Imaging GIST Response to Treatment

GIST Support International asked radiologist Haesun Choi, M.D. several questions about the techniques for monitoring tumor response to treatment using CT and PET scans. Dr. Choi is an Associate Professor of Diagnostic Radiology at M.D. Anderson Cancer Center in Houston, Texas. She and her colleagues have written numerous excellent papers about imaging techniques in the diagnosis of gastrointestinal stromal tumors and the evaluation of their response to treatment (see list at end of this page). Dr. Choi is a member of the board of directors of the Connective Tissue Oncology Society, at whose meetings GIST research is frequently presented.

You can also watch a webcast of Dr. Choi’s presentation at our 2009 GIST Information Support and Therapy Summit at this link. (After clicking the link, click Dr. Choi’s presentation in the table of the program schedule for the video.)

Here are Dr. Choi’s answers to our questions.

1. Your papers have pointed out that RECIST guidelines (Response Evaluation Criteria in Solid Tumors) fail to account for the different patterns of response shown by gastrointestinal stromal tumors during treatment with imatinib (Gleevec or Glivec). Can you briefly summarize the reasons why RECIST guidelines are not applicable to GISTs?

Response Evaluation Criteria in Solid Tumors (RECIST) are designed completely to depend on the change in tumor size evaluated at 2 months from the beginning of treatment. For example, partial response is defined as a decrease in the sum of longest dimensions of the target lesions by 30% and progression of disease an increase by 20%. In GISTs, the responding or progressing tumors do not necessarily show these levels of size changes. As you read in our article published in AJR (American Journal of Roentgenology 2004; 183:1619-1628), about 70% of the patients with responding tumors did not show the tumor size changes. RECIST categorized them into the group of stable disease.

2. In your December 2004 paper in AJR, you described how GISTs may sometimes increase in size during response to imatinib treatment. The paper’s figures illustrated this phenomenon in a peritoneal metastasis. Does this also occur in liver metastases? Does it ever occur in inoperable primary tumors? What is the current thinking on the possible reasons for the size increase?  [This paper is now free-access via PubMed: see the link at the end of this article and access the paper to see thes figures.]

a. Yes it occurs in the liver or soft tissue, such as abdominal wall, as well.

b. I do not particularly remember seing it in those with inoperable primary GISTs, but, in general, the pattern of response is very similar to those with resectable or unresectable primary tumors or metastasis. It is not uncommon to see those with responding inoperable primary disease without tumor size decrease at an early post-imatinib treatment stage, like at 8 weeks from the beginning of the treatment.

c. The tumor size can increase temporarily due to development of intratumoral hemorrhage while the tumors are responding to imatinib. Also, myxoid degeneration or necrosis, that are the results of response, is thought to increase the tumor size as well.

3. Can you describe the “nodule within a mass” phenomenon in GIST?

It is one of the quite common patterns of the initial presentation of recurrent GISTs.

Once the tumors respond to imatinib, the tumor typically becomes homogenous and hypodense. They look like cystic masses. These tumors then become continuously smaller or stabilized in their sizes, if they are continuously responding. At one point, some of these, treated hypodense tumors, develop a small solid nodule within them. This has been described as “nodule within a mass.” It can takes months for this small solid nodule to fill the entire treated tumor and to increase the tumor size.

The importance of this finding is in early detection of recurrent disease. As mentioned above, based on the current response evaluation criteria (WHO criteria, or RECIST), this recurrence will not be counted until the original treated mass actually grows months later. And most of these patients are asymptomatic.

4. Your research has found that decreasing tumor density is a favorable sign of GIST response to treatment. Is this true for all types of GISTs (primary, recurrent, metastases)? How does the presence of necrosis either before or during treatment affect the use of density measurements to monitor response to treatment? (In other words, would necrosis prevent a tumor from attaining the less dense, homogeneous appearance you describe?)

a. Yes. It applies to primary, recurrent, and metastatic disease.

b. The tumor density decrease is the result of development of myxoid degeneration that is post treatment “necrosis” in GISTs, and/or development of hemorrhage in the responding tumors.

In those with necrosis presented prior to the treatment, application of this sign can be difficult. (It can also be difficult to evaluate these lesions with PET scan, as well, because there is no visible metabolic activity.) But most of necrotic GISTs do have some solid component, predominantly at the edge. This solid component will follow the same principle.

5. Your papers have alluded to “overall tumor response” criteria being developed and tested in a multi-institutional effort. Can you estimate when results from this investigation may become available?

