Predictive Significance of Changes in Haemoglobin Level Following Imatinib Treatment in Patients with Advanced Gastrointestinal Stromal Tumour

By Burhan Safakoglu¹, Tulay Kus², Gokmen Aktas³

Affiliations

1. Department of Internal Medicine, School of Medicine, Gaziantep University, Gaziantep, Turkiye
2. Department of Medical Oncology, School of Medicine, Gaziantep University, Gaziantep, Turkiye
3. Department of Medical Oncology, Gaziantep Medical Point Private Hospital, Gaziantep, Turkiye

doi: 10.29271/jcpsp.2025.10.1255

ABSTRACT
Objective: To evaluate the predictive value of haemoglobin decrease following imatinib treatment in patients with advanced gastro- intestinal stromal tumours (GIST).
Study Design: An retrospective observational study.
Place and Duration of the Study: Department of Medical Oncology, Gaziantep Medical Point Private Hospital, Gaziantep, Turkiye, between July 2021 and February 2024.
Methodology: The study included 79 patients diagnosed with advanced GIST following imatinib therapy. Comprehensive clinical information, such as age, gender, tumour location, sites of metastasis, number of metastatic sites, and initial staging before treatment, was collected, and haematological parameters were measured both prior to and 1–3 months after starting imatinib. Changes in haematological parameters were analysed, and the relationship between clinical variables and both disease-free survival (DFS) and overall survival (OS) was assessed using the Kaplan-Meier method.
Results: Post-imatinib treatment, 35 patients (44.3%) experienced a decrease in haemoglobin levels. Those with reduced haemoglobin levels exhibited better progression-free survival (PFS) compared to those without a decline [43.0 months (31.5-54.5) vs. 22.0 months (14.9-29.1); p <0.001]. No significant predictive associations for PFS were identified with changes in other haematological parameters. Multivariate analysis revealed that only a decrease in haemoglobin values remained an independent predictive factor, even after adjusting for tumour localisation [2.5 (1.3-4.7); p = 0.004]. Conversely, concerning the overall survival (OS), neither the decrease in haemoglobin levels nor other haematological and clinical parameters demonstrated statistical significance following imatinib treatment.
Conclusion: Decreased haemoglobin levels following imatinib treatment is a predictive on-target toxicity in patients with advanced GIST.

Key Words: GIST, Imatinib treatment, On-target toxicity, Predictive marker, Anaemia.

INTRODUCTION

Imatinib mesylate acts as a competitive inhibitor that speci- fically targets certain tyrosine kinases, including BCR-ABL, KIT, and platelet-derived growth factor receptors (PDGFR). As a result, it shows considerable effectiveness in treating conditions such as Philadelphia chromosome-positive acute lymphoblastic leukaemia with the BCR-ABL fusion protein and gastrointestinal  stromal  tumour  (GIST)  that  express KIT.¹
 

The key molecular characteristics of GIST include gain-of- function mutations in the KIT proto-oncogene (75–85%), mutations in the PDGFR alpha (5–7%), and kinase-negative GIST (12–15%).^2,3 Immunohistochemistry (IHC) with an anti-KIT (CD117) antibody is a commonly used diagnostic technique for most GIST.⁴ Higher KIT expression intensity detected by IHC has been reported to correlate with better progression-free survival (PFS) following imatinib therapy.⁵ The identification of KIT expression through IHC is generally adequate for assessing the potential benefits of imatinib treatment across different GIST subtypes.⁴ Although traditional risk factors such as tumour size, rupture, necrosis, location, and mitotic index are established prognostic indicators for resected GIST, their relevance is less clear in cases that are unresectable or metastatic.^4,6 Notably, the type of KIT mutation is recognised as a significant prognostic factor for advanced GIST expressing KIT.⁴ The clinical benefit of imatinib in advanced GIST is well established, and recent trials have extended its role in the adjuvant setting.⁷

