Abstract
Mesenteric fibrosis (MF) surrounding a mesenteric mass is a hallmark feature of small intestinal neuroendocrine tumours (SI-NETs). Since this can induce intestinal obstruction, oedema and ischaemia, prophylactic resection of the primary tumour and mesenteric mass is often recommended. This study assessed the predictors for mesenteric metastasis and fibrosis and the effect of MF and palliative surgery on survival. A retrospective analysis of 559 patients with pathologically proven SI-NET and available CT-imaging data was performed. Clinical characteristics, presence of mesenteric mass and fibrosis on CT imaging and the effect of palliative abdominal surgery on overall survival were assessed. We found that MF was present in 41.4%. Older age, 5-HIAA excretion ≥67 μmol/24 h, serum CgA ≥121.5 μg/L and a mesenteric mass ≥27.5 mm were independent predictors of MF. In patients ≤52 years, mesenteric mass was less often found in women than in men (39% vs 64%, P = 0.002). Corrected for age, tumour grade, CgA and liver metastasis, MF was not a prognostic factor for overall survival. In patients undergoing palliative surgery, metastasectomy of mesenteric mass or prophylactic surgery did not result in survival benefit. In conclusion, we confirmed known predictors of MF and mesenteric mass and suggest a role for sex hormones as women ≤52 years have less often a mesenteric mass. Furthermore, the presence of MF has no effect on survival in a multivariate analysis, and we found no benefit of metastasectomy of mesenteric mass or prophylactic surgery on overall survival.
Introduction
Small intestinal neuroendocrine tumours (SI-NETs) are rare neoplasms with an incidence of approximately 0.2–1.2 per 100,000 individuals. However, due to the mostly indolent nature and increasing incidence, their prevalence is rising (Fraenkel et al. 2014).
Neuroendocrine tumours are able to produce and secrete bioactive amines and peptides that induce distinct clinical syndromes. SI-NETs are well known to cause the carcinoid syndrome, characterized by flushing and diarrhoea, via the release of mediators like serotonin (de Herder 2005, Halperin et al. 2017). Furthermore, SI-NETs are associated with fibrosis, most notably mesenteric and right-sided endocardial fibrosis. Although this association is well documented, the pathobiology remains largely elusive (Modlin et al. 2004).
Patients with mesenteric fibrosis (MF) often present with abdominal pain and cachexia by intestinal obstruction, oedema or ischaemia (Makridis et al. 1990). A mesenteric mass with radiating strands of soft-tissue on CT imaging is a pathognomonic feature of MF associated with SI-NET (Fig. 1A) (Pantongrag-Brown et al. 1995). To date, treatment of patients with SI-NETs and complaints due to MF is limited to surgery (Fig. 1B) (Makridis et al. 1990, Modlin et al. 2004). Furthermore, resection of lymph nodes and the primary intestinal tumour seems to increase survival, even in patients with extensive metastatic disease (Landry et al. 2013, Guo et al. 2017).
Appearance of mesenteric fibrosis associated with SI-NETs. (A) CT imaging showing mesenteric mass (asterisk) with typical desmoplastic reaction as radiating strands of soft-tissue. (B) Small bowel resection of the same patient showing mesenteric retraction due to centrally located mass.
Citation: Endocrine-Related Cancer 25, 3; 10.1530/ERC-17-0282
We have conducted the largest retrospective study to investigate the relationships between MF and clinical factors in order to elucidate potential pathogenic mechanisms. Furthermore, we have assessed the survival of patients with MF and the effect of palliative surgery on survival.
Methods
Patients
Medical histories of patients treated for SI-NET between 1993 and 2016 in the ENETS Centre of Excellence for Neuroendocrine Tumours, Erasmus MC and Erasmus MC Cancer Institute, Rotterdam, the Netherlands were analysed. SI-NETs were diagnosed on the basis of a combination of markers, imaging and histology according to current guidelines (Niederle et al. 2016). Patients with NET of unknown primary despite extensive work-up including nuclear imaging by targeting somatostatin receptors were excluded. Furthermore, patients with proven SI-NET were included if there was at least one CT scan available, and they were resident in the Netherlands during follow-up or continued follow-up in Erasmus MC. The disease characteristics, tumour grade, ENETS stage and presence of liver metastases were recorded at the time of diagnosis (Niederle et al. 2016).
Tumour markers
Serum chromogranin A (CgA, REF: <94 μg/L) was determined at the time of diagnosis, or if this was not available at first visit to our centre. Urinary 5-HIAA levels were determined in 24-hour urine samples (REF: <50 μmol/24 h). 5-HIAA levels <10 μmol/24 h were excluded because of probable erroneous sampling. The methods of tumour marker measurement were described previously (Zandee et al. 2016).
Imaging
Radiologic features were assessed by means of contrast-enhanced CT. An enlarged mesenteric lymph node of ≥10 mm on the short axis was considered pathologic. As there is no possibility to radiologically distinguish between an enlarged mesenteric lymph node and mesenteric tumour mass, both were classified as mesenteric mass. MF was defined as radiating soft-tissue strands in the mesentery. Furthermore, CT scans were evaluated for mesenteric infiltration, characterized by a ‘misty’ soft-tissue attenuation and thickening of the small bowel wall.
