PET Imaging for GIST


GIST Support International discussed PET scans for gastrointestinal stromal tumor (GIST cancer) with radiologist Annick D. Van den Abbeele, MD, who
was interviewed by GSI science committee member Nancy Berezin. 

Dr. Van den Abbeele is Chief of the Department of Imaging, Director of Nuclear Medicine/PET, and Founding Director of the Center for Biomedical Imaging in Oncology (CBIO) at Dana-Farber Cancer Institute. She also serves as Co-Director of the Tumor Imaging Metrics Core (TIMC) at Dana-Farber/Harvard Cancer Center and is an Associate Professor of Radiology at Harvard Medical School. Her research interests include the use of imaging as a biomarker of therapeutic response to targeted therapy in GIST and other cancers, the development of novel applications and quantitative metrics in cancer imaging, and the validation of imaging as a biomarker. 

Here are Dr. Van den Abbeele’s answers to our questions about her research. 

For a general description of PET scan technology, please refer to


1. What is the underlying difference in metabolism between normal tissues and most cancers that causes tumors to “light up” on PET scan?

All cells require sugar — in the form of glucose — to function. But the glucose uptake of tumor cells is much greater than that of normal cells due to enhanced expression of glucose transporter molecules on the cell surface, increased enzymatic function, and metabolic requirements. PET takes advantage of this basic feature of cancer metabolism. The patient is injected with an analog of glucose that is radiolabeled with a positron emitter (fluorine-18), called fluorodeoxyglucose or FDG.  FDG is transported into the cell and phosphorylated, but then gets trapped within the cell. This accumulation of the tracer provides a picture of tissue metabolic activity that is detected in three dimensions by a ring-shaped PET scanner. The scanner identifies “paired” gamma rays — emitted 180 degrees apart when the positron emitted by the radioisotope decay interacts with an electron — and uses them to localize the source of emission. Because active cancer cells use so much more sugar than healthy cells, we can arrive at a very accurate picture of the level of glucose avidity/tumor metabolic activity anywhere in the body at a specific point in time.


2. We have read that a small percentage of gastrointestinal stromal tumors (GISTs) are not PET avid prior to treatment with drugs such as imatinib (Gleevec).  What is implied about these non-PET-avid tumors? Does lack of PET avidity have any prognostic value?

First, it is important to emphasize that most GISTs are very FDG avid; the number of GISTs that are not PET avid is actually very small — less than 10%. It is also important to rule out confounding factors that may result in false-negative scans.  For example, the patient may have mistakenly eaten a meal prior to injection of radiolabeled glucose. This is why compliance with fasting is important prior to the scan. Also, scans performed in patients receiving steroid therapy, or in patients with a history of diabetes receiving insulin, need to be performed at particular times relative to the administration of the medication to avoid a false-negative scan.  All of these situations may affect tumor FDG uptake as a result of an alteration in glucose metabolism and biodistribution. If the tumor is exceptionally large, the contrast may also not achieve adequate penetration. So, we are reluctant to draw any conclusions regarding the prognostic implications of a non-PET-avid GIST. More work needs to be done to understand the genomic and phenotypic patterns of the rare GIST tumors that would be truly PET negative.  

3. PET scan results report the Standardized Uptake Value (SUV) of tumors.  What is the typical SUV for metabolically active normal tissues such as the liver?  What excess above normal SUV is needed to detect a GIST metastasis to the liver?

The Standardized Uptake Value (SUV), which is used for the semi-quantitative analysis of PET, represents the tissue activity concentration at a particular time point relative to the injected dose divided by the body weight. It is commonly used to assess changes in tumor metabolism relative to normal tissues, but great care has to be taken with regard to the technique used to calculate this number to make it reliable and reproducible and avoid pitfalls. Although the typical background SUV of the blood pool is in the range of 2.5, each patient’s normal background is slightly different, and normal background in the liver is higher than in the blood pool because the liver uses glucose for its normal function. Also, some tumors may actually have an SUV of less than 2.5, while other benign processes such as active inflammation may have an SUV above 2.5. So the key is not to rely on an absolute number, but to identify what is abnormal for the particular patient in his or her particular context and to use all the relevant clinical and imaging information on hand to provide the proper information to the patient and his or her provider. A high target-to-background ratio suggests a pathologic process — but not necessarily a malignancy. It could be an infectious or inflammatory process. The art of medicine lies in interpreting PET and other study results in the context of the patient’s history and physical findings.

