Extrahepatic ⁹⁰Y Complication; Gastric Ulcer Secondary to ⁹⁰Y Therapy for Liver Metastasis Despite Negative Preprocedural Imaging

Abstract

Transarterial radioembolization using yttrium-90 (⁹⁰Y) therapy has become a standard modality of treatment for primary and metastatic liver malignancies due to its high efficacy rate and relatively low risk of adverse effects compared to other forms of locoregional and systemic therapies. Non-target distribution of radio embolic beads and adjacent structure radiation are the two most common adverse effects. However, these are rarely encountered due to thorough imaging and mapping studies prior to ⁹⁰Y therapy. We present the case of a 66-year-old male who developed a radiation-induced gastric ulcer following ⁹⁰Y therapy with negative pre-procedural imaging and mapping who was retrospectively found to have an accessory artery from the left hepatic artery to the gastric antrum.

Introduction

Transarterial radioembolization (TARE) is an established therapy option for some patients with unresectable primary liver tumors or hepatic metastases. TARE is performed by injecting glass or resin-coated microspheres coated with yttrium-90 (⁹⁰Y) into the hepatic vasculature. The antitumoral effect of this treatment is due to the radiation emitted by ⁹⁰Y. Given the targeted approach, TARE has been shown to reduce the incidence of grade 3 or higher adverse events, cause less post-procedure abdominal pain, and reduce the length of stay compared to some other liver-directed and systemic therapies [1–3].

The risk of adverse events following TARE is generally low. Nontarget distribution (NTD) can occur in the setting of anatomical variants, collateral circulation, or microsphere reflux [1]. For this reason, mapping angiography using ^99mTc-macroaggregated albumin (^99mTc-MAA) should be performed prior to TARE therapy to detect any extrahepatic shunting to the lungs or gastrointestinal tract. A repeat mapping study is necessary if ⁹⁰Y treatment is to occur greater than 6 months after the initial mapping, such as in cases of retreatment [4]. One study demonstrated that for patients who underwent ^99mTc-MAA mapping with the use of single proton emission computerized tomography (SPECT/CT), the therapy plan was changed in 29% of cases [5]. If the patient is found to have a high lung shunt fraction of ^99mTc-MAA, TARE cannot be performed due to the risk of radiation-induced pneumonitis/radiation-induced lung disease (RILD). Nontarget GI distribution of radioembolization can lead to mucosal thinning, gastroenteritis, mucositis, and ulceration with possible progression to perforation [6]. If necessary, prophylactic embolization of extrahepatic vessels can be performed in order to avoid gastrointestinal embolization of ⁹⁰Y radioembolization at the time of treatment. Due to this pre-procedural planning and improvement in radioembolization techniques, the overall incidence of NTD and its resulting complications are low. The incidence of RILD is < 1% across prospective and retrospective studies [6]. The incidence of nontarget GI radioembolization secondary to TARE is also rare [7, 8].

We present a case of a 66-year-old man who underwent ⁹⁰Y TARE therapy for metastatic pancreatic neuroendocrine tumor (pNET) in the liver. On pre-procedural ^99mTc-MAA mapping, there was no evidence of gastric collateral uptake of the radiotracer. On post-procedural SPECT/CT, the patient was found to have extrahepatic ⁹⁰Y at the lesser curvature of the stomach, and in retrospect had a small accessory vessel from the left hepatic artery which supplied areas of the lesser curvature of the gastric antrum.

Case Report

We present the case of a 66-year-old man who presents for TARE ⁹⁰Y therapy for metastatic pancreatic neuroendocrine tumor (pNET) to the liver. The patient was initially diagnosed with grade 2 pancreatic gastrinoma with multiple liver metastasis after a hospitalization related to nausea, vomiting, GERD, and esophageal perforation. Prior to ⁹⁰Y therapy, ^99mTc-MAA mapping procedure was performed first with MAA injection into the right hepatic artery, then into the left hepatic artery. A limited intraprocedural CTA was performed as part of the mapping procedure. This was performed and interpreted by an interventional radiologist with 6 years of experience. The hepatic artery perfusion study SPECT/CT that was subsequently performed demonstrated satisfactory hepatic perfusion without evidence of significant extrahepatic perfusion (Fig. 1). The hepatic artery perfusion study with SPECT/CT was interpreted by a nuclear medicine physician with over 15 years of experience.

Fig. 1.

