A Rare Cause of Esophagopleural Fistula Due to Intensity Modulated Proton Therapy: A Case Report and Review of Literature

Abstract

Esophagopleural fistula (EPF), initially described in 1960, is an abnormal communication between the esophagus and the pleural cavity which can occur due to congenital malformation or acquired due to malignancy or iatrogenic treatment. The most common presenting symptoms are of a respiratory infection, such as fever, chest tenderness, cough and imaging findings consistent with pleural fluid consolidation. In this report, we present a 59-year-old man who exhibited shortness of breath, productive cough, and significant weight loss for two weeks. His medical history was significant for smoking related lung disease and pulmonary squamous cell carcinoma (SCC). His SCC (T4N0) was diagnosed six years prior to this presentation and was treated with chemoradiotherapy. The cancer recurred a year ago and he was treated with intensity modulated proton therapy (IMPT) and consolidation chemotherapy. During admission, he was found to have an EPF by CT scan after initially failing antibiotic treatment for suspected complicated pneumonia and pleural effusion. Multiple attempts of esophagopleural fistula closure were made using endoscopic self-expandable metallic stents and placement of an esophageal vacuum-assisted closure device. However, these measures ultimately failed and therefore he required an iliocostalis muscle flap (Clagett window) procedure for closure. Esophageal pulmonary fistulae should be suspected whenever patients have undergone thoracic IMPT and may present with acute pulmonary complications, particularly pneumonia refractory to antibiotic treatment. This case reviews the current literature, potential complications, and treatment options for esophagopleural fistulas.

Background

Esophagopleural fistula (EPF) is a rare subtype of esophagorespiratory fistulas (ERF). EPF is an abnormal communication between the esophagus and the pleural cavity. It can be observed in neonates secondary to congenital abnormalities and in adults, can be acquired secondary to surgery, malignancies, infections, trauma, and rarely from chemotherapeutic drugs or radiation therapy.^1–3 Patient’s typically present with non-specific respiratory symptoms including fever, sudden onset of dyspnea, productive cough, and imaging findings of lung consolidation, and pleural effusions. On imagining, a fistula is most commonly found on the right side due to its anatomical proximity and less common on the left side due to the presence of the aorta between esophagus and trachea.⁴.

The aim of this case report is to make readers aware of this rare, potentially life-threatening complication of treatment following thoracic radiation with intensity modulated proton therapy (IMPT).⁵ Our patients’ symptoms initially mimicked clinical and radiological findings consistent with pneumonia and pleural effusion and the treatment for EPF was delayed due the misdiagnosed initial presentation.

Case presentation

A 59-year-old man with a 15-pack-year smoking history and stage IIIa T4N0 unresectable squamous cell carcinoma (SCC-2012), presented in 2019 with chest pain, shortness of breath, productive cough and 25lb weight loss over a 2-week period. He complained of right sided chest pain, which was sharp, non-radiating, and aggravated with cough. He noticed non-bloody, foul smelling greenish-yellow phlegm along with progressive shortness of breath with minimal exertion. His medical history was significant for right lower lobe SCC status-post completion of weekly Paclitaxel, Carboplatin and thoracic radiation to 66Gy in 2013. Other significant past medical history included COPD, congestive heart failure and atrial fibrillation. Despite treatment in 2013 with interval improvement, he developed recurrence of his squamous cell carcinoma of the right lower lobe of the lung with concern for progression (right pleura/peripheral lung) on PET scan. Following this discovery, he received multiple cycles of Carboplatin and Paclitaxel and PET (1/24/19) demonstrating resolution of certain hypermetabolic foci, but persistent RLL, subpleural, and Internal Mammary Lymph node (IMN) disease which was treated with a course of IMPT to 61.20Gy directed to these lesions.

During his current hospitalization, physical evaluation revealed the patient was afebrile, tachycardic, normotensive and saturating well on room air. Upon auscultation, decreased breath sounds, and crackles were noticed on the right side of the chest. He had no palpable lymphadenopathy on general examination. Given his symptoms, a CT angiography of the chest was ordered which was negative for pulmonary embolism but revealed cavitary consolidation in the right middle and superior segment of the lower lobe along with chronic loculated pleural effusion (Figure 1). He was admitted for pneumonia and was started on broad-spectrum antibiotics, Vancomycin and Piperacillin-Tazobactam. Given unresolved symptoms despite empiric coverage for pneumonia, a repeat CT chest without contrast was performed which showed air-fluid levels in the right basilar pleural effusion and unchanged consolidation in the right lung from prior imaging (Figure 2). Decision to drain the pleural effusion was made using an Interventional Radiology (IR) guided 14 French pigtail catheter placed in the right side of the chest to mobilize the loculations and drain the effusion. Purulent discharge was noticed from the pigtail catheter and pleural fluid analysis revealed exudative effusion with pH of 6.3, glucose less than 20 mg/dl, LDH 575 units/L. Cultures were negative for acid fast bacilli for concerns of active tuberculosis, however were positive for polymicrobial oral flora. His pleural fluid outflow was 1L each day without any improvement.