Statistics
Analyses were performed using SPSS software (version 21 for Windows, SPSS). Data were presented as percentage with count or median and interquartile range (IQR; 25th–75th percentiles). Comparisons were performed for continuous data with unpaired t-test or Mann–Whitney U test as appropriate. For categorical data, the Fisher exact test was carried out. To increase clinical relevancy continuous variables were dichotomized. Cut-off values were obtained by the receiver-operating characteristic (ROC) curve and maximizing the Youden Index. Odds ratios (ORs) of predictive factors were determined using univariate and multivariate logistic regression. Survival curves were generated using Kaplan–Meier method, and the log-rank test was used to compare the survival difference between groups. Hazard ratios (HRs) and 95% confidence intervals (CI) were calculated by Cox regression. In multivariate Cox regression models variables were considered independent if the F-statistic had a probability of less than 0.05. A P value of <0.05 was considered statistically significant, and no corrections were made for multiple testing.
Results
Mesenteric fibrosis
A total of 559 patients with SI-NET were included in this retrospective analysis. Their clinical characteristics are shown in Table 1. Signs of MF on CT imaging were present in 41.4% of patients. As shown in Table 1, the majority of these patients have a mesenteric mass. However, 4.3% had no mesenteric mass of ≥10 mm on the short axis. In these cases, the fibrotic strands radiated from a central node, which itself was smaller than 10 mm.
Baseline clinical and radiologic characteristics.
| All patients (n = 559) | Mesenteric fibrosis (n = 232) | No mesenteric fibrosis (n = 327) | P value | |
|---|---|---|---|---|
| Age of diagnosis | 60.4 (52.1–68.1) | 62.7 (54.7–69.3) | 58.7 (50.6–66.5) | <0.001 |
| Gender | 0.036 | |||
| Male | 53.0% (n = 296) | 58.2% (n = 135) | 49.2% (n = 161) | |
| Female | 47.0% (n = 263) | 41.8% (n = 97) | 50.8% (n = 166) | |
| Tumour grade | 0.306 | |||
| Grade 1 | 48.5% (n = 271) | 44.4% (n = 103) | 51.4% (n = 168) | |
| Grade 2 | 25.9% (n = 145) | 28.4% (n = 66) | 24.2% (n = 79) | |
| Grade 3 | 2.7% (n = 15) | 2.2% (n = 5) | 3.1% (n = 10) | |
| Unknown | 22.9% (n = 128) | 25.0% (n = 58) | 20.8% (n = 68) | |
| ENETS stage IV | 76.0% (n = 425) | 85.8% (n = 199) | 69.1% (n = 226) | <0.001 |
| Liver metastases | 71.0% (n = 397) | 81.9% (n = 190) | 63.3% (n = 207) | <0.001 |
| CgA (μg/L) | 213.0 (91.3–770.3) | 314.5 (125.3–1002.3) | 159.0 (72.5–706.5) | <0.001 |
| 5-HIAA (μmol/24 h) | 124.25 (46.52–457.90) | 184.06 (75.53–595.26) | 75.73 (37.02–314.93) | <0.001 |
| Mechanical ileus in history | 18.1% (n = 101) | 23.3% (n = 54) | 14.4% (n = 47) | 0.008 |
| Presence of mesenteric mass | 65.3% (n = 365) | 95.7% (n = 222) | 43.7% (n = 143) | <0.001 |
| Size largest mesenteric mass (mm) | 29 (22–38) | 32 (24–40) | 25 (19–33) | <0.001 |
| Mesenteric infiltration | 16.8% (n = 94) | 26.3 (n = 61) | 10.1% (n = 33) | <0.001 |
| Small bowel thickening | 7.0% (n = 39) | 13.4% (n = 31) | 2.4% (n = 8) | <0.001 |
Numerical data are median with IQR in brackets. Categorical data are percentages with count in brackets. CgA, serum chromogranin A, normal range <94 μg/L; 5-HIAA, urinary 5-HIAA excretion, normal range <50 μmol/24 h. P value: mesenteric fibrosis vs no mesenteric fibrosis.
To determine potential predictors of MF, we selected clinical factors that significantly differed between patients with and without MF (Table 1). 5-HIAA ≥67 μmol/24 h (AUC 0.64 (95% CI: 0.59–0.68), P < 0.001), CgA ≥121.5 μg/L (AUC 0.61 (95% CI: 0.56–0.66)), mesenteric mass ≥27.5 mm (AUC 0.64 (95% CI: 0.58–0.70)), age of diagnosis ≥55.8 years (AUC 0.61 (95% CI: 0.51–0.66)), gender, ENETS stage IV, liver and mesenteric metastasis were all significant predictors of MF in univariate analyses (Table 2). Mechanical ileus, mesenteric infiltration and small bowel wall thickening were excluded from analysis, as they are generally the result of MF. In a multivariate analysis, independent predictors of MF were 5-HIAA ≥67 μmol/24 h, a mesenteric mass and a mass ≥27.5 mm. Age of diagnosis ≥55.8 years, ENETS stage IV, liver metastases, CgA ≥121.5 μg/L and gender were not independent predictors of MF (Table 2).
Predictive factors of mesenteric fibrosis and mass in patients with SI-NETs (n = 559).
