Computed Tomography in Radiotherapy Treatment Planning

Technical Aspects of Positron Emission Tomography/Computed Tomography in Radiotherapy Treatment Planning Paola G. Scripes, MS, and Ravindra Yaparpalvi, MS The usage of functional data in radiation therapy (RT) treatment planning (RTP) process is currently the focus of significant technical, scientific, and clinical development. Positron emission tomography (PET) using (18F) fluorodeoxyglucose is being increasingly used in RT planning in recent years. Fluorodeoxyglucose is the most commonly used radiotracer for diagnosis, staging, recurrent disease detection, and monitoring of tumor response to therapy (Lung Cancer 2012;76:344-349; Lung Cancer 2009;64:301-307; J Nucl Med 2008; 49:532-540; J Nucl Med 2007;48:58S-67S). All the efforts to improve both PET and computed tomography (CT) image quality and, consequently, lesion detectability have a common objective to increase the accuracy in functional imaging and thus of coregistration into RT planning systems. In radiotherapy, improvement in target localization permits reduction of tumor margins, consequently reducing volume of normal tissue irradiated. Furthermore, smaller treated target volumes create the possibility of dose escalation, leading to increased chances of tumor cure and control. This article focuses on the technical aspects of PET/CT image acquisition, fusion, usage, and impact on the physics of RTP. The authors review the basic elements of RTP, modern radiation delivery, and the technical parameters of coregistration of PET/CT into RT computerized planning systems. Semin Nucl Med 42:283-288 © 2012 Elsevier Inc. All rights reserved.

Introduction to Radiation Treatment Planning (RTP)

C

omputerized treatment planning, the virtual juxtaposition of radiation dose distributions on anatomical images, has become the standard of care in radiation oncology. Although ubiquitous in modern practice, this trend started in the late 1970s, when point doses calculated at certain depths within patients were the only available tools. From 2-dimensional dose maps on patient contours, computerized plans developed into complex dose distributions superimposed onto full 3-dimensional (3D) anatomical reconstructions based on computed tomography (CT) and magnetic resonance imaging (MRI) data sets. In radiation therapy (RT), specification of volumes and doses is required for consistent prescription, recording, and

Department of Radiation Oncology, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, NY. Address reprint requests to Ravindra Yaparpalvi, MS, Department of Radiation Oncology, Montefiore Medical Center, Albert Einstein College of Medicine, 111 East 210th Street, Bronx, NY 10467. E-mail: ryaparpa@ montefiore.org

0001-2998/12/$-see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1053/j.semnuclmed.2012.04.006

reporting of dose. The International Commission on Radiological Units and Measurements (ICRU) was founded in 1925 to develop quantities and units for radiation as well as to establish methods for their measurements and provide physical data required for this process. Three definitions for tumor (target) were proposed in ICRU Report 501: gross tumor volume (GTV), clinical target volume (CTV), and planning target volume (PTV) (Fig. 1). The GTV is the detectable extent of a tumor, for example, by imaging. The GTV is considered the volume of highest density of malignant cells. The CTV surrounds GTV that contains local subclinical spread. The density of the tumor cells is typically assumed to reduce with distance from the GTV. Unlike the GTV, CTV is determined anatomically not by imaging the tumor itself; a typical example is a nodal area at high risk close to the primary, without any evidence of tumor involvement on imaging. If tumor has been removed by surgery, no GTV exists and the CTV becomes the primary target in adjuvant RT (anatomical concept). The PTV is the geometrical volume around the CTV (geometrical concept). The PTV is the margin that lies beyond the CTV and accounts for movement of tissues (eg, patient respiration), variations in tissues that contain CTV (eg, full bladder), and variations in beam geometry. 283

P.G. Scripes and R. Yaparpalvi

284

fore, PTV can be considered as a 3D envelope in which the tumor and any macroscopic extensions reside. In addition to target volumes, the ICRU defines 2 concepts that indicate the volume in the patient is actually receiving dose: the treated volume is the volume enclosed by an isodose surface and specified by the radiation oncologist as being appropriate for treatment (such as 95% isodose); the irradiated volume is the tissue volume that receives a dose that is considered significant in relation to normal tissue tolerance. It is usually represented by the 20% isodose line.

