Serious gastric ulcer event after stereotactic body radiotherapy (SBRT) delivered with concomitant vinorelbine in a patient with left adrenal metastasis of lung cancer

To the Editor,

Hypofractionated stereotactic body radiation therapy (SBRT) is a form of high-precision radiotherapy delivery characterized by: a) reproducible immobilization to avoid patient movement during treatment sessions; (b) measures to account for tumor motion during imaging, treatment planning, and radiation delivery; c) use of dose distributions tightly covering the tumor, with rapid dose fall-off in surrounding normal tissues, in order to reduce toxicity; and d) use of extremely high doses of radiation, usually delivered in a small number of treatment fractions within a short period. The main aim of SBRT is to acquire better local control of the tumor by providing a higher dose of irradiation to a specified area during a short period. SBRT has been available for more than 10 years and is gaining clinical interest as a means of treating tumors in various organs, particularly for patients with stage I non-small cell lung cancer [1–6]. SBRT for oligometastases represents a new trend in radiation oncology [7,8]. It is also commonly accepted that SBRT is safe and frequently effective in such patients, including patients with adrenal gland metastases [9–11]. Although late effects in normal tissues after SBRT have been demonstrated [12–16], few papers have described the gastrointestinal late effects of SBRT. We describe and discuss herein a serious gastric ulcer event occurring after SBRT was delivered with concomitant vinorelbine in a patient with left adrenal metastasis of lung cancer.

Clinical case

A 54-year-old man consulted a clinic with symptoms of pneumonia. Thorough radiological examinations revealed left lung adenocarcinoma with multiple metastases to mediastinal lymph nodes, the brain, and the left adrenal gland. The lung cancer was staged at T4N3M1. Systemic chemotherapy (two cycles of 1000 mg/m² gemcitabine on Days 1 and 8 and area under the curve (AUC) =5 (1000 mg/body) carboplatin on Day 1 of a 21-day cycle) was administered, followed by five courses of weekly paclitaxel (1000 mg/m²), achieving partial response in the left lung lesions. In addition, brain metastases were treated using gamma knife therapy, and primary left lung and mediastinal lesions were treated with accelerated radiotherapy (69 Gy in 23 fractions, delivered in 5 fractions per week). These treatments proved successful, with only the left adrenal mass enlarging rapidly (Figure 1a). The adrenal lesion was 3 cm in diameter. Positron emission tomography (PET) showed high uptake of ¹⁸F-fluoro-2 deoxy-D-glucose only in the adrenal mass. The left adrenal lesion was thus judged as the only residual active metastasis, and the patient was referred to us for SBRT. Although other regimens of chemotherapy were considered, the patient wished to undergo SBRT as local intensive therapy after our explanation that SBRT was relatively safe and the adrenal metastasis was the only residual lesion, control of which might contribute to his survival. The patient was informed of the possibility of gastric toxicity resulting from treatment, but not of the potential for fatal progression of such toxicity. Long-term medical treatments including non-steroidal anti-inflammatory drugs (NSAIDs) were not administered before, during, or after SBRT.

Figure 1. (a) CT image of the left adrenal metastasis (arrow) before SBRT. (b) Dose distribution of SBRT. The white line indicated by arrowheads shows the 90% isodose line for the prescribed dose of 60 Gy in 10 fractions. (c) CT at 2 months after SBRT, showing partial response of the left adrenal metastasis. Deep ulceration was shown on the posterior wall (arrow) of the stomach, almost corresponding with the area in the 90% isodose line.