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| Age of diagnosis ≥55.8 years | ||||||
| Mesenteric fibrosis | 2.05 | 1.43–2.95 | <0.001 | 1.43 | 0.89–0.2.31 | 0.142 |
| Mesenteric mass | 2.16 | 1.51–3.10 | <0.001 | 1.93 | 1.28–2.90 | 0.002 |
| Male | ||||||
| Mesenteric fibrosis | 1.44 | 1.02–2.01 | 0.037 | 1.15 | 0.75–1.79 | 0.522 |
| Mesenteric mass | 1.82 | 1.28–2.58 | 0.001 | 1.68 | 1.14–2.48 | 0.009 |
| ENETS stage IV | ||||||
| Mesenteric fibrosis | 2.87 | 1.84–4.48 | <0.001 | 1.19 | 0.42–3.39 | 0.749 |
| Mesenteric mass | 2.63 | 1.76–3.92 | <0.001 | 0.77 | 0.34–1.76 | 0.531 |
| Liver metastases | ||||||
| Mesenteric fibrosis | 2.60 | 1.74–3.89 | <0.001 | 1.10 | 0.41–2.94 | 0.855 |
| Mesenteric mass | 2.64 | 1.81–3.86 | <0.001 | 1.64 | 0.75–357 | 0.216 |
| CgA ≥121.5 μg/L | ||||||
| Mesenteric fibrosis | 2.56 | 1.76–3.73 | <0.001 | 1.12 | 0.65–1.93 | 0.684 |
| Mesenteric mass | 2.33 | 1.62–3.43 | <0.001 | 0.97 | 0.61–1.54 | 0.910 |
| 5-HIAA ≥67 μmol/24 h | ||||||
| Mesenteric fibrosis | 3.28 | 2.21–4.87 | <0.001 | 1.96 | 1.15–3.36 | 0.014 |
| Mesenteric mass | 3.10 | 2.11–4.54 | <0.001 | 2.72 | 1.73–4.28 | <0.001 |
| Mesenteric mass | ||||||
| Mesenteric fibrosis | 28.57 | 14.61–55.84 | <0.001 | 11.49 | 5.58–23.63 | <0.001 |
| Mesenteric mass ≥27.5 mm | ||||||
| Mesenteric fibrosis | 8.94 | 6.01–13.30 | <0.001 | 3.01 | 1.90–4.76 | <0.001 |
CgA, serum chromogranin A, normal range <94 μg/L; 5-HIAA, urinary 5-HIAA excretion, normal range <50 μmol/24 h.
OR, odds ratio; 95% CI, 95% confidence interval.
As the presence of a mesenteric mass was a strong predictor of MF, we also evaluated predictive markers for the presence of a mesenteric mass. Since the optimal cut-offs for continuous factors predicting a mesenteric mass (5-HIAA 62 μmol/24 h, CgA 121.5 μg/L and age of diagnosis 56.7) approximated the values found for MF, the same cut-off values were used. In univariate analyses, significant predictors of a mesenteric mass were age of diagnosis ≥55.8 years, gender, liver metastases, ENETS stage IV, CgA ≥121.5 μg/L and 5-HIAA ≥67 μmol/24 h. Fitting these variables in a multivariate model, only age of diagnosis ≥55.8 years, 5-HIAA ≥67 μmol/24 h and male gender remained independent predictors of a mesenteric mass (Table 2).
Gender
Male patients had a significant higher occurrence of MF than women (Table 1). In accordance, median urinary 5-HIAA excretion is higher (male 142.41 μmol/24 h vs female 96.94 μmol/24 h, P = 0.001) and the presence of a mesenteric mass is more frequent (72% vs 58% P = 0.001). Noteworthy, male patients also less frequently underwent abdominal surgery (male 69% vs female 79%, P = 0.005). Other characteristics, including age of diagnosis and ENETS stage, did not differ significantly between male and female patients (data not shown). As shown in Table 2, gender was an independent predictor of the presence of a mesenteric mass. As the presence and the effect of sex hormones are age dependent, we divided the cohort in 4 equal age categories: ≤52.1, 52.2–60.4, 60.5–68.1, ≥68.1 years. In the first group with age of diagnosis ≤52.1 years, there were 70 female and 70 male patients, while 39% of the female patients vs 64% of the male patients had a mesenteric mass (P = 0.002). This significant difference in presence of a mesenteric mass was not found in other age groups and other characteristics did not significantly differ in this age group between genders (data not shown).
Effect of mesenteric fibrosis on survival
During a median follow-up time of 62.3 months (IQR 32.9–103.4), 208 patients died. Kaplan–Meier analysis demonstrated that patients with MF had a significantly shorter overall survival (P < 0.001) with a median survival of 102 months in the group with MF vs 174 months in the group without MF (Fig. 2). This was in accordance with a five-year survival rate of respectively 71% vs 80%.
Overall survival according to presence of mesenteric fibrosis on CT imaging (n = 559).
Citation: Endocrine-Related Cancer 25, 3; 10.1530/ERC-17-0282
To assess possible prognostic factors, we performed univariate analyses on known prognostic factors, age, CgA, tumour grade and ENETS stage (Niederle et al. 2016), and possible prognostic factors, urinary 5-HIAA excretion, presence of liver metastases (Janson et al. 1997), mesenteric mass, MF and gender. To enhance clinical utility, ROC curve analysis was used to determine optimal cut-offs for 5-HIAA (AUC 0.70 (95% CI 0.66–0.75)) and CgA (AUC 0.72 (95% CI 0.68–0.76)) for survival. This resulted in a cut-off for 5-HIAA of 215 μmol/24 h and for CgA of 310 μg/L. These cut-off values approximated the median value in patients with ENETS stage IV disease, which were 188 μmol/24 h and 298 μg/L, respectively. When assessed in univariate analyses, only gender and ENETS stage were not significant predictors of worse survival (Table 3).
