Gastrointestinal stromal tumors (GIST) and their management have become an important focus for clinical research.  At the 2007 meeting of the American Society of Clinical Oncology (ASCO), results of large Phase III trials evaluating therapy for adjuvant and metastatic disease were presented.  In addition, presentations of clinical trials testing new agents, potential biomarkers, and methods to assess disease response were reported.  

Low Dose Versus High Dose Imatinib: MetaGIST

A meta-analysis was presented of the two large Phase III trials comparing low dose to high dose imatinib.[1] The studies, performed in North America, Europe, Australia and Asia, were planned together and had similar trial designs except for the primary endpoints.  The North American study utilized overall survival (OS) as its primary endpoint whereas the European-Australasia study utilized progression free survival (PFS) as its primary endpoint.  This difference in endpoints resulted in 200 fewer patients accrued to the North American study.  A small, but statistically significant improvement in PFS was detected (HR 0.89, log rank p=0.04), however there was no difference in (HR 1.00, p=0.97).  

A multivariate analysis of factors associated with worse PFS and OS, irrespective of treatment, identified poor performance status, high neutrophil count at study entry, absence of KIT exon 11 mutation and male sex.  When looking at PFS only, the analysis revealed small bowel origin and low hemoglobin at study entry as significant factors.  For OS alone, the important traits identified were advanced age, low albumin and large lesions.  Data from the American-Finland study also reported no difference in response rates for tumors of varying size, but did confirm that time to tumor progression and overall survival decreased as tumor bulk increased.[2]  

The MetaGIST study also analyzed PFS and OS based on mutation type.  For both PFS and OS, patients with tumors carrying an exon 11 mutation had the best outcomes with a median PFS and OS of 26 and 60 months, respectively.  Wild type tumors followed with a median PFS and OS of 16 and 43 months, respectively.  Tumors containing exon 9 mutations had a median PFS of 13 months and the poorest overall median survival of 31 months.  The outcomes for those with other mutations were similar with a median PFS and OS of 11 and 34 months.  An analysis to determine if a particular subgroup benefited from the higher dose therapy, given the limited benefit in PFS noted, suggested that patients with exon 9 mutations were the only group that benefited from the higher dose therapy.  This benefit was only for PFS (p=0.015), but not for OS (p=0.071).  

Can Good Outcome Patients Discontinue Imatinib Therapy?

The French Sarcoma Group has been conducting a study (BFR-14) to determine if patients with recurrent or metastatic GIST who are treated with imatinib mesylate with a good outcome, defined as stable disease, partial or complete response, are candidates for discontinuing imatinib therapy.  Patients enrolled on this study were allowed to have surgery as part of their management.  The study initially randomized patients after 12 months of therapy to continue imatinib, or stop therapy.  The primary endpoints of the study were PFS and RECIST determined response rate, with secondary endpoints of OS, response rate to reinitiating of imatinib, quality of life assessment for patients continuing versus discontinuing imatinib, and the number of patients able to undergo radical resection after therapy with imatinib.  

Dr. Rios presented updated information from the initial stage of this trial, which previously had been reported demonstrating early progression in the STOP arm of the trial.[3] Of the initial 183 patients enrolled on the study, 98 reached the 12-month endpoint.  At that time, the response rates were: CR 18.4%, PR 29.6% and SD 27.6%; the CR rate is higher than in most prior studies, however patients were allowed to have surgical resection to achieve a CR on this study.  The randomized patients were balanced in terms of age, gender, site of primary tumor, response to therapy and other prognostic variables.  Follow up data was presented with a median follow up time of 39 months.  Of the 32 patients randomized to discontinue imatinib, 91% developed progressive disease at a median of six months.  This is in contrast to the patients in the continuous imatinib arm in which 17 of 26 patients randomized have had progression of their disease at a median of 29 months. The patients in the stop imatinib group were allowed to resume imatinib at the time of progression.  In addition, at the time of the initial analysis that demonstrated significant progression in the discontinuation arm, all patients randomized to the stop imatinib were allowed to restart treatment.  Only two patients in the discontinuation arm did not resume therapy, one of whom had evidence of progressive disease after stopping therapy.  Twenty-four of the 27 patients achieved tumor control on reinitiating of imatinib.  Overall survival in the stop versus continuous imatinib arms at three years is 71% and 73%, respectively.  There was no difference in the rate of relapse in patients who were in CR at the time of imatinib discontinuation compared to those with radiographic evidence of disease.  Although patients in CR had a suggestion of longer PFS compared to patients with PR or residual disease, there was no difference in the rate of progression in patients in CR.  