The Modern Linear Accelerator and Radiation Planning

Figure 1 Target Volume Definitions for a lung cancer patient (definitions based on ICRU 50 and 62 concepts). (Color version of figure is available online.)

The expansion from GTV ⫹ CTV to PTV involves a margin that depends on the uncertainty of being able to direct the radiation beam to the target. The sources of uncertainties in RT delivery are (a) motion and deformation of the target and (b) the ability to position the target reproducibly in the radiation beam. PTV will approximately be equal to CTV if there is very little scope for movement during treatment (more normal tissue spared). In a subsequent report (ICRU 62),2 the margin between the CTV and the PTV has been further subdivided into 2 more clearly described volume subsets: internal margin (IM) and setup margin. The PTV margins are typically applied without distinguishing between IM and setup margin. In principle, IM can be determined using timeresolved imaging, such as 4-dimensional (4D) CT. This allows for creation of internal target volume (such as in gynecological cases where in empty and full bladder scans are obtained to determine the extent of target motion). There-

Linear accelerators used for treatment delivery have become fully digital devices, with dynamic multileaf collimators (DMLCs) replacing the century old collimator jaws. DMLCs are composed of 80-120 pairs of computer controlled leadtungsten leaves that move in concert with the accelerator’s gantry and produce complex photon fluence maps known as intensity-modulated radiation therapy (IMRT), which replaced the homogeneous radiation beam. The variation of beam intensity through space allows for dose distributions that closely resemble the tumor target shape and spare normal tissues, increasing the therapeutic ratio of radiation tremendously. IMRT with “3D dose painting” is the current golden standard of curative radiation treatment (Fig. 2). Because IMRT produces inhomogeneous photon fluence maps rather than a homogeneous beam of radiation, it is effectively a composition of millions of pixel-size “beamlets” that will be deposited in tissue. For computerized radiation planning, this has effectively eliminated the possibility of virtually placing radiation beams on patients CT scans. Inverse treatment planning algorithms were then developed, where physician and physicist work in conjunction to establish “Dose Volume Histograms” for the target tumor and normal tissues, and the planning system optimizes the position of the linear accelerator gantry and the movement of the DMLC leaves to obtain the desired dose distributions. This allows for the unprecedented possibility of delivering different doses to subvolumes of the target tumor and for effec-

Figure 2 “Dose Painting” with intensity-modulated radiation therapy. From left to right: coronal, axial, and sagital views showing dose sculpting, giving radiation dose to tumor while sparing normal structures as spinal cord and parotids.

Technical aspects of PET/CT in RTP tively constraining the radiation dose to the normal tissues (Fig. 2). The increased precision of radiation planning and dose delivery has created a new paradigm where precision and localization are essential. Linear accelerators are now outfitted with onboard imaging devices, with which CT scan may be quickly obtained before the initiation of each radiation fraction, whereas Kilo Voltage (KV) fluoroscopy is used to accompany the appropriate positioning during delivery. Fiducial markers were developed to track tumors of difficult visualization, and a global positioning system-based implantable system allows for tracking prostate movement during a radiation session. The use of 4D CT scans for radiation planning with motion tracking has become routine in advanced radiation centers, while treatment delivery is handled by respiratory gating. Images from 4D CT scans can be transformed into composite target volumes that account for motion and deformation in maximum intensity projection reconstructions. In scientific visualization, a maximum intensity projection is a volume-rendering method for 3D data that projects in the visualization plane the voxels with maximum intensity that fall in the way of parallel rays traced from the viewpoint to the plane of projection.