4. During what time period post surgery can PET results be artificially elevated by the metabolism involved in wound healing?

FDG uptake may be seen along the surgical wound for the first week or two due to the inflammation and healing process, but it usually does not interfere with the reading and can be easily identified.

5. Some criteria for evaluating the response of GIST to treatment depend on a change in PET SUV.  Should every patient with metastatic GIST receive a baseline PET scan prior to treatment?

Yes. A baseline PET should be obtained prior to initiation of drug treatment if changes in SUV are used to monitor response to treatment.  We also recommend that a new baseline PET be obtained before initiation of a new therapeutic regimen. Should a biopsy be necessary, PET may also help guide the biopsy to the relevant part of a tumor mass. 


6. What are the goals of your research in imaging the response of GIST to drug therapies? 

Early identification of non-responders is certainly a major goal. No one wants to expose a patient to any treatment if he or she is not going to derive significant benefit from it.  I think that imaging has a role to play soon after initiation of drug therapy to allow for an earlier change to alternative therapy and to avoid unnecessary toxicity in patients who are not responding. I also believe that PET imaging has propelled science forward — particularly in patients with GIST — by giving us a simple, noninvasive, and rapid way to demonstrate biologic effects on the tumor as early as hours after a single dose of the drug, as we have shown with imatinib.  These profound metabolic changes within the tumor are consistent with response to therapy and precede significant changes in tumor size by weeks and months. 

We can also use PET to study a drug’s pharmacodynamic effect. If we can see what works and what does not early on, we can direct our energies and research dollars accordingly. The use of functional imaging with PET has already shortened the length of several clinical trials, with the result that new drugs became available to the public sooner than would otherwise have been the case, and development costs were re
duced. We demonstrated this with sunitinib and helped bring the drug to the market six months ahead of schedule (based on feedback from the company).  I also anticipate that imaging, and in particular functional imaging with PET, will play a major role in the future in the delivery of personalized medicine. I think that imaging can lead to significant cost savings by helping identify the right patient for the right drug. So the facilitation of clinical trials, drug development, and personalized medicine through the use of imaging would be other major goals.  


7. What PET criteria have you identified as best predicting a successful response to imatinib?

Metabolic changes tend to precede structural changes within the tumor and are predictive of both clinical and subsequent radiologic responses. Qualitative and quantitative imaging methods can be used to assess metabolic response to therapy.  A complete metabolic response by either qualitative or quantitative evaluation is highly predictive of good outcome.  Conversely, persistent metabolic activity suggests residual tumor, and recurrent tumor activity in an area that had previously shown response may be indicative of secondary resistance.  Several quantitative and semi-quantitative methods can be used to assess metabolic response, but these vary in complexity and reproducibility and are mainly used in the context of clinical trials. 


8. After what duration of tyrosine kinase (TK) inhibitor therapy do you recommend performing PET to assess response?  Would this be advisable if CT showed stability or shrinkage, or only if CT showed tumor size increase?

Because the metabolic response to drug treatment precedes the anatomic response, reductions in tumor activity may be seen in as little as 24 hours – long before physical shrinkage is visible on CT or MRI. Thus, PET is a really efficient tool for determining how well a drug is working in a particular patient. That said, access to CT might be easier than to PET, depending on where you live.  If the CT scan shows stability or shrinkage, that is probably all the information the oncologist needs to know to assess if the drug is working, and if you should stay on that drug. 

Where PET is really valuable is in resolving discrepancies between the CT scan and clinical findings: for example, the CT may show an increase in tumor growth but other clinical information suggests that the patient is actually doing well.  A negative PET scan will confirm that this is the case.  Conversely, tumors may look stable in size on CT but show small nodular development within the wall of the mass. These tumors may be harboring isolated “hot spots,” which show up vividly on PET and may be a sign of recurrence.  PET can also be very helpful in guiding the biopsy to these sites of active tumor metabolism and possible secondary resistance.

9. Are there any patients who should be followed with routine PET — as opposed to CT — scans, and what would be the criteria for such followup?

FDG-PET is an ideal imaging tool in GIST for staging the disease, assessing therapeutic response, evaluating primary and secondary resistance, and resolving discrepant results between CT and the clinical findings. For Medicare patients, the Centers for Medicare and Medicaid services have approved reimbursement of FDG-PET for initial treatment strategy, and PET scans that are not reimbursable by Medicare are eligible for entry in the National Oncologic PET Registry (NOPR).  Other third-party payers may or may not provide reimbursement for the scan.

