>> Hello, I'm Kirk Frey. I'm a Physician at University Michigan Hospital and I'm Chief of the Division of Nuclear Medicine. I've been asked to speak with you about the use of FDG-PET imaging in non-small cell lung cancer and this will be the first of four modules on this topic. The first module, this one will discuss the FDG-PET imaging of pulmonary nodules and primary lesions. These are my relevant disclosures. I don't believe any of these will have significant impact on the content that what we're going to discuss. Before beginning the content of this module, let's post a question, which will be addressed by the material that we're going to review. Lesion intensity in FDG-PET tumor imaging, a, distinguishes benign from malignant lesions. Or b, is independent of lesion size. Or c, is predictive of survival in malignant lesions. Or d, is unaffected by patient respiratory motion or respiration. So let's keep these in mind, as we review the following slides. So the learning objectives for this module are for participants to achieve an understanding of FDG tracer properties in PET tumor imaging in general and specifically in thoracic oncology. To learn and understand the technological aspects of thoracic PET. To understand how FDG-PET characteristics of malignant versus benign pulmonary lesions affect our diagnostic impressions and how the use of FDG-PET in patient management and prognosis offers unique advantages. So to begin, we need to understand the tracer properties of FDG labeled with fluorine 18 and it's use in positron tomography imaging. Fluorodeoxyglucose or FDG is injected in a patient intravenously. It arrives in tissues in the vascular space in the plasma, as 18F FDG. If then, enters the tissue via convection and capillary transport where is again, represented as 18 FDG, but now in the tissue space. FDG is then metabolized in most tissue to flourodeoxyglucose or FDG-6 phosphate. In some tissues such as liver, for instance, there is a significant metabolism further of the FDG-6 phosphate back to FDG. These are the descriptions of the rate constants in the picture above where K1 and K2 are primarily representing convection, that is blood flowing capillary transport between the vascular space and the tissue. K3 represents phosphorylation of FDG To FDG-6 phosphate, usually by the enzyme hexokinase. And when present, K4 represents the activity of glucose-6-phosphatase. Dephosphorylating FDG back to the parent tracer FDG. The FDG metabolic rate. In Kinetic terms is given by K1 times the fraction that is a trap in metabolic pool, which is K3 divided by the sum of K2+K3. In the majority of our used of FDG in oncology imaging, we do not routinely obtained time series of images after injection in patients. And instead, rely on a single static image approximately in hour after injection. In these images, uptake can be measured in semi-quantified in a unit refer to is this standardized uptake value or SUV. And this is simply at the time of imaging, the lesion activity divided by the injected radio tracer dose and divided by the patient's body weight. So you will see throughout the series of lectures, the use of the term SUV and this is its derivation and significance. In our laboratory, we follow a rather standardized protocol for use of FDG in PET thoracic imaging. The first patient should fast for four to six hours. We preview six. This minimizes the uptake of radio tracer in skeletal muscle and emphasizes its uptake in our target lesions, which are usually pulmonary lesions, lymph nodes and associated structures. It is important to obtain good IV access and to check the patient's blood glucose level. A significant hyperglycemia will reduce the tissue uptake of FDG and can result in suboptimal imaging. Approximately two hours before imaging, we administer a partial dose of oral contrast for which we use 1% barium sulfate and this is to highlight the gusto intestinal tract. At approximately an hour before imaging, we inject the FDG dose. And in our laboratory, we employ a dose of eight milicarries in an adult patient. 30 minutes before imaging, we administer the other half of the oral contrasts to ensure adequate contrast of the upper gastrointestinal system. The patient is then asked to empty the urinary bladder for comfort and positioned on the scanner table. Our imaging begins with performance of the CT topogram to identify the segmental anatomy of the patient and to define the CT field of view that we will image. We then perform the CT scan, which in our laboratory is a low dose CT obtained with approximately 50 MAS, 120 kVp with 5 millimeters slices in a pitch of 1. This is CT is reconstructive with iterative technique and provide usually a more satisfactory anatomic depiction of the patients thorax. PET imaging is done with respiratory gaining of the thorax and upper abdomen, and the image field of view that we employ for the whole patient involves of the head through the mid-thigh level to cover the areas of interest for the primary as well as possible metastatic lesions. So we mentioned with targeting imaging of thoracic lesions, the use of respiratory gating in PET as performed in our laboratory. This cartoon illustrates a typical patients respiratory cycling during an epoch of imaging with this curve, representing the thoracic volume. The red is inspiratory and the blue is the expiratory phase. The typical use of gating involves what's preferred to as phase-based gating where each of these respiratory cycles is divided into several time slices and they're reconstructed individually. In our laboratory, however, we use an alternative form of respiratory gating referred to as optimal gating. And to do this, we recognize that a majority or at least a very significant fraction of the patients status during the respiratory cycle is near towards the end of tidal respiration. We define a volume range between index expiratory and this near to end tidal respiratory thoracic volume, and this defiance what we refer to as the optimal gate. The case that occur when the patient thorax is in this range. Are added to the gated image, those that occur when the patient's respiratory volume exceeds this threshold are set aside. Well, let's see how this performs. So this is an example of the same patient study reconstructed in three different fashions. In this case, this is a sagittal view through the lower thorax, this is the patient's heart interior, posterior, superior and inferior walls and there are three pulmonary lesions in the nearby lung. This