With the development and widespread clinical use of precise radiotherapy techniques such as 3D conformal and conformal intensity modulation, the determination and outlining of precise target areas are receiving more and more attention. the use of PET technology in radiotherapy planning has, to a certain extent, changed the traditional concept of defining target areas by anatomical images, providing more valuable in vivo biological information for target area determination, especially in 2000. In particular, the emergence and development of PET/CT all-in-one machine in 2000 has eliminated the inconvenience of multiple scans and reduced the error of image alignment, allowing the organic combination and visualization of biofunctional and anatomical target areas, which can be more conveniently used for precise outlining of target areas in clinical radiotherapy planning, guiding the setting of high-dose areas in radiotherapy and helping to enhance the target area dose and reduce the normal tissue irradiation dose. Nevertheless, there is no unified standard for the actual target area outline, and there are many differences and confusions.
1. PET/CT scanning conditions and image fusion methods.
Currently, the tracer used in more than 90% of PET images is 18F-fluorodeoxy-glucose (FDG), which mainly reflects the level of glucose uptake and metabolism of cells. Due to the limitations of the small aperture of the PET scanner and the long scanning time at that time, it was not possible to complete the scans under simulated radiation therapy body conditions. Since 2000, commercially available PET machines are PET/CT all-in-one machines with hardware image fusion and large mechanical aperture, and CT images are used as attenuation correction for PET, thus significantly reducing the scan time (by 40%) and allowing PET/CT to be used as a radiotherapy simulation for positioning, with improved image quality accuracy and reduced fusion errors to less than 3 mm. However, PET/CT all-in-one machines are still mainly used for tumor diagnosis, and PET/CT scans for radiotherapy simulation are not yet covered by health insurance in Europe and the United States. Therefore, re-fusion of images with localized CT scans is still widely used. For lesions located in the brain and pelvis, which are less displaced by physiological activity, the error is smaller with rigid fusion and manual visual adjustment, usually around 3 mm. If the lesion is located in the thorax and abdomen, the displacement of the lesion is large due to the influence of periodic respiration and heartbeat. multi-row spiral CT scans in PET/CT can be completed in seconds at multiple levels and can be seen as instantaneous images, whereas PET scans are usually completed in minutes for a bedside scan that includes the range of activity of multiple respiratory cycles, and the areas of hypermetabolic foci they show can be seen as lesions The area of hypermetabolic foci shown can be seen as the area of homogenization of activity following the effects of physiological respiration and heartbeat. On the one hand, this would result in an enlarged hypermetabolic focal area, on the other hand, it would result in a lower Standardized Uptake Value (SUV), especially in the peripheral part of the lesion, and a blurred image, which would lead to a greater error in the outline of the PET target area and the integration with the CT image. Therefore, the ideal approach is to perform PET/CT scans under respiratory gating or to use the latest generation of 4DPET/CT scans. In the absence of the above, abdominal plate compressions can also be used to limit respiratory motion or to choose to scan under shallow and calm respiratory conditions. The different effects of physiological cyclic motion should be taken into account when analyzing and outlining the respective target areas of PET and CT, especially when the PET lesion exceeds the CT lesion, and the outward expansion distance should be considered appropriately reduced when the outward expansion edge diameter forms the ClinicalTargetVolume (CTV). Or when choosing the threshold value for outlining the SUV of the PET target area, appropriate reduction of the threshold value for automatic outlining should also be considered. Since there are no exhaustive studies and uniform standards in this area, each research and treatment center should be cautious according to their own experience.
2. Who should outline the target area?
Studies of interobserver differences in PET/CT target area outlining have yielded inconsistent results; Caldwell et al. reported that for non-small cell lung cancer target area outlining, PET/CT significantly reduced the difference in outlining target areas between three radiotherapists (compared to CT), and Ciernik et al. and Syed et al. reached similar conclusions in studies of head and neck tumors. However, Riegel et al.’s study of outlined target areas in 16 head and neck tumors showed a significant difference in mean PET/CT target volume between 4 physicians (p=0.0002). The reasons for the differences were analyzed: physicians’ understanding and preference were different when they found inconsistencies in the tumor areas shown between different imaging modalities, some took to outline the overlapping parts of the target areas of both PET and CT imaging modalities, while others favored the tumor extent demonstrated by a single image, and others had no choice but to compromise the approach, which were important factors contributing to the errors between different observers. In contrast, the difference between 2 radiologists and 2 nuclear medicine physicians was not significant, provided that strict target area outlining protocols were in place to guide them. It is worth mentioning that all four physicians involved in this study have more than 10 years of experience and have a deep knowledge of CT and PET diagnostics. Therefore, the accumulated knowledge of PET diagnostics is a core factor in reducing the target area errors of radiation oncologists and nuclear medicine physicians. When CT simulation was first widely used in oncologic radiation therapy planning, radiation oncologists were not used to and good at outlining the target area contours and the structural extent of normal tissues, whereas now they do not feel much difficulty because the development of 3D conformal radiotherapy techniques has forced them to learn more about anatomical diagnostics, but it is relatively difficult to elaborate and interpret PET, and the high contrast uptake may be due to tumor hypermetabolism, but may also arise from artifacts or normal physiologic processes such as brain tissue, heart, urinary tract and gastrointestinal system, and is also frequently seen in areas of post-surgical inflammatory response and radiotherapy. Furthermore, the SUV varies somewhat with the patient’s body mass, body surface area and activity of the injected nuclide, and the range of normal values is still not well established. In view of the above factors, at this stage, it is suggested that nuclear medicine physicians should assist radiotherapists to jointly outline the target area.