One hypothetical risk of discontinuing therapy and reinitiating treatment with progression is an increased risk of developing resistance to imatinib.  In this study, there was no statistically significant difference in the time to secondary resistance to imatinib in either group.  One benefit to discontinuing therapy may be a better quality of life because of decreased drug side effects; using the QLQC30, there was no difference in global quality of life at six months following randomization to continued imatinib versus stopping imatinib.  This analysis may have been confounded by recurrence in the patients in the STOP arm.  In addition, toxicities to imatinib tend to wane with time.  

Dr. LeCesne reported on the continuation of the BFR14 study, which was amended to allow continued accrual and a randomization of patients completing 36 months of therapy to continuation versus discontinuation of imatinib.[4] Again this randomization was discontinued because of increased rate of progression of patients on the discontinuation arm at a median of six months.  The percent of patients progressing at 6, 9 and 12 month in the group that discontinued imatinib was 47, 64, and 85% respectively, with no deaths.  There was no difference in the PFS of patients that discontinued imatinib at 12 versus, 36 months. Patients that continued to be progression free at three years were most commonly women with small bowel primary tumors that had metastasized to the liver.  The BFR-14 study plans to continue enrollment and randomization of patients to continuation versus stopping with imatinib at 60 months of therapy to determine if there are any long-term survivors on imatinib who can discontinue therapy.  Tumor samples are being collected and stored for mutational analysis, which will allow analysis of the molecular subtypes that benefit from long-term treatment.  

Adjuvant Imatinib: ACOSOG Z9001

Given the efficacy of imatinib in the metastatic disease setting, it is a reasonable hypothesis that adjuvant imatinib will decrease the risk of recurrent disease in patients with completely resected GIST.  Dr. DeMatteo and colleagues reported on ACOSOG Z9001, a randomized double blind trial of imatinib or placebo in patients with GIST tumors 3 cm or larger.[5] The study accrual was terminated in April of 2007 when the Data Safety Monitoring Board determined that the treatment arm had a statistically significant improved risk in PFS.  The study had accrued 644 patients from 230 sites.  The median age in the two groups was 58-59, with slightly more men in the placebo cohort (54% versus 46%).  The range in tumor sizes was not significantly different.  Sixty-seven percent of patients completed one year of imatinib versus 71% in the placebo group.  There were 14.5% of patients that required a dose reduction of imatinib versus 4% in the placebo group.  Of the patients that did not complete the 12 months of therapy, the most common reason in the imatinib cohort was toxicity, whereas in the placebo group it was tumor recurrence; equivalent numbers withdrew their consent or discontinued study participation for other reasons.  

Toxicities were more frequent in the imatinib cohort, however there were still approximately 15% of patients reporting grade 3 toxicities and 1-2% reporting grade 4 toxicities in the placebo group.  These toxicities were typically abdominal pain, whereas the grade 3 and 4 toxicities for the imatinib group were more typically hematologic toxicities.  At a median of 1.3 years follow up, there is no difference in the overall survival of the two groups.  However when comparing PFS, 97% of the patients receiving imatinib were free of recurrence compared to only 83% in the placebo group.  In an unplanned analysis, PFS was analyzed based on tumor size: 3-6 cm, 6-10 cm, and greater than or equal to 10 cm in size.  Although the hazard ratio for the smallest tumors was 0.44 (0.14-1.4), this was not statistically significant (p=0.15).  In the two cohorts with intermediate and large sized tumors, there was a statistically significant improvement in PFS (HR= 0.37, (0.17-0.81); p=0.01 for 6-10 cm, and HR=0.19, (0.09-0.41); p<0.001).  Analyzing the PFS curves, it appears that there is a significant benefit in the first 18 months following entry onto study, but then the recurrence rate in patients receiving imatinib increases and the slope of this curve approximates the recurrence seen in the placebo arm.  This observation will have to be followed given the relatively few patients that have long-term follow up. 

