Esophageal Gastrointestinal Stromal Tumor Enucleation
Esophageal Gastrointestinal Stromal Tumor Enucleation
A total of 19 patients were identified who had a histologically proven diagnosis of an E-GIST. There were 5 men (26.3%) and 14 women (73.7%), with a median age of 61 years (24–88 years). The most common presenting symptom was dysphagia (n = 6), whereas 7 patients were diagnosed incidentally—3 during endoscopy and 4 after unrelated radiological investigation.
Three patients did not undergo surgical resection because of the presence of hepatic metastases at the time of diagnosis (n = 2) or advanced age (n = 1; 88 years old). For those 3 patients, the tumor was located in the proximal (n = 1) or middle third (n = 2) of the esophagus, with 2 of these 3 patients having evidence of mucosal ulceration on diagnostic endoscopy. They were all treated by TKI and 1 benefited from an endoscopic stenting.
Sixteen patients underwent surgical resection—8 had an esophagectomy and 8 had a tumor enucleation. Eleven of the 16 patients undergoing surgery were female and 15 were American Society of Anaesthesiologists (ASA) score I or II (Table 1). Six tumors were located in the proximal third, 9 in the middle third, and 1 in the distal third of the esophagus. All resected patients underwent a diagnostic endoscopy and staging computerized tomography scan; 5 patients underwent an endoscopic ultrasound and 2 a positron emission tomography scan. No resected patients had evidence of distant metastasis on staging computerized tomography scan. Of the 16 resected patients, 4 had evidence of mucosal ulceration on endoscopic examination. Two patients diagnosed with a locally advanced tumor (median diameter 85 mm vs 50 mm for nonlocally advanced tumors) underwent neoadjuvant treatment with TKI.
In view of the historical absence of recommendations regarding the surgical management of E-GISTs, the decision whether an enucleation or esophagectomy was performed was made by each individual surgeon in conjunction with the local multidisciplinary team. Eight patients underwent esophagectomy and 8 underwent tumor enucleation, with excision of the surrounding muscularis and without mucosal resection. Seven patients undergoing esophagectomy had a 2-stage operation by open laparotomy and thoracotomy with an intrathoracic anastomosis, and 1 patient had a 3-stage esophagectomy with cervical anastomosis. Of the 8 patients undergoing surgical enucleation, 5 were completed thoracoscopically (median tumor size, 42 mm; range, 18–45 mm), 1 thoracoscopic resection was converted to open thoracotomy due to a 40-mm tumor involving a large circumference of the esophageal wall, 1 enucleation was performed by a laparoscopic transhiatal approach for a 18-mm tumor, and 1 enucleation was performed by planned open thoracotomy for a 20-mm tumor. No patient was converted intraoperatively from a planned enucleation to an esophagectomy. The median duration of hospital stay was significantly longer in patients undergoing esophageal resection (11.5 days; range, 8–32 days) than in those undergoing enucleation (5.5 days; range, 2–15 days; P = 0.013) as was the median duration of operation—400 minutes (range, 300–420 minutes) versus 110 minutes (range, 60–180 minutes) (P = 0.025) (Table 1).
A total of 6 patients had a complicated postoperative recovery (Table 1): 2 patients having a pulmonary embolus, 2 patients a documented pneumonia, 1 patient a chylothorax, and 1 patient an anastomotic leak. There were 2 postoperative deaths, both occurring in patients undergoing esophagectomy, one after massive pulmonary embolism and one because of acute respiratory distress syndrome after extensive pneumonia. All 4 patients undergoing an adjuvant therapy were treated with a TKI—the 2 patients receiving a TKI after esophagectomy did so for 12 months and the 2 patients being so treated after enucleation continued a TKI for a period of 4 and 12 months. There was no statistical difference regarding tumor location, ASA score, administration of neoadjuvant treatment, or the mitotic index of tumors dependent on surgical technique (P > 0.05) (Table 1). No patients undergoing an enucleation and 4 having an esophagectomy had evidence of mucosal ulceration on endoscopy (P = 0.077). As expected, extended resections to adjacent organs were performed only in patients having an esophagectomy, with 1 patient undergoing an en bloc resection of the tail of the pancreas, spleen, and diaphragm.
