Association of recurrence with patient and tumor factors in gastrointestinal stromal tumors of small intestine: A single‑institutional retrospective study

Introduction

Gastrointestinal stromal tumors (GISTs) are mesenchymal tumors that commonly occur in the gastrointestinal tract. They originate from the precursor cells of Cajal mesenchymal cells in the muscle plexus. GISTs are rare tumors, with an incidence of approximately eight cases per million per year (1). They occur throughout the gastrointestinal tract from the lower esophagus to the anus. The most common site is the stomach (60%), followed by the small intestine (35%) and rectum, esophagus, omentum, and mesentery (less than 5% each) (2).

The incidence of small intestinal GISTs is increasing, which may reflect advances in radiologic and endoscopic techniques and an increase in the frequency of imaging (3,4). Small intestinal GISTs have a higher recurrence rate compared to gastric GISTs (5–7), and the 5-year disease-free survival rate is 44–89% (3,5,8). Miettinen et al proposed a risk classification system based on the differences in recurrence rates for each tumor location (2). However, owing to the rarity of GISTs, data on GISTs of the small intestine and other intestinal regions are lacking (2,7). The aim of this study was to identify the factors involved in the recurrence of small intestinal GISTs using data obtained from patients treated at our institution.

Patients and methods

Patients

A retrospective review was conducted of adult patients who underwent surgery for small intestinal tumors at Chiba University Hospital between January 2002 and December 2021. Among these, patients who were histologically diagnosed with primary small intestinal GIST, had undergone preoperative imaging to assess for distant metastasis, and received curative resection were included in this study. Patients who underwent palliative surgery, had concurrent active malignancies, or lacked sufficient clinicopathological data were excluded. Based on these criteria, 26 patients with 27 lesions were included in the present study. Data on demographic details, clinical history, preoperative evaluation including blood tests and imaging studies, treatment details, histopathological examination results, clinical course, and survival follow-up were extracted from the patients' medical records.

Ethical approval

This study was approved by the Ethics Committee of Chiba University Graduate School of Medicine (Chiba, Japan; approval no: 3043). Consent for this study was provided on an opt-out basis by 26 patients.

Statistical analysis

The Fisher's exact test was used for categorical variables, and the Wilcoxon test was used for continuous variables, as appropriate. The cut-off values for continuous variables were set using the receiver operating characteristic (ROC) curve, as the values vary depending on the report and there are no clear standard values. The outcome was defined as recurrence. Univariate predictors of recurrence were determined using a nominal logistic regression analysis. Correlations between the factors were estimated using the pairwise method. Statistical analyses were performed using JMP Pro version 18.1.1 (JMP Statistical Discovery LLC., Cary, NC, USA), and statistical significance was set at P<0.05.

Results

Characteristics of patients

The patients' characteristics are listed in Table I. The median patient age was 59 years (15 male and 11 female). The localization of the tumor was as follows: 10 lesions in the duodenum, 11 lesions in the jejunum, and five lesions in the ileum, with one lesion in which the tumor was large and the localization was unclear. Sixteen patients had clinical symptoms at the time of their first visit, and the main symptoms were gastrointestinal bleeding, abdominal pain, and abdominal distension. Positron emission tomography-computed tomography (PET-CT) data were available for 21 lesions, with a median maximum standardized uptake value (SUVmax) of 4.16. Fourteen lesions were classified as moderate or high risk according to Miettinen's recurrence risk classification, including nine lesions in the jejunum, four lesions in the ileum, and one large lesion that spanned the jejunum and ileum. One patient was treated with imatinib preoperatively, and 12 received postoperative adjuvant chemotherapy with imatinib. Recurrence occurred in nine patients, six of whom were administered adjuvant chemotherapy. The sites of recurrence were the liver, lymph nodes, and peritoneum. Hepatic resection was performed in four patients with solitary liver metastasis. The median observation period for all patients was 72 (11–208) months.

Table I. Clinicopathological features of all patients.

Table I.

Clinicopathological features of all patients.

