2024 Volume 71 Issue 8 Pages 809-815
When a neuroendocrine tumor with abundant blood flow is located in the pancreatic tail, it is difficult to distinguish it from accessory spleen. The patient was a 71-year-old woman who was admitted with impaired consciousness and hypoglycemia, raising suspicion of insulinoma. The selective arterial calcium injection test suggested a lesion in the pancreatic tail. Contrast-enhanced computed tomography and magnetic resonance imaging (MRI) showed a mass in the splenic hilum; however, its continuity with the pancreas was unclear. Contrast-enhanced MRI using super paramagnetic iron oxide (SPIO) showed no SPIO uptake in the splenic hilar mass. SPIO contrast-enhanced MRI is considered useful for differentiating pancreatic endocrine tumors from paraspleen tumors.
Neuroendocrine tumors (NETs) are rare tumors in which somatostatin receptor (SSTR) expression is higher in the cell membrane than in the normal pancreatic tissue [1]. Some organs, such as the spleen, adrenal gland, and liver, physiologically express SSTR and may show imaging characteristics similar to those of somatostatin receptor scintigraphy (SRS).
In contrast, the parasympathetic spleen, present in approximately 10% of the population [2], is a mass of splenic tissue resulting from defective fusion of the splenic primordia and is often observed near the splenic hilum or pancreatic tail. In particular, when an accessory spleen is present in the pancreatic tail, it is difficult to differentiate from pancreatic NET, and surgery is often performed with the diagnosis of a pancreatic tumor, only being confirmed as an accessory spleen after surgery [3].
Super Paramagnetic Iron Oxide (SPIO) is a liver-specific contrast agent that is specifically taken up by reticuloendothelial cells (Kupffer cells), mainly in the liver and spleen. Therefore, SPIO contrast-enhanced magnetic resonance imaging (MRI) is considered superior for the detection of metastatic liver tumors.
In this study, we encountered a case of insulinoma in which SPIO contrast-enhanced MRI was useful in differentiating a splenic hilar mass with unknown continuity with the pancreas from an accessory spleen.
Patient: 71 years old, female
Chief complaint: impaired consciousness
History: Cataract (52 years old), hepatitis C (58 years old), latent hypothyroidism (60 years old)
Family history: None noted
Life history: No smoking, no drinking
Current medical history: At the age of 67, her daughter pointed out that she was forgetful, and from the age of 69, her symptoms worsened, and she became unable to plan menus for meals. The patient was as usual until she went to bed the day before coming to our hospital, but from 7:00 on the morning of the day of her arrival, she responded to her husband’s calls with unintelligible responses other than “yes,” and was brought to our hospital as an emergency.
The patient was 155 cm tall, weighed 50 kg, and had a BMI of 20.8 kg/m2. Her consciousness was based on the Glasgow Coma Scale E3V3M5, and she was actively moving her limbs with her eyes closed while trying to get up. Body temperature was 37.1°C, blood pressure was 117/59 mmHg, heart rate was 94/min, and transcutaneous oxygen saturation was 99% (room air). No ocular deviations or obvious paralysis were observed.
Blood and CSF findings (Table 1a): The fasting blood glucose level was 111 mg/dL, indicating no hypoglycemia, and the HbA1c level was 4.1%, which was low. There were no obvious electrolyte abnormalities and thyroid function was within the reference range. Cerebrospinal fluid (CSF) findings showed elevated erythrocyte and protein levels; however, the CSF was colorless and without xanthochromy, suggesting peripheral blood contamination.