Sunitinib Continuous Daily Dosing

Sunitinib is the established second-line treatment for patients with GIST.  Initial trials evaluated varying doses and schedules with the pivotal trial conducted utilizing sunitinib 50 mg daily for four weeks followed by two weeks off therapy.  In GIST, this schedule has disadvantages because of the known rebound phenomenon off of targeted tyrosine kinase therapy that has been documented by PET scan as well as anecdotally in some patients with progressive disease.  Because of this, a more continuous dosing schedule was investigated with results of a Phase II, multi-center trial presented by Dr. George.[6]  

The study sought to assess the safety and tolerability of sunitinib 37.5 mg daily in patients that were refractory or intolerant to imatinib, determine the clinical benefit rate (CR, PR, SD>24 weeks as defined by RECIST), overall response rate, time to progression, progression free survival, overall survival, and determine the tolerability of dosing medication in the morning compared to the evening.  Sixty patients were randomized in a 1:1 fashion to morning versus evening dosing.  Patients tolerating 37.5 mg daily could be dose escalated at the discretion of the treating physician to 50 mg daily.  Patients requiring a dose reduction were reduced to 25 mg daily. The majority of patients treated on this study had progressed on prior imatinib therapy (90%) with a median length of therapy on imatinib of 100 weeks.  Eighteen percent of the patient population had progressive disease as their best response to imatinib.  Two patients had their dose escalated to 50 mg daily, however one required the dose to be decreased to 37.5 mg.  Of the 14 patients requiring a dose reduction to 25 mg daily, half of them were able to have their dose re-escalated to 37.5 mg, suggesting ongoing therapy is associated with developing tolerance to side effects.  Fifty three percent of the patients have discontinued therapy, with 72% stopping treatment as a result of progressive disease.  The clinical benefit rate in this study is 52%, with 12% of patients achieving PR.  Of note, 42% of the patients were continuing on study greater than 12 weeks, and the majority of patients on study at 12 weeks went on to continue on treatment for 24 or more weeks.  Only 8% of the patients had progressive disease as their best response to sunitinib.  When comparing PFS of the daily dose to the 4/2 schedule, there is a suggestion that this schedule may improve PFS.  The PFS for this study was 34 weeks compared to 25 weeks for patients receiving sunitinib on the original schedule in the Phase III trial, however the patient populations may not be comparable.  Certainly, this study does not suggest that the continuous schedule is inferior to the standard 4/2 schedule.  

No conclusions can be drawn on OS, as the median overall survival in this study has not been reached.  Side effects reported in more than 20% of patients that appear to be more common in the continuous dosing schedule were abdominal pain, asthenia, vomiting, hypertension, hand foot syndrome, and stomatitis.  When comparing morning to evening dosing, there was greater diarrhea, asthenia, fatigue, vomiting and hypertension with morning dosing, but more frequent abdominal pain in the evening dosing group.  There was a 10-15% increased risk of anemia and neutropenia with continuous dosing compared with intermittent dosing, but no difference when comparing morning and evening dosing; the minority of these toxicities were grade 3 and 4.  Pharmacokinetic sampling demonstrated no evidence of drug accumulation.  

Surgery in the setting of Tyrosine Kinase Inhibitor Therapy  

Imatinib Experience

Gronchi and collaborators reported on their experience resecting patients with metastatic GIST who are stable on imatinib for at least six months.[7] Patients were only taken to the operating room if they were thought to be completely resectable based on available imaging.  In a four-year period, 29 patients underwent surgery having received a median of 17 months of imatinib therapy.  Ninety percent underwent macroscopically complete surgery.  The median postoperative hospitalization time was nine days. Complications seen postoperatively were ascites (2%), elevations of ALT/AST irrespective of whether the surgery included liver resection (48%), post-operative pneumonia (38%), and major surgical complications in 14%.  The median PFS was 69% at two years.  Given the morbidity observed, these authors suggested surgery in this patient group should be considered investigational given no known survival advantage.  