Prognostic factors for overall survival.
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| HR | 95% CI | P value | HR | 95% CI | P value | |
| Age of diagnosis | 1.07 | 1.05–1.08 | <0.001 | 1.07 | 1.05–1.09 | <0.001 |
| Gender | 1.08 | 0.82–1.42 | 0.580 | NS | ||
| Tumour grade | ||||||
| Grade 1 | Reference | Reference | ||||
| Grade 2 | 2.52 | 1.76–3.61 | <0.001 | 2.17 | 1.50–3.14 | <0.001 |
| Grade 3 | 6.05 | 3.15–11.63 | <0.001 | 4.85 | 2.29–10.25 | <0.001 |
| ENETS stage at diagnosis | NS | |||||
| Stage I | Reference | |||||
| Stage II | 0.28 | 0.02–4.53 | 0.372 | |||
| Stage III | 0.52 | 0.07–3.99 | 0.530 | |||
| Stage IV | 2.08 | 0.29–14.90 | 0.465 | |||
| CgA ≥310 μg/L | 2.87 | 2.17–3.79 | <0.001 | 1.90 | 1.28–2.80 | 0.001 |
| 5-HIAA ≥215 μmol/24 h | 2.22 | 1.68–2.95 | <0.001 | 1.50 | 1.02–2.21 | 0.042 |
| Liver metastases | 3.42 | 2.20–5.33 | <0.001 | 2.64 | 1.44–4.85 | 0.002 |
| Mesenteric mass | 1.62 | 1.19–2.20 | 0.002 | 0.73 | 0.45–1.17 | 0.185 |
| Mesenteric fibrosis | 1.78 | 1.35–2.35 | <0.001 | 1.47 | 0.98–2.19 | 0.060 |
NS, non-significant in univariate analysis; CgA, serum chromogranin A normal range <94 μg/L; 5-HIAA, urinary 5-HIAA excretion, normal range <50 μmol/24 h; HR, hazard ratio; 95% CI, 95% confidence interval.
The predictors of worse survival, age of diagnosis, tumour grade and CgA >310 μg/L, 5-HIAA >215 μmol/24 h, MF, liver and mesenteric metastasis, were further assessed in a multivariate model (Table 3). The known factors (age, tumour grade and CgA) were confirmed as independent prognostic markers, as well as liver metastasis and urinary 5-HIAA excretion >215 μmol/24 h. However, MF and a mesenteric mass were not independent prognostic factors for overall survival in this multivariate model (Table 3).
Therapy
As patients with MF had more advanced disease (Table 1), almost all patients were treated with somatostatin analogues (SSAs) (MF 92.2% vs no MF 74.6%, P < 0.001). Since the study was performed in a tertiary referral centre, the majority of patients already received SSA therapy at the first visit to our centre. Therefore, the effect of starting SSAs on 5-HIAA excretion and MF could not be evaluated. Since patients with MF had predominantly ENETS stage IV disease (Table 1), we selected patients with ENETS stage IV (n = 425) to assess the effect of MF on surgical management of SI-NETs. The majority of patients in our cohort underwent abdominal surgery, and the characteristics of the abdominal operations are shown in Table 4. Patients with MF had less often abdominal surgical procedures and underwent more frequent first abdominal surgery >6 months after diagnosis. However, when they did undergo surgery, there was no difference in the number of procedures as compared to patients without MF (range 1–5, P = 0.683).
Characteristics of abdominal surgery in patients with ENETS stage IV.
| All patients (n = 425) | Mesenteric fibrosis (n = 199) | No mesenteric fibrosis (n = 226) | P value | |
|---|---|---|---|---|
| Abdominal surgery | 67.8% (n = 288) | 57.8% (n = 115) | 76.5% (n = 173) | |
| Emergency surgery | 12.5% (n = 53) | 15.1% (n = 30) | 10.2% (n = 23) | <0.001 |
| Elective surgery | 55.3% (n = 235) | 42.7% (n = 85) | 55.4% (n = 150) | |
| Small bowel resection | 19.1% (n = 81) | 17.6% (n = 35) | 20.4% (n = 46) | <0.001 |
| Ileocaecal resection | 13.9% (n = 59) | 8.0% (n = 16) | 19.0% (n = 43) | |
| Right-sided hemicolectomy | 16.7% (n = 71) | 10.6% (n = 21) | 22.1% (n = 50) | |
| Other | 5.6% (n = 24) | 6.5% (n = 13) | 4.8% (n = 11) | |
| No abdominal surgery | 32.2% (n = 137) | 42.2% (n = 84) | 23.5% (n = 53) | <0.001 |
| Resection of mesenteric mass* | 16.7% (n = 50) | 21.6% (n = 41) | 8.2% (n = 9) | 0.001 |
| Surgery <6 months after diagnosis | 56.7% (n = 241) | 42.7% (n = 85) | 69.0% (n = 156) | 0.002 |
| Indication for initial surgery | ||||
| Curative | 16.3% (n = 47) | 11.3% (n = 13) | 19.7% (n = 34) | <0.001 |
| Palliative; symptomatic | 42.4% (n = 122) | 53% (n = 61) | 35.3% (n = 61) | |
| Palliative; prophylactic | 28.5% (n = 82) | 29.6% (n = 34) | 27.7% (n = 48) | |
| Not reported | 12.8% (n = 37) | 6.1% (n = 7) | 17.3% (n = 30) |
P value: mesenteric fibrosis vs no mesenteric fibrosis.