3. Selection of target area outline threshold: What SUV threshold should be chosen to define normal and tumor tissues in order to reflect the tumor range most realistically?
SUV is the most common semi-quantitative index for determining the benignity and malignancy of a lesion, and it is the unit for determining the degree of increased lesion activity. It is important to understand its influencing factors and limitations when applying it to radiotherapy planning target area outlining.
There are a number of factors that affect the accuracy of SUV measurement, the main ones being.
(1) The type of tissue cells, and physiological or pathological changes are the most important factors determining cellular glucose uptake.
(2) Blood glucose level: high blood glucose competitively inhibits 18F-fluorodeoxyglucose FDG uptake and reduces SUV. blood glucose above 200 g/L should be postponed in patients, preferably without insulin before FDG injection; otherwise, increased intramuscular FDG uptake leads to increased background and a lower target area/ background ratio.
(3) SUV is time-dependent, and the SUV of malignant lesions usually reaches a higher level 90 min after injection, within a certain time frame, the longer the time, the higher the SUV. Therefore, the image acquisition time after FDG injection must be standardized.
(4) Size of the area of interest: Due to the limitation of PET image resolution, partial volume effect will lead to underestimation of the true SUV of the lesion shown. when the mass size is only 1.5 times the size of the image resolution, the measured SUV is only 60% of the actual SUV, and only when the tumor size reaches 4 times the PET resolution, the difference between the maximum SUV shown and the actual SUV will be The difference between the maximum SUV and the actual SUV is less than 5% only when the tumor size is 4 times the PET resolution.
(5) Body mass or body surface area: The measured SUV should be normalized to the FDG uptake in the diseased tissue in the area of interest based on the total FDG injection dose and the patient’s body mass in order to make it comparable between patients.
(6) Image reconstruction mode: filtered inverse projection reconstruction underestimated the true radioactive counts by 20%, and its underestimation was much larger than that of the stacked-subset reconstruction method (underestimation by 5%); the number of stacked-subset maximum expectation maximization (OSEM) reconstruction also significantly affected the SUV, from 5 to 40 stacked generations, the maximum The maximum SUV increases gradually by 28% from 5 to 40 times. In addition, the accuracy of the measured SUV is affected by such factors as target motion that will cause the measured SUV to be lower than the actual SUV, FDG injections that are missed subcutaneously or remain in the injection vessel, the spatial resolution of the system, incorrect scan detector collimation correction and dose calibration, and the heterogeneity of the tumor itself. Because SUV is only a semi-quantitative index, and many influencing factors make it less accurate, and there is some variation when it is used to define benign and malignant tissues, a relatively uniform threshold range cannot be defined so far. Nevertheless, SUV is still a major parameter for PET imaging for target area outlining in radiotherapy planning.
According to the literature, the commonly used methods for PET target area outlining are.
(1) visual discrimination method, which means that the tumor hypermetabolic range is best outlined by the naked eye based on the experience of nuclear medicine physicians or radiotherapists, and there is a large observer variation in this outlining method. However, since only the density of the PET image is currently retained after transferring the image from the PET/CT image workstation to the radiotherapy planning system, and its inherent SUV data is lost, visually resolved outlining of the PET target area is still a common clinical method. A similar method, known as the halo outlining method, utilizes the halo phenomenon around the PET chromogenic image and outlines the PET target area along the halo in a simple manner. 19 patients with NSCLC were analyzed by Ashamalla et al. This method is extremely simple and practical to outline the GrossTumorVolume (GTV), and the inter-observer error is very small.
