Preoperative Imatinib Treatment in Patients With Advanced Gastrointestinal Stromal Tumors: Patient Experiences and Systematic Review of 563 Patients

Patient characteristics

This retrospective study was approved by the institutional review boards of the Renji Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China. A consecutive series of 22 patients (Group A) with locally advanced and/or metastatic GIST receiving IM treatment before surgery from May 2008 to January 2012 were reviewed. In addition, 29 patients who had a high risk of recurrence and underwent initial tumor resection during the same period constituted the control (Group B). Emergency cases were excluded in the present study. In the former group, the diagnosis of GIST was confirmed histologically through a core needle biopsy, which was obtained endoscopically or transabdominally before the onset of IM treatment. IM treatment started at 400 mg once daily orally. The response to treatment was evaluated according to the Response Evaluation Criteria in Solid Tumors (RECIST)¹³ criteria with computed tomography imaging examinations performed every 2 to 3 months. When there was no further reduction or increase in tumor size between 2 consecutive radiographic images or the surgeon deemed that further treatment would play minimal impact to the surgical procedure, the patients underwent surgical resection. The primary goal of the surgery was radical cure, with the aim of a 2-cm clear margin when possible. According to the NCCN guidelines,⁵ all patients started and continued adjuvant IM treatment after a median time of 13 postoperative days, except 1 patient in Group B refused to receive continued IM treatment because of cost issues. Patients with localized primary GISTs were generally treated for approximately 1 year. For patients with metastasis, intraoperative tumor rupture, or failed R0 resection, adjuvant therapy continued unless resistance occurred. Sunitinib was applied as an alternative if there was disease progression or side effects with IM.

Clinical and surgical parameters

The clinical parameters that we recorded included sex, age, pathologic classification, extent of disease at initial presentation, dose and duration of tyrosine kinase inhibitor (TKI) treatment, surgical outcome, and disease status at last follow-up. Patient demographic data are listed in Table 1. Preoperative IM efficacy was assessed by dividing the patients into 2 clinical categories. The first category, stable disease (SD), included patients initially presenting with primarily unresectable disease or metastatic GIST who achieved a drug response significant enough to render their disease completely resectable. The second category included patients who, during the treatment with IM, had progressive disease (PD), defined as growth of at least 1 extra tumor by radiologic images.

Table 1 Clinical and demographic characteristics of the two study groups

View Fullscreen

The primary goal of surgery was to remove all sites of tumors with function-preserving as much as possible. The type of the surgical resection was determined from the operative reports and the pathology records. If a visible tumor was not completely resected or if the margins were grossly involved, the resection was considered R2. If positive margins were confirmed by the postoperative pathologic examinations, the resection was coded as R1. If all the lesions were completely resected with microscopically tumor-free margins, the resection was considered R0.

The clinical follow-up of the cohort study was updated in August 2012, with a median follow-up of 25 months (range, 8–44 months) after the surgery. PFS was measured from the date of surgery to the initiation of documented progression of residual disease, recurrent disease, or death associated to the tumor. Overall survival (OS) was defined as the length of time from the time of the initial treatment to the associated death. All the times were reported in months.

Pathologic analysis

All the GIST specimens were confirmed immunohistochemically according to the immunohistochemistry examination and CD117, CD34, or PDGFRA positivity. Mutational analyses of KIT exons 9, 11, 13, and 17 and PDGFRA exons 12 and 18 were performed with denaturing high-performance liquid chromatography (dHPLC), and bidirectional direct sequencing in all tumors treated by IM.

Systematic review and data extraction

A systematic literature review was conducted using the EMBASE, Medline, and PubMed databases to detect relevant English-language articles (published up to November 2012). The following search terms were used: “gastrointestinal stromal tumor,” “surgery,” and “imatinib.” All eligible studies were retrieved, and their bibliographies were checked for other relevant publications. Only published studies with full-text articles were included. Two authors independently searched and extracted the data. Any differences were resolved by mutual agreement.

To be included in our analysis, studies had to meet the following criteria: (1) they assessed the efficacy of preoperative IM treatment associated with surgery in patients with advanced GIST, regardless of its status as unresectable, recurrent, or metastatic; (2) they were clinical trials; and (3) they had sufficient data for estimating a risk ratio (RR) with a 95% confidence interval (CI).