It is widely recognised that on-target side effects associated with tyrosine kinase inhibitors (TKIs) can serve as predictive biomarkers for treatment efficacy in various cancer types. Imatinib, in addition to its potent inhibitory effect on BCR-ABL, also targets KIT, implicated in early haematopoiesis. This off-target effect results in myelosuppression, which has been established as a poor prognostic biomarker for lymphoblastic leukaemia.^5,6

However, the scenario differs for patients with GIST undergoing imatinib treatment. As KIT is the intended target for GIST, myelotoxicity becomes an on-target toxicity associated with imatinib. Therefore, the occurrence of myelosuppression following imatinib treatment may serve as a significant prognostic or predictive factor in this context. This study aimed to evaluate whether post-treatment myelosuppression can be considered a predictive indicator for patients with GIST who were receiving imatinib  therapy.

METHODOLOGY

This retrospective study was conducted at the Department of Medical Oncology, Gaziantep Medical Point Private Hospital, Gaziantep, Turkiye, between July 2021 and February 2024. The study included 79 patients diagnosed with advanced GIST. Ethical approval was granted by the Gaziantep University Faculty of Medicine Ethics Committee (Approval No. 2021/386), and the research adhered to the ethical principles outlined in the  Declaration  of  Helsinki.

Diagnosis of GIST was confirmed through pathological evaluation, which included histopathological analysis and the detection of KIT (CD117) or discovered on GIST-1 (DOG1) positivity via IHC. To focus on the predictive value of myelosuppression (an adverse effect linked to KIT), the study included only patients with GIST expressing KIT as determined by IHC.

Patients with advanced-stage GIST who had undergone at least one month of imatinib treatment, with complete blood count assessments performed before treatment and within three months afterwards, were included. Ten patients were excluded due to incomplete blood tests. Data on peripheral blood were collected one month prior to and one to three months after imatinib treatment. Haematological parameters were classified as follows: haemoglobin levels of <9 g/dL at baseline, a decrease of 0.8 g/dL from baseline after treatment, or no change/an increase post-treatment; neutrophil levels with a decrease of 1,000mL after treatment, or no decrease/ unchanged; lymphocyte levels with a decrease of 500mL after treatment, or no decrease/unchanged; platelet levels with a decrease of 50,000mL after treatment, or no decrease/ unchanged.

Patients without computed tomography (CT) or 18-fluoro-deoxyglucose positron emission tomography (18F-FDG PET-CT) scans at treatment initiation and at response assessment were excluded. Additional exclusions applied to individuals with chronic inflammatory diseases, active infections, or gastrointestinal bleeding either at baseline or during the first three months of the treatment. Patient demographics were recorded, including age, gender, tumour location, sites of metastasis, number of metastatic sites, and initial stage (unresectable locally advanced disease or stage IV). Information on whether patients received second- or third-line treatments was also documented. The study tracked the date of diagnosis, progression, and the last follow-up or death.

Patients underwent regular follow-up assessments using CT or 18F-FDG PET-CT scans every three months. Progression was evaluated according to the Response Evaluation Criteria in Solid Tumours, version 1.1 (RECIST 1.1), or the Positron Emission Tomography Response Criteria in Solid Tumours (PERCIST).

Quantitative variables were reported as means with standard deviations (SDs) and as medians with ranges, while qualitative variables were expressed as frequencies (percentages). Normality plots, statistical tests, and histograms were used to assess normal distribution. The PFS was calculated from diagnosis to the onset of metastasis, disease progression, or death. Overall survival (OS) was determined from diagnosis to metastasis, death, or the date of the last follow-up visit. The Kaplan–Meier method was used to estimate PFS and OS and to compare the impact of clinicopathological factors and myelosuppression on survival during univariate analysis. Subsequently, the Cox proportional hazards model was utilised for multivariate analysis to identify independent predictive and prognostic factors for PFS and OS. All potential predictive factors with a p-value of <0.10 on univariate analysis were included in the multivariable Cox regression analysis. A p-value of 0.05 or less was deemed statistically significant, with all statistical analyses conducted using SPSS version 22.0 software (SPSS, Chicago, IL, USA).