Preprocedural 99mTC-MAA mapping SPECT/CT demonstrates hepatic perfusion without extrahepatic perfusion

One month later, the patient received ⁹⁰Y TheraSphere selective radiation therapy of the right hepatic lobe. SPECT/CT following the procedure showed delivery of radio-emboli to the right lobe and no evidence of shunting to the lungs or visceral organs. The patient tolerated the therapy well despite experiencing mild fatigue for a week following the procedure.

The following month, approximately 2 months after the initial mapping study, the patient presented for left hepatic lobe ⁹⁰Y therapy. Selective administration of ⁹⁰Y into the left hepatic lobe via the left hepatic artery for segments 2, 3, and 4 was performed. Hybrid angio-CT was performed as part of the procedure. Cone beam CT was not performed as it is not considered routine for ⁹⁰Y therapy at our institution. Subsequent SPECT/CT demonstrated radio-emboli not only in the left hepatic lobe but also in the lesser curvature of the gastric antrum (Fig. 2). At this time, retrospective review by interventional radiology of a prior CT angiography (CTA) demonstrated a tiny accessory artery extending from the left hepatic artery to the lesser curvature of the stomach (Fig. 3). In addition, retrospective review of the same day intraoperative fluoroscopic images also revealed the accessory artery supplying the lesser curvature of the stomach (Fig. 4).

Fig. 2.

Post-procedural 90Y SPECT/CT of the upper abdomen demonstrates radio-emboli administration to the left hepatic lobe and lesser curvature of the gastric antrum

Fig. 3.

Interventional radiology suit CT images demonstrate a small accessory artery extending from the left hepatic artery to the lesser curvature of the stomach

Fig. 4.

Intraoperative fluoroscopic image demonstrates a tiny artery extending from the left hepatic artery towards the lesser curvature of the stomach

Two days later, a previously scheduled esophago-gastro-duodenoscopy (EGD) was performed to evaluate the patient’s concurrent esophageal stricture. In addition to esophageal stenoses, the EGD revealed a non-bleeding superficial gastric ulcer in the lesser curvature of the gastric antrum (Fig. 5). The lesion was 30 mm in the largest dimension. Biopsies were obtained and pathology was consistent with erosive chemical gastropathy. The histology was negative for evidence of Helicobacter pylori infection and the patient was not using non-steroidal anti-inflammatory drugs (NSAIDs). Given this new finding compared to an EGD several months before ⁹⁰Y therapy, the pathology results, and the evidence of radio-emboli in the gastric antrum via SPECT/CT, this finding most likely represents a radiation-induced ulcer.

Fig. 5.

Photograph from the first EGD 2 days after radioembolization demonstrates a superficial gastric ulcer along the lesser curvature of the gastric antrum

Over the next several weeks following the procedure, the patient experienced worsening nausea, vomiting, weight loss, and poor appetite, which led to his readmission approximately 2 weeks after the left hepatic lobe ⁹⁰Y therapy. At that time a repeat EGD was performed which showed that the previously seen ulcer had cratered and was now friable with necrotic appearing edges but was non-bleeding (Fig. 6). After stabilizing, the patient was discharged with proton pump inhibitors, sucralfate, and instructions to follow up in 2 months.

Fig. 6.

Photograph from the second EGD 2 weeks after radioembolization demonstrates cratering of the ulcer, and friability with necrotic appearing, non-bleeding edges

Discussion

Administration of ⁹⁰Y therapy is done with either resin-based beads (SIR-sphere) or glass-based beads (TheraSphere). The glass-based TheraSphere is associated with less adverse events, however, local and regional complications still occur. Local complications include cholecystitis (1.5%), hepatic encephalopathy (3.9%), and hepatic failure (6.9%) [9]. Extrahepatic complications happen in two ways; radiation to adjacent structures and NTD. The risk of adjacent structures being affected is low given the mean and maximum path lengths of ⁹⁰Y therapy to be just 2.5 and 10 mm, respectively [10]. Adjacent structures, such as the bowel, diaphragm, and stomach can be affected by the formation of adhesions; however, this rarely creates a clinically significant issue unless the patient requires resection surgery later in life [11, 12]. In a study evaluating ⁹⁰Y radioembolization of left hepatic tumors within 1 cm of the stomach (n = 97), less than half of patients reported any abdominal pain, and no gastrointestinal ulceration was reported in any patients [13].