Variables	Groups	Value
Age, years	Median (min-max)	59 (42–81)
Sex, n	Male/female	15/11
Localization, n	D/J/I/J-I	10/11/5/1
Clinical symptom, n	Yes/no	16/10
Tumor size, mm	Median (min-max)	47 (10–300)
SUVmax	Median (min-max)	4.16 (0–24)
WBC, /µl	Median (min-max)	5,800 (2,200–11,600)
Hb, g/dl	Median (min-max)	12.3 (9.4–15.7)
TP, g/dl	Median (min-max)	7.1 (4.6–8.1)
Alb, g/dl	Median (min-max)	4.3 (2.2–4.9)
CRP, mg/dl	Median (min-max)	0.1 (0.0–3.7)
NLR	Median (min-max)	2.3 (1.1–8.8)
PLR	Median (min-max)	212.5 (74.4–589.6)
OPNI	Median (min-max)	51.2 (30.5–58.6)
CEA, ng/ml	Median (min-max)	1.5 (0.4–5.7)
CA19-9, U/ml	Median (min-max)	8.3 (0.1–95.8)
Operative procedure, n	PpPD/PR/lap-PR	2/23/1
Mitotic count, /50HPF	Median (min-max)	4 (0–62)
Ki-67 index, %	Median (min-max)	5.0 (1.0–30.8)
Risk classification (Miettinen), n	None/low/moderate/high/NA	4/8/4/10/1
Adjuvant chemotherapy, n	Yes/no	12/14
Recurrence, n	Yes/no	9/17
Recurrent organsa, n	Liver/lymph node/peritoneal dissemination	6/2/2
Observation period, months	Median (min-max)	72 (11–208)

a Including duplicates. D, duodenum; J, jejunum; I, ileum; SUVmax, maximum standardized uptake value; WBC, white blood cell count; Hb, hemoglobin; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; TP, total protein; Alb, albumin; OPNI, Onodera's prognostic nutritional index; CRP, C-reactive protein; CEA, carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9; PpPD, pylorus-preserving pancreatoduodenectomy; PR, partial resection; lap-PR, laparoscopic partial resection; HPF, high power field; NA, not available.

Comparison of patients with recurrence and no recurrence of tumors

Patient backgrounds, tumor factors, and laboratory values of recurrence and no-recurrence cases are listed in Table II. No consistent trends were observed in terms of age, sex, or the presence of clinical symptoms (P=0.571, 0.217, and 0.399, respectively). The tumor locations were significantly different between the recurrence group and the no-recurrence group (non-duodenal tumors: 100 (9/9) vs. 44% (8/18); P=0.009). Regarding tumor factors, the recurrence group had larger tumor sizes (70 vs. 40 mm; P=0.010), higher SUVmax (8.8 vs. 3.3; P=0.032), and higher Ki-67 index (12.8 vs. 4.2; P=0.019) than those in the no-recurrence group. In addition, the recurrence group had lower hemoglobin level (11.4 vs. 13.2 g/dl; P=0.046), higher neutrophil-to-lymphocyte ratio (NLR; 4.4 vs. 1.9; P=0.040), and higher platelet-to-lymphocyte ratio (PLR) (269.8 vs. 161.8; P=0.036) than those in the no-recurrence group. No differences were observed in the levels of carcinoembryonic antigen (CEA) and CA19-9, commonly used tumor markers in gastrointestinal cancer, between the two groups (P=0.979 and 0.872, respectively). The median observation period for the recurrence group was 120 (11–208) months, and the period to recurrence was 66 (4–176) months. The median observation period for the no-recurrence group was 64 (41–101) months.

Table II. Patient characteristics with or without recurrence.

Table II.

Patient characteristics with or without recurrence.

Variables	Group	Recurrence (n=9/9)a	No-recurrence (n=17/18)a	P-value
Age, years	Median (min-max)	59 (43–75)	59 (42–81)	0.571b
Sex, n	Male/female	7/2	8/9	0.217c
Localization, n	D/J, I, J-I	0/6, 2, 1	10/5, 3, 0	0.009c
Clinical symptom, n	Yes/no	7/2	9/8	0.399c
Tumor size, mm	Median (min-max)	70 (42–300)	40 (10–115)	0.010b
SUVmax	Median (min-max)	8.8 (3.3–18.7)	3.3 (0.0–24.0)	0.032b
Mitotic count, /50HPF	Median (min-max)	9 (3–23)	3 (0–62)	0.073b
Ki-67 index, %	Median (min-max)	12.8 (2.0–30.0)	4.2 (1.0–30.8)	0.019b
WBC, /µl	Median (min-max)	6,400 (3,700–9,600)	5,600 (2,200–11,600)	0.571b
Hb, g/dl	Median (min-max)	11.4 (9.6–14.9)	13.2 (9.4–15.7)	0.046b
TP, g/dl	Median (min-max)	6.7 (5.8–7.7)	7.2 (4.6–8.1)	0.137b
Alb, g/dl	Median (min-max)	4.0 (3.0–4.8)	4.3 (2.2–4.9)	0.449b
OPNI	Median (min-max)	46.1 (34.7–58.6)	52.0 (30.5–58.3)	0.131b
CRP, mg/dl	Median (min-max)	0.1 (0.0–0.7)	0.1 (0.0–3.7)	0.735b
NLR	Median (min-max)	4.4 (1.4–8.8)	1.9 (1.1–8.5)	0.040b
PLR	Median (min-max)	269.8 (115.9–589.6)	161.8 (74.4–429.3)	0.036b
CEA, ng/ml	Median (min-max)	1.5 (0.7–3.7)	1.5 (0.4–5.7)	0.979b
CA19-9, U/ml	Median (min-max)	8.6 (0.1–24.7)	8.0 (0.1–95.8)	0.872b
Adjuvant chemotherapy, n	Yes/no	5/4	6 / 11	0.411c
Observation period, months	Median (min-max)	120 (11–208)	64 (41–101)	0.004b
Period to recurrence, months	Median (min-max)	66 (4–176)	-	-