Figure 1.

CTA chest demonstrating right middle (A) to lower lobe (B) cavitating consolidation with loculated pleural effusion

Figure 2.

CT chest demonstrating air-fluid level in the right basilar pleural effusion and unchanged consolidation in the right lung

Repeat chest CT with oral contrast was performed after 10 days of chest tube placement, which newly revealed a 12mm lower esophagopleural fistula (EPF) that was 35cm from the incisors (Figure 3). Endoscopic argon plasma coagulation (APC) was performed encircling the proximal fistula site in the esophagus followed by endoscopic suturing of the proximal end of the fistula. A self-expandable metallic stent (SEMS) was then placed in the esophagus (Figure 4). His oral feeding was immediately stopped and esophagogram was performed on the same day which revealed no leak. Subsequent drainage from the chest tube gradually decreased over the next 3 days. He was subsequently discharged home on a clear liquid diet, however during following up EGD, there was persistence of the fistula. Three more attempts of fistula closure were done with endoscopic suturing and a SEMS placement without a success. The reason for the failure was thought to be a large mediastinal space that was connected to the esophageal fistula. This led to placement of esophageal vacuum-assisted closure (EVAC) device with continuous suction (Figure 5). In spite of 60 days of EVAC treatment with a total of nineteen interventions which were performed every 3–4 days for a sponge change, his fistula remained unclosed. After failure of endoscopic treatment, he underwent an initial attempted free flap closure of the esophageal defect using the latissimus dorsi approach. This flap failed due to venous thrombosis. A repeated flap closure using the iliocostalis muscle (Clagget window approach) was then performed. Esophagogram followed by EGD demonstrated healed esophageal-pleural fistula after two weeks of repeat flap closure and the patient was discharged home safely (Figure 6).

Figure 3.

CT chest demonstrating a fistulous connection between the lower esophagus with the air and debris filled right-sided pleural cavity

Figure 4.

Endoscopic placement of self-expandable metallic stent (SEMS) in the middle third of the esophagus for esophago-pleural fistula

Figure 5.

Endoscopic placement of esophageal vacuum-assisted closure (EVAC) for persistent esophago-pleural fistula

Figure 6.

Esophagogram showing contrast extending outside of the proximal margin of the stent on the left, pooling between the stent and esophageal wall. No extravasation outside the esophageal lumen, demonstrating the closure of esophago-pleural fistula after surgical repair with muscle flap

Discussion & Conclusions

ERFs represent an atypical communication between the esophagus and respiratory system. The most common subtype of ERF is a tracheoesophageal fistula, while an esophageal pulmonary fistula is more rare.⁶ Non-malignant causes of an ERF include infections, inflammatory mediastinal diseases, such as Crohn’s or sarcoidosis, and traumatic complications following endoscopic or bronchoscopic procedures.^7–10 Among the most common malignancies leading to development of an ERF are lung, esophageal, and mediastinal carcinomas.⁷ Specifically, 0.2% of lung carcinomas are reported to cause EPF.¹¹ Chemotherapeutic agents such as carboplatin, irinotecan and VEGF inhibitors (bevacizumab) have also shown an association with the development of a fistula however a clear underlying mechanism of pathogenesis is still unknown.¹²

Fistula formation is a rare, and established late complication of radiation therapy that may be seen in a variety of tumor subsites.^13–15 Radiation causes cell injury by means of DNA damage which, with increasing doses, can overwhelm the body’s repair mechanisms leading to cell death. The administration of concurrent chemotherapy hinders these repair processes which further augments toxicity.¹⁶ In the re-irradiation setting, the risk of toxicity is amplified due to an increased cumulative dose to organs-at-risk leading to fibrosis, inflammation, and impaired wound healing, all of which can contribute to fistula formation.^17,18 Proton therapy is often used in the re-irradiation setting, as in our patient, due to its ability to deposit radiation dose with enhanced sparing of surrounding normal tissues compared to conventional photon-based radiation.¹⁹ It is also considered in the setting of large treatment volumes, targets abutting critical organs at risk, or when photon-based radiation is not feasible due to violation of established dose constraints. A patient that receives proton therapy, especially for re-irradiation, should therefore be heavily scrutinized for potential toxicities given these indications for its use. Unfortunately, long-term prospective trials comparing patient outcomes of proton to photon therapy are still under investigation and are not expected to be completed until 2024.^20,21

