For GIST, it has been reported that the stomach (60–65%) is the most frequent primary site, followed by the small intestine (20–25%), and other sites are much less common [1, 18]. In this study, the proportion of primary sites in the small intestine and other sites (52.6%) was greater than that in the stomach (37.8%), as all members of the population were surgically treated and categorized as having an intermediate or high risk of relapse, and the prognosis of non-gastric cancer is expected to be worse than that of gastric primary [19] with a higher risk of recurrence [20]. The etiology of the periorbital edema caused by IM remains unclear. The most accepted view currently involves platelet-derived growth factor receptor (PDGFR), which is one of the targets of IM. PDGFR is highly expressed in the skin around the eyes and maintains interstitial fluid homeostasis by regulating the tension between the extracellular matrix structures and cells [21, 22]. PDGFR inhibition leads to decreased interstitial pressure, increased capillary permeability, and fluid extravasation [23, 14]. There is also a complex network of connections between the lower eyelid skin and infraorbital wall, forming an enclosed space, which may allow for easier fluid collection and edema formation [22]. The incidence of IM-induced periorbital edema was similar to that reported in other studies; however, the ratio of more severe periorbital edema was much higher in this study [5, 14]. This may be attributed to the fact that the population in this study was postoperative, showing an intermediate or high risk of recurrence, and physiological statuses differed from patients in other studies. Additionally, the present study focused on late toxicity instead of early toxicity, as it has been reported that approximately half of the patients developed grade ≥ 2 toxicities attributed to TKIs after the first cycle [24].
Periorbital edema can be resolved by discontinuing IM, but this does not appear to be a reasonable option for most patients and is not recommended in situations where IM may markedly extend the recurrence-free survival of patients [25, 26] and when therapeutic alternatives are limited. The risk of relapse increased after discontinuation, as shown in the BRF 14-trial, which highlighted that most patients relapsed or progressed after interruption of IM treatment [27]. In the event of a relapse, all previous efforts were lost. For mild edema, there is no need to take any steps, but to weigh and monitor regularly to detect any unusual weight gain or severe fluid retention [28]. In cases of grade 2 periorbital edema that poses difficulties in daily life, symptomatic treatment may be available, such as salt restriction and topical steroids. Diuretics or surgical intervention may be feasible approaches for grade 3 periorbital edema cases [22]. A researcher from Canada reported a patient with type 2 diabetes mellitus (T2DM) and GIST who developed grade 3 periorbital edema immediately after IM treatment. This greatly affected the patient’s life. However, it was only two days after replacing metformin with canagliflozin that the periorbital edema appeared markedly alleviated, and completely subsided after one week [29]. Canagliflozin is a new antihyperglycemic drug that inhibits glucose reabsorption in the proximal tubule by inhibiting sodium-glucose cotransporter-2 (SGLT-2). There are no prospective clinical trials of a certain scale to confirm this effect and clarify the mechanism, but an SGLT-2 inhibitor may be a good option for GIST patients with T2DM who are taking IM.
More severe periorbital edema did not correlate with C0/D (P > 0.05), as described in a previous study [15]. The present study also indicates that genetic polymorphisms of ABCG2 rs2231142 and SLC22A1 rs683369 may influence the trough plasma levels of IM. Genetic polymorphisms of SLC22A1 rs1867351, EGFR rs2072454, and CYP1A2 rs11636419 have been found to play roles in IM-induced grade ≥ 2 periorbital edema. ABCG2 functions as a xenobiotic transporter and may play a major role in multi-drug resistance. ABCG2 plays a role in the transportation of many drugs, including IM [30], and has attracted considerable interest. The present study found that polymorphisms in ABCG2 rs2231142 significantly affected IM transport. ABCG2 rs2231142 is a non-synonymous coding SNP that converts the originally encoded glutamine to another amino acid, and may have significantly affected ABCG2 activity, thereby affecting the plasma concentration of the drug. Many endogenous small organic cations, as well as a wide range of drugs and environmental toxins, are eliminated by organic cation transporters encoded by SLC22A1. This was confirmed by the fact that polymorphisms in SLC22A1 were linked to grade ≥ 2 periorbital edema and IM trough plasma levels in this study. Gene polymorphisms can affect toxicity by influencing drug exposure or the ability of transporters to eliminate drug toxicity. Its impact on IM toxicity in this study was similar to that reported in a 2017 study [15].
EGFR, an oncogene, encodes a receptor for members of the epidermal growth factor family and plays an essential role in cell proliferation. This study describes the influence of EGFR rs2072454 on adverse reactions of IM for the first time. In the present study, the rs2072454 CC genotype represented an adjusted OR of 2.85 (95% CI = 1.10–7.40, P = 0.032) for grade ≥ 2 periorbital edema compared to T carriers (CT + TT). The receptor, expressed in the ocular structure, is also a receptor for IM [31], and toxicity may be a consequence of IM inhibition in this region. CYP1A2 encodes a member of the cytochrome P450 superfamily of enzymes localized to the endoplasmic reticulum. Polymorphisms in CYP1A2 were identified to be associated with more severe periorbital edema, and the gene was also found to be related to the need for dose reduction due to IM toxicity [4]. This was thought to result from the inconsistent metabolic rates of IM between different genotypes. However, other genes may compensate for this.
The sample size of this study was limited and the results may be biased. Additional research with larger samples size and prospective clinical trials are required to confirm the findings of this study. Additionally, the identification of haplotype-tagged SNP (htSNP) genotypes guarantees that all common variants can be captured in gene-drug response correlation studies, whereas single SNP-based correlation studies are difficult to achieve. Nonetheless, this study revealed several SNPs associated with individual treatments, which will facilitate research to identify better treatments and help to optimize individualized medical treatment.