Perioperative FLOT chemotherapy plus surgery for oligometastatic esophagogastric adenocarcinoma: surgical outcome and overall survival

All patients with symptoms suggesting the presence of esophageal or gastric cancers were taken through diagnostic and staging work up according to German S3 guidelines [6]. This includes a thorough medical history and physical exam, as well as upper GI endoscopy with several biopsies, endoscopic ultrasound if technically possible and a CT Thorax/Abdomen to exclude distant metastases. Diagnostic laparoscopy with peritoneal biopsies and PET-CT scans were added in selected cases of suspected peritoneal carcinomatosis or distant metastases, otherwise distant metastases were diagnosed by staging CT. The management pathway was chosen according to TNM staging. For patients with locally advanced EGAC (pT3 or pT4), re-staging was carried out after the completion of neoadjuvant chemotherapy, in order to plan surgical management.

Perioperative chemotherapy consists of four cycles prior to surgery (over 8 weeks) and further four cycles post-surgery, with each cycle lasting 2 weeks. The FLOT regime consists of infusions of 5-FU 2600 mg/m² (24 h), leucovorin 200 mg/m² (2 h), oxaliplatin 85 mg/m² (2 h) and docetaxel 50 mg/m2 (1 h) every 2 weeks [5].

Surgery was usually carried out between 4 and 6 weeks after the completion of the neoadjuvant cycles of chemotherapy, with few selected patients undergoing surgery at a later point in time. Surgery was chosen according to tumor location and size. Routinely, esophagectomy plus proximal gastrectomy with two-field lymphadenectomy was performed for esophageal or junctional adenocarcinoma (AEG I + II), whilst patients with AEG III tumors (in some selected cases also AEG II underwent transhiatal extended gastrectomy with lower mediastinal and modified DII-lymphadenectomy. Total or subtotal gastrectomy plus modified DII-lymphadenectomy was performed for patients with gastric cancer. In extended tumors, the surgical approach was adapted as necessary. Resectability of the primary and metastases depended on the location and was determined by an interdisciplinary team. It was then carried out accordingly e.g. as liver resection, adrenalectomy or peritonectomy for peritoneal carcinomatosis. Hyperthermic intraperitoneal chemotherapy (HIPEC) was additionally performed for patients with limited peritoneal metastases using Cisplatin (75 mg/m²) and Doxorubicin (15 mg/m²). No peritoneal lavage cytology was performed for patients in this cohort. Routine postoperative standard histopathological workup and staging was performed.

Statistical analysis was performed using IBM SPSS statistics, Version 23. Categorical variables were put in absolute and relative frequencies; differences were evaluated by Chi-Square or Fisher’s exact test as appropriate. Quantitative values were expressed as medians with range and differences were measured using the Mann–Whitney-U test. Multivariate analysis was performed through forward logistic regression model, with relative risk and a 95% confidence interval. The Kaplan–Meier method was used to evaluate survival, with a long-rank test for the comparison of subgroups. A p-value < 0.05 was considered statistically significant.

Out of 48 patients, 31 patients (65%) were diagnosed with gastric cancer and 17 patients (35%) with esophageal adenocarcinoma. Median follow up was 13 months (11 months for deceased patients and 17 months for all others). All patients, except for one (pT2), had pT3 or pT4 tumors at initial staging. Comorbidities, including cardiac, pulmonary, renal and hepatic disease were present in 64% of patients (n = 31). Other patients’ characteristics are summarized in Table 1. 83% of patients completed four cycles of neoadjuvant FLOT chemotherapy (n = 40), whereas the adherence to postoperative chemotherapy was less, with only 31 patients (65%) completing their adjuvant treatment. Potentially resectable metastases were present in all patients either at the time of initial diagnosis or during preoperative staging. Figure 1 shows patient selection through a flowchart of diagnosis and staging (Fig. 1).