a Number of cases/lesions.

b Wilcoxon test.

c Fisher's exact test. D, duodenum; J, jejunum; I, ileum; SUVmax, maximum standardized uptake value; HPF, high power field; WBC, white blood cell count; Hb, hemoglobin; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; TP, total protein; Alb, albumin; OPNI, Onodera's prognostic nutritional index; CRP, C-reactive protein; CEA, carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9.

Factors associated with recurrence

To investigate the association between recurrence and clinicopathological characteristics in patients with small intestinal GISTs, a univariate nominal logistic analysis was performed. The results showed that jejunal lesions [odds ratio (OR)=7.8; 95% confidence interval (CI): 1.13–67.4; P=0.023], tumor size ≥60 mm (OR=11.38; 95% CI: 1.9–101.71; P=0.007), SUVmax ≥8.4 (OR=32.5; 95% CI: 3.25–851.4; P=0.002), mitotic counts ≥4/50 HPF (OR=9.43; 95% CI: 1.25–199.01; P=0.028), Ki-67 index ≥10% (OR=24.5; 95% CI: 3.13–537.24; P=0.001), NLR ≥3.7 (OR=16.33; 95% CI: 2.57–159.05; P=0.002), PLR ≥227 (OR=11.78; 95% CI: 1.9–101.71; P=0.007), and Onodera's prognostic nutritional index (OPNI) ≤51 (OR=6.42; 95% CI: 1.13–53.39; P=0.035) were significantly associated with postoperative recurrence (Table III). Because the number of events was small, a multivariate analysis was not performed. The sensitivity and specificity of each factor are listed in Table IV. The mitotic counts and Ki-67 indices had high sensitivity, whereas the SUVmax and NLR values had high specificity.

Table III. Univariate analysis of recurrence.

Table III.

Univariate analysis of recurrence.

Variables	Groups	OR	95% CI	P-value
Age, year	≤48	2.14	0.22–21.12	0.491
	>48	Ref.
Sex	Male	3.94	0.70–31.99	0.123
	Female	Ref.
Localization	Jejunum	7.80	1.13–67.40	0.023
	Others	Ref.
Tumor size, mm	≥60	11.38	1.90–101.71	0.007
	<60	Ref.
SUVmax	≥8.4	32.50	3.25–851.40	0.002
	<8.4	Ref.
Mitotic count, /50HPF	≥4	9.43	1.25–199.01	0.028
	<4	Ref.
Ki-67 index, %	≥10	24.5	3.13–537.24	0.001
	<10	Ref.
WBC, /µl	≥6,200	3.67	0.70–22.95	0.125
	<6,200	Ref.
Hb, g/dl	≤11.4	6.00	0.90–54.26	0.065
	>11.4	Ref.
NLR	≥3.7	16.33	2.57–159.05	0.002
	<3.7	Ref.
PLR	≥227	11.78	1.90–101.71	0.007
	<227	Ref.
OPNI	≤51	6.42	1.13–53.39	0.035
	>51	Ref.
CEA, ng/ml	≥1.8	0.71	0.12–3.76	0.694
	<1.8	Ref.
CA19-9, U/ml	≥2.9	0.5	0.018–13.75	0.642
	<2.9	Ref.

[i] OR, odds ratio; CI, confidence interval; D, duodenum; SUVmax, maximum standardized uptake value; HPF, high power field; WBC, white blood cell count; Hb, hemoglobin; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; OPNI, Onodera's prognostic nutritional index; CEA, carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9.

Table IV. Prediction accuracy of each factor.