Positron Emission Tomography (PET)/CT and Radiation Planning The role of PET3-6 scanning in radiation oncology treatment planning has simultaneously evolved. Initially, patients with disease outside the proposed radiation beam were spared ineffective local treatment. This was followed by the realization that better outlining of tumor anatomy could make PET scanning the tool that would help make the promises of IMRT for normal tissue sparing a safe reality. The quest to treat the entire tumor and nothing more than tumor goes hand in hand with the prowess of PET scanning in many oncologic diseases. IMRT dose painting and different levels of uptake within a target tumor make PET and IMRT highly complementary planning methodologies. As evidence accumulates that tumor do not behave passively in response to fractionated radiation sessions, risk adaptive treatment and dynamic adaptive treatment planning is becoming extremely important. New PET probes are being

285 developed specifically for this purpose, and studies should clarify the timing and reliability of on-treatment scanning. The initial thought that the interference of radiation with fluorodeoxyglucose update was a liability is being transformed into a major line of scientific discovery, which may result in an innovative role for PET scanning during the course of RT to be used in risk adaptive radiation planning.

Image Acquisition and Registration CT is the primary imaging modality in RT planning. The CT images are obtained in a RT simulation procedure with the patient lying on a flat table top aided by immobilization devices and positioning lasers. CT images provide for both tumor delineation as well as the electron density data necessary for accurate dose calculations in RT treatment planning process. All other imaging modalities (such PET, magnetic resonance imaging) are considered as secondary images. The secondary images in RT planning will have to be registered (fused) to the primary planning CT scan. The fusion between both PET and CT provides anatomic information to improve the tumor localization and characterize sites of radiotracer uptake (Fig. 3). The fusion process can be executed automatically by a hybrid PET-/CT-dedicated RTP scanner. In cases where the PET and RTP CT images are acquired on separate scanners, a registration module in the RTP computer system can be used to fuse images. Although the hybrid PET/CT scanner appears to have the advantage of producing coregistered images, the bore size on these scanners may prove to be the limiting factor in accommodating sometimes bulky immobilization devices that are necessary for patient positioning in RT treatments. By contrast, fusing separately obtained CT and PET images in RTP planning system could be limited by window/ level visualization of the 2 image sets. Two methods of image registration are commonly used: rigid and deformable type. Positioning a patient in the RT treatment position during the diagnostic staging PET scan acquisition on a flat couch insert, improves the accuracy of rigid registration of staging PET and RTP CT scans.7 In the deformable registration, potential differences in image data sets, such as those caused by differences in anatomical positioning, are reduced by estimating the spatial relationship

Figure 3 Positron emission tomography (PET)/computed tomography (CT) fusion. From left to right: CT image, PET image, and CT/PET images fused.

P.G. Scripes and R. Yaparpalvi

286

Figure 4 PET/CT fusion role in delineation of tumor. Left: CT image showing gross tumor. Right: PET image showing disease uptake. Middle: PET/CT fusion correlating gross tumor with uptake in PET.

between the volume elements of the image sets. Deformable registration has been shown to improve the accuracy of registration of a staging PET/CT and RTP CT scans in head and neck cancer patients.8

Clinical Considerations The introduction of the combined PET/CT imaging modalities into routine clinical RT practice has promise to be of great clinical significance in the accurate delineation of RT target volumes (Fig. 4). Excellent reviews on the subject of PET-/ CT-guided RTP are available in literature.9-11 Currently PET-/ CT-guided RTP is being increasingly applied in the treatment of cervix,12,13 head and neck,14,15 and lung cancers.16-18 The main hallmark of RT is precise and reproducible immobilization. Head and neck masks, body vacuum bags (Vac Loc; CIVCO Medical Solutions, Kalona, IA), and several body frames are always used. When PET/CT fusion is used for treatment planning, it is important that the immobilization device fits in the PET/CT bore and that patients are comfortably placed in the apparatus to minimize uptake caused by nonresting positions. Considerations for specific disease sites are important for developing technical routines that will allow for best PET/CT fusion for radiation planning. In head and neck cancers, the proximity of the brain, the normal salivary gland and lymphoid tissue uptake has to be considered. Immobilization devices, such as masks and shoulders, are relatively easy to position on the PET/CT table. For lung and other thoracic malignancies, the use of gated PET/CT acquisition is becoming more popular. For abdominopelvic malignancies, good bowel preparation and proper motion analysis is a fundamental step. With increasing concern for diagnostic radiation exposure, especially in young patients submitted to curative patients, disease site-specific repeat PET/CT are being considered for risk adaptive radiation planning purposes.