The largest tumor undergoing enucleation measured 65 mm, whereas the largest tumor being excised by esophagectomy was 250 mm. The median size of enucleated tumors was 40 mm (range, 18–65 mm), whereas the median size of tumors for which esophagectomy was performed was 85 mm (range, 55–250 mm), (P = 0.001) (Table 1), reflecting that tumor size is a significant factor when considering the safety and feasibility of enucleation. All specimens underwent standard pathological preparation and immunohistochemical analysis to make the diagnosis of a GIST. Only one patient, who underwent an esophagectomy for an 80-mm E-GIST, was noted to have breach of the tumor capsule (Patient 3Eso phagectomy, Table 2) intraoperatively, and this was confirmed both on macroscopic and microscopic examination of the tumor specimen. No other patient had either macroscopic or microscopic evidence of capsular disruption, and no difference was noted in the circumferential resection margins depending on the operative technique (P = 0.809). Tissues from all tumors were analyzed by immunohistochemistry, and their staining patterns are summarized in Table 3. Tumors were defined as staining positively if greater than 10% of tumor cells stained positively. All tumors in this series, which stained positively for CD117, showed CD117 positivity in more than 50% of tumor cells. Similarly for the 14 CD34 positive tumors, all stained strongly (>50% of tumor cells) for CD34.
After a median follow-up of 6.4 years, of 14 of the 16 resected patients alive after hospital discharge, 2 patients in the esophagectomy group exhibited disease recurrence (Table 4). One death occurred secondary to disease progression, also in the esophagectomy group. Importantly, neither recurrences nor deaths were observed during follow-up after tumor enucleation.
A preoperative biopsy was performed in 9 patients—6 undergoing esophagectomy and 3 undergoing surgical enucleation (Table 5). All patients with evidence of mucosal ulceration (Patients 1, 3, 4, and 7 undergoing esophagectomy—Table 2) underwent biopsy with tissue forceps. Fine-needle aspiration cytology was performed at the time of endoscopic ultrasound for 5 lesions without mucosal ulceration (3 patients undergoing enucleation and 2 undergoing esophagectomy—Table 2). All 5 FNAs used a 22-gauge needle being passed through the channel of the endoscope, and under real-time endoscopic ultrasound guidance the target lesion was punctured with negative pressure applied by the syringe. In all 5 cases, the sufficiency of samples was immediately assessed by a cytopathologist. Biopsy confirmed the diagnosis of an E-GIST in 7 patients (77.8%), with one preoperative biopsy being suggestive of a squamous cell carcinoma and one biopsy result was nondiagnostic. The only patient in whom a thoracoscopic enucleation was converted to an open thoracotomy had not undergone a preoperative biopsy. Oncological resection was not compromised by the performance of preoperative biopsy, with no increased risk of breach of the tumor capsule (P = 1.000) or circumferential resection margin positivity (P = 0.331), nor was performing a diagnostic biopsy related to increased postoperative morbidity (P = 0.145) (Table 5). Overall disease recurrence, and local, regional, and metastatic recurrence showed no correlation with the performance of a diagnostic biopsy (P > 0.05).
Four patients, all of whom underwent esophagectomy, had mucosal ulceration evident on preoperative endoscopy in tumors measuring 55, 80, 80, and 100 mm. Mucosal ulceration was more frequent in larger tumors. In the 8 patients undergoing esophagectomy, there was a trend for mucosal ulceration to be related with a higher mitotic index (P = 0.082), but it did not correlate with circumferential resection margin positivity. The sole patient to present with regional recurrence after esophagectomy (patient 7Eso phagectomy, Table 2) had a large tumor (100 mm), with the high mitotic index (>10 mitoses/50 hpf) and mucosal ulceration.