PET technology and combined PET/CT scanners (which allow for superior functional and anatomic assessment in one setting) are becoming more widely available, and I believe that there could be an increased use of this technology over time.  At the present time, however, PET is ordered primarily when early assessment of response to therapy is needed for clinical reasons, to resolve discrepant results between CT and clinical findings, and in the context of clinical trials. CT is mostly used in all the other contexts. If the prognostic and predictive values of FDG-PET in GIST are confirmed, there may be an option to redesign the scheduling and type of imaging studies in the future in such way as to reduce radiation exposure and health care costs, while providing the relevant information for the management of patients with GIST. We just don’t have the data to recommend doing routine PET scans at this time.

10. What criteria have you found useful in quickly identifying patients with primary resistance to imatinib?  For patients who do not have access to PET scans, can CT criteria be used with equal effectiveness?

Primary resistance to imatinib can be easily identified with FDG-PET shortly after initiation of treatment.  Most of these patients will, however, likely also show an increase in tumor size and/or new lesions that can be readily seen on CT.  New lesions equate with progressive disease, regardless of the imaging methodology used. An increase in size of an existing mass on CT that also shows increased FDG uptake on FDG-PET is consistent with active disease. 

11. What PET criteria are useful in detecting secondary resistance later in the course of therapy with imatinib or sunitinib?

We look for a re-emergence of metabolic activity – a “hot spot” – against a background of quiescence. It is very common in GIST to see secondary resistance develop in isolated areas within the existing tumor bed.  PET can pinpoint these as targets for future intervention.

12. Many patients are concerned about the radiation exposure they receive during repetitive CT and PET scans. Could you comment on this issue, and also on the question of whether – particularly in patients who have no evidence of disease after successful surgery – MRI could be used for routine surveillance instead?

I am unaware of any reports of excess secondary cancers among GIST patients receiving follow-up scans at this time. I also think that it is important for patients to be aware of the fact that the imaging community has always devoted a great deal of attention to minimizing radiation risk. We even have an acronym for it: “ALARA,” or “As Low As Reasonably Achievable.”  It is the principle that we go by as we aim to deliver the best diagnostic test while doing everything possible to reduce exposure to ionizing radiation to both the patients undergoing the study and the technologists performing these scans.  This process includes strict review of each study to confirm that it is medically necessary, that the proper scan is ordered, and that the imaging protocol used to acquire this scan uses the lowest amount of radiation needed.  We also collect radiation exposure data from the scanners themselves, and we are working with the equipment manufacturer to determine the actual amount of radiation received and to store this information in the patient’s medical record.

With regard to the second point, contrast-enhanced MRI can certainly be used to follow up GIST, particularly for patients with liver metastases and/or rectal GIST. Contrast-enhanced CT is generally more available and is currently the preferred imaging modality for routine follow up of GIST. The time it takes to do a CT scan is shorter than that required for MRI, and the CT scan can be performed in patients with pacemakers or certain other metal implants who would not be appropriate candidates for MRI. 


Dr. Van den Abbeele graciously concluded: "Thank you very much for asking these excellent questions about PET imaging in GIST and for giving me the opportunity to answer them.  Let me also take this opportunity to thank all the patients and the teams of clinicians and researchers who have contributed so much to the new paradigms in canc
er therapeutics and imaging, and for the renewed hope that they are providing to every patient with cancer. "  


Free-access papers by Dr. Van den Abbeele that you can read

Van den Abbeele AD.
The lessons of GIST–PET and PET/CT: a new paradigm for imaging.
Oncologist. 2008;13 Suppl 2:8-13. PubMed PMID: 18434632

Demetri GD, Heinrich MC, Fletcher JA, Fletcher CD, Van den Abbeele AD, Corless CL, Antonescu CR, George S, Morgan JA, Chen MH, Bello CL, Huang X, Cohen DP, Baum CM, Maki RG.
Molecular target modulation, imaging, and clinical evaluation of gastrointestinal stromal tumor patients treated with sunitinib malate after imatinib failure.
Clin Cancer Res. 2009 Sep 15;15(18):5902-9. PubMed PMID: 19737946.

Holdsworth CH, Badawi RD, Manola JB, Kijewski MF, Israel DA, Demetri GD, Van
den Abbeele AD.
CT and PET: early prognostic indicators of response to imatinib mesylate in patients with gastrointestinal stromal tumor.
AJR Am J Roentgenol. 2007 Dec;189(6):W324-30. PubMed PMID: 18029844.