PET‐based GTV definition is the future of radiotherapy treatment planning

In treatment planning, the gross tumor volume (GTV) is most commonly defined using CT, sometimes supplemented by MRI. It has been suggested that in the future such volumes will likely be defined primarily with the use of positron emission tomography (PET). This is the claim debated in this monthˈs Point/Counterpoint. Arguing for the Proposition is Salahuddin Ahmad, Ph.D. Dr. Ahmad received his Ph.D. degree in Physics from the University of Victoria, Canada in 1981 followed by postdoctoral training in Medical Physics from UT MD Anderson Cancer Center. He has been a faculty member at Rice University and Baylor College of Medicine, and the Chief Physicist at the Houston VA Medical Center. He is currently Director of Medical Physics and the Medical Physics Residency Program and Full Professor in the Department of Radiation Oncology at the University of Oklahoma Health Sciences Center. He is an Editorial Board Member of the JACMP and Medical Dosimetry, Bangladesh Liaison for the AAPM, a Fellow of the ACMP, and an ABR Diplomate. Arguing against the Proposition is Slobodan Devic, Ph.D. Dr. Devic obtained his B.Sc., M.Sc., and Ph.D. degrees in Physics from the University of Belgrade, Belgrade, Serbia, and subsequently completed a Medical Physics Residency at McGill University, Montreal, Canadain 2002. Since then he has worked as a Medical Physicist at McGill University, first, in the Department of Medical Physics, Montreal General Hospital and, currently, in the Radiation Oncology Department, Jewish General Hospital, where he is an Assistant Professor. His major research interests are PET/CT-based treatment planning for lung and rectal carcinomas, new treatment techniques for GI cancers, and GafChromic film dosimetry. Dr. Devic is a Member of the Medical Physics Editorial Board. PET is a functional imaging method that has become widely used in radiation simulation over the last decade. The most critical component of radiotherapy treatment planning (RTP) is delineation of the GTV, which is essential to deliver a high dose to the malignant tissue, while keeping the dose to surrounding tissue low. FDG-PET is also routinely used for diagnosis and staging of several types of cancers including nonsmall cell lung cancer (NSCLC). Conventional CT simulation for GTV delineation introduces uncertainties into RTP because of difficulties in determining tumor margins, especially in NSCLC with atelectasis, pleural effusion, pneumonitis, or normal tissue displacement, and its limitations in identification of tumor-involved local lymph nodes. Also, neither CT nor MRI is well suited for distinguishing if hilar and/or mediastinal lymph nodes are involved. These factors contribute to marked variability in GTV delineation among even experienced radiation oncologists. In lung cancer staging, FDG-PET has proven to have greater sensitivity and specificity than CT or MRI. PET data complement anatomic data provided by CT, help distinguish tumor from normal anatomy and consequently lead to a more consistent delineation of the GTV with the reduction of inter- and intraobserver variation. PET image quality improvements and tools for accurate segmentation and quantification have been developed recently. Among various segmentation methods, the gradient-based technique best estimated true tumor volume, out-performing threshold-based techniques in accuracy and robustness for delineation of primary tumor volumes in NSCLC.1 For the quantitative implementation of PET data into GTV delineation, the differences in reproducibility of segmentation technique-based contoured volumes, however, should be kept in mind.2 Respiratory motion of lung tumors during PET image acquisition causes artifacts affecting quantification of FDG uptake and determination of tumor size. Respiratory gating using 4D-PET, such as 4D CT, can show us where the tumor is at a given time and how it moves. This capability removes image degradation associated with the partial volume effect, allowing us to see involved lymph nodes and other tumor activity where it may have been blurred out in 3D-PET. 4D-PET provides biological characterization of tumors, e.g., regions of hypoxia or necrosis, in sharp detail. This provides greater treatment efficacy, improves tumor edge definition, defines the full physiologic extent of moving tumors, and thus improves RTP for lung tumors. In addition, reduction of blurring from free-breathing images may reveal additional information regarding regional disease.3 Use of PET for GTV definition in RTP, which has been validated for lung, head, neck, and brain tumors, was found to be beneficial in colorectal cancer,4 and influenced CT-based RTP for locally recurrent nasopharyngeal carcinoma by changing target volume definition.5 Recently, PET/CT derived tumor volumes were found smaller than those derived by CT, with nodal GTV contours changed in 51% of patients.6 PET signals may now be used to define a subvolume in a CT-derived GTV to deliver escalated doses inside the GTV for more radio-resistive areas. In summary, PET for GTV delineation appears beneficial and has tremendous promise to impact RT planning and treatment.7 Over the last decades, radiotherapy treatment planning has been based on an anatomical object, namely the gross tumor volume (GTV), from which the clinical target volume (CTV) and the planning target volume (PTV) are derived. Because of its superb spatial reproducibility and the ability to provide information on electron density, computed tomography (CT) was, and still is, the backbone of three-dimensional radiotherapy treatment planning (RTP). Since the inception of the PET/CT scanner as a single imaging modality, numerous studies have reported on possible changes to the GTV as defined on anatomical images.7,8 Following the initial attempts for target thresholding,7,9–11 numerous variations of the PET-based GTV definition approaches were developed over the years.12,13 However, an undeniable drawback of such approaches is that they tend to create a single PET-based target volume to replace the traditional CT-based GTV. Thus far clear guidelines on