Laboratory data at admission
| Hematology | |||
| WBC | 7,100/μL | LDL-cho | 94 mg/dL |
| RBC | 4.45 × 106/μL | HDL-cho | 88 mg/dL |
| Hb | 13.5 g/dL | TG | 58 mg/dL |
| Plt | 176 × 103/μL | HbA1c | 4.1% |
| Biochemistry | CSF examination | ||
| TP | 7.3 g/dL | color tone | colorless |
| Alb | 4.7 g/dL | appearance | slight turbidity |
| CK | 370 U/L | xanthochromy | (–) |
| AST | 23 IU/L | cell count | 3/mm3 |
| ALT | 12 IU/L | RBC | 1,000/mm3 |
| LDH | 256 U/L | TP | 41 mg/dL |
| glucose | 111 mg/dL | glucose | 64 mg/dL |
| UA | 5.0 mg/dL | ||
| BUN | 18.6 mg/dL | ||
| Cre | 0.89 mg/dL | ||
| Na | 141 mEq/L | ||
| K | 4.1 mEq/L | ||
| Cl | 103 mEq/L | ||
| Ca | 9.3 mg/dL | ||
Endocrinological data
| (Day 2) | Morning fasting | (Day 3) | Morning fasting |
| glucose | 13 mg/dL | GH | 6.35 ng/mL |
| IRI | 2 μU/mL | Somatomedin C | 117 (56–172) ng/mL |
| TSH | 3.928 μU/L | LH | 28.9 mIU/mL |
| FT4 | 1.24 ng/dL | FSH | 44.6 mIU/mL |
| FT3 | 2.49 pg/mL | estradiol | ≤10 pg/mL |
| anti-TPO antibody | <9 IU/mL | ACTH | 25.8 pg/mL |
| anti-thyroglobulin antibody | 38 U/mL | Cortisol | 11.2 μg/dL |
| Prolactin | 12.9 ng/mL | ||
| (Day 3) | during hypoglycemia | glucose | 117 mg/dL |
| glucose | 65 mg/dL | C-peptide | 6.3 ng/mL |
| IRI | 6 μU/mL | anti insulin body | <0.4 U/mL |
| Gastrin | 90 pg/mL | ||
| Glucagon | 8.3 pg/mL | (Day 6) | |
| intact-PTH | 19 pg/mL |
Clinical course: The patient was admitted to the neurology department for suspected nonconvulsive seizures. After sedation, she was able to move both upper limbs without differences in GCS E2V2M4. On the second day (after 20 h of continuous infusion without glucose), there was no obvious improvement in her level of consciousness, even after the discontinuation of propofol. Blood tests showed that her serum blood glucose level was 13 mg/dL, and her level of consciousness improved to Glasgow Coma Scale E4V5M6 after administration of 50% intravenous glucose. The patient was considered to have impaired consciousness due to hypoglycemia and was transferred to our department for further investigation. At the time of hypoglycemia, serum insulin was 2 μU/mL and IRI/PG (immunoreactive insulin/plasma glucose, Fajans index [4]) was 0.15 < 0.3. There was no obvious decrease in anterior pituitary hormone levels (Table 1b). A central venous catheter was placed on the third day and high-calorie supplemental fluid (17.5% glucose concentration) was administered to maintain blood glucose levels between 60 and 130 mg/dL. The IRI/PG remained below 0.3 in the range of 0.09–0.24, but Whipple’s triad was observed and serum insulin was within the sensitivity of measurement when blood glucose was below 45 mg/dL, suggesting insulinoma as a possible diagnosis [5]. A 72-hour fasting test was not performed because of poor nutritional status.
A 75-g OGTT showed no reactive hypoglycemia (Table 2a). In the C-peptide suppression test, rapid-acting insulin was administered at a dose of 0.1 U/kg/h (4 units of humulin R + 10 mL of fresh food) by continuous IV administration [6], and blood glucose and CPR were measured before and every 10 minutes after the start of the test. C-peptide was not suppressed to less than 50% of the previous value and was judged positive (Table 2b).
75g OGTT
| time (min) | 0 | 30 | 60 | 90 | 120 | 180 | 240 | 300 |
|---|---|---|---|---|---|---|---|---|
| PG (mg/dL) | 110 | 179 | 183 | 181 | 151 | |||
| SMBG (mg/dL) | 98 | 184 | 187 | 177 | 140 | 183 | 110 | 124 |
| IRI (μU/mL) | 12 | 164 | 122 | 127 | 78 | |||
| CPR (ng/mL) | 2.9 | 9.4 | 10.3 | 10.6 | 9.1 |
C-peptide suppression test
| time (min) | –30 | 0 | 30 | 60 |
|---|---|---|---|---|
| PG (mg/dL) | 61 | 56 | 41 | 29 |
| CPR (ng/mL) | 2.2 | 2.0 | 1.9 | 1.3 |
Computed tomography (CT) scan of the head: There was no obvious bleeding.
MRI of the head: No acute infarcted lesions, chronic ischemic changes, or stenosis of the right internal carotid artery were noted.
Full abdominal CT: A mass measuring approximately 14 mm was observed in the splenic hilum. The mass was darkly stained from the early stage and was slightly more strongly and uniformly enhanced than the pancreatic parenchyma, even in the equilibrium phase. The density of the mass in the arterial and equilibrium phases was similar to that of the spleen, suggesting the presence of an accessory spleen. However, the mass was stained darkly in the arterial phase and was uniform in the interior; therefore, continuity with the pancreas was unclear (Fig. 1A, B).
Splenic hilar mass of 14 mm darkly stained from the early stage and is mildly, more strongly, and uniformly enhanced than the pancreatic parenchyma, even in the equilibrium phase (arrow).
A. Full abdominal contrast-enhanced CT (arterial phase)
B. Full abdominal contrast-enhanced CT (equilibrium phase)
The splenic hilar mass showed mild signal reduction compared to the spleen on sshT2W1 and mild signal enhancement in T1W1 opposed phase (arrow).