To minimize the adrenal respiratory motion during irradiation [17], the patient was trained in voluntary self-breath-holding during the inspiration phase using a respiratory indicator [18]. Upper abdominal computed tomography (CT)-scan under voluntary self-breath-holding was performed and a plan was established using CT images with the help of a 3-dimensional treatment-planning computer. Planning target volume (PTV) was determined as the gross tumor volume (GTV) of the left adrenal mass plus the personal internal margin with an additional margin of 2 mm to compensate for intra-session reproducibility and to provide a safety margin. Precise reproducibility of tumor position in this patient under voluntary self-breath-holding was measured on repeated CT images. Measurements for this patient showed the variability in tumor position due to respiration as the internal margin 1.5 mm cranio-caudally, 1.2 mm antero-posteriorly and 0.7 mm in a right-left direction. Since tumor position was adjusted to the planned position before every session using the CT images taken in the vicinity of the tumor, set-up error was considered likely to be small. Ten different non-coplanar static beams were used for irradiation. The isocenter was set at the center of the GTV. The radiation port was made with dynamic sliding multileaves adjusted with 3-mm margins around the border of the PTV. A total dose to the isocenter of 60 Gy in 10 fractions over 12 days was delivered using a 10-MV x-ray. Dose was calculated using a superposition algorithm with heterogeneity correction. Targets delineations and isovalue lines written on CT are shown in Figure 1b. Ninety-five percent of the PTV was within the 95% isodose line to the prescribed dose. According to the linear-quadratic model, the biologically effective dose (BED, α/β =10 Gy) at the isocenter was approximately 96 Gy. The dose constraints of normal tissue were defined for the intestine and spinal cord. For the intestine, volumes with dose >52.5 Gy and >43.2 Gy in 10 fractions (BED =144.4 Gy and 105.0 Gy, respectively, α/β =3 Gy) were restricted within 10 ml and 100 ml, respectively. For the spinal cord, maximum dose was restricted to under 36 Gy in 10 fractions (BED =79.2 Gy, α/β =3 Gy). These criteria represented a modification of the dose constraints provided in the protocol of the Japanese Clinical Oncology Group (JCOG)-0403 study, a prospective study of SBRT for stage I NSCLC. In the present case, gastric volumes with doses >52.5 Gy and >43.2 Gy in 10 fractions were 2.9 ml and 11.4 ml, respectively. Maximum doses to the gastric wall and spinal cord were 61 Gy and 15 Gy in 10 fractions, respectively (BED =185.0 Gy and 22.5 Gy, respectively, α/β =3 Gy). A dose-volume histogram of the stomach is shown in Figure 2.

Figure 2. Dose-volume histogram of the stomach.

SBRT was performed along with concurrent administration of vinorelbine (25 mg/m²) on Days -10, -3, +4, +11, +23, and +30 from the start of SBRT. Treatment was delivered using a unit comprising a linear accelerator (EXL-15DP; Mitsubishi Electric, Tokyo, Japan) coupled to a self-moving gantry-CT scanner (Hi-Speed DX/I; GE Yokogawa Medical Systems, Tokyo, Japan) and sharing a common couch, as a so-called “CT-on-rails” [3].

Before every radiotherapy fraction, the beam isocenter was visually adjusted with in-room CT images of 2-mm thickness taken under voluntary self-breath-holding to correspond to the planned isocenter. Examples of the daily CT images acquired prior to beam delivery for image guidance in setting the isocenter of SBRT are shown in Figure 3. Forms of the stomach changed slightly from day to day, but position of the posterior wall bordering the left adrenal grand on all days seemed to remain constant.

Figure 3. Examples of the daily CT images acquired prior to beam delivery for image guidance in setting the isocenter of SBRT. (a) First treatment. (b) Fourth fraction. (c) Eighth fraction. Forms of the stomach changed slightly from day to day, but the position of the posterior wall bordering the left adrenal grand on all days seemed to remain constant.

No acute side effects were recorded during the treatment, which was well tolerated. CT-scan showed partial response to the adrenal metastasis three months after SBRT (Figure 1c). The patient complained of acute epigastralgia with bloody anal discharge 2.5 months after SBRT and gastric fiberscopy showed a deep ulcer on the dorsal wall of the upper gastric body (Figure 4). This gastric ulcer was seen on CT-scan at three months after SBRT (Figure 1c) and corresponded with the high dose area (Figure 1a). Serum hemoglobin levels decreased rapidly and blood transfusions were frequently performed. Although the symptoms of gastric ulcer initially improved slightly after fasting with administrations of various medications, including gastric antacids, a mucosa-protective agent, H2-blockers, and proton-pump inhibitors, the ulcer gradually progressed. The patient died of massive bleeding from the ulcer 5.5 months after end of SBRT. At that time, he complained of sudden severe abdominal pain and peritoneal stimulating symptoms were observed, so gastric wall perforation combined with peritonitis was considered to have occurred.

