2025 Volume 47 Issue 4 Pages 201-207
We report a case of mediastinal abscess formation following endoscopic ultrasound-guided tissue acquisition (EUS-TA) performed for the diagnosis of mediastinal lymphadenopathy. A 50-year-old man, who had undergone surgery for lung cancer and received adjuvant chemotherapy, presented two months after completing chemotherapy with subcarinal lymph node enlargement on computed tomography (CT) and fluorodeoxyglucose (FDG) uptake on positron emission tomography (PET)-CT. The size of the lymph node increased more than twofold over the following month, as revealed by evaluation with EUS. EUS-TA was then performed, confirming lymph node metastasis from lung cancer. Six days later, contrast-enhanced CT revealed a hypodense lesion with marginal enhancement surrounding the punctured lymph node, leading to the diagnosis of a mediastinal abscess. The patient’s clinical condition improved with conservative treatment using intravenous antibiotics. The abscess formation was primarily outside the lymph node, suggesting that the puncture needle may have inoculated bacteria into a poorly vascularized area. The patient’s immunocompromised state following chemotherapy and the inadequate blood supply within the rapidly enlarging lymph node likely contributed to the abscess formation. Mediastinal abscess formation following EUS-TA for lymphadenopathy is rare, and current guidelines do not recommend prophylactic antibiotics. However, as demonstrated in this case, endoscopists may encounter this complication when multiple risk factors are present.
The first report of endoscopic ultrasound-guided tissue acquisition (EUS-TA) in clinical practice was for pancreatic tumors in 1992 [1]. Since then, EUS-TA has gained worldwide acceptance and is now widely performed not only for pancreatic lesions but also for extra-luminal lesions accessible through the digestive tract, such as those in the mediastinum. Complications associated with EUS-TA for mediastinal lesions include mediastinitis and mediastinal abscess; however, the risk is higher for cystic lesions and rare for solid lesions, including lymphadenopathy [2]. Consequently, the American Society for Gastrointestinal Endoscopy (ASGE) and the European Society of Gastrointestinal Endoscopy (ESGE) guidelines recommend prophylactic antibiotics for EUS-TA of cystic lesions but not for solid lesions or lymph nodes, due to the very low incidence of infectious complications [2, 3].
Here we report a case of a mediastinal abscess associated with EUS-guided fine-needle biopsy (EUS-FNB) of a non-cystic mediastinal lymph node in a patient with postoperative lung cancer. We also review the literature on infectious complications following EUS-TA of non-cystic mediastinal lesions.
A 50-year-old man with no specific medical history, including diabetes mellitus (HbA1c 5.4%), underwent lingular segmentectomy and left lower lobectomy for squamous cell carcinoma of the lung. Postoperative pathology revealed mediastinal invasion with hilar lymph node metastasis, staged as T4N1M0. The patient was treated with cisplatin and docetaxel as adjuvant chemotherapy for 4 months. Two months after completion of adjuvant chemotherapy, contrast-enhanced computed tomography (CT) revealed a new lymphadenopathy with a diameter of 10mm in the subcarinal region, exhibiting heterogeneous enhancement (Figure 1A). 18F-fluorodeoxyglucose positron emission tomography CT (FDG-PET/CT) showed FDG uptake at the same site, but no other uptake (Figure 1B, C). One month after CT imaging, we performed EUS-TA for the subcarinal lymph node using an Olympus GF-UCT260 scope (Olympus, Tokyo, Japan). On EUS, the enlarged subcarinal lymph node was detected as a hypoechoic well-circumscribed solid mass with no obvious cystic component (Figure 2A). The diameter was 24 mm, larger than the CT one month earlier. Three passes with a 22-gauge FNB needle (SharkCore™ FNB, Covidien) were performed using the slow-pull method without any immediate complications (Figure 2B), and specimen adequacy was determined by macroscopic on-site evaluation. The patient received levofloxacin prophylaxis for 3 days after the procedure.
He was discharged the day after the procedure with no evidence of infection. Pathology revealed squamous cell carcinoma, consistent with lymph node metastasis from lung cancer (Figure 2C).
Two months after completion of adjuvant chemotherapy for lung cancer, A: CT showed a lymphadenopathy with a diameter of 10mm in the subcarinal region (red arrows). B and C: FDG-PET/CT showed FDG uptake at the same site, but no other uptake.
A: The enlarged subcarinal lymph node was detected as a hypoechoic well-circumscribed solid mass with no obvious cystic component. The diameter was 24 mm. B: EUS-FNB was performed and the puncture needle was seen in the lymph node. C: Hematoxylin and eosin staining of a specimen obtained from subcarinal lymph node with EUS-FNB (×100). Pathology revealed squamous cell carcinoma.