C. Contrast-enhanced MRI of the upper abdomen (sshT2W1)
D. Contrast-enhanced MRI of the upper abdomen (T1W1 opposed phase)
Hotspot consistent with a splenic hilar mass. No obvious liver metastasis noted (arrow).
E, F. Octreotide scan
No obvious SPIO uptake in the splenic hilar mass (arrow).
G. SPIO contrast-enhanced MRI (T2*WI before contrast)
H. SPIO contrast-enhanced MRI (T2*WI after contrast)
Contrast-enhanced MRI of the upper abdomen: The splenic hilar mass showed low signal intensity on T1WI and high signal intensity on T2WI compared with the pancreatic parenchyma and was stained dark in the early contrast phase, showing a contrast pattern similar to that of the spleen (Fig. 1C, D).
Octreotide scan: A hotspot consistent with a splenic hilar mass was observed (Fig. 1E, F).
Based on the results of a full abdominal contrast CT scan, upper abdominal MRI, and octreotide scan, we suspected the possibility of an accessory spleen, an intrapancreatic accessory spleen, or an extrapancreatic NET; however, it was difficult to differentiate between them, and we decided to add a SPIO contrast-enhanced MRI scan to distinguish if it was an accessory spleen or not.
SPIO-enhanced MRI: No obvious SPIO uptake was observed in the splenic hilar mass on T2*WI, and the parasympathetic spleen was negative, suggesting the possibility of a NET in the pancreatic tail (Fig. 1G, H).
Endoscopic Ultrasound-Fine Needle Aspiration (EUS-FNA): An endoscopic FNA was performed on the splenic hilar mass. Immunostaining revealed chromogranin A(+), synaptophysin(+), CD56(+), an MIB1 index of 0.6%, and insulin(+), and insulinoma was diagnosed.
In the selective arterial calcium injection test (hereinafter referred to as SACI test), calcium gluconate hydrate was administered at a dose of 0.025 mEq/kg per site. The dorsal pancreatic artery showed an increase of more than 300 U/mL, whereas the intrinsic hepatic, splenic, and inferior pancreaticoduodenal arteries showed no significant increase (Fig. 2).
SACI test results. Dorsal pancreatic artery stimulation increased insulin levels.
Based on these findings, the mass in the splenic hilum was diagnosed as a pancreatic tail insulinoma of pancreatic origin, and laparoscopic pancreaticoduodenectomy, splenectomy, and cholecystectomy were performed on the 52nd day. Surgical findings revealed a tumor protruding toward the anorectal side of the pancreatic tail near the spleen, which was thought to be a primary pancreatic tumor, although fatty tissue and blood vessels were present between the pancreas and the tumor. Pathological results showed chromogranin A(+), synaptophysin(+), CD56(+), an MIB1 index of 3.8%, insulin(+), NET G2 (fission image 0/50HPF), and pT1N0M0 (Fig. 3). Postoperatively, her preprandial blood glucose remained at 80–110 mg/dL. Diffuse NET was considered based on Algorithm 18 of the Guidelines for the Clinical Practice of Neuroendocrine Tumors of the Pancreas and Gastrointestinal Tract (NEN), a diffuse NET was considered.
Immunohistochemical staining
A. Chromogranin staining (×400)
B. Insulin staining (×400)
C. HE staining (×100)
Insulinomas are the most common functional NETs derived from pancreatic islet β-cells, but they are very rare, with 1–4 new cases per million people per year, and the age of onset is predominantly in the 50s, with a slight female predominance. Ninety percent of insulinomas are benign, 90% are solitary, >90% occur within the pancreas, and 90% are less than 2 cm in size [7-10]. The psychiatric symptoms of hypoglycemia include confusion, behavioral changes, personality changes, and visual disturbances. Symptoms may be misinterpreted as psychiatric, cardiac, or neurological diseases before a tumor is detected [11]. In the present case, one year before diagnosis, a head MRI Voxel-based Specific Regional analysis system for Alzheimer’s disease (VSRAD) showed mild atrophy with a VOI atrophy of 1.31, but the patient had disorientation, memory impairment, and decreased awareness of her disease. The patient was diagnosed with Alzheimer’s disease. The patient had a preoperative Mini-Mental State Examination (MMSE) score of 17/30, which improved to 24 six months after surgery, and her cognitive decline could have been due to hypoglycemia.
Hypoglycemia can promote dementia. Previous studies have described cases of improved or residual cognitive function after insulinoma surgery [12-14]. A previous study reported that lower initial blood glucose and higher levels of insulin and C-peptide were related to worse tumor grading, during a 72-h-fast test, in patients with insulinoma [15]. However, the relationship between insulin sensitivity and cognitive function in patients with insulinoma has not been reported. In addition, it has been reported that the lowest blood glucose value and longest period of hypoglycemia were not associated with cognitive decline in patients with insulinoma [12]. Therefore, a more detailed investigation is required to reveal the relationship between insulin sensitivity, prolonged hypoglycemia, and cognitive function, in patients with insulinoma.