is an ungated acquisition including all of the activity that was recorded while the patient's thorax was scanned. In the next frame is the typical phase base gating where the respiratory cycle was in this case divided into eight time frames. And this is the reconstruction of one of those eight frames. Notice that many of the features in the field of view have apparently higher resolution than in the ungated image but the fact that only one-eighth of the total activity is represented in this reconstruction, makes image noise and other aspects of image quality substantially degraded. If we instead show the optimally gated re-construction, so this static picture that includes only those events that occurred near to the end title respiratory cycle. Notice that unlike the phase gated, the noise characteristics in this image are very favorable, compared to the ungated image, the image is crispier or sharper because it doesn't have the respiratory blur effect. And notice that the apparent intensity in several of these lesions is greater than the ungated image. The effect of this form of gating was investigated in our laboratory when we implemented it originally. To do this, we took 18 subjects who had hypermetabolic lesions in their thoracic gated field of view. We measured the lesion intensity as a peak intensity in the lesion, in both gated, that is optimally gated and ungated images. The results were that the gated to ungated maximum intensity of lesions averaged 1.19. That is about 20% higher in the gated images and this was highly statistically significant. The range of values in the individual subjects and individual lesions was from 0.99 to 1.59 so there were virtually no lesions whose metabolic activity was decreased by the use of optimal gating and the range included lesions that gained as much as 60% in peak activity in the gated images. Let's move on to some examples of actual patients and lesions. This is the result of a pet scan obtained in a subject for characterization of a pulmonary nodule. On the right of the images you see a whole body representation of the maximum intensity throughout the scanned body parts. And you can see here in the left lung, there is a small focus of increased activity and this is the nodule in question. On the left side of the images, are either transaxial in this first column, or coronal image slices through the lesion. In these series, the top images are the metabolic images. The next are CT images obtained at the same time in lung windows. In the bottom is the fusion of those two where the the color represents the PET intensity and the grey scale is the CT. And you can see in this case that there is indeed a nodule situated here in the left lung, near to the aorta, which demonstrates significant, metabolic activity, this is a region and a scan, we would characterized has positive. So most laboratories when analyzing and reporting FTG path studies who characterized a lesion as positive when it's peak metabolic activity exceeds that in vascular background. And here, you can see the lesion clearly exceeds the aortic arch vascular blood pool. Here's another patient, again referred for characterization of a pulmonary nodule. And in this case, reviewing the whole body projection images, we don't see a clear depiction of the lung lesion. Looking at the transaxial images through the lesion we can see in the CT scan, there is a pulmonary nodule located here in the right lung. But in the metabolic imaging of that region, there is no correlative PET signal. This would be characterized or classified as a negative lesion. That is its metabolic activity is certainly not above vascular background and it's probably not even above remote lung background metabolic activity. Does the positive or negative nature of a lung lesion characterize it as neoplastic or benign? And the answer is no. So we know from experience, gain thus far that there are non FDG-avid lung neoplasms. These often include bronchoalveolar adenocarcinoma, carcinoid tumors, and a minority of squamous cell carcinomas. And we'll see in the subsequent slides that this amounts to less than 5% of squamous cell carcinomas, and usually corresponds to those who they well-differentiated histology. Now, in ALM publish an important paper in 2002 in radiology, detailing the overall performance of FDG PET imaging related to characterizing the primary lesions in a long fields of view. In this study, they took almost 200 patients characterized with FDG PET and they correlated the PET findings with the biopsy or resection histopathology in each of these lesions. You can see they included both non-small cell lung cancer which was the focus of our lecture today, but also had a small member of small cell lung cancer patients. I will focus our attention on the non–small cell lung cancer patients for the time being. Well, you can see here is that of the lesions identified 183 were PET positive. Of these, they were all relatively small-sized primary lesions, characterized as T1 tumor size. But you can see that eventually the patients had a range of cancer staging, reflecting spread to lymph nodes and to distant metastatic sites. You can see the majority of these cancers were have FTG positive, however, there were a few that were characterized as FTG negative, and notice that they were all limited stage one, that is stage one A tumors in this regard. Further investigating the specifics of what was found at resection in these patients, you can see that the number of FDG negative scans in the right most column here, and in parentheses reflected as a percentage of these pathologies. Significantly represented carcinoid tumors and adenocarcinoma of the bronchioloalveolar cell subtype. You can see in the line just above this, however, there were 2 patients, 4% of the total, who had squamous cell carcinomas that were characterized as FDG PET negative. So this gives us our less than 5% false negative predictive value regarding squamous cell carcinoma and PET imaging. The other cell types had very limited, if any, significant negative or false negative performance. And furthermore, it was observed by these investigators that the intensity in the primary lesion was predictive of overall patient survival. So here in this graph up to more than 80 months of follow up. You can see in the solid line that the survival of patients who were characterized has lesion PET negative. Aside from the small morbidity and mortalities associated with a resection of the lesion, they survived with excellent