The patient presented to his primary care physician with a fever six days after the EUS-FNB, and was prescribed levofloxacin. Due to persistent fever thereafter, he presented to our department on the 8th day after the EUS-FNB. On arrival at our department, the patient’s body temperature was 37.9°C, his blood pressure was 130/80 mmHg, SpO2 97% in room air, and he had clear respiratory sounds in both lungs without rales. Laboratory data revealed severe inflammation (white blood cell count 12,600/µl and C-reactive protein 14.54 mg/dl) and an elevated procalcitonin level (35.2 ng/ml). No pathogens were detected in blood cultures. Contrast-enhanced CT of the chest clearly showed a enlargement of the subcarinal lymph node compared to the previous CT and the appearance of a hypodense area within the lymph node. In addition, a new hypodense lesion with marginal enhancement was observed outside the subcarinal lymph node (Figure 3). We diagnosed a mediastinal abscess caused by EUS-FNB, and the patient was admitted to our hospital.
Contrast-enhanced CT showed enlargement of the known subcarinal lymph node and the appearance of a hypodense area within the lymph node (red arrow). A new hypodense lesion with marginal enhancement was also observed on the right side of the subcarinal lymph node (yellow arrowheads). A: Axial view. B: Axial view of the foot side of A. C: Coronal view.
After consultation with the surgeon, conservative treatment with intravenous antibiotics (meropenem 3.0 g/day) was initiated. There was a gradual decrease in the inflammatory response and resolution of fever on the 5th day of hospitalization. On the 10th day of hospitalization, the levels of inflammatory markers had normalized and the abscess outside the subcarinal lymph node had shrunk on contrast-enhanced CT (Figure 4), and meropenem was discontinued (Figure 5). The patient was discharged on the 12th day of hospitalization, and started chemotherapy for lung cancer on the 55th day after EUS-FNB. The abscess did not recur after starting chemotherapy.
Contrast-enhanced CT showed no obvious change in the findings of the known subcarinal lymph node (red arrow), but the abscess on the right side of the subcarinal lymph node had shrunk (yellow arrowheads). A: Axial view. B: Axial view of the foot side of A. C: Coronal view.
BT: body temperature, CRP: C-reactive protein, MEPM: meropenem, WBC: white blood cell.
A wide range of mediastinal lesions are indicated for EUS-TA, including mediastinal tumors, lymph node metastases, malignant lymphomas, sarcoidosis, and tuberculosis. Catalano et al reported a successful diagnosis in 90% of patients and an impact on subsequent treatment in 87% of patients with mediastinal lymphadenopathy of unknown etiology [4]. In our case, which was postoperative for lung cancer, and given the observed FDG uptake, the subcarinal lymphadenopathy was most likely due to the postoperative recurrence of lung cancer. However, both infection and reactive lymphadenopathy can also show FDG uptake on PET [5]. Moreover, the diagnostic accuracy of EUS-guided fine-needle aspiration (EUS-FNA) for lymphadenopathy has been shown to exceed that of PET for diagnosing lung cancer [6].
Another endoscopic tissue acquisition method for mediastinal lesions is endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA). EBUS can generally visualize the anterior region of the trachea and main bronchus, while EUS is better suited for delineating the paraoesophageal space, aortopulmonary window, and subcarinal region [7]. A Japanese nationwide survey found that the incidence of infectious complications with EBUS-TBNA was similar to that of EUS-FNA for mediastinal lesions, at 0.19% versus 0.17%, respectively [8, 9]. In a study of 104 patients with suspected mediastinal lymph node metastases from lung cancer, no infectious complications were observed following EUS-FNA [10].
However, cystic lesions and sarcoidosis of the mediastinum have been reported to carry a small but not non-negligible risk of infectious complications [2, 11]. Wiersema et al reported a 14% infection rate after EUS-FNA for cystic lesions in the mediastinum, pancreas, and rectum [12]. Necrotic lymph nodes are also considered more susceptible to infection after EUS-FNA, likely due to the difficulty the immune system faces in reaching poorly vascularized lesions [13]. In our case, the heterogeneous enhancement of the lymph node on contrast-enhanced CT likely indicated focal necrosis or vascular abnormalities in the malignant lymph nodes, such as caliber fluctuations, sinusoid formation, or arteriovenous shunts. Necrosis was quite possible, given that the lymph node enlargened rapidly, although no clear findings of focal necrosis were observed on EUS. One study evaluating lymph node metastasis of lung cancer using EBUS, reported interobserver variability in diagnosing the presence or absence of necrosis [14]. This suggests that in some cases, it may be difficult to diagnose focal necrosis based on ultrasound alone, and additional evaluation, such as contrast-enhanced MRI or EUS, may be helpful for diagnosing focal necrosis.