Sunitinib Experience

Raut and colleagues presented their experience with surgery following therapy with sunitinib.[8] They retrospectively analyzed patients that underwent surgical resection while on sunitinib and compared them to patients taken for surgery while receiving imatinib.  The rate of complications was compared.  Twenty-six patients on sunitinib and 46 on imatinib were assessed.  Patients on sunitinib had their medication discontinued a median of five days prior to surgery and resumed a median of 33 days postoperatively.  They reported that 50% of procedures were associated with complications, in contrast with 38% in patients that were on imatinib, however there was a similar rate of dehiscence, leaks and or fistulas in the two groups. 

Bio-Markers of Tumor Biology

Romero, et al. reported on 33 completely resected high or intermediate risk GIST patients and explored the correlation between KIT exon 11 deletions affecting codons 557-558 with mitotic count and mitoses.[9] The tumor size ranged from 1.9-22 cm with a median of 10 cm, with mitoses ranging from 1-32/50 HPF with a median of 2.  There were 11 tumors with exon 11 deletions.  Samples with exon 11 deletions were associated with a higher median mitotic rate of 8 compared with 2 for the other tumors.  The investigators reported no correlation between mutation type and size of tumor.  The authors plan on following up these patients to determine if this difference in proliferative potential affects long-term outcome.  

Bio-Markers of Therapeutic Response

Cioffi and co-authors presented data evaluating patients with advanced GIST that remain free of progression on imatinib for greater than three years.[10] Of 266 patients participating in two prospective randomized trials initiated in 2001, they identified 31 patients.  Clinical factors associated with long-term PFS were male gender (61%), small bowel origin (60%), liver involvement (80%), synchronous metastases at study inclusion (58%), and baseline hemoglobin of greater than 10 grams/dL.  In addition, 15 patients were studied for mutations in their tumors.  Fourteen had exon 11 mutations, most commonly with an in frame deletion between 550-558.  There was one patient with an exon 9 mutation.  

McAullife and colleagues analyzed 54 GIST specimens for expression of stem cell factor (SCF) from patients with surgically resected tumors who received 400-800 mg imatinib as adjuvant therapy.[11] Using immunohistochemistry, the expression of SCF was graded from none to strong and correlated with PFS and OS.  They found no correlation.  They also analyzed 15 baseline biopsies from patients treated with imatinib in a neoadjuvant/adjuvant setting for expression of phosphorylated-KIT and CD31.  These analyses indicated the presence of phosphorylated KIT in the tumor-associated endothelial cells and suggested the possibility of KIT/SCF cross talk within the tumor microenvironment.  

Blackstein and colleagues reported on the use of soluble KIT as a serum marker of response in patients receiving sunitinib therapy.[12] A prior report from Bono and colleagues demonstrated in patients responding to imatinib that serum KIT levels decreased in responding patients where as the ligand for KIT, stem cell factor, increased in concentration.[13] Data from patients treated on the Phase I and II trials of sunitinib, suggested that a decrease in serum KIT levels correlated with a longer PFS.  To further assess this finding, serum samples were collected on all study patients of the Phase III trial of sunitinib versus placebo at baseline, day 1, 14 and 28 during cycle 1, days 1 and 28 in cycles 2-4, and day 1 of cycle 6.  There was a statistically significant decrease in the sKIT in all samples starting cycle 1 day 28, p<0.001. Patients with a decreased sKIT on day 28 of cycle 2 had an improved outcome in both patients receiving placebo or sunitinib, with the converse of an increase in sKIT being associated with a poorer prognosis.  The median time to progression in patients with a decrease in sKIT on day 28 of cycle 2 was 34.3 weeks versus 16.1 weeks in those with a rise in the sKIT, (HR= 0.512, p<0.001).  This was confirmed and the association perhaps strengthened evaluating the same data for the sKIT level on day 1 of the third cycle.  TTP in patients with decreased sKIT was now 40.1 versus 16.1 weeks in those with elevated levels (HR=0.416, p<0.001).  Overall survival using sKIT on C2D28 and C3D1 was 92.6 weeks versus 55.4 weeks in those randomized to sunitinib when stratified based on decreased or increased sKIT (HR=0.366 with a range from 0.216-0.621, p=0.0002 for C2D28 and HR=0.381 with a range from 0.148-0.98, p=0.46 for C3D1).  This large data set supports sKIT as a potential biomarker of response, but whether it will supplant more traditional methods of response assessment is to be determined.  