Target Motion Management The problem of target motion is such an important issue that it is the subject of a separate article that follows this one by V. Bettinardi, M. Picchio, N. Di Muzio, and M.C. Gilardi. However, we are including a brief overview of the subject as a preface to that article and for the sake of completeness of this one.

When one is working clinically with thoracic tumors and disease sites influenced by respiratory motion, accurately localizing targets on PET and CT images remain challenging. Respiration can introduce artifacts in CT images caused by the interaction between the axial images acquisition and the motion of the tumor and healthy tissues. These artifacts may potentially deteriorate the quality of CT images and, therefore, the accuracy of diagnosis and/or localization of lesions. For PET, respiratory motion also causes artifacts that ultimately results in degraded image quality and quantification. In PET imaging, the data are usually collected for 3-7 minutes per bed position (field of view) and, therefore, are time averaged over many breathing cycles. Therefore, respiratory motion will result in blurring of the lesion, consequently underestimating the corresponding SUV (specific uptake value) and overestimating the lesion volume. As the SUV is the most clinically used quantitative parameter to stratify lesions as malignant or benign, to stage disease, and monitor their response to treatment, an underestimation of SUVs can result in inaccurate diagnoses, if not potential misdiagnoses.19 Nevertheless, PET has shown an increase in both sensitivity and specificity over CT in lung cancer.20 The use of PET in radiotherapy can potentially improve the quality of radiotherapy planning, minimize the risk of geographic misses, and spare unnecessary toxicity to normal tissues not affected with tumor cells. However, careful attention must be paid to the artifacts introduced by breathing motion can lead to overestimation of target volumes and, therefore, an unnecessarily large radiation dose to the normal tissues. Another cause for reduced SUVs is the spatial mismatch between PET and CT, which results in inaccurate attenuation correction. The mismatch occurs because of the difference in the image acquisition times between PET and CT. A CT image is collected at a distinct phase of the respiratory cycle, whereas a PET image is a time-averaged image over many breathing cycles. The misalignment between the PET and CT can compromise the interpretation of PET images, resulting in potential mislocalization of the lesion and inaccurate quantitation of the SUV values.21 This emphasizes the need for methods to correct for the respiratory motion effects in all 3 imaging modalities: CT, PET, and PET/CT. Respiratory motion correction in CT is mainly done by either deep inspiration breath-holding technique (DIBH) or 4D CT imaging. The 4D CT scan provides images in all phases of the breathing cycle.22 Although DIBH