A study was excluded if (1) the outcomes of interest were not reported or impossible to calculate; or (2) there was considerable overlap between authors, centers, or patient cohorts evaluated in another published study. If patient or material was reported more than once by the same institution, the most recent article was included in our analysis.

Statistical analysis

Institutional experience

In Group A, the median duration of IM treatment prior to surgery was 6 months (range, 2–14 months). All 22 patients received 400–600 mg IM per day and had partial response, with primary tumors decreasing in mean diameter from 12.0 ± 4.9 cm to 6.9 ± 3.1 cm by the time of surgery. All the patients tolerated the prescribed therapy, with only mild complications in 10 (45.5%) patients. These complications included fatigue, rash, mild edema, and gastrointestinal upset. All the patients had a complete resection without tumor spread, and only 2 (9.1%) of them had a positive resection margin confirmed by pathologic diagnosis after operation. Multi-organ resection was needed in 4 (18.2%) patients, including 2 (9.1%) patients with hepatic metastasis in which 1 patient with initial hepatic metastasis was diagnosed with hepatic relapse 18 months after the resection and switched to sunitinib treatment. She ultimately died in the 32nd month after the surgery.

The clinical and demographic characteristics of Group A were compared with Group B in Table 1. The 2 groups had no statistically significant differences in sex, age, tumor pathologic nor genetic mutation features. In addition, the median time of postoperative TKI therapy was similar in both groups. In Group B, 5 patients (17.2%) had tumor spread during the operation, and 14 patients (48.3%) underwent multi-organ resection, including the spleen, colon, liver, or uterine adnexa. These 2 proportions were significantly different compared with Group A (P = 0.04 and 0.026, respectively). The proportion of patients not achieving R0 resection in Group B, included 7 patients (24.1%) with R1 resection and 1 patient (3.4%) with R2 resection, was also greater than that in Group A but not statistically significant (P = 0.099). Furthermore, there was no case of perioperative death (i.e., in-hospital mortality) in the cohort study. Postoperative complications occurred in <15% of the patients, with no significant differences between the 2 groups.

As shown in Fig. 1, tumor progression after the surgery was observed in 1 patient (4.5%) in Group A and 8 patients (27.6%) in Group B, resulting in 3-year estimated PFS of 94.4% and 61.4%, respectively (P = 0.045). Figure 1 displays the Kaplan-Meier curve for OS, measured from the date of initial treatment. At the last follow-up, 1 (4.5%) patient had died of the disease in Group A, whereas in Group B, 3 (10.3%) patients had died of the disease, and 1 (3.4%) patient had died due to an unexplained cause. There was no difference in 3-year OS between the 2 groups (83.3% and 81.6%, respectively).

View Figure

• Download as Powerpoint
• Download Figure

Univariate analysis of the pathologic features and the treatment variables, which could potentially influence RFS and OS, detected intraoperative tumor spread, curability, and preoperative IM treatment as prognostic factors correlating with PFS (P = 0.009, 0.003, and 0.033, respectively). Patients who were treated by the preoperative IM, who achieved R0 resection and who were without the intraoperative tumor rupture had better PFS, with hazard ratios of 1.318, 1.805, and 2.174, respectively. No factors were associated with adverse OS. Multivariate analysis yielded no significant outcomes due to the relatively small number of patients in each cohort.

Systematic review and meta-analysis

From our computerized search and after extensive cross-checking, 86 relevant articles were extracted and reviewed by 2 independent reviewers. Thirteen articles were ultimately eligible for inclusion in this analysis, and data were extracted from these studies.^15–27 The details of individual studies are given in Table 2. Because these articles were nonrandomized studies, the DerSimonian–Laird test (random-effects model) was used to estimate the bias in the selected articles.

Table 2 Descriptions and summary of studies eligible for the meta-analysis

View Fullscreen

Compared with recurrent/metastatic GIST, the patients with primarily unresectable GIST had superior resection status and PFS. Incomplete resection (R1/R2) was significantly less frequent in the locally unresectable GIST subgroup than the recurrent/metastatic GIST (RR, 0.39; 95% CI, 0.20–0.75; P = 0.005; Fig. 2, top). Correspondingly, the patients with recurrence or metastasis had a higher risk of progression within 2 years after the surgery (RR, 0.35; 95% CI, 0.07–1.74; P = 0.20; Fig. 2, bottom).