*Percentage of patients with mesenteric mass on first available CT scan and resection afterwards.
To determine factors that influence the likelihood of undergoing surgical treatment, we performed univariate and multivariate analyses of clinically relevant factors in patients with ENETS stage IV disease (Table 5). We were mostly interested in the effect of MF and as patients with MF had in >95% of the cases a mesenteric mass, we excluded this variable from the multivariate model. The multivariate analysis showed that patients with MF were less likely to have abdominal surgery independent of age, sex, tumour grade and the presence of liver metastases.
Predictive factors for undergoing abdominal surgery in patients with SI-NET ENETS stage IV at diagnosis (n = 425).
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| Age | 0.97 | 0.95–0.99 | 0.004 | 0.96 | 0.93–0.98 | 0.002 |
| Male | 0.63 | 0.42–0.96 | 0.029 | 0.57 | 0.34–0.96 | 0.035 |
| Tumour grade | ||||||
| Grade 1 | Reference | Reference | ||||
| Grade 2 | 0.42 | 0.26–0.69 | 0.001 | 0.43 | 0.26–0.72 | 0.001 |
| Grade 3 | 0.42 | 0.13–1.40 | 0.157 | 0.41 | 0.12–1.44 | 0.164 |
| Liver metastases | 0.22 | 0.06–0.73 | 0.013 | 0.14 | 0.10–1.25 | 0.107 |
| Mesenteric mass | 0.15 | 0.08–0.28 | <0.001 | N/A* | ||
| Mesenteric fibrosis | 0.42 | 0.28–0.64 | <0.001 | 0.56 | 0.34–0.93 | 0.025 |
*Excluded from multivariate analysis to avoid collinearity with mesenteric fibrosis. OR, odds ratio; 95% CI, 95% confidence interval.
Effect of surgery on survival
To assess the effect of surgery on survival in patients with SI-NETs, we categorized patients according to indication for surgery, i.e. curative, palliative prophylactic and palliative for symptomatic control. The majority of patients who underwent surgery with curative intent (n = 131) has ENETS stage III disease (51.1%, n = 67); however, 35.1% (n = 46) had already metastasized disease, ENETS stage IV disease at diagnosis. In a small percentage (6.1%, n = 8), the surgery was irradical and another 9.2% (n = 12) had recurrent disease within 1 year. The median disease-free survival was 117 months (95% CI: 83.6–150.4), with five-year disease-free survival of 64.2% and 10 year of 48.1%. Importantly, we found that after 10-year disease-free survival (n = 20), 60% of these patients (n = 12) developed recurrent disease with disease-free time ranging up to 300 months after initial curative surgery. The median overall survival of patients operated with curative intent was 183.5 months (95% CI: 129.1–237.9) with a five-year survival of 87.1%. Conversely to surgery with curative intent, if palliative resection of the primary tumour or metastasectomy in the context of metastasized disease prolongs survival is often debated as a benefit on overall survival has not been shown unequivocally (Niederle et al. 2016). Therefore, we fitted the previous multivariate model of survival (including age, tumour grade, CgA >310 μg/L, 5-HIAA >215 μmol/24 h and the presence of liver metastases) on survival of patients with ENETS stage IV disease (n = 425). In these patients, the presence of liver metastases (HR 1.09 (95% CI: 0.44–2.71), P = 0.858) and 5-HIAA >215 μmol/24 h (HR 1.45 (95% CI: 0.99–2.15), P = 0.060) were no longer significant predictors of survival. Further survival analyses have been fitted on a multivariate model consisting of age of diagnosis, tumour grade and CgA >310 μg/L. To investigate the effect of prophylactic palliative surgery in metastasized disease, we additionally excluded patients who underwent surgery with curative intention or for symptomatic control. The selected 217 patients had a median follow-up time of 57.0 months (IQR 29.0–89.7), during which 90 patients died. The median survival of 99 months (95% CI: 83.9–113.9) for non-operated patients was significantly shorter than the 147 months (95% CI: 122.77–170.63, P = 0.019) for operated patients. This corresponded with five-year survival rates of 67% vs 78%, respectively. However, when added to a multivariate model with the above-mentioned independent predictors (age, tumour grade and CgA >310 μg/L), prophylactic abdominal surgery was no independent prognostic factor (HR 1.34 (95% CI: 0.72–2.49), P = 0.358). Patients who received prophylactic surgery were younger and more often female than non-operated patients (median age of diagnosis 58.9 years vs 63.2 years, P = 0.002, and 54.9% female vs 38.5% male P = 0.02). Furthermore, MF and a mesenteric mass were less frequently present in operated patients (MF in 41% vs 61% in non-operated, P = 0.004, and mesenteric mass in 63% vs 91% in non-operated, P < 0.001). Even so, also in this population, MF and mesenteric mass were not independent predictors of overall survival (HR 1.09, P = 0.76, HR 1.52, P = 0.392, respectively) when corrected for age, tumour grade and CgA.