prognosis for the remainder of their follow up. Whereas, patients who had FTG positive lesions, that were ressected showed a continuing decline in survival throughout this period. And these were statistically significantly different survival curves. In this study, false negative PET scans were seen as I said in approximately 5% of patients with squamous cell carcinoma. Those patients with a negative PET scan who subsequently prove to have cancer have a significantly longer survival than do patients with a positive scan and cancer. All patients with a negative PET scan who subsequently proved to have cancer in this study had stage 1A disease at the time of tissue diagnosis. None of these cases were upstaged by the use of interval anatomic imaging follow-up to decide on lesion resection. The overall predictive value of FDG-PET lesion intensity and patient survival has been replicated by several authors following this initial report. Higashi in Japan, had reported that not only was proportional survival but the proportion with disease pre-survival was significantly better in patients. Who had a maximum uptake in their lesion of less then five compared to those with a lesion uptake of greater than five. And similarly, survival data in FDG uptake reported by Nair et al., in this paper. It shows that, again using a cutoff threshold, in this case a 10, that patients with high FDG activity, have overall worse prognosis and less effective survival. Now here is another lesion, an example of the patient referred for metabolic imaging of their lung nodule. And you can see on the right in the maximum intensity projection view there's an intense focus of activity here in the left lung. And you can see in the sliced views this is a pleural based apparently predominantly solid lesion with intense FDG activity compared to mediastinal vascular background. Here is a follow up scan in this same patient six months later and these slices transaxially and coronally are taken through the location of the original legion. Note that there are no surgical clips and there are no anatomic abnormalities in this patient's left lung. Turns out that this patient underwent bronchoscopy after their FDG head study which demonstrated absence of malignant cells and presence of bacteria, specially coxiella. The patient was treated for six to eight weeks with intravenous antibiotics and this is the result of their cavitary pneumonia treatment outcome. This case illustrates that many pulmonary lesions may be intensely hypermetabolic but not necessarily neoplastic. FDG positive lesions that represent benign diseases include at least these. So sarcoid and histoplasmosis, that is idiopathic or infectious granulomatous disease or other granulomatous infectious diseases are all often FDG very conspicuous. Pleural inflammation and reaction especially patients undergoing talc pleurodesis or have intense FDG activity associated with this reaction. Following radiation therapy, there can be radiation pneumonitis. This increases FDG activity and may persist long after the termination of radiation. And finally, there are some examples of pulmonary thromboembolism, especially early in the course of the embolic disease, as well, may have increased FDG activity. Final point about the use of FDG PET/CT in primary thoracic lesion characterization is illustrated by this recent paper from the Journal of Nuclear Medicine. In this paper, investigators studied the use of FDG-PET for the targeting of lesion biopsy in those patients who needed to undergo a needle biopsy rather than tolerate a lesion resection in case that lesion was benign. They studied a large number of patients and characterized the true positive versus false negative biopsy results as shown in the table on the right. The most significant indicator of a false negative biopsy was an anatomic disparity between the biopsy needle tip in the lesion versus the peak FDG intensity. The top row, A, B and C, you see illustrated irrelatively solid appearing tumor with uniform FDG activity in it and the crosshairs represent sight that was biopsied. This was a true positive malignant lesion. The bottom row of images, A, B, and C, depict a lesion which has cavitary characteristics anatomically. But we're using anatomic targeting to place the biopsy needle in the middle of the lesion and you can see the crosshairs in C indicate that this targeted area of relatively low metabolic intensity. And this was a false negative biopsy of a cavitary primary lung carcinoma. So at this point, let's return to the question we posed at the beginning o the lecture. Remember, we are trying to identify the best answer to the following statement. Lesion intensity in FDG-PET tumor imaging is or does, a distinguish benign from malignant lesions. B is independent of lesion size; c, is predictive of survival in malignant lesions; and d, is unaffected by patient respiration. The images or the data that we've discussed should allow you to identify c as the best answer to this statement. A is incorrect, we've seen several examples of FTG activity that failed to distinguish benign from malignant regions. Including increased activity in several benign lesions and the possibility of allow level of negative FDG activity in malignant lesions. We didn't discuss specifically B, however this is false. Lesion size is not important necessarily in detecting hyper metabolism and indeed some lesions that are at or below the resolution of the PET scan may appear hypermetabolic. C is correct. We've shown from several studies that the intensity of the primary lesion in an initial FDG PET scan correlates very significantly with overall patient survival and with progression free survival. And then d is also incorrect. We know that lesion intensity is indeed affected by respiratory motion which can be minimized or significantly corrected by the use of respiratory gating in PET imaging. So there's several take home points for this lecture. First, most lung cancers are indeed FDG avid. Second, some benign processes also are FDG avid. Non-FDG- avid lung cancers tend to be less aggressive. And it reasonable to manage patients with FDG negative lesions by serial anatomic followup to identify those that are slowly growing and less aggressive tumors. FDG intensity is predictive of over all survival in patients with malignancy. And FDG can help direct the use of needle biopsy to enhance diagnostic yield in patients who require needle biopsy rather than lesion resection as part of their initial management.