Early studies in simulation models showed that the 40%-50% SUVmax. range was closer to the physical size of the FDG vial contained in the simulation model, and thus the 40%-50% SUVmax. threshold was widely used for the PET-defined gross tumor volume (PETdefinedgrosstumorvolume). Ciernik et al. selected 40% or 50% of the maximum SUV of the lesion as the threshold value to outline the target area by studying the physical phase of PET in a body model with FDG containers of different sizes with different activity levels, which can truly reflect the size of the actual container, but the problem is that the distribution of FDG-loving tumors in vivo is not possible. However, the problem is that the distribution of FDG-loving tissues in vivo is not as uniform and regular as that of the body model, and there are often gradients and various morphological variations, despite this, choosing 40% or 50% of the maximum SUV to outline the PET-GTV is still widely used clinically. Biehl et al. showed that the choice of SUVmax. percentage threshold was negatively correlated with the size of the CT target area, and the SUVmax. percentage threshold to achieve a PET target area similar to the CT-GTV was 15% ± 6%, respectively. SUVmax. (tumor diameter ≥ 5 cm); 24% ± 9% SUVmax. (tumor diameter 3-5 cm); 42% ± 2% SUVmax. (tumor diameter ≤ 3 cm). The individualized calculation formula was: %SUVmax.=59×log[(CT-GTV)-18].
(3) SUV=2.5 absolute value outline method: A threshold value of 2.5 for the maximum SUV was chosen to outline PET-GTV. this threshold value is commonly used clinically to differentiate benign and malignant tumors as a diagnostic value, and is the one SUV that most nuclear medicine physicians feel more comfortable in elaborating and defining non-small cell lung cancer as positive. it is also currently used as a threshold value to outline PET-GTV. but for some other tumors, such as lymphoma and head and neck tumors, the SUV threshold that defines a positive tumor requires further study and remains an open question to date.
(4) Target area/ background SUV ratio method: Because of the differences in background SUV values among normal tissues and organs, such as lung SUV is usually less than 1, while liver may be greater than 3, a constant ratio value (usually 3) is obviously not applicable to the outlining of tumor target areas at all sites.
(5) Various other mathematical formulae methods: Since the %SUVmax. threshold varies with lesion SUVmax. which is measured in association with lesion size and uneven FDG metabolic distribution, Black et al. suggested the 31 + 59/(meanSUV) percent threshold method for outlining GTV; while Biehl et al. suggested 59 × log (CT-GTV)-18 formula to calculate the percent threshold.
The above listed SUV thresholds for outlining PET-GTV are set using CT-GTV as a reference and measure, with the aim of designing an SUV threshold to outline a GTV that matches the CT-GTV. The FDG-PET image itself represents the activity level of glucose metabolism of malignant tumor cells, and thus indirectly represents the density and active intensity of malignant tumor cell aggregation, and therefore differs from the anatomical tissue structure changes shown by CT. The above methods and thresholds for outlining the PET target area do not ensure that they distinguish between tumor and normal tissues, and are only semi-quantitative in nature due to many complex factors affecting PETSUV. At present, it is not very clear which of the above outlined methods can most accurately reflect the true extent of tumors in vivo, especially the extent of subclinical lesions defined as the periphery of solid tumors,? Therefore, it is necessary to investigate the correspondence between PET images and pathological specimens. To date, only two reports have been found in the literature, in which Daisne et al. compared the PET-GTV, CT-GTV, and MRI-GTV of 29 oropharyngeal, hypopharyngeal, and laryngeal cancers with the corresponding 9 post-surgical pathology specimens registered for laryngeal cancer, and found that CT, MRI, and PET all significantly overestimated the size of the actual lesion in the naked-eye pathology specimens. In contrast, the PET-GTV outlined by the lesion/background threshold method was the closest to the size of the pathological specimen. It was also found that the CT, MRI, PET outlined target areas and surgical specimens did not overlap with each other by about 10-20% of the volume, which shows that the current imaging tools are not perfect and cannot be sure that the abnormal areas represent the tumor areas. Nevertheless, the authors did not recommend the integrated target area under multiple image fusion for radiotherapy planning, which would lead to larger target areas and more overestimation of the actual tumor size. In a study of 22 cases of esophageal cancer at Shandong Cancer Hospital, Yu Jinming et al. compared in vivo surgical specimens with different threshold values of fluoro-L-thymidine PET/CT (Fluoro-L-ThymidinePET/CT,FLT-PET/CT) and FDG-PET/CT to outline the length of esophageal lesions.FLT-PET/CT used seven threshold methods to outline the length of the target area. Visual method; SUV1.3; SUV1.4; SUV1.5; 20% SUVmax.;25% SUVmax.;30% SUVmax.FDG-PET/CT used 3 threshold methods to outline the length of the target area, visual method, SUV2.5, 40% SUVmax. It was found that FLT-PET/CT, SUV1.4 and FDG-PET/CT, SUV2.5 threshold methods were used to outline the length of the target area. CT, SUV2.5 threshold outlined the length of target area of esophageal cancer was found to be the closest to the length of in vivo surgical specimens. The above two studies comparing with surgical pathology specimens were both comparisons with intraductal malignancies, but the changes in spatial location and structure of the specimens after surgery, as well as changes in tissue structure such as regression after specimen processing, make the post-surgical specimens uncertainly deviate from the original in vivo tumor structure. clinical studies comparing PET threshold outlined GTV with surgical specimens of other solid malignancies have more difficult to solve practical difficulties.