As delayed surgery can result in developing symptoms, which necessitates surgery with worse outcome, we investigated if undergoing palliative surgery for symptomatic control vs prophylactic palliative surgery is a prognostic factor. In our cohort, 276 patients received palliative surgery, of which 83 died during a median follow-up of 62.7 months (IQR 33.4–106.1). Palliative surgery was prophylactic in 34.4% (n = 95), in 52.2% (n = 144), palliative surgery was performed because of symptoms (abdominal pain in 10.5%, obstruction or ischaemia in 38.8%, and cachexia in 2.2%) and in 13.4% (n = 37) the intent was not clearly reported. Patients undergoing prophylactic palliative surgery had a median survival of 152.2 months (95% CI: 80.8–223.6). This is significantly longer than the median survival of 137.1 months (95% CI: 80.3–193.9, P = 0.012) of non-operated patients. However, when added in a multivariate model to independent predictive factors of survival (age, tumour grade and CgA >310 μg/L), prophylactic palliative surgery was no longer an independent predictor of survival (HR 0.62 (95% CI: 0.35–1.10), P = 0.10).
Finally, to assess the effect of metastasectomy of the mesenteric mass on survival, we selected the patients with a mesenteric mass who underwent surgery after the first available CT scan or those who did not undergo abdominal surgery. Of these 244 patients, 110 (45.1%) underwent abdominal surgery of which 19.1% (n = 21) was performed with curative intent, 37.3% (n = 41) palliative prophylactic and in 43.6% (n = 48) palliative for symptomatic control. The frequency of successful resection of mesenteric mass differed significantly between surgical indication from 90.5% (n = 19) in patients with curative intent to 65.0% (n = 26) in prophylactic and 52.1% (n = 25) in symptomatic palliative surgery (P = 0.009). The median follow-up time of these 244 patients was 50.4 months (IQR 19.1–83.4), during which 91 patients died. The median survival of 81.6 months (95% CI: 43.4–119.8) in patients receiving metastasectomy was not significantly different as compared to 100.2 months (95% CI: 89.84–110.6, P = 0.485) of patients not undergoing metastasectomy. In addition, the number of abdominal surgical procedures did not differ between both groups (range 1–5, P = 0.219). Furthermore, when focusing on patients receiving palliative prophylactic surgery with a mesenteric mass on CT image preoperatively (n = 41), we find no effect of successful resection of mesenteric mass (n = 14) on survival compared to patients with residual mesenteric disease (n = 26) (HR 2.46 95% CI: 0.63–9.67, P = 0.197).
Discussion
Small intestinal neuroendocrine tumours are slow-growing tumours, which are characterized by their ability to secrete bioactive amines and peptides and induce associated syndromes such as carcinoid syndrome, carcinoid heart disease and MF. We have conducted the largest retrospective study so far in 559 patients with a median follow-up time of 62.3 months to assess the effect of MF and palliative surgery on survival. Because of this large patient cohort, we were able to assess predictors of MF and to search for novel insights into the pathogenic mechanisms of mesenteric fibrosis and metastasis.
In our cohort, 41.4% of all patients showed the hallmark of radiating soft-tissue strands which earlier research has shown to be correlated with profound fibrosis in the mesenteric fat (Pantongrag-Brown et al. 1995). Most of the patients with MF had a mesenteric mass and in the small percentage (4.3%) without a mesenteric mass, the fibrotic strands radiated from a central node which itself was smaller than 10 mm on the short axis.
In accordance with other studies, increased urinary 5-HIAA excretion and larger mesenteric mass were independent predictors of MF (Rodriguez Laval et al. 2017). Additionally, in our cohort, increased urinary 5-HIAA excretion was an independent predictor for the presence of a mesenteric mass as well, affirming the above-mentioned link between secretion of biogenic molecules and mesenteric fibrosis and metastasis.
Interestingly, we found that gender was also an independent predictor of a mesenteric mass. As tumour grade, ENETS stage, serum CgA level and the presence of liver metastases did not differ between male and female patients, this seems not to be an effect of tumour aggressiveness or burden. In addition, only in the youngest quartile, patients diagnosed before 52 years of age, the prevalence of a mesenteric mass differed significantly between sexes. In this age group, women had in 39% a mesenteric mass vs 64% of the men. As this age cut-off correlates with the age before menopause in women, it is suggestive of a potential relationship between sex hormones and metastatic patterns. Further research is needed to investigate the role of gender and sex hormones in SI-NETs and the pathogenesis of mesenteric metastasis and fibrosis (Viale et al. 1992, Estrella et al. 2014).
Mesenteric mass and the associated MF results in considerable disease burden in patients with SI-NETs as it can cause small bowel obstruction, ischaemia and perforation. Accordingly, in our study, MF was associated with a decreased five-year survival. However, MF did not remain a significantly negative prognostic factor when corrected for known prognostic factors (Niederle et al. 2016). This suggests that factors such as age, tumour aggressiveness expressed as tumour grade and tumour burden measured as serum CgA mostly determine the prognosis (Kolby et al. 2004, Capdevila et al. 2014, Niederle et al. 2016) and the decreased survival in patients with MF reflects the presence of more advanced disease.