Technical aspects of PET/CT in RTP significantly reduces respiratory tumor motion, it requires active patient cooperation and may not be well tolerated by patients with already reduced lung capacity. By contrast, the major drawback of 4D CT is the increased radiation dose to the patient to acquire repeated images to the same axial slice at all different breathing cycles.23 In PET, the correction for motion artifacts caused by respiration is mainly performed through three protocols: gated PET (4D PET), respiratory-correlated dynamic PET, and DIBH PET. The respiratory-gated PET is done similarly to the 4D CT scan, where data are acquired into discrete bins in synchrony with the breathing cycle. Different respiratory tracking systems have been investigated to monitor respiratory motion and they include pressure sensor elastic belt placed around the abdomen, spirometer measuring flow of respired air, thermoprobe measuring the temperature of respired air, and infrared stereovision system tracking the motion of thoracic markers (such as the real-time position management respiratory gating system (RPM, Varian Medical Systems, Palo Alto, CA) or active breathing control.23,24 Gated PET has shown to reduce lesion blurring, increase lesion SUV, and improve lesion detect ability.25 Respiratorycorrelated dynamic PET is a method to perform separation of the respiratory phases without making use of gating PET. Although not requiring hardware to track the respiratory motion, it does require much more computation time than gated PET. In the DIBH technique, patients are required to take a deep breath and to hold it for about 9-20 seconds, while the PET data are acquired. The disadvantage of this method is that some patients do not tolerate holding their breath well for this amount of time, and in the case the patient failed to hold the breath, additional scans need to be acquired.26 Accurate correction of motion artifacts in PET/CT imaging is a product of a combination of the techniques for respiratory motion management in PET and CT scans and includes 4D PET/CT, DIBH PET/CT, attenuation correction in PET images by average CT, and motion-corrected PET reconstruction. In 4D PET/CT, both CT and PET are acquired with respiratory motion tracking, and both scans need to spatially match at each phase of the breathing cycle. The 4D CT is acquired and sorted into 10 groups, according to their corresponding phase of the breathing cycle. The PET is also acquired using gating and then both 4D CT and 4D PET are correlated according to the respiratory phase. The 4D PET/CT has shown to reduce blurring in the images, improve the accuracy in PET/CT coregistration, and increase the measured SUV.27,28 DIBH PET/CT has also shown to significantly reduce motion artifacts, enabling better target localization, as well as to increase SUV values.29 Similarly, improvements in spatial matching of PET and CT have been demonstrated with the attenuation correction in PET images by average CT technique, consequently increasing tumor SUVs. However, this method does not correct for motion.30 Despite all the techniques available to reduce motion artifacts in PET scans, gated PET exhibits reduced statistics because of longer acquisition time compared with clinical PET. Consequently, several approaches have been

287 studied to improve image statistics by combining counts from all gated PET bins at same time preserving temporal resolution.31,32

References 1. International Commission on Radiation Units and Measurements: Prescribing, recording and reporting photon beam therapy. Report 50. Bethesda, MD, ICRU, 1993 2. International Commission on Radiation Units and Measurements: Prescribing, recording and reporting photon beam therapy. Report 62 (Supplement to ICRU Report 50). Bethesda, MD, ICRU, 1999 3. Lin M, Ambati C: The management impact of clinically significant incidental lesions detected on staging FDG PET-CT in patients with non-small cell lung cancer (NSCLC): An analysis of 649 cases. Lung Cancer 2012;76:344-349 4. Subedi N, Scarsbrook A, Darby M, et al: The clinical impact of integrated FDG PET-CT on management decisions in patients with lung cancer. Lung Cancer 2009;64:301-307 5. Ong S, Schoder H, Lee NY, et al: Clinical utility of 18F-FDG PET/CT in assessing the neck after concurrent chemoradiotherapy for locoregional advanced head and neck cancer. J Nucl Med 2008;49:532-540 6. Quon A, Fischbein NJ, McDougall IR, et al: Clinical role of 18F-FDG PET/CT in the management of squamous cell carcinoma of the head and neck and thyroid carcinoma. J Nucl Med 2007;48:58S-67S 7. Nestle U, Walter K, Schmidt S, et al: 18F-deoxyglucose positron emission tomography (FDG-PET) for the planning of radiotherapy in lung cancer: High impact in patients with atelectasis. Int J Radiat Oncol Biol Phys 1999;44:2593-2597 8. Hwang AB, Bacharach SL, Yom SS, et al: Can positron emission tomography (PETP or PET/CT) or PET/Computed Tomography (CT) acquired in a nontreatment position be accurately registered to a headand-neck radiotherapy planning CT? Int J Radiat Oncol Biol Phys 2009; 73:578-584 9. Ford EC, Herman J, Yorke E, et al: 18F-FDG PET/CT for image-guided and intensity-modulated radiotherapy. J Nucl Med 2009;50:16551665 10. Zaidi H, Vees H, Wissmeyer M: Molecular PET/CT imaging-guided radiation therapy treatment planning. Acad Radiol 2009;16:11081133 11. Macapinlac HA: Clinical applications of positron emission tomography/computed tomography treatment planning. Semin Nucl Med 2008;38:137-140 12. Dolezelova H, Slampa P, Ondrova B, et al: The impact of PET with 18 FDG in radiotherapy treatment planning and in the prediction in patients with cervix carcinoma. Results of pilot study. Neoplasma 2008;55:437-441 13. Lin LL, Mutic S, Low DA, et al: Adaptive brachytherapy treatment planning for cervical cancer using FDG-PET. Int J Radiat Oncol Biol Phys 2007;67:91-96 14. Paulino AC, Koshy M, Howell R, et al: Comparison of CT- and FDGPET-defined gross tumor volume in intensity-modulated radiotherapy for head-and-neck cancer. Int J Radiat Oncol Biol Phys 2005;61:13851392 15. Henriques de Figueiredo N, Barret O, Demeaux H, et al: Comparison between CT- and FDG-Pet defined target volumes for radiotherapy planning in head and neck cancers. Radiother Oncol 2009;93:49-482 16. Erdi YE, Rosenzweig K, Erdi AK, et al: Radiotherapy treatment planning for patients with non-small cell lung cancer using positron emission tomography (PET). Radiother Oncol 2002;62:51-60 17. Greco C, Rosenzweig K, Cascini GL, et al: Current status of PET/CT for tumour volume definition in radiotherapy treatment planning for nonsmall cell lung cancer (NSCLC). Lung Cancer 2007;57:125-134 18. Bradley J, Thorstad WL, Mutic S, et al: Impact of FDG-PET on radiation therapy volume delineation in non–small-cell lung cancer. Int J Radiat Oncol Biol Phys 2004;59:78-86 19. Pevsner A, Nehmeh SA, Humm JL, et al: Effect of motion on tracer activity determination in CT attenuation corrected PET images: A lung phantom study,”. Med Phys 2005;32:2358-2362