The diagnosis is based on Whipple’s triad and confirms that serum insulin is within the sensitivity of measurement, even when blood glucose is less than 45 mg/dL [5]. In this case, the IRI/PG never exceeded 0.3. However, their sensitivity and specificity are not high [16].
Localization of insulinomas is important because it avoids blind resection, allows for more rapid surgery, and reduces the associated mortality [17]. CT, MRI, and EUS were used for the localization diagnosis. The disease is characterized by hypervascularity, typically with marked dark staining in the parenchymal phase of the dynamic study pancreas [18], it is difficult to determine the localization only by imaging studies, and the SACI test is useful.
On the other hand, when a hypervascularized mass is found in the pancreatic tail, differentiation from the intrapancreatic or splenic hilar adnexa is problematic; in this case, contrast CT, contrast MRI, and octreotide scans were not able to differentiate the mass from the adnexa. When contrast-enhanced CT and MRI could not differentiate the two, 68Ga-DOTATATE PET/CT, 99mTc-impaired red blood cell scintigraphy (99mTc-HDRBC SPECT/CT), and 99mTc hepatosplenic tin colloid scintigraphy have been reported to be useful for differentiating the two [19]. However, because 68Ga-DOTATATE PET/CT functions by binding to SSTR2, it is difficult to distinguish it from the paraspleen of the pancreatic tail. 99mTc-HDRBC accumulates in the splenic tissue based on the uptake of denatured red blood cells by the spleen to remove abnormal red blood cells, but it can be a false negative when the lesions are small, and respiratory-dependent excretion may result in false-negative results [19]. 99mTc hepatosplenic tin colloid scintigraphy, such as 99mTc-HDRBC, is difficult to contrast with the surrounding tissues when the hypersplenic region itself is small or when the intrasplenic reticular system in the hypersplenic region has poor function. Even a hypersplenic region as small as 10 mm may not be separated from the spleen, which is a disadvantage compared with MRI. SPIO has the disadvantage of inferior spatial and tissue resolution compared to MRI.
SPIO is a liver-specific contrast agent, and after administration, 80% is taken up by Kupffer cells in the liver, 8–9% by endoreticular cells in the spleen, and the remainder by the bone marrow and lymph nodes [20]. The accumulated SPIO disrupts the local magnetic field and causes strong T1 and T2 shortening effects, resulting in a decrease in the T2 signal. In intrapancreatic accessory spleens, similar to the spleen, a decrease in the T2WI signal was observed after SPIO administration. SPIO contrast-enhanced MRI is a radiation-free imaging method compared with nuclear medicine and is useful for differentiating small tumors with high spatial and tissue resolutions.
In many cases in which intrapancreatic subsplenomegaly was difficult to differentiate by contrast CT or MRI were pancreatic tail lesions, although there have been reports of cases in which subsplenomegaly was finally diagnosed by SPIO contrast-enhanced MRI [21], there have been no reports of its usefulness in differentiating between pancreatic NETs and subsplenomegaly.
Localization of an insulinoma by EUS is often diagnosed when the tumor is located in the head or body of the pancreas [22, 23].
The sensitivity of EUS is reported to be 92.6% for the pancreatic head, 78.9% for the pancreatic body, and 40% for the pancreatic tail; however, the accuracy of EUS localization is low for insulinomas with many lesions in the pancreatic tail [24]. Therefore, the present case was not a case in which EUS localization was useful in differentiating between the adnexal spleen and pancreatic tail. Therefore, SPIO contrast-enhanced MRI may be useful for differentiating cases in which an accessory spleen or pancreatic tail mass is suspected, as in the present case [7].
However, when the mass is a pancreatic epidermoid cyst originating from the intrasplenic portion of the pancreas, caution should be exercised because it is difficult to distinguish it from an adenosplenic mass, even with SPIO contrast-enhanced MRI, in cases in which an epithelial cyst (cystic component) and an intrasplenic component (substantial component) are seen, and the substantial component becomes thin and capsular.
In conclusion, SPIO contrast-enhanced MRI is useful for differentiating splenic hilar masses, whose continuity with the pancreas remains unclear.
Endocrine tumors of the pancreas and splenic hilar adenosplenomegaly are sometimes difficult to differentiate, and noninvasive SPIO contrast-enhanced MRI is useful for differentiating them.
We thank Dr. Satoru Yasukawa, Department of Pathology, Kyoto Second Red Cross Hospital, for his assistance with pathological diagnosis.
All authors have no conflicts of interest to disclose regarding this paper.
Informed consent was obtained.