In conclusion, the choice of different thresholds or methods to define PET biological target areas has a great impact on the biological target volume, and arbitrarily raising or lowering the thresholds will eventually lead to serious consequences of underdose in tumor areas and overdose in normal tissues, but if it is clear that the purpose of PET-GTV for radiotherapy planning is only for guiding the arrangement of radiotherapy high-dose areas, the above various PET target area outlining methods may be a good choice. The results of long-term clinical follow-up in the future, especially the study of local recurrence patterns, will finally establish the feasibility of biological target area delineation principles.
4. Principles of target area establishment for the difference of PET and CT target area range: How to determine the target area range when the tumor ranges shown by PET and CT are not consistent?
In most cases, the number, location and extent of tumors shown by PET and CT do not exactly match, and there are usually three situations: firstly, PET can find tumor boundaries not shown by CT; secondly, PET finds abnormal hypermetabolic areas in areas other than tumors shown by CT or distant metastatic sites; thirdly, PET finds biologically active hypermetabolic areas within the scope of tumors shown by CT. .
The vast majority of the literature studying the impact of PET or PET-CT on radiotherapy planning does not describe specific guidelines for outlining the target area when PET and CT do not match in a localized area, which indirectly reflects the dilemma of clinical practice in outlining the target area. In studies of non-small cell lung cancer, it is difficult to determine the actual location and size of the tumor target area by CT in the presence of pulmonary atelectasis, pleural fluid, or obstructive pneumonia, and relies heavily on PET to outline the target area, which is usually a reduced target volume, with the result that the volume and dose to vital organs such as the lung, esophagus, heart, and spinal cord are reduced. At this stage, it is not possible to determine the extent of the tumor pathology target area in vivo by imaging, so the following approach to determine the extent of the target area is to add the structural target area [CT-definedgrosstumorvolume (CT-GTV)] and the biofunctionally defined metabolic hypertrophic area (PET-GTV) to define the anatomical-structural-biofunctional-metabolic combined target area. In a study of 40 head and neck tumor cases, Paulino et al. found that the CT-GTV was larger than the PET-GTV in 30 cases, including 7 cases that were as much as 5-fold larger; the PET-GTV was larger than the CT-GTV in 7 cases, with a maximum ratio of 2.5-fold. If the intensity planning was done based on CT-GTV, 10 cases of PET-GTV received less than 95% of the prescribed dose, and the minimum prescribed dose of 95% PET-GTV was less than 75% in 5 cases, and PET-GTV was not included in the high dose area in about 25% of the cases, so a composite GTV of CT-GTV and PET-GTV added together was recommended for intensity planning.
In the case that the PET target area is located within the CT target area, there are several questions to ponder. To what extent do PET hypermetabolic foci actually represent the actual extent of the living tumor? Is it the case that abnormal CT regions outside of PET hypermetabolism are devoid of malignant tumor cells?The ICRU50 and 62 reports have not yet taken into account the influence of functional molecular metabolic imaging targets when defining the GTV concept, and abnormal anatomical structure regions are taken for granted as tumor regions by default. Therefore, in the absence of conclusive evidence at present, it is a more reasonable choice to define PET hypermetabolic area as radiotherapy thrust target area, as opposed to CT target area.
For detecting local uncontrolled, residual and recurrence, it is difficult to distinguish them from post-treatment structural changes by CT showing anatomical tissue density and structural changes, while FDG-PET has the outstanding advantage of relatively high sensitivity and specificity. In this special case, it is not necessary to ignore the abnormal area shown by CT and only outline the PET active area for the re-treatment, because the normal tissue around the target area has a lower tolerated dose, and the tighter target area is more suitable to achieve the tumor suppression or lethal dose of re-treatment.