The cornerstone of treatment for patients with abdominal complaints due to mesenteric metastasis and fibrosis remains surgery with bowel resection and guidelines recommend surgery for locally or advanced metastasized SI-NET in the case of symptoms or possible curation (Makridis et al. 1990, Öhrvall et al. 2000, Niederle et al. 2016). In our cohort, the majority of patients have local or distant metastasized disease, also the patients undergoing surgery with curative intent. In these patients, we find consistent with literature a favourable 5-year survival rate of 87.1% (Niederle et al. 2016). However, 51.9% of patients operated with curative intent have recurrent disease after 10 years, highlighting the difficulty of achieving curation in patients with advanced SI-NETs.
Compared to clear guidelines in case of possible curation or abdominal symptoms, the place of prophylactic palliative resection of primary tumour and mesenteric mass in distantly metastasized disease remains a matter of debate as the benefits of surgery on survival remain controversial (Niederle et al. 2016). Based on retrospective analyses, which found survival benefit of palliative resection of primary tumours, early prophylactic surgery is often recommended to avoid complications. However, this benefit could have resulted from biased patient selection, as proper correction for known prognostic markers was often not performed. Moreover, a recent prospective cohort study found no survival benefit of prophylactic surgery (Daskalakis et al. 2017). To investigate the effect on survival of palliative surgery in patients with metastasized SI-NET, we selected patients with ENETS stage IV disease at diagnosis. Five-year survival rates for patients receiving palliative surgery were higher compared to non-operated patients. However, when corrected for age of diagnosis, tumour grade and serum CgA, prophylactic palliative surgery did not result in better survival rates. Also, we found no survival benefit or reduction in number of surgical procedures in patients who received metastasectomy of the largest mesenteric mass. Furthermore, when we assessed survival in a multivariate analysis with known independent prognostic factors, prophylactic palliative surgery in an asymptomatic stage resulted in no survival benefit compared to palliative surgery for symptom control.
However, it is also important to note that although overall patients with advanced SI-NET may not benefit from prophylactic surgery, some patient populations might (Wu et al. 2017). In order to be able to identify patients who might benefit from prophylactic surgery, more insight is needed on the development of MF (Norlen et al. 2012, Niederle et al. 2016). The robust correlation between 5-HIAA excretion and mesenteric metastasis and fibrosis suggests a pathogenic relation (Modlin et al. 2004, Rodriguez Laval et al. 2017). It can therefore be hypothesized that medical therapy such as SSAs and serotonin synthesis inhibitors should aim to fully normalize serotonin production in order to minimize development of MF, although this should be evaluated in prospective studies. Additionally, SI-NETs are generally slow-growing tumours, and it is possible that certain patients without MF or with asymptomatic MF will never develop obstructive or ischaemic complaints. Furthermore, intestinal resection with metastasectomy can result in significant postoperative morbidity due to short-bowel syndrome, adhesions and bile-salt diarrhoea. Therefore, as there seems to be no potential survival benefit of prophylactic surgery in the overall population of patients with advanced SI-NET, we would generally advise a watchful wait-and-see approach and in case of symptoms, growth of tumour or MF to consider palliative surgery in a tailor-made approach.
This study has several limitations. It is a retrospective study and performed in a tertiary referral centre. Patients often already received their first (medical) treatment before referral, making the evaluation of initial serum CgA and urinary 5-HIAA excretion biased. Also, it is possible that a selection of patients is not referred, such as curatively operated patients, or patients with rapidly progressive disease and poor clinical condition. Despite efforts to adjust for known prognostic factors, a biased patient selection can still occur. Due to conflicting findings on the effect of palliative surgery for SI-NETs, this study emphasizes again the need for randomized controlled trials.
Conclusion
We have confirmed known predictors of MF in patients with SI-NETs and presented that these are congruent with the predictors of a mesenteric mass. We also found that in patients aged ≤52 years, female gender is associated with lower incidence of a mesenteric mass. However, in our cohort, MF is not an independent prognostic factor for overall survival. In addition, we found no general benefit of palliative resection of mesenteric mass or prophylactic surgery on survival.
Declaration of interest
A Blažević, W T Zandee, G J H Franssen, M F van Velthuysen, L J Hofland, R A Feelders: no disclosures to report. J Hofland has received speaker and travel fees from Ipsen and Novartis. W W de Herder has joined advisory boards for Ipsen and Novartis.
Funding
Study was funded by an unrestricted fund of IPSEN.