Six English case reports describe infectious complications following EUS-FNA for mediastinal lymphadenopathy, excluding sarcoidosis or cystic lesions [13, 15–19]. Seven cases, including ours, are summarized in the Table. The primary disease in five cases was malignancy, including two cases of lung cancer. Of these, five patients developed mediastinal abscesses or mediastinitis, one developed an esophageal-mediastinal fistula, and one developed infective endocarditis. Among the five cases of mediastinal abscess or mediastinitis, three were treated with antibiotics alone, while two required drainage or surgery. Prophylactic antibiotics were administered in our case only. The median time to the onset of infection symptoms was 2 days, with our case being the longest at 6 days. The needle size was 22 gauge in all cases, with FNB used in our case only. The median number of punctures was 3. Only one other case, which involved an esophageal-mediastinal fistula, mentioned the use of the suction technique in EUS-TA. Excluding the tuberculosis cases, pathogens were identified in two cases, both of which involved oropharyngeal bacteria. No common background factors predisposing patients to infectious complications were identified in any of the cases. In our case, the patient’s recent chemotherapy may have contributed to immune compromise and abscess formation.
| Author | Age / Sex | Location of LN | Diameter of LN (mm) | Needle gauge | Number of punctures | Prophylactic antibiotics | Infectious complication | Number of days to infection onset | Final diagnosis | Pathogen | Treatment |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Will U et al [15] | 76 / M | N.D. | 20 | 22 | 2 | none | Mediastinal abscess | N.D. | HCC | N.D. | EUS-D |
| Pai KR et al [16] | 40 / M | Posterior mediastinum | 40 | 22 | 2 | none | Mediastinal abscess | 1 | Teratoma | N.D. | Percutaneous drainage Thoracotomy |
| Aerts JG et al [13] | 68 / M | Subcarinal LN | N.D. | 22 | 4 | none | Mediastinal abscess | 2 | Small cell carcinoma | N.D. | Antibiotics |
| Savides TJ et al [17] | 78 / M | AP window | 10 | 22 | N.D. | none | Mediastinitis Osteomyelitis | 5 | Benign | Gemella morbillorum | Antibiotics |
| van Fraeyenhove F et al [18] | 72 / M | N.D. | N.D. | N.D. | N.D. | none | Endocarditis | 2 | Renal cell carcinoma | Streptococcus salivarius | Antibiotics |
| von Bartheld MB et al [19] | 26 / M | Subcarinal LN | 25 | 22 | 4 | N.D. | Mediastinal-esophageal fistulae | 2 | Tuberculosis | Mycobacterium tuberculosis | Antibiotics |
| Our case | 50 / M | Subcarinal LN | 24 | 22 | 3 | LVFX | Mediastinal abscess | 6 | Lung cancer | Negative | Antibiotics |
AP window: aortopulmonary window, EUS-D: endoscopic ultrasound-guided drainage, HCC: hepatocellular carcinoma, LN: lymph node, LVFX: levofloxacin, M: male, N.D.: not described
Several studies on bacteremia and infectious complications following EUS-FNA have reported positive blood cultures for various pathogens, including coagulase-negative staphylococcus and gram-negative bacilli [18]. One study found oropharyngeal bacteria in 35% of TBNA needle washings [20]. In that study, while bacteremia was confirmed in three cases, no clinical signs of infection were observed, suggesting that while contamination from oral flora is possible, clinically significant infections remain rare. In our case, however, an infection still occurred despite the administration of prophylactic antibiotics. It is likely that oropharyngeal bacteria were transmitted to the mediastinum through the puncture needle. Oropharyngeal bacteria often include anaerobic bacteria, and the prophylactic effect may have been insufficient, since levofloxacin has limited activity against anaerobic bacteria. Quinolone antibiotics with stronger activity against anaerobic bacteria are preferable in patients with an immuno-compromised state following chemotherapy, such as in our case. Antibiotics such as grepafloxacin or sitafloxacin would be more effective for infection prevention in these patients [21]. Furthermore, as revealed by the imaging findings of the mediastinal abscess, the abscess in our case formed primarily outside the punctured lymph node. It is possible that the abscess within the lymph node perforated through its capsule, but it is also possible that the needle tip penetrated the target lymph node and inoculated bacteria into the surrounding tissue.
In conclusion, EUS-TA for mediastinal lesions is highly useful for diagnosis, but there remains a risk of infectious complications, although they are rare. In our case, the patient’s immuno-compromised state following chemotherapy may have contributed to the formation of a mediastinal abscess, highlighting the importance of infection prevention in patients with immune suppression. The selection of prophylactic antibiotics for EUS-TA in immuno-compromised patients warrants further investigation. Furthermore, to minimize the risk of infection after EUS-TA for mediastinal lesions, it is crucial to accurately confirm the position of the needle tip during the procedure and to consider the potential presence of focal necrosis in rapidly growing malignant lesions.
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
Appropriate written informed consent was obtained from the patient for publication of this case report and accompanying images.
The authors received no financial support to produce this manuscript.
The authors declare no Conflict of Interest for this manuscript.