how to incorporate PET data into the RTP process have not emerged from either clinical or phantom studies. MacManus14 suggests that the "best judgment" of the radiation oncologist is the guideline to be followed for GTV definition using PET in patients with lung cancer, while Nestle et al.15 stated that "… at this time we can only rely on the qualitative visual approach interpreted by a well-trained nuclear medicine specialist." However, such conclusions deny the specific role of quantitative physiological information contained in functional images such as FDG-PET that could have a role in radiotherapy treatment planning through definition of biological target volumes (BTVs). Instead of replacing the CT-based anatomical information with PET-based functional data, another approach would be to integrate both datasets in such a way that they complement each other. While such harmonization of PET and CT data into the radiotherapy treatment planning process is becoming evident, methods to actually put this complementation in place are not apparent yet. One of the possible scenarios, elaborated by Ling et al.,16 would be creation of BTVs embedded within the previously defined gross-tumor volume. According to his recommendations, a GTV defined by inherently low spatial resolution functional imaging such as PET should not be a surrogate for a CT-based GTV. The frame for one or several BTVs should be gross tumor volume defined with CT, which has superior spatial resolution and reproducibility and, if possible, with MRI for enhanced soft tissue contrast. Once this general frame is delimited, different functional sub-volumes (tiles) can be added to the mosaic. The incorporation of regions with increased FDG uptake as, for example, the glycolytic BTV,17 within the CT-based GTV, may lead to escalated radiotherapy doses being delivered to specific parts of the tumor in order to improve the probability of cure. I agree with Dr. Devic that CT is the backbone of RTP and guidelines to incorporate PET data are not clear. However, PET identifies additional gross disease and detects significant tumor extension outside GTV delineation on CT that needs to be incorporated to enhance GTV precision. Objective approaches to tumor segmentation have been developed where tumor is either defined based on the standardized uptake value (SUV) and auto contouring all areas with a value at or above that marker, or defined as the area enclosed in the 40%–50% intensity level relative to the tumor maximum. Recently, a gradient-based algorithm is being incorporated to aid PET-based GTV definition. I also agree with Dr. Devic that optimal definition of GTV is dependent on integration of multimodality imaging. CT provides morphological information whereas PET offers information about metabolism, physiology, and molecular biology of tumor tissue, improving GTV delineation. PET/CT is becoming a routine imaging tool for radiation oncology because of its combined benefits of improved staging and tumor delineation provided by PET, and high-resolution 3D anatomic display by CT. Together these allow better treatment with precise targeting. The continuing innovations in PET/CT technology, along with increasing availability of these scanners, is rapidly becoming an accepted and routine clinical tool in RTP for anatomic and biologic tumor targeting. Recently, integrated or hybrid PET-CT scanners have become available. These have been shown to be superior to CT alone, or a combination of PET and CT acquired separately. In conclusion, PET images used for RTP improve GTV delineation and offer additional information about tumor biology and physiology. Improvements in PET image quality, segmentation, and quantification are essential components for future biologically based image-guided radiotherapy, the aim of which is to modify dose distributions to particular regions according to voxel intensities identified by functional imaging. Dr. Ahmadˈs arguments are related primarily to the use of PET in redefinition of CT-based CTV (inclusion of proximal nodes) and not GTV. While this role is undeniable, contemporary clinical practice only uses PET data to localize positive nodes or suspicious proximal tissue to be included into CTV definition, which is still performed on CT images. However, replacement of CT-based GTV by PET-based GTV is far from being widely accepted clinical routine, with published data showing no clear clinical importance of such a change. My colleague also points out the importance of image acquisition gating for RTP, and rightfully concludes that it will have the same impact on both PET and CT image quality and interpretation, and I have no problem with that. Taking into account the specific image properties of contemporary medical imaging modalities, the frame of the biological PTV should be ideally defined with both CT, having superb spatial resolution and spatial reproducibility, and MRI, owing to superb soft tissue contrast characteristics. While the CT and MRI provide an overall spatial uncertainty of the order of 2 mm, the spatial resolution of the current commercially available PET/CT systems is more than double this.18 Once the frame of the mosaic (GTV) is set on the radiation treatment planning easel, different subvolumes (tiles) can be added. Tumor and normal tissue physiology, together with the nature of the radiopharmaceutical used, must be taken into account for both quantitative PET signal interpretation and its incorporation into the treatment planning process. It should be remembered that the single intensity value for a given voxel is based on the catabolic activity of more than 105 cells and it is unrealistic to expect that the domains of the subvolumes with different physiological characteristics will be defined by sharp boundaries within the tumor volume. Even if it would be possible to define certain target subvolumes within the predefined GTV, these would be characterized only by the relatively greater abundance of a certain type of cell or metabolic condition. This would not imply exclusion of other biological situations coexisting within the same volume.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.