RESULTS

In total, 79 patients diagnosed with advanced GIST were included in this retrospective analysis. The mean age was 65.35 ± 14.2 years, with 28 patients (35.4%) being female. Among the participants, 26 (32.9%) had gastric GIST, while 37 (46.8%) had non-gastric gastrointestinal tract tumours. The remaining patients presented with retroperitoneal GIST. Approximately 27.9% of patients had locally advanced disease, while the majority had metastatic disease, primarily affecting the liver (53.2%), peritoneum (6.3%), bone (6.3%), and lung/adrenal regions (3.8%). Of these, 12 (15.2%) patients had metastases in two sites, while 46 (58.2%) patients had metastases in a single site.

Out of the 54 patients who experienced progression during imatinib treatment, 28 (35.4%) received the second-line treatment with sunitinib, and 19 (24.1%) were deemed eligible for the third-line treatment with regorafenib.

Haemoglobin levels below 9g/dL were observed in 9 (11.4%) patients, while a decrease in haemoglobin levels was noted in 35 patients (44.3%) following imatinib treatment.

Table I: Effect of clinical and haematological parameters on PFS: univariate and multivariate analyses.

Clinical parameters	Univariate analysis		Multivariate analysis
	PFS, months (95% CI)	p-values	Hazard ratio (95% CI)	p-values
Age (years)	-	-	-	-
<65	25.0 (3.0-47.0)	0.452	-	-
≥65	27.0 (16.6-37.4)	-	-	-
Gender	-	-	-	-
Female	22.0 (17.4-26.6)	0.981	-	-
Male	30.0 (20.4-39.7)	-	-	-
Stage	-	-	-	-
Locally advanced	39.0 (24.6-53.4)	0.975
Metastatic	-	-	-	-
Metastatic site	-	-	-	-
Locally advanced	39.0 (24.6-53.4)	0.738	-	-
Liver dominant	24.0 (15.8-32.3)	-	-	-
Peritoneal metastasis	43.0 (1.0-98.8)	-	-	-
Bone metastasis	30.0 (12.8-47.2)	-	-	-
Lung/adrenal metastasis	68.0 (unreached)	-	-	-
Number of metastatic sites	-	-	-	-
1	26.0 (14.8-37.2)	0.593	-	-
≥2	12.0 (5.8-18.2)	-	-	-
Primary tumour localisation	-	-	-	-
Gastric	39.0 (12.2-65.8)	0.022	-	0.224
Non-gastric GIST	33.0 (21.4-44.6)	-	-	-
Retroperitoneal-intraabdominal	9.0 (5.1-12.9)	-	-	-
Laboratory parameters	-
Haemoglobin levels	-	-	-	-
Decreasing	43.0 (31.5-54.5)	<0.001	1 (reference)	-
Stable or increasing	22.0 (14.9-29.1)	-	2.512 (1.336-4.723)	0.004
<9 gr/dL at the beginning	3.0 (1.0-5.9)	-	11.2 (4.29-29.2)	<0.001
Lymphocyte levels	-	-	-	-
Decreasing	33.0 (11.0-54.9)	0.239	-	-
Stable or increasing	26.0 (14.3-37.8)	-	-	-
Neutrophils	-	-	-	-
Decreasing	30.0 (20.3-39.7)	0.705	-	-
Stable or increasing	26.0 (10.1-41.9)	-	-	-
Platelets	-	-	-	-
Decreasing	39.0 (28.2-49.9)	0.56	-	-
Stable or increasing	30.0 (16.7-43.2)	-	-	-
Primary tumour localisation and change in haemoglobin level were included in a multivariable Cox-regression analysis. PFS: Progression-free survival; GIST: Gastrointestinal stromal tumour; CI: Confidence interval.