288 20. Toloza EM, Harpole L, McCrory DC: Noninvasive staging of non-small cell lung cancer: A review of the current evidence. Chest 2003;123: 137S-146S 21. Erdi YE, Nehmeh SA, Pan T, et al: The CT motion quantitation of lung lesions and its impact on PET-measured SUVs. J Nucl Med 2004;45: 1287-1292 22. Yorke E, Mageras GS: Deep inspiration breath hold and respiratory gating strategies for reducing organ motion in Radiation treatment. Semin Radiat Oncol 2004;14:65-75 23. Nehmeh SA, Erdi YE: Respiratory motion in positron emission tomography/computed tomography: A review. Semin Nucl Med 2008;38: 167-176 24. Martínez-Möller A, Bundschuh R, Riedel M, et al: Comparison of respiratory sensors and its compliance for respiratory gating in emission tomography. J Nucl Med 2007:426 25. Nehmeh SA, Erdi YE, Ling CC, et al: Effect of respiratory gating on reducing lung motion artifacts in PET imaging of lung cancer. Med Phys 2002;29:366-371

P.G. Scripes and R. Yaparpalvi 26. Nehmeh SA, Erdi YE, Rosenzweig KE, et al: Reduction of respiratory motion artifacts in PET imaging of lung cancer by respiratory correlated dynamic PET: methodology and comparison with respiratory gated PET. J Nucl Med 2003;44:1644-1648 27. Nehmeh SA, Erdi YE, Pan T, et al: Four-dimensional (4D) PET/CT imaging of the thorax. Med Phys 2004;31:3179-3186 28. Lamb JM, Robinson C, Bradley J, et al: Generating lung tumor internal target volumes from 4D-PET maximum intensity projections. Med Phys 2011;38:5732-5737 29. Nehmeh SA, Erdi YE, Meirelles GS, et al: Deep-inspiration breath-hold PET/CT of the thorax. J Nucl Med 2007;48:22-26 30. Pan T, Mawlawi O, Nehmeh SA, et al: Attenuation correction of PET images with respiration-averaged CT images in PET/CT. J Nucl Med 2005;46:1481-1487 31. Klein G, Reutter R, Huesman R: Four-dimensional affine registration models for respiratory-gated PET. IEEE Trans Nucl Sci 2001:756-760 32. Menke M, Atkins M, Buckley K: Compensation methods for head motion detected during PET imaging. IEEE Trans Nucl Sci 1996:310-317