Results
Characteristics of Overall and Resected Populations
A total of 19 patients were identified who had a histologically proven diagnosis of an E-GIST. There were 5 men (26.3%) and 14 women (73.7%), with a median age of 61 years (24–88 years). The most common presenting symptom was dysphagia (n = 6), whereas 7 patients were diagnosed incidentally—3 during endoscopy and 4 after unrelated radiological investigation.
Three patients did not undergo surgical resection because of the presence of hepatic metastases at the time of diagnosis (n = 2) or advanced age (n = 1; 88 years old). For those 3 patients, the tumor was located in the proximal (n = 1) or middle third (n = 2) of the esophagus, with 2 of these 3 patients having evidence of mucosal ulceration on diagnostic endoscopy. They were all treated by TKI and 1 benefited from an endoscopic stenting.
Sixteen patients underwent surgical resection—8 had an esophagectomy and 8 had a tumor enucleation. Eleven of the 16 patients undergoing surgery were female and 15 were American Society of Anaesthesiologists (ASA) score I or II (Table 1). Six tumors were located in the proximal third, 9 in the middle third, and 1 in the distal third of the esophagus. All resected patients underwent a diagnostic endoscopy and staging computerized tomography scan; 5 patients underwent an endoscopic ultrasound and 2 a positron emission tomography scan. No resected patients had evidence of distant metastasis on staging computerized tomography scan. Of the 16 resected patients, 4 had evidence of mucosal ulceration on endoscopic examination. Two patients diagnosed with a locally advanced tumor (median diameter 85 mm vs 50 mm for nonlocally advanced tumors) underwent neoadjuvant treatment with TKI.
Surgical Approach
In view of the historical absence of recommendations regarding the surgical management of E-GISTs, the decision whether an enucleation or esophagectomy was performed was made by each individual surgeon in conjunction with the local multidisciplinary team. Eight patients underwent esophagectomy and 8 underwent tumor enucleation, with excision of the surrounding muscularis and without mucosal resection. Seven patients undergoing esophagectomy had a 2-stage operation by open laparotomy and thoracotomy with an intrathoracic anastomosis, and 1 patient had a 3-stage esophagectomy with cervical anastomosis. Of the 8 patients undergoing surgical enucleation, 5 were completed thoracoscopically (median tumor size, 42 mm; range, 18–45 mm), 1 thoracoscopic resection was converted to open thoracotomy due to a 40-mm tumor involving a large circumference of the esophageal wall, 1 enucleation was performed by a laparoscopic transhiatal approach for a 18-mm tumor, and 1 enucleation was performed by planned open thoracotomy for a 20-mm tumor. No patient was converted intraoperatively from a planned enucleation to an esophagectomy. The median duration of hospital stay was significantly longer in patients undergoing esophageal resection (11.5 days; range, 8–32 days) than in those undergoing enucleation (5.5 days; range, 2–15 days; P = 0.013) as was the median duration of operation—400 minutes (range, 300–420 minutes) versus 110 minutes (range, 60–180 minutes) (P = 0.025) (Table 1).
A total of 6 patients had a complicated postoperative recovery (Table 1): 2 patients having a pulmonary embolus, 2 patients a documented pneumonia, 1 patient a chylothorax, and 1 patient an anastomotic leak. There were 2 postoperative deaths, both occurring in patients undergoing esophagectomy, one after massive pulmonary embolism and one because of acute respiratory distress syndrome after extensive pneumonia. All 4 patients undergoing an adjuvant therapy were treated with a TKI—the 2 patients receiving a TKI after esophagectomy did so for 12 months and the 2 patients being so treated after enucleation continued a TKI for a period of 4 and 12 months. There was no statistical difference regarding tumor location, ASA score, administration of neoadjuvant treatment, or the mitotic index of tumors dependent on surgical technique (P > 0.05) (Table 1). No patients undergoing an enucleation and 4 having an esophagectomy had evidence of mucosal ulceration on endoscopy (P = 0.077). As expected, extended resections to adjacent organs were performed only in patients having an esophagectomy, with 1 patient undergoing an en bloc resection of the tail of the pancreas, spleen, and diaphragm.