As you have read in this article, “overall tumor response” criteria are thought to be the best but these are subjective and require some experience to apply them. Based on these results, we have developed objective CT response criteria, using a combination of tumor size and density. We have found that these criteria are highly accurate in predicting the long-term survival based on a trial in about 100 patients at MDACC. The preliminary results have been presented at ASCO, 2003. We are hoping to carry these results to a multi-institutional trial.

6. Many patients have their scans read by community radiologists who do not see many GISTs. How can a patient ensure that the community radiologist has access to the latest information about imaging GISTs? How can a community radiologist learn to recognize the nodule-within-a-mass and other GIST-relevant phenomena?

They will have to educate themselves. Multiple articles on this issue are now available. Also, there will be a pictorial essay published in Radiographics in March 2006, which will describe the entire radiological scope of the GIST. This should help as well.

7. When does a patient truly need a PET scan to guide treatment decisions (as opposed to the PET providing “nice-to-have” supplementary information)?

PET has clearly shown advantage in evaluating response of imatinib treated GISTs.

It is recommended whenever CT findings are equivocal and whenever there is discrepancy between the clinical findings and CT.

Also, the National Comprehensive Cancer Network (NCCN) currently recommends considering the use of PET in those with marginally resectable or resectable GIST, with risk of significant morbidity to determine therapeutic options as early as possible by observing the response within 2-4 weeks from the beginning of the imatinib treatment (JNCCN Vol. 2, supplement 1, May 2004).

8. Please explain what tri-phasic CT scans involve and summarize the advantage of tri-phasic CT technique for identifying and monitoring liver metastases.

The triphasic CT scan is an enhanced CT technique to evaluate liver lesions. In general, “routine enhanced CT” includes only one phase that is a portal venous phase. The triphasic technique acquires images at 3 different time points following administration of the contrast. These time points include early and late arterial phases, and portal venous phase. Accurate timing is very important to take advantage of the technique as explained below.

GISTs are, in general, highly vascular tumors, and show up better on the arterial phase CT images and can be enhanced to the same degree of the normal liver on portal venous phase. This makes the metastatic GISTs in the liver invisible on the portal venous CT images.


References and Further Reading

Here are links to PubMed abstracts of some of Dr. Choi’s papers on GIST.  These include her more recent papers proposing new "Choi criteria" for evaluating response of GIST to treatment with imatinib (Gleevec).  Some of these papers are free-access, meaning you can download them to read or print through PubMed.  Click the title to link to the PubMed abstract.

Choi H.
 Response evaluation of gastrointestinal stromal tumors.
Oncologist. 2008;13 Suppl 2:4-7. Review.
PMID: 18434631  Free access via PubMed and The Oncologist.

Benjamin RS, Choi H, Macapinlac HA, Burgess MA, Patel SR, Chen LL, Podoloff DA, Charnsangavej C.
 We should desist using RECIST, at least in GIST.
J Clin Oncol. 2007 May 1;25(13):1760-4.
PMID: 17470866  Free access

Choi H, Charnsangavej C, Faria SC, Macapinlac HA, Burgess MA, Patel SR, Chen LL, Podoloff DA, Benjamin RS.
 Correlation of computed tomography and positron emission tomography in patients with metastatic gastrointestinal stromal tumor treated at a single institution with imatinib mesylate: proposal of new computed tomography response criteria.
J Clin Oncol. 2007 May 1;25(13):1753-9.
PMID: 17470865  Free access

Hong X, Choi H, Loyer EM, Benjamin RS, Trent JC, Charnsangavej C.
 Gastrointestinal stromal tumor: role of CT in diagnosis and in response evaluation and surveillance after treatment with imatinib.
Radiographics. 2006 Mar-Apr;26(2):481-95. Review.
PMID: 16549611   Free access via Pubmed.

Choi H.
 Critical issues in response evaluation on computed tomography: lessons from the gastrointestinal stromal tumor model.
Curr Oncol Rep. 2005 Jul;7(4):307-11. Review.
PMID: 15946591

Choi H, Charnsangavej C, de Castro Faria S, Tamm EP, Benjamin RS, Johnson MM, Macapinlac HA, Podoloff DA.
 CT evaluation of the response of gastrointestinal stromal tumors after imatinib mesylate treatment: a quantitative analysis correlated with FDG PET findings.
AJR Am J Roentgenol. 2004 Dec;183(6):1619-28.
PMID: 15547201  Free access via PubMed.