Patients on sunitinib are developing resistance as is seen in patients receiving imatinib.  Heinrich and colleagues presented data on patients who have developed resistance on sunitinib analyzing tumor samples for mutations associated with resistance and then confirming resistance in an in vitro model system.[14] Two patients with primary refractory GIST treated with sunitinib on the 4/2 schedule underwent surgical resection for progression on sunitinib.  Nine progressing tumor samples from two patients were analyzed.  These revealed additional mutations in exon 17 as well as a novel mutation in exon 16.  In the same patients, two other non-progressing lesions were found to contain a secondary mutation in exon 14.  In vitro, using Chinese hamster ovary cells transfected with various constructs, sunitinib was shown to decrease phosphorylation of KIT (a marker of drug efficacy in GIST) in samples containing exon 11 mutations with exon 13 or exon 14 mutations, which are insensitive to imatinib.  Cells with exon 11 and the novel exon 16 mutation were equally sensitive to imatinib and sunitinib, whereas those with exon 11 with either exon 17 or 18 mutations were equally resistant to both imatinib and sunitinib.  

Radiologic Assessment Tools

Response assessment in GIST in patients utilizing tyrosine kinase therapies, as in many other diseases with targeted therapies, is not well described using standard RECIST criteria.  Tumors with progression by size can be shown by PET scans to have had a dramatic metabolic response.  In addition, small nodules representing an outgrowth of a resistant clonal population are not considered progression.  Bulusu and colleagues report on their experience with the CHOI criteria for response assessment in patients with GIST treated with imatinib.[15] This criteria utilizes the change in size as well as the change in tumor density to make determinations about disease response to therapy.  Response rates using the criteria were found to be higher than those determined by RECIST (88 versus 16 at 3 months).  However, with longer follow up, response rates become more similar.  Limitations defined by these investigators included calcification, hemorrhage or perforation within the lesions which affected the density assessments.  

Imatinib Toxicities  

There has been concern about the potential of imatinib causing cardiac toxicity based on a mouse model study suggesting damage to cardiac myocytes.  In adddition several cases of patients with chronic myelogenous leukemia developing congestive heart failure have been reported.  Khakoo and colleagues presented data on 219 patients treated on clinical trials with imatinib between December 2000 and May 2006.[16] They identified 18 patients (8.22%) with potential cardiac adverse events, with a median age of 65.5 and 11 of which were males.  The onset of symptoms was at a median of 173.5 days, with the majority receiving 400 mg daily.  Thirteen of these patients had prior history of risk factors associated with coronary artery disease, including hypertension, diabetes, coronary artery disease, hypertrophic obstructive cardiomyopathy, peripheral vascular disease, hyperlipidemia and cardiac arrhythmia.  Radiographically none of the patients had evidence of pulmonary edema.  Eight of these patients had studies to determine ejection fraction, only one of which revealed a LVEF<50%, which had been greater than 50% at baseline.  This group reported an incidence of 1 out of 219 patients with congestive heart failure, for an incidence of 0.4%. In Dr. Verweij’s discussion of this abstract, he reported on two additional data sets, that of the EORTC Phase III trial as well as the Novartis database, all of which contain less than 0.5% incidence of possible cardiac toxicity.  

New Agents  

Masatinib

Masatinib mesylate was tested in the front-line setting in patients with metastatic and unresectable disease.[17] This agent has been shown to have greater activity and selectivity than imatinib in vitro in wild type GIST as well as those with KIT mutations in the juxtamembrane region.  Importantly the agent also has activity against PDGFR.  A prior Phase I study demonstrated safety and a complete response in a patient that was intolerant to imatinib.  In this multicenter non-randomized trial, 21 patients have been enrolled and treated with 7.5 mg/kg/day.  With a median follow up of 7.6 months, 16 of 21 patients have been on study for greater than 8 weeks.  Ten patients have had a PR (48%), with an additional eight patients achieving SD (33%), and two patients demonstrating progressive disease at eight weeks (10%).  One patient with stable disease has subsequently progressed.  Grade 3 toxicities have been primarily gastrointestinal and skin related.  The activity is suggestive of an active agent for GIST, however if this drug will be better than imatinib in the front-line setting is not clear.  In addition, if there is greater activity for wild type GIST this may be an important new agent.  Mutational testing on tumors from this study is planned.  