Figure 4. Findings from gastrofiberscopy at 2.5 months after SBRT. A large, deep ulcer with necrosis was found on the posterior wall of the upper gastric body. The location of the ulcer was almost in accordance with the area in the 90% isodose line.

Discussion

To the best of our knowledge, this represents the first report of radiation-induced fatal gastric ulcer forming after SBRT for an adrenal lesion. Some reports have examined SBRT for adrenal metastases [19,20], but no optimal SBRT dose for local control has yet been defined. We fixed the prescription dose as 60 Gy in 10 fractions based on our numerous experiences with SBRT for stage I non-small cell lung cancer, although this dose was higher than the reported SBRT doses for adrenal metastases.

Late effects in normal tissues are easily enhanced with the large fraction size applied in SBRT because the α/β ratios are smaller than those of tumors. Concerning normal tissue tolerance of the stomach to conventional external beam radiation therapy with a fraction size of 2 Gy, the general consensus is that the maximum tolerated dose causing no unacceptable complications for whole-stomach irradiation is in the region of 40–45 Gy, and a stomach dose of 35 Gy increases the risk of ulcer complications [21]. No definitive radiation dose constraints for irradiation with large fractions to part of the stomach have been defined. As a complication involving the stomach after SBRT for cancer of an adjacent organ, Hoyer et al. reported that four patients suffered from severe mucositis or ulceration of the stomach or duodenum and one of these patients experienced non-fatal ulcer perforation of the stomach after SBRT of 15 Gy on three occasions within 5–10 days, among 22 patients with locally advanced and surgically non-resectable pancreatic carcinoma [22]. No detailed analysis of dose-volume to the stomach or duodenum was described in that report. Kopek et al. reported that six of 27 patients produced duodenal or pyloric severe ulceration after SBRT of 45 Gy in 3 fractions for unresectable cholangiocarcinoma and duodenal radiation exposure was higher in patients developing moderate to high-grade gastrointestinal toxicity with the difference in mean maximum dose to 1 cm³ of duodenum reaching statistical significance [23]. According to the protocol of the aforementioned JCOG-0403, the dose constraint for organs at risk (OAR) are provided for planning OAR volume (PRV) with inclusion of a 5 mm additional margin outside each organ to compensate for intra-session reproducibility and some setup margin, while no clear evidence for such limits has been accumulated. In the present case, the dose constraint was provided for the volume of stomach itself without a margin because we considered that setup error and organ motion might have reduced the volume irradiated by a high dose. Although the dose of gastric volume had not interfered with its dose constraint in the present case, adverse effects on the stomach might have been enhanced by the use of chemotherapy [24], especially concomitant vinorelbine. Long-term medical treatment especially regarding NSAIDs, which sometimes result in ulceration of the stomach, had not been prescribed for the patient.

We had not informed the patient of a potential risk of fatal gastric toxicity since we had assumed that the irradiated dose to the stomach wall would not be sufficient to cause severe gastric complications. It was chiefly because the dose of gastric volume had not interfered with its dose constraint. In addition, we had thought the dose would be scattered by the inter- and intrafractional organ motion produced by peristalsis. However, in this case, the position of the posterior wall bordering the left adrenal grand did not change sufficiently even though the actual form of the stomach changed slightly from day to day (Figure 3). The notion that stomach wall motion could differ according to position is important, and the size and pattern of stomach wall motion should be examined individually before treatment. In order to separate the stomach wall and the high dose area, suction of air into the stomach before every irradiation may be effective.

Finally, though we had given the patient SBRT as a relatively safe and possibly meaningful treatment option to improve survival, this might not have been appropriate given the presence of numerous systemic metastases. Although SBRT is generally safe and frequently effective in oligometastatic patients, including patients with adrenal gland metastases, and this represents the first reported case of these fatal side effects, we should be aware that severe ulceration producing fatal bleeding can occur after SBRT combined with chemotherapy, if the stomach wall is included in the high-dose area.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.