Tumor Size, Resection Margins, and Histopathology
The largest tumor undergoing enucleation measured 65 mm, whereas the largest tumor being excised by esophagectomy was 250 mm. The median size of enucleated tumors was 40 mm (range, 18–65 mm), whereas the median size of tumors for which esophagectomy was performed was 85 mm (range, 55–250 mm), (P = 0.001) (Table 1), reflecting that tumor size is a significant factor when considering the safety and feasibility of enucleation. All specimens underwent standard pathological preparation and immunohistochemical analysis to make the diagnosis of a GIST. Only one patient, who underwent an esophagectomy for an 80-mm E-GIST, was noted to have breach of the tumor capsule (Patient 3Eso phagectomy, Table 2) intraoperatively, and this was confirmed both on macroscopic and microscopic examination of the tumor specimen. No other patient had either macroscopic or microscopic evidence of capsular disruption, and no difference was noted in the circumferential resection margins depending on the operative technique (P = 0.809). Tissues from all tumors were analyzed by immunohistochemistry, and their staining patterns are summarized in Table 3. Tumors were defined as staining positively if greater than 10% of tumor cells stained positively. All tumors in this series, which stained positively for CD117, showed CD117 positivity in more than 50% of tumor cells. Similarly for the 14 CD34 positive tumors, all stained strongly (>50% of tumor cells) for CD34.
Oncological Outcomes
After a median follow-up of 6.4 years, of 14 of the 16 resected patients alive after hospital discharge, 2 patients in the esophagectomy group exhibited disease recurrence (Table 4). One death occurred secondary to disease progression, also in the esophagectomy group. Importantly, neither recurrences nor deaths were observed during follow-up after tumor enucleation.
Impact of Preoperative Biopsy
A preoperative biopsy was performed in 9 patients—6 undergoing esophagectomy and 3 undergoing surgical enucleation (Table 5). All patients with evidence of mucosal ulceration (Patients 1, 3, 4, and 7 undergoing esophagectomy—Table 2) underwent biopsy with tissue forceps. Fine-needle aspiration cytology was performed at the time of endoscopic ultrasound for 5 lesions without mucosal ulceration (3 patients undergoing enucleation and 2 undergoing esophagectomy—Table 2). All 5 FNAs used a 22-gauge needle being passed through the channel of the endoscope, and under real-time endoscopic ultrasound guidance the target lesion was punctured with negative pressure applied by the syringe. In all 5 cases, the sufficiency of samples was immediately assessed by a cytopathologist. Biopsy confirmed the diagnosis of an E-GIST in 7 patients (77.8%), with one preoperative biopsy being suggestive of a squamous cell carcinoma and one biopsy result was nondiagnostic. The only patient in whom a thoracoscopic enucleation was converted to an open thoracotomy had not undergone a preoperative biopsy. Oncological resection was not compromised by the performance of preoperative biopsy, with no increased risk of breach of the tumor capsule (P = 1.000) or circumferential resection margin positivity (P = 0.331), nor was performing a diagnostic biopsy related to increased postoperative morbidity (P = 0.145) (Table 5). Overall disease recurrence, and local, regional, and metastatic recurrence showed no correlation with the performance of a diagnostic biopsy (P > 0.05).
Impact of Mucosal Ulceration
Four patients, all of whom underwent esophagectomy, had mucosal ulceration evident on preoperative endoscopy in tumors measuring 55, 80, 80, and 100 mm. Mucosal ulceration was more frequent in larger tumors. In the 8 patients undergoing esophagectomy, there was a trend for mucosal ulceration to be related with a higher mitotic index (P = 0.082), but it did not correlate with circumferential resection margin positivity. The sole patient to present with regional recurrence after esophagectomy (patient 7Eso phagectomy, Table 2) had a large tumor (100 mm), with the high mitotic index (>10 mitoses/50 hpf) and mucosal ulceration.