View Figure

• Download as Powerpoint
• Download Figure

The difference in surgical findings between the SD and PD subgroups was also significant. All the results in the SD subgroup were superior to the PD subgroup, including resectability rate (RR, 0.32; 95% CI, 0.18–0.55; P < 0.0001) (Fig. 3, top), PFS (RR, 0.45; 95% CI, 0.37–0.56; P < 0.00001; Fig. 3, middle) and OS (RR, 0.13; 95% CI, 0.04–0.42; P = 0.0008; Fig. 3, bottom).

View Figure

• Download as Powerpoint
• Download Figure

Furthermore, the incidence of PD during the IM treatment was greater in recurrent/metastatic GIST patients, and for this subset of patients, there have been more controversies on the effects of the surgery. Therefore, the differences in resectability rates and PFS between SD and PD patients with recurrent/metastatic GIST were analyzed in our series. As shown in Fig. 4, surgery had a minor effect on recurrence or metastasis in the PD subgroup, with a higher rate of bulky residual disease remaining after the surgery (RR, 0.32; 95% CI, 0.17–0.61; P = 0.0005) and a higher risk of progression within 2 years after the surgery (RR, 0.44; 95% CI, 0.35–0.55; P < 0.00001) than the SD subgroup.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

In our study, preoperative IM therapy benefitted the patients with advanced GIST more, in terms of fewer intraoperative tumor ruptures and multivisceral resections, than those without prior IM therapy. The PFS of these patients was also superior to those who underwent surgery initially. However, 2 distinct clinical scenarios that may produce different prognoses for postimatinib surgery were included in Group A of our clinical case-series analysis: first, locally advanced tumors or tumors in “difficult” locations, and second, presence of metastasis. Because Group A had a low patient number, a meta-analysis was then conducted in which we further divided advanced GIST patients into an unresectable and a metastasis subgroup or a stable disease and a progressive disease subgroup to identify which subset of advanced patients could benefit from postimatinib surgery. For patients with primary disease, neoadjuvant IM decreases the size of responsive tumors leading to less morbidity in surgery with organ- or function-preservation and to render previously inoperable tumors operable. Our meta-analysis shows that preoperative IM treatment had a positive effect in advanced GIST patients. However, compared to the satisfactory results of primary GIST, the outcomes of metastatic GIST were poor. In our systemic analysis, we found a higher risk of recurrence within 2 postoperative years in the metastatic group (P = 0.20), as well as fewer complete resections (P = 0.005). However, some studies have indicated that some metastatic patients still benefit from surgery, even though IM treatment or incomplete resection was performed.^20,28 A control study between unresected patients who received IM treatment alone and patients who received IM treatment followed by resection also showed that resection of residual tumors after preoperative IM treatment in patients with metastatic GIST is feasible.²⁹ We think the difference between our study and these studies is primarily related to the relatively limited number of patient in our study. Therefore, currently, IM is still the standard treatment of choice for patients with metastasis, and surgery is not curative but only secondary to IM treatment. Particularly for metastatic GIST not controlled by preoperative IM treatment, it is evident that surgery offers poor results.^26,27 Therefore, these patients should be treated with high doses of IM, according to the tolerance, or with second-line agents such as sunitinib; surgery is indicated only for palliative reasons to control complications, such as pain, bleeding, bowel obstruction, or rupture of tumor masses. It is worth noting that the patients with advanced GISTs still gain the opportunity for surgical resection after sunitinib treatment.^30,31

