Patients who were diagnosed with liver metastases from GISTs at Sun Yat-Sen University Cancer Center (Guangzhou, China) between January 2010 and December 2022 were identified from the institutional medical records. Ethical approval for the present study was granted by the Institutional Review Board of Sun Yat-sen University Cancer Center. The inclusion criteria encompassed the following conditions: i) Pathological confirmation of the GIST as the primary tumor site and removal of the tumor; ii) diagnosis of liver metastases from GIST through post-operative pathology, biopsy, imaging and clinical data; and iii) radiofrequency ablation or hepatectomy being used as a treatment for liver metastases. Exclusion criteria were as follows: i) Patients <18 years of age; ii) the presence of other malignancies; iii) the existence of other distant metastases from GIST; and iv) incomplete patient data. Liver metastases identified concomitantly with the primary tumor diagnosis or within the initial 6 months were classified as synchronous metastases. By contrast, metachronous metastases were classified by the detection of liver metastases >6 months after the primary tumor diagnosis. Comprehensive data, including patient demographics, clinicopathological features and other pertinent information, were meticulously reviewed.

A multidisciplinary team composed of gastrointestinal surgeons, hepatobiliary surgeons and interventional radiologists, based on a comprehensive assessment of the patient's condition, made the decision to perform either RFA or a hepatectomy. In cases involving patients with a large volume of liver metastases or those located near critical structures, the multidisciplinary team favored liver resection. Additionally, for patients at an unacceptable risk of liver dysfunction post-resection, the preference leaned towards RFA. If RFA or hepatectomy were deemed technically viable but no decision was made during a multidisciplinary discussion meeting, the patient and the surgeons would confer before making a final choice. For patients selected to undergo hepatectomy or RFA, the procedures were determined following a comprehensive evaluation. Percutaneous RFA was performed with the patient under local anesthesia and conscious sedation, guided by ultrasound imaging. The radio frequency system was chosen based on the operator's thorough assessment of the amount and location of liver metastases. The procedural objective for targeted RFA was to achieve an ablative margin measuring at least 0.5–1.0 cm. Surgical margins were categorized as follows: i) Microscopically complete (R0), indicating no observable tumor cells; ii) macroscopically complete with microscopic residual tumor cells; or iii) macroscopically incomplete.

Patients usually went to the hospital for regular follow-up every 3 months, including physical examinations, laboratory tests and imaging evaluations. Imaging examinations included contrast-enhanced multiphase computed tomography or magnetic resonance imaging, while laboratory tests encompassed a complete blood count, liver and renal function assessments, and tumor marker analyses, among others. The key parameters of interest were PFS time and time to imatinib treatment failure (TTF). Specifically, PFS time was characterized as the duration spanning from the occurrence of RFA or hepatectomy for liver metastases to the point of disease progression or mortality due to any cause. TTF was defined as the duration from the initiation of imatinib treatment for liver metastases from GISTs to the point at which treatment discontinuation occurred due to therapeutic ineffectiveness, intolerable toxicity or other factors influencing treatment failure. The period from the date of hepatic resection or RFA for liver metastases until the final follow-up or death was referred to as the OS time.

Nominal data were analyzed through Pearson's χ² test or Fisher's exact test, as deemed appropriate. To depict survival trends, the Kaplan-Meier method was employed, and differences between groups were ascertained via the log-rank test. For examining the potential prognostic factors and their association with PFS time, variables with a P-value of <0.2 or those considered potentially associated with PFS in the univariate analysis were included in the subsequent multivariable Cox proportional hazards model. All statistical analyses were performed using R version 4.2.2 (R Core Team). A significance threshold was established at a two-tailed P-value of <0.05.

The present study was ultimately comprised of 43 patients with liver metastases from GISTs treated between January 2010 and December 2022. A comparative analysis of baseline clinicopathological characteristics between the RFA group (n=18) and the hepatectomy group (n=25) is detailed in Table I. Notably, there were no statistically significant differences in the baseline attributes between these two groups. The patient age range was 19–68 years, with a median age of 50.5 years. The majority of patients (60.5%) were male.

Regarding the primary GIST site, 25 patients (58.1%) had tumors not located in the stomach, while 18 patients (41.9%) had tumors in the stomach. Liver metastases occurred beyond 6 months after the removal of the original tumor in 21 patients (48.8%; metachronous), whereas 22 patients (51.2%) initially presented with metastatic GISTs (synchronous). Among the analyzed patients, mutational analysis was conducted in 27 patients, revealing KIT 11 mutations in 17 patients (63.0%), KIT 9 mutations in 5 patients (18.5%) and wild-type variants in 5 patients (18.5%).

Following the diagnosis of liver metastases from GISTs, all patients underwent TKI treatment tailored to the nature of their condition. Initially, 41 patients were administered first-line imatinib therapy, while 2 patients received second-line sunitinib treatment following their initial diagnosis of liver metastases from GISTs. In the current study, the treatment approaches were diversified. Among the participants, 25 patients (58.1%) underwent hepatectomy combined with TKI therapy, while 18 patients (41.9%) received RFA combined with TKI treatment (Table I). Within the hepatectomy group, 15 patients (60.0%) underwent resection and received perioperative TKIs, whereas 10 patients (40.0%) underwent hepatic resection and exclusively received postoperative TKI therapy. Notably, all resections within the hepatectomy group achieved R0 status. Conversely, the RFA group experienced entirely ablative RFA procedures to eliminate target liver metastases.

With a median follow-up duration of 38.1 months (range, 8.5–131.6 months), the 1-, 3- and 5-year PFS rates for all patients with liver metastases were 66.5, 38.2 and 33.9%, respectively. Specifically, patients with hepatic metastases exhibited 1-, 3-, and 5-year TTF rates of 95.1, 70.5 and 54.7%, respectively. Recurrence was observed in 11 patients (61.1%) in the RFA group and 12 patients (48.0%) in the hepatectomy group. Notably, patients who underwent hepatic resection alongside TKI therapy demonstrated significantly improved PFS and TTF times compared with those who received RFA combined with TKI therapy (median PFS, 42.7 months vs. 14.3 months, P=0.034; median TTF, 71.1 months vs. 38.0 months, P=0.041) (Fig. 1). Furthermore, no statistically significant disparity in OS time was observed between patients who underwent hepatectomy and those who underwent RFA (median OS, not reached vs. not reached, P=0.120) (Fig. 2). There was no statistically significant difference in PFS and TTF times between patients who received hepatectomy combined with postoperative TKI and those who received hepatectomy combined with perioperative TKI (median PFS: 29.5 months vs. not reached, P=0.520; median TTF, 66.4 months vs. 71.1 months, P=0.430) (Fig. 3). Table II outlines the outcomes of univariate and multivariate analyses for a comprehensive view of potential prognostic factors affecting PFS. These analyses identified the hepatectomy group as being independently associated with longer PFS time (HR, 0.379; 95% CI, 0.159–0.899; P=0.028).