Patient’s with EPF present with fever, sudden onset of dyspnea, chest tenderness, respiratory infection and imaging findings of cavitary lesions associated with pleural effusions. Initially, it may be difficult to suspect an underlying fistula based on non-specific symptoms. A sour odor from the exudative pleural effusion, presence of food material and growth of normal oral flora from the pleural fluid is pathognomonic for an abnormal communication between esophagus and pleural space.⁹ Imagining modalities such as Esophagography with water-soluble Gastrografin are helpful by demonstrating the extravasation of oral contrast into the pleural space. However, during the earlier stages of fistula development, esophagogram may be falsely negative. If there is high clinical suspicion, CT scan of the chest either with or without IV contrast should be performed. CT is the most sensitive and noninvasive test that can detect small quantities of air/fluid and can be performed rapidly.^10,22

Conservative management of an ERF by intravenous infusions, narcotics, antibiotics, and occasional nasogastric intubation has not shown survival benefit in adults and resulted in death from complications of aspiration, asphyxiation, starvation and sepsis.²³ Most cases require treatment with endoscopic injection of glue, clipping, suturing, stenting, or surgical repair for the closure of the fistula.²⁴ Covered and retrievable Self-expandable metallic stents (SEMS) are considered safe and durable and provide immediate relief in the management of malignant fistulous and palliative dysphagia.²⁵ SEMS are introduced into the esophagus using an endoscope with direct visualization of the fistula to ensure proper seal. In one case series, SEMS have shown a 55.6% clinical success rate in closure of EPF however, fistula persistence or reopening is possible within 15 days following stent placement therefore frequent follow up and replacement of stents as necessary is advised.²⁶ Endoscopic vacuum-assisted closure (EVAC) therapy has also been shown to be successful in closure of upper GI tract anastomotic leaks.²⁷ This technique involves placement of a sponge into the lumen which is connected to a nasogastric tube via a negative pressure system. This gradually decreases the cavity size which augments healing. In a retrospective study, vacuum assisted closure for the management of post-esophagectomy anastomotic leakage was shown to have a significantly lower mortality rate of 12%, when compared to 50 % in surgically repaired and 83% in stented patients.²⁸ A recent retrospective study of 123 patients with acquired malignant and benign causes of ERF compared non-surgical (endoscopic) versus surgical intervention. The study revealed, patients with malignant ERF treated surgically survived longer than patients undergoing nonsurgical treatment (hazard ratio=5.6, P=0.005). In contrast, reintervention was more common in those who underwent nonsurgical treatment (hazard ratio=2.3, P=0.03).²⁹ Currently, the mainstay for the initial treatment of ERF remains stenting with possible suturing.^30,31 In complex or refractory cases, surgical closure repair may be necessary using a musculocutaneous flap creation approach and a vacuum assisted closure is used in conjunction to endoscopic or surgical interventions in patients with non-healing esophageal fistulas.^32,33

In this case, we highlight the importance of early recognition of EPF and the management steps necessary in a complex post chemoradiation and intensity-modulated proton therapy patient presenting with non-specific pulmonary symptoms. We hypothesized that the patient’s second course of radiation with IMPT may have contributed to the development of EPF in our patient. Due to delayed healing, our patient required multiple stent procedures in addition to EVAC therapy. After failed attempts, the patient had to undergo a definitive invasive flap procedure for closure. This case also highlights the importance of suspecting EPF in patients with a history of thoracic radiotherapy with IMPT presenting with symptoms of pneumonia and acute respiratory distress which is refractory to empiric antibiotic coverage. In this case, initial radiological imaging did not reveal EPF and was noticed only after refractory improvement. The importance of obtaining either videofluoroscopy or contrast enhanced CT chest imaging is imperative for diagnosis and treatment management. Overall, as IMPT gains more traction in the realm of radiation therapy, the association between EPF and IMPT warrants recognition and efforts for investigation of patient outcomes and complications related to IMPT should be continued.

Abbreviations

ERF

Esophago-respiratory fistula

EPF

Esophago-pleural fistula

SCC

Squamous cell carcinoma

IMPT

Intensity modulated proton therapy

EGD

Esophagogastroduodenoscopy

SEMS

Self-expandable metallic stents

EVAC

Endoscopic vacuum assisted closure

IMN

Internal mammary lymph nodes