References
Capdevila J, Meeker A, Garcia-Carbonero R, Pietras K, Astudillo A, Casanovas O & Scarpa A 2014 Molecular biology of neuroendocrine tumors: from pathways to biomarkers and targets. Cancer and Metastasis Reviews 33 345–351. (https://doi.org/10.1007/s10555-013-9468-y)
Daskalakis K, Karakatsanis A, Hessman O, Stuart HC, Welin S, Tiensuu Janson E, Oberg K, Hellman P, Norlen O & Stalberg P 2017 Association of a prophylactic surgical approach to stage IV small intestinal neuroendocrine tumors with survival. JAMA Oncology [epub]. (https://doi.org/10.1001/jamaoncol.2017.3326)
de Herder WW 2005 Tumours of the midgut (jejunum, ileum and ascending colon, including carcinoid syndrome). Best Practice and Research Clinical Gastroenterology 19 705–715. (https://doi.org/10.1016/j.bpg.2005.05.007)
Estrella JS, Ma LT, Milton DR, Yao JC, Wang H, Rashid A & Broaddus RR 2014 Expression of estrogen-induced genes and estrogen receptor beta in pancreatic neuroendocrine tumors: implications for targeted therapy. Pancreas 43 996–1002. (https://doi.org/10.1097/MPA.0000000000000203)
Fraenkel M, Kim M, Faggiano A, de Herder WW, Valk GD & Knowledge N 2014 Incidence of gastroenteropancreatic neuroendocrine tumours: a systematic review of the literature. Endocrine-Related Cancer 21 R153–R163. (https://doi.org/10.1530/ERC-13-0125)
Guo J, Zhang Q, Bi X, Zhou J, Li Z, Huang Z, Zhang Y, Li M, Chen X, Hu X, et al. 2017 Systematic review of resecting primary tumor in MNETs patients with unresectable liver metastases. Oncotarget 8 17396–17405. (https://doi.org/10.18632/oncotarget.14156)
Halperin DM, Shen C, Dasari A, Xu Y, Chu Y, Zhou S, Shih YT & Yao JC 2017 Frequency of carcinoid syndrome at neuroendocrine tumour diagnosis: a population-based study. Lancet Oncology 18 525–534. (https://doi.org/10.1016/S1470-2045(17)30110-9)
Janson ET, Holmberg L, Stridsberg M, Eriksson B, Theodorsson E, Wilander E & Oberg K 1997 Carcinoid tumors: analysis of prognostic factors and survival in 301 patients from a referral center. Annals of Oncology 8 685–690. (https://doi.org/10.1023/A:1008215730767)
Kolby L, Bernhardt P, Sward C, Johanson V, Ahlman H, Forssell-Aronsson E, Stridsberg M, Wangberg B & Nilsson O 2004 Chromogranin A as a determinant of midgut carcinoid tumour volume. Regulatory Peptides 120 269–273. (https://doi.org/10.1016/j.regpep.2004.03.017)
Landry CS, Lin HY, Phan A, Charnsangavej C, Abdalla EK, Aloia T, Nicolas Vauthey J, Katz MH, Yao JC & Fleming JB 2013 Resection of at-risk mesenteric lymph nodes is associated with improved survival in patients with small bowel neuroendocrine tumors. World Journal of Surgery 37 1695–1700. (https://doi.org/10.1007/s00268-013-1918-8)
Makridis C, Oberg K, Juhlin C, Rastad J, Johansson H, Lorelius LE & Akerstrom G 1990 Surgical treatment of mid-gut carcinoid tumors. World Journal of Surgery 14 377–383; discussion 384–375. (https://doi.org/10.1007/BF01658532)
Modlin IM, Shapiro MD & Kidd M 2004 Carcinoid tumors and fibrosis: an association with no explanation. American Journal of Gastroenterology 99 2466–2478. (https://doi.org/10.1111/j.1572-0241.2004.40507.x)
Niederle B, Pape UF, Costa F, Gross D, Kelestimur F, Knigge U, Oberg K, Pavel M, Perren A, Toumpanakis C, et al. 2016 ENETS Consensus Guidelines update for neuroendocrine neoplasms of the jejunum and ileum. Neuroendocrinology 103 125–138. (https://doi.org/10.1159/000443170)
Norlen O, Stalberg P, Oberg K, Eriksson J, Hedberg J, Hessman O, Janson ET, Hellman P & Akerstrom G 2012 Long-term results of surgery for small intestinal neuroendocrine tumors at a tertiary referral center. World Journal of Surgery 36 1419–1431. (https://doi.org/10.1007/s00268-011-1296-z)
Öhrvall U, Eriksson B, Juhlin C, Karacagil S, Rastad J, Hellman P & Åkerström G 2000 Method for dissection of mesenteric metastases in mid-gut carcinoid tumors. World Journal of Surgery 24 1402–1408. (https://doi.org/10.1007/s002680010232)
Pantongrag-Brown L, Buetow PC, Carr NJ, Lichtenstein JE & Buck JL 1995 Calcification and fibrosis in mesenteric carcinoid tumor: CT findings and pathologic correlation. American Journal of Roentgenology 164 387–391. (https://doi.org/10.2214/ajr.164.2.7839976)
Rodriguez Laval V, Pavel M, Steffen IG, Baur AD, Dilz LM, Fischer C, Detjen K, Prasad V, Pascher A, Geisel D, et al. 2017 Mesenteric fibrosis in midgut neuroendocrine tumors: functionality and radiological features. Neuroendocrinology [epub]. (https://doi.org/10.1159/000474941)
Viale G, Doglioni C, Gambacorta M, Zamboni G, Coggi G & Bordi C 1992 Progesterone receptor immunoreactivity in pancreatic endocrine tumors. An immunocytochemical study of 156 neuroendocrine tumors of the pancreas, gastrointestinal and respiratory tracts, and skin. Cancer 70 2268–2277. (https://doi.org/10.1002/1097-0142(19921101)70:9<2268::aid-cncr2820700910>3.0.co;2-x)
Wu L, Fu J, Wan L, Pan J, Lai S, Zhong J, Chung DC & Wang L 2017 Survival outcomes and surgical intervention of small intestinal neuroendocrine tumors: a population based retrospective study. Oncotarget 8 4935–4947. (https://doi.org/10.18632/oncotarget.13632)
Zandee WT, Kamp K, van Adrichem RC, Feelders RA & de Herder WW 2016 Limited value for urinary 5-HIAA excretion as prognostic marker in gastrointestinal neuroendocrine tumours. European Journal of Endocrinology 175 361–366. (https://doi.org/10.1530/EJE-16-0392)