In the meta-analysis study, we also showed that patients who were responsive to or had stable disease after initial IM treatment could benefit from the surgery, with a higher rate of complete curability and prolonged PFS and OS compared to those who did not undergo the subsequent surgery, regardless of whether they had primarily unresectable, recurrent or metastatic GIST. In contrast, surgery had a limited role in patients with advanced GIST who initially regressed under IM treatment but then developed progression in later course. The analysis indicated that the type of surgical resection, that is, its “quality” (R0, R1, or R2), depending on the type of response to IM, and these differences have prognostic implications. A recent study of 35 patients at the MD Anderson Cancer Center¹⁸ found that patient survival in those who achieved R0 resection was 100% at the 30th months, compared to 80% at 12 months in those who achieved R1 or R2 resection. Other reports^15,23 also found a link among IM response, surgery quality and survival. However, a study¹⁹ indicated that selected patients with local progression also benefit from incomplete resection in overall survival if tumor nodules that progress on IM therapy are removed. We think that it is difficult to evaluate the degree of progression accurately only on the basis of radiologic evidence. Some tumor nodules shown by CT scan as local progression frequently have micrometastases that is difficult to be identified by current imaging technology. In addition, univariate analysis in the present study demonstrated that patients with R1/R2 resection or with intraoperative tumor rupture had a poor PFS. The exact value of partial resection, especially for the patients in progression, is uncertain and should be evaluated in future studies. At present, incomplete resection or “debulking surgery” is preferable only in selected patients to alleviate symptoms.

Given the collective evidence reviewed previously, we suggest that the IM treatment is capable of creating best timing for surgical resection leading directly to the patient's prognosis. However, the optimal time to treat patients with IM before undergoing surgery has not been established. Theoretically, the surgery should be performed after the maximum response and before the development of secondary resistance. This is usually approximately 6 to 12 months.^20,32 However, secondary KIT mutations have been identified only 10 months after neoadjuvant IM therapy.³³ Hence, some authors argue that it is not always necessary to wait for the maximum response; especially for patients with primarily unresectable GIST, surgery should be performed as soon as there is sufficient shrinkage.³⁴ In our study, the median IM preoperative treatment time was 6 months (range, 2–14 months). All the cases received early surgical treatment, and the outcomes were satisfactory. For metastatic patients, more time should be allowed to achieve the maximum IM benefit. In these patients, a treatment response assessment by PET or CT is key in choosing the optimal time for surgery. Cessation of tumor shrinkage on successive imaging or CT scans or the formation of new nodules or tumor growth after a period of shrinkage may represent the beginning of secondary resistance and should prompt evaluation for surgery.

Bonvalot et al¹⁶ reported that 3% of the patients treated with the neoadjuvant IM therapy-developed complications, particularly in relation to the rupture of large tumor masses that became necrotic under treatment. This is largely due to the edema of the tumor tissues caused by the drug leading to higher friability. In these cases, emergency surgery seems to be associated with increased mortality and postoperative morbidity, and most of the patients cannot obtain complete en bloc resection. The poor outcomes of the emergency surgery have been confirmed by other reports.^15,16 Therefore, elective surgery should be considered for patients at higher risk of complications during pharmacologic debulking. One argument often brought forward against surgery in metastatic or recurrent GIST is that potential complications outweigh the rather minor clinical benefit. In our series, however, all the patients tolerated the prescribed preoperative IM therapy without severe complications, and the rate of surgical morbidity was comparable with that of the control group as well. This is consistent with the results of a similar series.²² Therefore, there is no strong reason to assume that preoperative IM treatment would lead to the increased surgical morbidity.

Considering that postoperative IM treatment has been approved for adjuvant treatment in patients with a medium or high risk of recurrence by the FDA⁵ and ESMO,⁶ all advanced patients who undergo post-imatinib surgery, whether or not they achieve a complete cytoreduction, should be given postoperative IM. Blay et al³⁵ presented the data of a randomized trial evaluating continuous vs. intermittent use of IM in metastatic GIST and found that continuous IM produced a better progression-free survival. In 2 studies we reviewed,^17,19 postoperative recurrence or progression was not rare in patients who discontinue further IM treatment after the complete resection. In light of these findings, continuous treatment with IM may be optimal for many patients in the clinical setting, especially those with intraoperative tumor rupture or a failure to achieve R0 resection.

In conclusion, surgical excision is suitable for advanced GIST patients whose debulking has been controlled by IM, especially patients with primary advanced tumors, patients in whom the surgeon estimates that an R0 resection may be difficult to achieve, or patients in whom R0 resection would be easier to achieve if tumor cytoreduction were possible. Among recurrent/metastatic patients, elective surgical resection may be favorable in patients who have IM-responsive disease, while surgery is generally not indicated for those resistant to IM treatment.