For PET-positive and CT-negative or PET-negative and CT-positive lesions, the target area is determined in two cases. For lesions that cannot be identified by CT, such as: structural disorder after surgery; residual or recurrence of lesions determined in fibrosis altered structure after radiotherapy; certain head and neck tumors that CT does not show the extent of lesions, such as carcinoma of the floor of the mouth, soft and hard palate, etc.; lung cancer with lobar opacification; residual and recurrence of some brain tumors after surgery; residual and recurrence after other special treatments, such as whether there is residual after liver cancer intervention, residual of lesions after radiofrequency ablation of tumors area. Only the PET hypermetabolic foci should be outlined as the baseline target, and then radiotherapy should be administered by expanding a certain edge diameter.
For other cases, the trade-off should be determined according to the respective sensitivity and specificity of PET and CT in the diagnosis of the corresponding disease. If the sensitivity of PET is quite high, it means that there are very few false negatives of PET, and even if there is abnormality in CT, it may not be outlined in the high-dose target area. If the sensitivity of PET for mediastinal lymph node metastasis of lung cancer is more than 95%, even if CT shows enlarged lymph nodes and PET is negative, the enlarged lymph nodes should not be outlined as high-dose target area, or it is prudent to outline the abnormal part of CT as preventive irradiation target area and give lower irradiation dose. If the specificity of PET is quite high, it means that there are very few false positives for PET. Even if CT shows normal and PET has hypermetabolic foci, it is more important to trust PET and outline it as a high-dose target area. For example, for the judgment of esophageal cancer lymph nodes, PET specificity is higher, therefore, when PET shows high metabolism in the lymph node area, it should be outlined into the high-dose irradiation area.
The effect of PET on target area outlining has been studied in only a small subset of malignancies, and to date, there are no convincing and very exhaustive studies of the effect of PET on radiation treatment planning for head and neck malignancies, and greater enthusiasm has been placed on the effect of PET on target area outlining in lung cancer. gould et al. used multiple search engines to examine eligible studies published in English and non-English prior to 2003 on The median sensitivity and specificity of CT was 61% and 79%, respectively, while that of PET was 85% and 90%, respectively; when CT showed enlarged mediastinal lymph nodes, the median sensitivity and specificity of PET was 100% and 78%, respectively, while when CT showed no enlargement of mediastinal lymph nodes, the sensitivity and specificity of PET was 82% and 90%, respectively. The sensitivity and specificity of PET were 82% and 93% when CT showed enlarged mediastinal lymph nodes, leading to the conclusion that 18F-FDG-PET was more accurate than CT for staging mediastinal lymph nodes in non-small cell lung cancer, and that the sensitivity of PET was higher when CT showed enlarged mediastinal lymph nodes, leading to the inference that the absence of mediastinal lymph node metastasis should be considered more favourably when PET was negative even though CT showed enlarged mediastinal lymph nodes, because the negative predictive value of PET for mediastinal The negative predictive value of PET for mediastinal lymph nodes is quite high and more reliable, thus it is not necessary to include CT showing enlarged mediastinal lymph nodes in the target area outline or to give only a relatively low irradiation dose that seems reasonable. It was also found that when CT showed no enlarged mediastinal lymph nodes, PET showed a high false-positive rate (about 25%), i.e., if the PET-positive lymph node target area was included, a significant portion of the wrongly expanded target area would be irradiated. However, Bradley et al. also found that PET could detect approximately 40% of CT-negative mediastinal lymph nodes as positive presentations. Furthermore, a prospective study of PET-positive mediastinal lymph node irradiation only by DeRuysscher et al. showed a reduced probability of tumor recurrence outside the planned target area. The results of these studies give reason to believe that at this stage it is reasonable to include PET-positive mediastinal lymph nodes in the irradiation target area.
However, there is not enough evidence to determine the principle of target area management in the management of residual lesions with PET negativity on CT after chemotherapy. For example, if a mass remains in the mediastinum after chemotherapy for lymphoma but PET is negative, it may be a fibrotic nodule, or it may be a temporary suppression of tumor cell activity without complete inactivation. In conclusion, the current statistics are not perfect and the exact correlation between PET negativity and long-term survival is still not very certain.
In conclusion, there are still a series of problems in the outline of PET/CT for precise radiotherapy target area, from the way of PET/CT scanning, the image fusion of different sites, the display and judgment of PET and CT for various types of tumors in different sites, and the selection of different thresholds and different outline methods all directly affect the determination of tumor target area size. Therefore, each tumor treatment center should have its own guideline specification. Therefore, each tumor treatment center should have its own guideline specification to facilitate the summing up of experience and gradually reach consensus in the future.