The first description of the esophageal NECs was reported by Mckeown in 1952 . Since then, the publicity of the disease was widened especially after the WHO definition. Though a bulk of literature has been published, most of them were case reports. And data summarizing the characteristics of the disease were still badly scarce. Compared to several published single or multicenter studies on the subject [3, 4, 9,10,11,12], our study is the largest and longest-running single center clinicopathologic study worldwide.
NECs contain neuroendocrine cells that secrete bioactive substances and the WHO definition for NET includes positive endocrine markers such as chromogranin A. In this study, we have analyzed 53 cases of NECs of the esophagus confirmed by the tissue pathology after surgery. Consistent with other studies [4, 13], our results showed that the widely accepted positive immunostaining should be synaptophysin, CD56 and chromogranin A. CK8/18 and CKpan could be candidate markers. Besides, p63, p40 and CK5/6 have been demonstrated to be potent markers of the squamous cell carcinoma or adenocarcinoma. P53 loss has been found nearly universal in poorly differentiated NECs , and it was a limitation that our center didn’t detect p53 in the surgical specimens. Interestingly, only one patient whose biopsy sample by endoscopy indicated neuroendocrine carcinoma, others reported esophageal squamous carcinoma or esophageal adenocarcinoma, which meant a high misdiagnosis from biopsy (data not shown).
Like other malignancies of the esophagus, the vulnerable patients were elderly males. Dysphagia was the most common symptom, followed by weight loss. Tumors were frequently found in the middle esophagus, and this could be the reason for the phenomenon of synchronous esophageal squamous carcinoma. It has been reported that esophageal squamous carcinomas were most frequently present in the middle esophagus [15, 16]. Protruding type with ulceration in the center was the obvious features of the pure NECs, which was different from the elevated type of the MiNECs. It could be attributed that the squamous component often overlies NECs .
Our study was a retrospective study analyzing all eligible patients from 2002 to 2018. The follow-up period was 16 years, and the unexpected thing was that some patients with pure NECs have lived longer than 5 years, since NECs are aggressive and the prognosis reported in literature is poor [4, 9,10,11,12, 17]. Specifically, there were several patients with localized pure NECs or NECs with small part of squamous cell carcinoma achieved complete response after resection. The longest survival time was more than 10 years. Several case reports also demonstrated that patients could achieve complete response after surgery despite the aggressive behavior of the disease [18,19,20]. The shared feature of these patients was that tumors were localized and could be completely removed.
The median survival year for patients with pure NECs was 3.53 years, and 7 years for patients with MiNECs, which was different from some studies. Kanakasetty et al. have summarized data in their center and concluded that the median survival time for pure NECs patients without metastasis was 18.25 months and only 6.5 months for patients with lymph node positive . In the study by Egashira et al., the median survival time of patients with non-metastasis was 17 months and as short as 8.5 months for patients with tumors outside loco-regional boundaries . Reports from western countries showed that the median survival for small cell carcinoma of the esophagus was 11 months . However, the study by Huang et al. showed a better prognosis for patients after resection. The median survival time for high-grade neuroendocrine carcinomas was 28.5 months, which was approximate to ours. Lymphovascular and organ invasion were found to be of prognostic value [4, 9, 11, 13]. And Lee et al. reported that tumor size (more than 2.0 cm) showed prognostic significance . Different from these results, we found that there were no prognostic factors for patients with pure NECs. Once the patient was definitely diagnosed, his median survival time was 3.53 years, and the 5-year survival rate was 46%, independent from any factor such as tumor size. For patients with MiNECs, age and pathologic stage were found to be significantly associated with prognosis. Patients older than 65 or pathologic stage higher than IIIA had worse prognosis. Overall, the 5-year survival rate for patients with MiNECs was 60%. The heterogeneity of the prognosis could be attributed to many factors, and the main one should be the small sample size of all of the studies. Others could be the quality of pathologic reports or the inclusion criteria of patients or treatment modality.
Due to lack of sufficient data about the clinicopathological characteristics of esophageal NECs, there has no established treatment yet. Surgery, chemotherapy and radiotherapy are recommended [22, 23]. In our center, for operable patients, surgery was the first choice. Even for inoperable NECs, palliative surgery was conducted. Adjuvant chemotherapy was always followed if the patient’s condition was allowed. In the study, except for seven patients who lost to follow-up, the remaining patients received chemotherapy after resection. There was one patient receiving radiotherapy. Although the co-existence of squamous cell carcinoma with NECs makes complete response by chemotherapy difficult, our study indicated that adjuvant chemotherapy for operable patients provided a better prognosis, which was also reported in other studies [1, 3, 24]. Nevertheless, detailed information about adjuvant therapy was omitted because most patients received chemotherapy in local hospitals. Cisplatin/etoposide and cisplatin/irinotecan were two major regimens for those patients in our center, which was also confirmed in other studies [25,26,27]. Furthermore, recent evidence from sequencing provides potential therapeutic strategies on the basis of genetics and epigenetics .
A major limitation of the work is that it was a retrospective study, and therefore, some data such as chemotherapy information were omitted and pathologic reports were not standardized. As a result, immunohistochemical profile of every case was not consistent. However, a prospective trial cannot be envisioned because of the rarity of the disease. The overall sample size is reasonably large for the rare disease. Besides, our retrospective study has a relatively low rate of loss to follow-up (13.2%) in the follow-up period of 16 years. Overall, the validity of our data is strengthened by the fact that the study included patients who were diagnosed by resection specimens rather than clinical characteristics or biopsy samples. Regardless, the findings of the study require larger studies to confirm and update.