62% of patients (n = 30) had peritoneal carcinomatosis, from which 84% arose from gastric cancer (n = 26). Distant lymph node metastases (lymph nodes outside of DI-II resection area) were found in 7 patients, other distant metastases sites were hepatic (n = 7), adrenal (n = 3) and pulmonary (n = 1). Complete remission of metastases after preoperative chemotherapy was achieved in 12 patients (Fig. 1). Out of these 12 patients (ypM0), 8 did not show evidence of metastatic lesions, both during the preoperative staging and intraoperatively, and 4 patients with macroscopically suspected peritoneal carcinomatosis showed no histologic evidence of tumor cells after peritonectomy. Resection of metastases was performed if distant metastases could be detected intraoperatively and surgical resection seemed feasible. Thus, in 68% of patients (n = 33) simultaneous resection of metastases was performed. For 11 patients (23%) metastases were not feasible to resect, defined here as R2. HIPEC was additionally carried out in 15 patients with peritoneal metastases after complete cytoreductive surgery.

Postoperative complications occurred in 48% of patients (n = 23) [42% after gastrectomy (n = 13) and 59% after esophagectomy (n = 10)]. Surgical complications include anastomotic leaks (n = 2) wound infection (n = 5), chylothorax (n = 3) and haemorrhage (n = 2), whilst medical complications were mainly pulmonary, such as pleural effusions (n = 4), pneumonia (n = 8) and the need for reintubation (n = 2). 16 patients (33%) experienced complications of grade I-II and seven patients had major complications of grade III-V (15%). Two patients died after esophagectomy: one patient due to postoperative arrosive bleeding from the splenic artery and one due to rapid progressive pleural carcinomatosis, and one patient after gastrectomy, due to an anastomotic leak followed by septic shock. Average length of hospital stay was 12 days (range 7–94), with an average length of stay on ICU/IMC of 4 days (range 2–22 days). Treatment data is summarized in Table 2.

Overall 5- year survival of patients with oligometastatic EGAC was 18%, with a median survival of 15 months after surgery. Tumor recurrence occurred in 51% (19 of 37 patients) without residual macroscopic tumor after surgery, in a median time interval of 6 months (1–15). Most recurrences were distant metastases (peritoneal carcinomatosis n = 5, hepatic n = 6, multiple distant n = 1). Post-recurrence treatment was individualized to the patient and included surgery, radiotherapy, palliative chemotherapy or best supportive care. According to the treatment used, the rate of overall survival will differ. Patients with gastric cancer and esophageal adenocarcinoma had 5-year survival rates of 25% and 10% respectively (p = 0.213). Tumor regression grading according to Becker et al. [12], showed that with 1a regression (no residual tumor, n = 6), patients had a 60% survival at 5 years. Patients with regression grades of 1b and 2 only had 11% and grade 3 and above only 12% 5- year survival (p = 0.012, Table 3, Fig. 2). Median survival was 21 months and 9 months, respectively. The survival rate was shown to be 60% for patients with T0 staging, compared to 27% for T1/T2 and 0% for T3/T4 tumors (n = 6, n = 15, n = 27, respectively, p = 0.047). The most significant finding was demonstrated by the difference in 5-year survival rate between patients with non-detectable tumor postoperatively (ypM0) and patients with detectable oligometastases. Here, patients with postoperative ypM0 (n = 12) had a 48% 5-year survival rate, with a median survival of 47 months, in contrast to only 11% at 5 years for patients with detectable tumor cells (ypM1, n = 36), with a median of 12 months (p = 0.012). Furthermore, the overall survival of patients with ypM0 is comparable to patients without metastatic disease at primary diagnosis (cM0), with 48% and 51% respectively (p < 0.001; Fig. 3). There is no significant difference in overall survival between patients with resected metastases, and those without resection (9% vs 17% p = 0.427; Fig. 4). The location of metastases, resection margin or nodal status (Fig. 5) had no significant correlation for 5-year survival (p = 0.945, p = 0.273, p = 0.062). Postoperative T-stage showed an influence on 5-year survival (p = 0.047). Multivariate analysis regarding T-, N- and M-staging, R-status or regression grade showed that the absence of metastases (ypM0) is an independent predictor of overall survival (p = 0.018; RR 3.612, KI 12.46–10.470). Results are summarized in Table 3.