, Smith RP, Andrade RS. Advances in image-guided radiation therapy-the role of PET/CT. Med Dosim. 2006;31:3-11.
6. Workman RB, Coleman RE. PET/CT Essentials for Clinical Practice. 1st ed. New York, NY: Springer; 2006;3-12.
7. Odero DO, Hartley JR, Shimm DS. Positron emission tomography and radiation therapy computerized treatment planning systems. J Mech Med Biol. 2008;8:235-250.
8. Berson AM, Stein NF, Riegal AC, et al. Variability of gross tumor volume delineation in head-and-neck cancer using PET/CT fusion, part II: the impact of a contouring protocol. Med Dosim. 2009;34:30-35.
9. Khan FM. Treatment Planning in Radiation Oncology. 2nd ed. Baltimore, MD: Lippincott Williams and Wilkins; 2007:120,177.
10. Palmer MB. Post-RT PET study for distal esophagus reoccurrences and the feasibility of dose escalation. Paper presented at: Radiation Oncology Symposia for Therapists and Dosimetrists; May 2008; Daytona Beach, FL.
11. Washington CM, Leaver D. Principles and Practice of Radiation Therapy. 3rd ed. St. Louis, MO: Mosby Elsevier; 2010:126-130.
12. Berrang TS, Truong PT, Popescu C, et al. 3D ultrasound can contribute to planning CT to define the target for partial breast radiotherapy. Int J Radiat Oncol Biol Phys. 2009;73:375-383.
13. Boda-Heggemann J, Mennemeyer P, Wertz H, et al. Accuracy of ultrasound-based image guidance for daily positioning of the upper abdomen: an online comparison with cone beam CT. Int J Radiat Oncol Biol Phys. 2009;74:892-897.
14. Grosu, AL, Lachner R, Wiedenmann N, et al. Validation of a method for automatic image fusion (Brainlab System) of CT data and 11C-methionine-PET data for stereotactic radiotherapy using a linac: first clinical experience. Int J Radiat Oncol Biol Phys. 2003;56:1450-1463.
15. Pelizzari CA. Image processing in stereotactic planning: volume visualization and image registration. Med Dosim. 1998;23:137-145.
16. Sharpe M, Brock KK. Quality assurance of serial 3D image registration, fusion, and segmentation. Int J Radiat Biol Phys. 2008;71:S33-S37.
17. Mutic S, Dempsey JF, Bosch WR, et al. Multimodality image registration quality assurance for conformal three-dimensional treatment planning. Int J Radiat Onc Biol Phys. 2001;51:255-260.
18. Earley L. Squamous cell carcinoma of the head and neck [case study]. Alta Bates Comprehensive Cancer Center.
19. Earley L. Squamous cell carcinoma of the anal canal [case study]. Alta Bates Comprehensive Cancer Center.
20. Earley L. Squamous cell carcinoma of the right oropharynx [case study]. Alta Bates Comprehensive Cancer Center.
21. McKenzie C. Small-cell lung cancer [case study]. University of Florida Proton Therapy Center.
22. McKenzie C. Supratentorial primitive neuroectodermal tumor [case study]. University of Florida Proton Therapy Center.
|What did you think of this article?
The Role of Image Fusion in Medical Dosimetry
|»||Comment From: handsinthedirt||» Posted on: 03/10/2010 5:07 AM|
|WOW! This is my 4th test on this website...Extremely well written, but an extreme overload of information for just 1 credit.|
|»||Comment From: leggetts||» Posted on: 06/16/2010 18:52 PM|
|»||Comment From: Nancy Palmisano||» Posted on: 10/03/2010 19:30 PM|
|A lot of good information|
|There are 8 total comments: View All Comments|