Patients characteristics
A total of 83 patients with recurrence after curative resection of EC (71 males and 12 females) meeting the inclusion criteria were included. The median age of the patients was 56 years (range: 37–79 years). A total of 11 (13.3%), 45 (54.2%), and 27 (32.5%) tumors were located in the upper, middle, and lower third, respectively. The number of patients with pathologically confirmed primary tumors classified as pT1, pT2, pT3, and pT4 was 5 (6%), 16 (19.3%), 47 (56.6%), and 15 (18.1%), respectively. The number of patients with pathologically confirmed regional lymph nodes (pN) classified as pN0, pN1 (1–2 positive LNs), pN2 (3–6 positive LNs), and pN3 (7 or more positive LNs) were 27(32.6%), 21 (25.3%), 25 (30.1%), and 10 (12%), respectively. The number of patients with tumor cell differentiation in G1, G2, and G3 was 6 (7.2%), 59 (71.1%), and 18 (21.7%), respectively. The number of patients treated with two- and three-field lymphadenectomy was 26 (31.3%) and 57 (68.7%), respectively. Nine patients received neoadjuvant chemotherapy, and 54 patients received adjuvant chemotherapy.
Recurrence pattern of OLR and non-OLR
The sites of recurrence in the OLR group (n=35; 42.2%) were the lungs (n=18; 51.5%), liver (n=6; 17.1%), bones (n=4; 11.4%), and brain (n=7; 20%). The median survival times (MSTs) after recurrence for metastases located in the lung only, liver only, bone only, and brain only were 30, 13, 15, and 16 months, respectively. The 3-year OS rates after recurrence for the lung, liver, bone, and brain only were 38.9, 0, 25, and 14.3%, respectively. The MST and 3-year OS of patients with OLR with liver oligometastasis were significantly worse than those of patients with lung oligometastasis (13 months and 0% versus 30 months and 38.9%, respectively; P=0.009). Statistical analysis showed no significant difference in MST and 3-year OS in brain oligometastasis (P=0.056) and bone oligometastasis (P=0.644) compared with lung oligometastasis (Table 1).
Conversely, 30 patients (36.1%) in the non-OLR group had a recurrence in 23 organs with regional recurrence and/or regional lymph node metastases and 18 (21.7%) without regional recurrence and/or regional lymph node metastases. The MST after recurrence for patients with EC in the non-OLR group with and without regional recurrence and/or regional lymph node metastases was 11 and 16 months, respectively. The survival curves after recurrence in patients with EC with non-OLR according to regional recurrence and/or regional lymph node metastases are shown in Figure 1a (1-, 2-, and 3-year OS of 46.3, 14.6, and 9.8% versus 57.1%, 14.3%, and 0%, respectively; P=0.829). In this study, regional recurrence and/or regional lymph node metastasis in the non-OLR group were not associated with long-term survival after recurrence.
OS after the recurrence of OLR and non-OLR
The MST after recurrence for patients with EC in the OLR and non-OLR groups was 17 and 8 months, respectively. For patients with EC in the OLR group (n=35), the 1-, 2-, and 3-year OS rates after recurrence were 71.4%, 42.9%, and 25.7%, respectively. For patients with EC in the non-OLR group (n=48), the 1-, 2-, and 3-year OS rates after recurrence were 47.9%, 14.6%, and 8.3%, respectively. Based on these results, OS after recurrence was significantly better for patients with EC in the OLR group than in those in the non-OLR group (P=0.030) (Figure 1b).
To identify the factors affecting long-term survival after recurrence in the entire cohort, we performed univariate analysis and demonstrated that tumor location (P=0.005), pathological lymph node (pN) stage (P<0.001), and OLR (P=0.030) were significantly associated with OS after recurrence. Multivariate analysis confirmed that the pN stage (P=0.003), OLR (P=0.003), and DFI (P=0.010) were independent prognostic factors (Table 2).
OS after recurrence in patients with OLR or non-OLR according to the treatment method
Fifty-seven patients received two or more cycles of CT alone after recurrence, and 26 patients received CCLT after recurrence. The local treatment for patients with EC after recurrence was surgical resection (n=2, 7.7%), SBRT or conventional radiotherapy (n=19, 73.1%), or radiofrequency ablation (n=4, 19.2%). The MST after recurrence for patients with EC in the OLR group treated with CT and CCLT was 17 and 32 months, respectively. For patients with EC in the OLR group treated with CT (n=26) or CCLT (n=9), the 1-, 2-, and 3-year OS rates after recurrence were 60.9%, 30.4%, and 13.0% versus 91.7%, 66.7%, and 50%, respectively. Based on these results, OS after recurrence was significantly better for patients with EC in the OLR group treated with CCLT than in those treated with CT (P=0.003) (Figure 2a). For patients with EC in the non-OLR group treated with CT (n=34) or CCLT (n=14), the 1-, 2-, and 3-year OS rates after recurrence were 38.2%, 11.8%, and 8.8% versus 71.4%, 21.4%, and 7.1%, respectively. From these results, the OS after recurrence was not significantly different between patients with EC in the non-OLR group treated with CT and those treated with CCLT (P=0.225) (Figure 2b).
To identify the factors affecting OS after recurrence in patients with OLR, we performed univariate analysis and demonstrated that the treatment method (P=0.003) and pN stage (P<0.001) were significantly associated with OS after recurrence. Multivariate analysis confirmed that the degree of differentiation (P=0.019), treatment method (P=0.016), and pN stage (P=0.015) were independent prognostic factors (Table 3).