Table IV.

Prediction accuracy of each factor.

Variables	Sensitivity, %	Specificity, %	AUC
Jejunal lesion	75.0	70.6	-
Tumor size ≥60 mm	77.8	76.5	0.82
SUVmax ≥8.4	71.4	92.3	0.82
Mitotic count ≥4/50HPF	85.7	64.7	0.73
Ki-67 index ≥10%	87.5	76.5	0.81
NLR ≥3.7	77.8	82.4	0.75
PLR ≥227	77.8	76.5	0.76
OPNI ≤51	77.8	64.7	0.69

[i] AUC, area under the curve; SUVmax, maximum standardized uptake value; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; OPNI, Onodera's prognostic nutritional index.

Correlations among factors associated with recurrence

Fig. 1 shows the correlations between continuous variables. Scatter plots for each pair of variables are displayed in the lower left of the matrix, while the corresponding correlation coefficients are shown in the upper right and presented as a heat map. Among the tumor factors (tumor size, SUVmax, mitotic counts, Ki-67 index), strong correlations were observed between tumor size and SUVmax (R=0.79; P<0.0001) and between mitotic counts and Ki-67 index values (R=0.74; P<0.0001). Among the patient factors (NLR, PLR, OPNI), NLR and PLR showed a strong correlation (R=0.79; P<0.0001), and both were negatively correlated with OPNI (R=−0.66 and −0.72; P<0.001 and P<0.0001, respectively). Regarding the relationship between tumor factors and patient factors, NLR showed a moderate correlation with tumor size (R=0.65; P=<0.001) and Ki-67 index (R=0.46; P=0.021), whereas PLR and OPNI exhibited only weak correlations with tumor factors (R ranging from 0.10 to 0.29 for PLR, and −0.38 to −0.28 for OPNI).

Figure 1. Scatter plot matrix and correlation coefficients for clinicopathological factors in patients with small intestinal GIST. Strong correlations were observed between tumor size and SUVmax (R=0.79; P<0.0001) and between mitotic counts and Ki-67 index (R=0.74; P<0.0001). The patient factors showed moderate to strong correlations with each other. Regarding the relationship between tumor factors and patient factors, the NLR showed a moderate correlation with tumor size (R=0.65; P<0.001) and Ki-67 index (R=0.46; P=0.021), whereas the PLR and OPNI were weakly correlated with tumor factors (R ranging from 0.10 to 0.29 for PLR, and −0.38 to −0.28 for OPNI). NLR, neutrophil-to-lymphocyte ratio; GIST, gastrointestinal stromal tumors; SUVmax, maximum standardized uptake value; PLR, platelet-to-lymphocyte ratio; OPNI, Onodera's prognostic nutritional index.

Discussion

In the present study, we reviewed the outcomes of patients with small intestinal GISTs who underwent curative resection at our institution. Of the 26 patients included in this study, recurrence occurred in nine patients (34.6%), which is comparable to the rate observed in previous studies (5). Although some GISTs metastasize within 1–2 years or less than a year of development of the initial tumor, in some cases, metastasis may occur after a long delay, such as 42 years after the initial tumor development (7). The median period to recurrence in this study was more than five years, indicating the need for long-term follow-up. In addition, approximately half of the patients with recurrent tumors in our study received adjuvant chemotherapy with imatinib, which may be related to the long time to recurrence. Adjuvant chemotherapy is recommended to be administered for three years after surgery for patients with high-risk GISTs after radical resection; however, the duration of adjuvant chemotherapy administration remains uncertain. Ongoing randomized clinical trials are evaluating the efficacy of administering adjuvant therapy for more than three years (1).

Tumor location, tumor size, and mitotic counts are important prognostic factors, and risk classifications that include these factors, such as the Miettinen or modified Fletcher (mFletcher) classification, are commonly used. These factors were also found to be associated with recurrence in this study. Ki-67, a nuclear marker of actively proliferating cells, is an indicator of cell proliferation and mitotic counts (9); a strong correlation was observed between the Ki-67 index and mitotic counts. Furthermore, the Ki-67 index was more sensitive and specific for recurrence than the mitotic counts. Because the Ki-67 index may show less variability between observers than that shown by mitotic counts (10), use of the Ki-67 index may be a potential alternative to the use of mitotic counts for analysis of recurrence.