As shown in Table I and II, the cohort that experienced a reduction in haemoglobin levels following imatinib treatment demonstrated significantly better PFS compared to those who did not [43.0 months (31.5-54.5) vs. 22.0 months (14.9-29.1), p <0.001; Table I]. No significant predictive associations for PFS were identified with changes in other haematological parameters. Upon evaluating all clinical parameters, tumour location emerged as the sole significant predictor for PFS. GIST originating from the gastric and extra-gastric gastrointestinal regions demonstrated superior PFS compared to retroperitoneal and abdominal GIST. Multivariate analysis, adjusting for tumour localisation, reaffirmed that only a decrease in haemoglobin values remained an independent predictive parameter for PFS [2.512 (1.336-4.723), p = 0.004; Table I]. However, concerning OS, neither the decrease in haemoglobin levels [hazard ratio (HR) 95% CI: 1.785 (0.892-3.570), p = 0.102 for decreasing vs. increasing/stable groups] nor other haematological and clinical parameters showed statistical significance after imatinib treatment (Table II). Notably, the survival of the patient group with haemoglobin levels below 9g/dL was statistically worse (Table II).

DISCUSSION

This study identified imatinib-induced decreases in haemoglobin levels as an independent predictive biomarker of improved clinical outcomes. To the best of the authors’ knowledge, this analysis represents the first of its kind.

The observed association supports the hypothesis that the development of cytopenia, specifically a decrease in haemoglobin, which occurs as an on-target toxicity, may serve as a viable biomarker for imatinib response in this patient population. Patients experiencing a decrease in haemoglobin levels demonstrated a median PFS more than two-fold longer than those with stable haemoglobin levels. Furthermore, in a multivariate Cox proportional hazards model, the decrease in haemoglobin levels remained an independent predictive parameter for disease progression even after adjusting for tumour localisation. Notably, other blood parameters did not exhibit predictive value. This discrepancy may be attributed to the rapid impact of acute inflammatory conditions on leucocyte, lymphocyte, neutrophil, and platelet values, unlike haemoglobin levels, which can be influenced by chronic inflammatory processes.

Table II: Effect of clinical and haematological parameters on OS.

Clinical parameters	Univariate analysis		HR 95% CI p-values
	PFS, months (95% CI)	p-values
Age (years)	-	-	-
<65	54.0 (13.4-94.6)	0.14	-
≥65	46.0 (32.9-59.0)	-	-
Gender	-	-	-
Female	46.0 (32.9-59.1)	0.94	-
Male	50.0 (34.0-65.9)	-	-
Stage	-	-	-
Locally advanced	39.0 (27.5-92.8)	0.82	-
Metastatic	46.0 (34.8-57.2)	-	-
Metastatic site	-	-	-
Locally advanced	Non reached	0.565	-
Liver dominant	-	-	-
Peritoneal metastasis	-	-	-
Bone metastasis	-	-	-
Lung/adrenal met.	-	-	-
Number of metastatic sites	-	-	-
1	46.0 (34.5-57.5)	0.88	-
≥2	40.0 (3.0-76.9)	-	-
Primary tumour localisation	-	-	-
Gastric	54.0 (44.3-67.7)	0.157	-
Non-gastric GIST	52.0 (24.5-79.5)	-	-
Retroperitoneal-intra abdominal	32.0 (6.7-57.3)	-	-
Laboratory parameters Haemoglobin levels	-	-	-
Decreasing	64.0 (12.1-115.9)	-	1 (reference)
Stable or increasing	46.0 (28.6-63.4)	<0.001	1.785; 0.892-3.570; p = 0.102
<9 gr/dL at the beginning	3.0 (1.0-5.9)	-	10.59; 4.30-26.1; p <0.001
Lymphocyte levels	-	-	-
Decreasing	52.0 (43.3-60.6)	0.805	-
Stable or increasing	46.0 (37.5-54.5)	-	-
Neutrophils	-	-	-
Decreasing	52.0 (32.2-71.8)	0.72	-
Stable or increasing	46.0 (31.3-60.7)	-	-
Platelets	-	-	-
Decreasing	52.0 (39.0-64.9)	0.65	-
Stable or increasing	46.0 (21.7-70.3)	-	-
The Cox regression analysis was used to assess hazard ratio of change in haemoglobin levels. OS: Overall survival; GIST: Gastrointestinal stromal tumour; CI: Confidence interval.

Anaemia is a well-recognised adverse event of imatinib therapy, and erythropoietin support has been explored in this context.⁸ Additionally, the finding that a decrease in haemoglobin level serves as a positive predictor, while an increase is expected with disease control, further reinforces the accuracy of this hypothesis. Intriguingly, this study revealed that although a decrease in haemoglobin level is predictive for PFS, it is not prognostic for OS. This distinction suggests that changes in haemoglobin, reflective of KIT- related on-target toxicity, specifically predict a response to imatinib, a KIT inhibitor. However, these changes do not predict a treatment response to other TKIs with multi-kinase activity, underscoring the precision of the hypothesis.

The oncology community has long recognised that toxicity arising from the suppression of a target pathway can serve as a valuable indicator of the inhibition of that specific pathway. A prominent example is the well-established association between hypertension (HT) and hypothyroidism, which are consequences of vascular endothelial growth factor (VEGF) pathway inhibition, and positive prognostic parameters in renal cell carcinoma (RCC).^9,10 Given that the VEGF pathway is central to the pathogenesis of RCC, HT, and hypothyroidism, on-target toxicities linked to this pathway serve as predictive indicators. This pattern is evident across various cancer types, with erlotinib, an epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI), serving as another notable example. In patients treated with erlotinib, the effectiveness of the therapy is significantly enhanced in those who experience skin toxicities, such as rash or diarrhoea, as these are considered on-target toxi- cities.^11-13

However, it is important to note that not all on-target toxi- cities carry the same predictive significance across different cancers. For example, the VEGF pathway plays a central role in the development of RCC. In this context, HT and hypothyroidism—both on-target toxicities related to VEGF TKI treatment—are strong predictors in RCC.¹⁴ In contrast, in hepatocellular carcinoma (HCC) and colon cancer, where multiple pathways contribute to disease progression beyond VEGF receptor overactivation, the occurrence of HT following VEGF TKI treatment does not exhibit the same predictive power, as seen in RCC.¹⁴

Sorafenib, an oral multikinase inhibitor that targets VEGF receptors, PDGFR receptors, FLT3, KIT, and the Raf kinases- MAPK/ERK pathway, exemplifies this. While HT is a common adverse effect, hand-foot skin reaction (HFSR) emerges as a more significant predictive toxicity for patients with HCC treated with sorafenib. This highlights that HFSR, influenced by the inhibition of multiple pathways, provides greater predictive value than HT in this specific context.^15,16

While on-target toxicities often serve as indicators of TKIs benefit, it is crucial to recognise that off-target toxicities can also predict the efficacy of these inhibitors. The half-maximal inhibitory concentration (IC50) levels, at which the pathways responsible for pathogenesis are inhibited, play a pivotal role in this context. An illustrative example is the development of HT in GIST, which predicts the efficacy of sunitinib treatment. In GIST, the proangiogenic growth factor VEGF has a lesser role in pathobiology compared to its prominence in RCC. Despite differences in tumour growth mechanisms and sunitinib's antitumour activity between GIST and RCC, HT has been established as a predictive biomarker not only for RCC but also for GIST.¹⁷ This can be attributed to the significantly higher IC50 of sunitinib for the VEGF receptor compared to the KIT receptor.¹⁸ The development of HT, indicative of VEGF receptor, implies that the drug concentration has exceeded the IC50 level for KIT receptor inhibition. Limited data on the relationship between on-target toxicity and treatment benefits in GIST may make this study particularly significant in establishing HT as a predictive biomarker in GIST treated with sunitinib.¹⁶ In this study, the authors showed for the first time that a decrease in haemoglobin levels—an on-target toxicity related to KIT—acts as an independent predictive marker in patients with GIST receiving imatinib treatment. Conversely, in the treatment of chronic lymphocytic leukaemia (CLL) with imatinib, anaemia induced by the drug has been recognised as a negative prognostic factor, primarily because the main target in CLL therapy is BCR-ABL.⁶ Consequently, understanding the primary driver of the relevant cancer and determining whether the toxicity is an on-target toxicity are crucial factors in interpreting the potential prognostic value of treatment-related toxicity. Beyond clinical predictors, pharmacokinetic monitoring of imatinib exposure has been investigated as a tool to refine treatment individualisation.¹⁹

The primary limitation of this study is the insufficient genetic analysis for the subtypes of the KIT mutation. While the pre- sence of the KIT Exon 11 mutation is regarded as the most favourable prognostic subtype compared to Exon 9 and PDGFR mutations, it is important to note that most KIT- staining GIST typically harbour the KIT Exon 11 mutation. Despite the absence of specific subtype analysis, the potential impact of this omission on the study's outcomes is considered minor. Other notable limitations include the small sample size and the retrospective design of the study. The modest sample size may limit the generalisability of the findings to a wider population, while the retrospective nature introduces inherent limitations, as it relies on historical data and may be subject to biases in data collection. Emerging therapeutic approaches, including novel TKIs and immuno- therapy-based strategies, may further improve outcomes in GIST.²⁰

CONCLUSION

A decrease in haemoglobin levels emerges as a robust predictive parameter in patients with advanced-stage GIST following imatinib treatment.

ETHICAL  APPROVAL:
Ethical approval was granted by the Gaziantep University Faculty of Medicine Ethics Committee (Approval No. 2021/386). The research also adhered to the ethical principles outlined in the Declaration of Helsinki.

PATIENTS’ CONSENT:
Informed consent was obtained from all patients.

COMPETING INTEREST:
The authors declared no conflict of interest.

AUTHORS’ CONTRIBUTION:
BS, TK, GA: Concept, design, and drafting of the manuscript.
BS: Data collection.
TK, GA: Analysis and interpretation.
All authors approved the final version of the manuscript to be published.

REFERENCES

1. Joensuu H, Dimitrijevic S. Tyrosine kinase inhibitor imatinib (STI571) as an anticancer agent for solid tumours. Ann Med 2001; 33(7):451-5. doi: 10.3109/07853890109002093.
2. Chen P, Zong L, Zhao W, Shi L. Efficacy evaluation of imatinib treatment in patients with gastrointestinal stromal tumours: A meta-analysis. World J Gastroenterol 2010; 16:4227-32. doi: 10.3748/wjg.v16.i33.4227.
3. Zong L, Chen P. Prognostic value of KIT/PDGFRA mutations in gastrointestinal stromal tumours: A meta-analysis. World J Surg Oncol 2014; 12:71. doi: 10.1186/1477-7819-12-71.
4. Miettinen M, Lasota J. Gastrointestinal stromal tumours: Pathology and prognosis at different sites. Semin Diagn Pathol 2006; 23(2):70-83. doi: 10.1053/j.semdp.2006. 09.001.
5. Chirieac LR, Trent JC, Steinert DM, Choi H, Yang Y, Zhang J, et al. Correlation of immunophenotype with progression-free survival in patients with gastrointestinal stromal tumours treated with imatinib mesylate. Cancer 2006; 107(9): 2237-44. doi: 10.1002/cncr.22226.
6. Sneed TB, Kantarjian HM, Talpaz M, O'Brien S, Rios MB, Bekele BN, et al. The significance of myelosuppression during therapy with imatinib mesylate in patients with chronic myelogenous leukemia in chronic phase. Cancer 2004; 100(1):116-21. doi: 10.1002/cncr.11863.
7. Blay JY, Penel N, Schiffler C, Chabaud S, Perol D, Le Cesne A. Six years duration of adjuvant imatinib improves disease-free survival in GIST with a high risk of relapse. Ann Oncol 2025; 36(1):120-1. doi: 10.1016/j.annonc.2024.09.018.
8. Duffaud F, Even C, Ray-Coquard I, Bompas E, Khoa-Huynh T, Salas S, et al. Recombinant erythropoietin for the anaemia of patients with advanced gastrointestinal stromal tumours (GIST) receiving imatinib: An active agent only in non-progressive patients. Clin Sarcoma Res 2012; 2(1):11. doi: 10.1186/2045-3329-2-11.
9. Rini BI, Cohen DP, Lu DR, Chen I, Hariharan S, Gore ME, et al. Hypertension as a biomarker of efficacy in patients with metastatic renal cell carcinoma treated with sunitinib. J Natl Cancer Inst 2011; 103(9):763-73. doi: 10.1093/jnci/ djr128.
10. Bianchi L, Rossi L, Tomao F, Papa A, Zoratto F, Tomao S. Thyroid dysfunction and tyrosine kinase inhibitors in renal cell carcinoma. Endocr Relat Cancer 2013; 20(5):R233-45. doi: 10.1530/ERC-13-0201.
11. Perez-Soler R, Chachoua A, Hammond LA, Rowinsky EK, Huberman M, Karp D, et al. Determinants of tumour response and survival with erlotinib in patients with non- small-cell lung cancer. J Clin Oncol 2004; 22(16):3238-47. doi: 10.1200/JCO.2004.11.057.
12. Ding K, Pater J, Whitehead M, Seymour L, Shepherd FA. Validation of treatment induced specific adverse effect as a predictor of treatment benefit: A case study of NCIC CTG BR21. Contemp Clin Trials 2008; 29(4):527-36. doi: 10.1016/j.cct.2008.01.004.
13. Fiala O, Hosek P, Pesek M, Finek J, Racek J, Stehlik P, et al. Serum concentration of erlotinib and its correlation with outcome and toxicity in patients with advanced-stage NSCLC. Anticancer Res 2017; 37(11):6469-76. doi: 10.21873/anticanres.12102.
14. Giampieri R, Prete MD, Prochilo T, Puzzoni M, Pusceddu V, Pani F, et al. Off-target effects and clinical outcome in metastatic colorectal cancer patients receiving regorafenib: The TRIBUTE analysis. Sci Rep 2017; 7:45703. doi: 10. 1038/srep45703.
15. Marisi G, Cucchetti A, Ulivi P, Canale M, Cabibbo G, Solaini L, et al. Ten years of sorafenib in hepatocellular carcinoma: Are there any predictive and/or prognostic markers? World J Gastroenterol 2018; 24(36):4152-63. doi: 10.3748/wjg. v24.i36.4152.
16. Di Costanzo GG, de Stefano G, Tortora R, Farella N, Addario L, Lampasi F, et al. Sorafenib off-target effects predict outcomes in patients treated for hepatocellular carcinoma. Future Oncol 2015; 11(6):943-51. doi: 10.2217/fon.14.291.
17. George S, Reichardt P, Lechner T, Li S, Cohen DP, Demetri GD. Hypertension as a potential biomarker of efficacy in patients with gastrointestinal stromal tumour treated with sunitinib. Ann Oncol 2012; 23(12):3180-7. doi: 10.1093/ annonc/mds179.
18. Kumar R, Crouthamel MC, Rominger DH, Gontarek RR, Tummino PJ, Levin RA, et al. Myelosuppression and kinase selectivity of multikinase angiogenesis inhibitors. Br J Cancer 2009; 101(10):1717-23. doi: 10.1038/sj.bjc.660 5366.
19. Ran P, Tan T, Li J, Yang H, Li J, Zhang J. Advanced gastrointestinal stromal tumour: Reliable classification of imatinib plasma trough concentration via machine learning. BMC Cancer 2024; 24(1):264. doi: 10.1186/s12885-024- 11930-6.
20. Jones RL, Golcic M. Recent advances in the systemic treatment of gastrointestinal stromal tumours. Cancer Biol Med 2023; 20(10):701-5. doi: 10.20892/j.issn.2095-3941.2023. 0302.