Gastric ulceration secondary to ⁹⁰Y therapy is more commonly seen in NTD cases, with etiologies including reflux of radio-embolic particles or accessory vasculature [1]. Accessory vasculature is typically discovered with pretreatment imaging and MAA mapping demonstrating extrahepatic perfusion. The data concerning the sensitivity of pretreatment angiography and MAA mapping for NTD is limited. Symptomatic NTD to the gastrointestinal system is estimated to be between 1.9 and 3.2%. However, in a recent systematic review, gastric ulceration was seen in only 0.1% of patients (n = 883) following TheraSphere therapy. SIR-sphere is more commonly associated with ulceration, though the incidence is still relatively low at 3.1% [9]. There are a few case reports in addition to the case presented here in which gastric ulceration was reported after ⁹⁰Y radioembolization [14, 15]. In contrast to these reports, our case demonstrates evidence of extrahepatic radioembolization on postprocedural SPECT/CT. In addition, our case demonstrates sequalae of gastric radioembolization at 2 days post-procedurally, an earlier timeline than previously documented. The cases described by South et al. presented with symptoms at ranging between 1 and 5 months after ⁹⁰Y treatment [14]. The case described by Yim et al. presented at 4 days post-procedurally [15]. Collectively, these case reports demonstrate that the time to symptom onset from gastric radioembolization is variable, ranging from 2 days to 5 months. It is possible that a patient’s individual risk factors play a role in an earlier presentation, such as in the case presented here in which the patient had a history of gastrinoma.

When extrahepatic radioembolization is detected by nuclear medicine physicians, we recommend clear communication with the proceduralist who performed the therapy, as well as with the oncologic team managing the patient. Subsequent management of the extrahepatic ⁹⁰Y radioembolization would be managed by the patient-facing physicians, and it is, therefore, important for the entire team to be aware of the increased risk for future gastric ulceration. If a patient becomes symptomatic, there should be a low threshold for consultation to gastrointestinal specialists and further investigation with EGD to confirm the diagnosis. Management of a gastric ulceration secondary to radioembolization depends on the severity of the gastric injury and the patient’s presenting symptoms. In mild cases, conservative management is often utilized, which includes pain management, administration of gastric acid suppressors, antiemetic medications, nutritional support, and close observation. For higher acuity cases such as bleeding ulcers, endoscopic interventions may be necessary. In the most severe cases, surgical intervention may be necessary to address life-threatening injuries such as perforation or large ulcers. Surgical options include resection or repair.

The presented case reflects the potential risk of NTD of ⁹⁰Y despite negative preprocedural ^99mTc-MAA mapping and emphasizes the importance of post-procedural SPECT/CT to confirm target and non-target embolization. Utilizing the Medical Internal Radiation Dose (MIRD) formula, the absorbed dose of ⁹⁰Y in the gastric antrum was estimated to be 141.6 Gray. Per the Nuclear Regulatory Commission guidelines, delivery of a radiation dose to an organ is a reportable event if the dose equivalent to the organ is greater than 50 rem and occurred due to administration of a dose of the radioactive drug differing by more than 20% of the prescribed dosage, or administration of the wrong radioactive drug, administration by the wrong route, or to the wrong subject [16].

Although the contribution of the dose equivalent to the stomach to the total effective dose equivalent can be estimated to be approximately 852 rem using a tissue weighting factor of 0.06, this case is not considered a reportable medical event given there was no misadministration or medical error during the procedure [17]. Our authors suspect that the ^99mTc-MAA mapping study was negative due to the administration of the ^99mTc-MAA in the left hepatic artery distal to the retrospectively discovered accessory artery. Although gastric ulceration in the setting of ⁹⁰Y is rare, this patient was likely at increased risk of developing an ulcer with NTD of ⁹⁰Y due to the concurrent gastrinoma and longstanding GERD.

Author Contribution

CS: writing original draft, editing, submission.

HL, ST, MS, DT, BN, KF: editing, data collection, supervision, visualization.

DR: statistical analysis.

All authors were involved in study conception and design, all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data Availability

Data sharing no applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Ethics Approval

Per our institutional guidelines for single case reports, this study was exempt from IRB review and/or ethics committee approval.

Consent for Publication

Consent for publication not applicable.

Competing Interests

The authors declare no competing interests.

Conflict of Interests

Connor Shea, Hannah Lamberg, Sevcan Turk, Mamadou Sanogo, Danielle Turgeon, Broko Nojkov, Kirk Frey and David Raffel declare no competing interests.