The SUVmax in GISTs has been reported to be significantly high in the high-risk groups categorized according to the Fletcher and mFletcher classifications (11–13). In this study, the SUVmax showed a high specificity for recurrence (92.3%), suggesting that SUVmax may be used to accurately identify patients with a high risk of recurrence. In addition, SUVmax enables preoperative evaluation, whereas tumor resection is necessary for accurate evaluation of the mitotic counts and Ki-67 index. This may be useful for determining the indications and efficacy of preoperative adjuvant chemotherapy. Furthermore, it provides information on the necessity of postoperative adjuvant chemotherapy prior to surgery, which is helpful in determining the treatment schedule and informing patients.

The prognosis of GIST is associated not only with tumor characteristics, but also with host immune and inflammatory responses; however, the underlying mechanism has not been elucidated (14–17). This study showed that the patient factors NLR, PLR, and OPNI were significantly associated with recurrence. This suggests that the immune and nutritional status of patients are involved in tumor progression. These factors could be modified by drugs that affect white blood cell and platelet counts, such as steroids, granulocyte colony-stimulating factor preparations, and thrombopoietin receptor agonists, or by nutritional interventions. If immune and nutritional status influence tumor progression, these factors could become novel therapeutic targets. Furthermore, NLR was moderately correlated with tumor factors, such as tumor size and Ki-67 index, suggesting that tumor progression itself may also alter the host's immune status. These findings may provide a basis for future studies investigating whether immune status and nutritional status play a role in the management of GIST.

Most GISTs, approximately 85–90%, contain genetic mutations in KIT or platelet derived growth factor receptor (PDGFRA) (7). The vast majority of KIT mutations are found in exon 11 (66–71%), exon 9 (13–15%), exon 13 (1–3%), and exon 17 (1–3%) (18,19). GISTs with KIT exon 9 mutations have a poorer prognosis than tumors with exon 11 mutations and are resistant to imatinib (19). Furthermore, KIT exon 9 mutations are almost specific to intestinal GIST, and the poor prognosis of small intestinal GISTs have been suggested to be caused by the high frequency of exon 9 mutations (2,19). PDGFRA mutations are found in 30% of KIT wild-type GISTs and are associated with resistance to imatinib (7). Genetic mutation information was not obtained in this study, and the association between the recurrence risk factors identified in this study and genetic mutations is uncertain. Further investigation of the association between clinical factors and genetic factors is needed.

This study had several limitations. First, as this was a retrospective study, it was prone to selection and information biases. Second, the sample size included in the study was relatively small. The small number of recurrence events (n=9), which was insufficient to allow for a reliable multivariate analysis. Therefore, the independence of each factor could not be evaluated. Third, the observation period for the group with non-recurrent tumors was short, and cases of recurrence may not have been accurately identified. Fourth, the lack of information on genetic mutations prevented analysis of their association with the identified risk factors. Fifth, as this was an observational study, and the potential for intervention of the identified risk factors could not be evaluated. Sixth, since this was a single-institutional study, the generalizability of our findings may be limited. However, larger prospective studies are required to validate these findings.

In conclusion, in this study, we investigated the factors associated with recurrence in patients with small intestinal GISTs who underwent curative resection. Our findings suggest that tumor location, tumor size, SUVmax, mitotic counts, Ki-67 index, and inflammatory/nutritional marker levels, such as NLR, PLR, and OPNI levels, are significantly associated with recurrence. In addition, NLR was correlated with tumor size and Ki-67 index. The identification of novel risk factors in addition to known risk factors may enhance current risk classification and allow for more accurate risk stratification of small intestinal GISTs. It may also provide information on the indications and management of perioperative adjuvant chemotherapy. Furthermore, elucidating the fundamental mechanisms underlying the association between inflammation/nutritional indicators and tumor behavior may clarify the oncological characteristics of GIST and contribute to the establishment of novel treatment strategies.

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

YN, YM, GO, KH, MM and HM contributed to the conceptualization of this study. YM, GO, KH, MM and HM conducted project administration. YN, YM, TT, TM, YK, RO, KO and AH performed data collection and analysis. GO and KH confirm the authenticity of all the raw data. YN wrote the manuscript with support from KO and AH. All authors provided critical feedback and read and approved the final manuscript.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Chiba University Graduate School of Medicine (Chiba, Japan; approval no. 3043). Consent for this study was provided on an opt-out basis by 26 patients.

Patient consent for publication

Patient consent for the publication of their clinical data was obtained on an opt-out basis.

Competing interests

The authors declare that they have no competing interests.

Authors' information

Yuri Nishioka, ORCID: https://orcid.org/0009-0003-1479-5123; Yasunori Matsumoto, ORCID: https://orcid.org/0000-0002-6239-6691.

References