Nilotinib

Nilotinib is a second-generation inhibitor of ABL with clinical activity against imatinib resistant chronic myelogenous leukemia.  In vitro the agent also has activity against KIT and PDGFR and has shown therapeutic efficacy in cell lines that are both imatinib-sensitive and resistant.  The Phase I study evaluating nilotinib as a single agent in patients with imatinib refractory or intolerant GIST, and in combination with imatinib in patients with imatinib refractory disease was presented.[18] Fifty-three patients with GIST received the drug alone (n=18) or in combination with imatinib (n=35).  The majority of the patients (74%) had progressed on second-line therapy. The drug was well tolerated with dose limiting toxicities of hyperbilirubinemia or skin rash in the patients receiving 400 mg twice daily of both agents.  The maximum tolerated dose of the combination was nilotinib 400 mg BID with imatinib 400 mg daily.  One patient on nilotinib alone, who was imatinib intolerant, had a partial response, a second patient in the combination arm (nilotinib 400 BID and imatinib 400 D) had a partial response.  PFS at six months ranged from 71-44% in the different treatment cohorts, and when assessing patients who had received prior sunitinib the six  month PFS was 50-75%; there was no apparent dose response when evaluating PFS.  Pharmacokinetic sampling did not reveal any interactions between the two agents.  Given its tolerability and favorable side effect profile, further testing of this agent is warranted.  

IPI504

A novel approach to therapy in advanced GIST refractory to tyrosine kinase inhibitors is treatment utilizing heat shock protein inhibitors.  This strategy is supported by in vitro data that HSP inhibitors can cause loss of KIT kinase expression on cells and also loss of phosphorylation of KIT.  These effects occur preferentially in cells with mutated KIT.  Demetri and colleagues presented early data from a Phase I trial evaluating IPI504, an intravenous compound that inhibits hsp 90.[19] IPI504 has been tested in 21 patients with advanced GIST previously treated with both imatinib and sunitinib.  The agent was initially delivered intravenously on days 1, 4, 8 and 11 and repeated every 21 days; the agent is now being tested as a twice weekly infusion because of the limited toxicity observed.  In addition, PET imaging, which has suggested clinical activity because of a decline in SUV readings, demonstrated a flare of metabolic activity during the time of drug.[20] To date, no radiographic responses have been observed utilizing standard criteria.  The preliminary results in this agent are encouraging and warrant further testing in advanced GIST.  

Conclusions

Data presented at this year’s ASCO was encouraging.  Analysis of the large Phase III trials evaluating imatinib in patients with advanced GIST has defined patients who may benefit the most from imatinib, and in particular suggested that the sub-group of tumors with exon 9 mutations are the only patients that may benefit from high dose therapy. Data from the BFR14 study indicates that patients who are not resistant or intolerant to imatinib do not benefit from discontinuation of imatinib from the standpoint of progression free survival.  In addition, there is not a decline in quality of life reported by patients that continued on therapy compared with those who discontinued imatinib. Results also reported at ASCO indicate that imatinib may have a role in the adjuvant setting of the management of GIST, but longer follow-up is warranted to define the impact on long-term PFS in this setting. Investigators also reported that surgical resection following therapy with either imatinib or sunitinib may be an option but given the morbidity observed, should be considered investigational.  

For patients resistant to or intolerant of imatinib, results were also hopeful.  Sunitinib demonstrates activity on a continuous dosing schedule. Mature results of the data presented at ASCO are necessary to determine sunitinib’s effect on overall survival. Lastly, early data presented at ASCO indicates that three new agents: nilotinib, masatinib, and the heat shock protein inhibitor, IPI504, may be active agents and further studies are warranted in advanced GIST.  

To successfully complete this CME activity, please read the disclosures and the additional article on metastatic colorectal cancer from ASCO 2007 before you take the post test:

Advances in the Treatment of Metastatic Colorectal Cancer: ASCO 2007
Lee Schwartzberg MD, FACP, The West Clinic, Memphis, TN

Click Here to take the Post Test and Evaluation

REFERENCES: