Involved-field irradiation in definitive chemoradiotherapy for locally advanced esophageal squamous cell carcinoma

Esophageal cancer (EC) is the fifth most common cancer and the forth leading cause of cancer deaths in China. Different from the western countries, squamous-cell carcinoma accounts for 95% of all Chinese EC patients [1]. EC is notorious for its lymph node (LN) metastases, which LN involvement is an early process and also skip metastases are common. Nodal spread of esophageal tumors may be extensive at initial clinical presentation. More than 50% EC cases are diagnosed at locally advanced stage with obvious enlargement node, long lesion and/or serious esophageal invasion. For these patients, the results of a phase III randomized trial (Radiation Therapy Oncology Group [RTOG] 85–01) comparing chemoradiotherapy (CRT) with radiation alone have made definitive CRT to one standard treatment option [2]. But in practice, the radiation field of EC has reached no global consensus till now. In RTOG 85–01, the range of clinical target volume (CTV) was from the supraclavicular region to the gastroesophageal junction. But in RTOG 94–05, 5-cm proximal and distal margins and a 2-cm lateral margin from the borders of the gross tumor volume (GTV) were recommended [3]. The supraclavicular nodes were included only when the tumor was located in the cervical esophageal area. Some recent studies showed that the involved-field irradiation (IFI, nodal target volume included only the malignant nodes) may be feasible for some sub-group patients with esophageal squamous-cell carcinoma (ESCC). Kawaguchi found that IFI did not result in significant incidence of regional LN failure in clinical stage I thoracic EC patients [4]. For patients of EC receiving definitive radiotherapy, some researchers employed three-dimensional conformal radiotherapy (3D-CRT) without intentional elective node irradiation (ENI) and the rate of isolate out-of-field nodal failure was only 2-8% [4‐6]. Moreover, some published reports indicated that serious toxicities would occur in at least 50% of patients with EC receiving concurrent chemoradiation therapy (CCRT) if ENI was adopted [3, 7]. Thus, IFI may result in reduced incidence of treatment toxicities which enable more patients to tolerate the CCRT.

To date, there have been few reports on the exclusive use of IFI for locally advanced ESCC patients. The present study sought to retrospectively document the failure patterns and survival, and evaluate the feasibility of IFI for this specific population.

In this study, we investigated treatment results and patterns of first site failure after IFI for locally advanced ESCC. We found that in-field failure and distant metastasis remained the predominant failure patterns in these cases. Among the 76 patients with failure, the rate of out-of-field regional LN failure alone was 12.5%. Zhao et al. have evaluated the results of IFI for EC patients with 3D-CRT. The rate of in-field recurrence was 44%, but only 8% cases were isolated out-of-field nodal recurrences [6]. Uno et al. reported that a radiotherapy PTV including only clinically-involved lesion for patients with EC aged 75 and older, resulted in no isolated LN recurrence [10]. The concern was that the persistence of disease was the greatest cause of treatment failure (despite therapy), and 26% of patients receiving combined therapy experienced local failure. Published data has reported that regional LN failure rates ranged from 5% to 15% for ENI in EC patients receiving radiotherapy [11]. In conclusion, for ENI or IFI,the regional LN failure was not the main pattern of recurrence in these advanced stage ESCC patients.

In present study, we also found a longer median survival time in patients with regional failure, but there was no statistic difference in OS time between different failure patterns. It was mainly caused by the overlap of failure sites in this population. According to our data, more than half (58.3%) cases were companied with other sites of failure in regional recurrence group. When we analyzed the survival with solitary failure pattern, a longest median survival time was found in the group of regional failure alone. Our further analysis showed that the patients with in-field or distant failure had worse survival. But the OS for patients with and without out-field failure showed no statistical significant difference (P = 0.665). These data may indicate that, compared with regional failure, the in-field and distant failure negatively impacted survival more significantly. The relatively high incidence of in-field recurrence and distant metastasis may mask regional nodal failures because many of the patients have died before their regional nodal failure became clinically apparent or threat to life. So the relatively high regional control acquired with ENI may not translate into the benefit of OS, especially for advanced stage cases.

And on the other hand, if only the solitary regional nodal failure occurred, a salvage treatment may provide better survival. In our study, 40% (4/10) of the cases with solitary regional failure experienced more than 24 months survival period after the salvage treatment. Another two studies also supported the efficacy of the salvage treatment for solitary regional node recurrence. Yano et al. reported the result of treatment for 35 patients with the cervical LN recurrence, and concluded that substantial survival could be attained by salvage local-regional treatment if it was a solitary node recurrence [12]. Watanabe et al. performed salvage cervical lymphadenectomy followed by adjuvant chemotherapy for all five patients with recurrence limited within the cervical nodes, and observed no disease relapse in a median follow-up period of 54 months. The median survival period after the salvage treatment was 60 months [13]. So the salvage treatment after regional LN failure alone may mitigate its adverse effect on survival.

Although all the patients included in our study were in locally advanced stage, we could still see a few patients with long-term survival, which indicated a good response to CCRT. Previous studies have demonstrated that only patients with histological complete response can acquire survival benefits [14, 15]. Unfortunately, most EC patients were resistant to chemoradiation, with only 20%–40% pathologic CR rate after definitive CCRT in advanced stage patients [16]. As high as 11–26% of those patients did not exhibit any morphological response, leading to a dismal prognosis with a median survival of only 9 months [17]. In our opinion, for those patients of complete responder, ENI may be appropriate to eliminate micrometastases in regional LN, which might get longer PFS. And for those non-responders or non complete responders,the ENI seems unnecessary,if the primary lesions could not be well controlled. So additional evaluations may be warranted to assess the sensitivity of patients to CCRT and then individualize treatments, thereby sparing patients unnecessary toxicity from ineffective therapy.

However, even though the in-field failure was the most common site (53.75%)of initial failure, patients in our study also experienced high rate (41.25%)of systemic failure. Thus, efforts to increase local control may not necessarily translate into improved survival unless systemic therapies also improved. The intensification of chemotherapy through the addition of induction chemotherapy had been in trial to strengthen the systemic therapy, but no consensus response was obtained until now [18, 19]. To improve both local and distant control in patients with EC, new regimens must be developed. Perhaps only when the high local failure rate is addressed, the value of this approach can get fully realized.

In the present study, the failure rate of more than 90% in the whole patient group and the survival data (18.8% for 3y) appeared worse than preceding reports. Some factors may have contributed, in part, to the high failure rate and worse overall survival in our study. First, all of the patients included in our study were in locally advanced stage with wide local tumor extension and/or clinically obviously node metastases, which was the most important cause for bad prognosis. Second, EUS, PET scans and other functional images were not available for all of the cases in the trial, which might lead to diagnostic underestimation, have impacted the target volume to IFI. The microscopic disease in normal sized nodes and LN enlargement caused by benign conditions, limit the diagnostic accuracy of CT for nodal enlargement caused by EC. According to a meta-analysis, the sensitivity and specificity of CT for regional lymph node metastases were 0.50 and 0.83 in thoracic tumor, respectively [20]. In the future study, if the IFI was used, the more accurate diagnostic technique should be performed to avoid the missing of involved node. Third, ENI should be really effective for providing regional control of LN micrometastases and leading to a longer PFS, especially for responding cases who had a relatively long survival.

In summary, our study found that in-field and distant failure remained the major failure patterns in patients with locally advanced ESCC treated with IFI and impacted the survival time more significantly, while the solitary nodal failure of out-of-field was acceptable and the salvage treatment for solitary regional failure may improve outcome. The omission of elective nodal irradiation did not sacrifice OS, to some extend, which suggested the feasibility of IFI for the locally advanced cases. However, this was a retrospective study with a relatively small sample size, which may limit the generalizability of our findings. Further observations with large-scale, multi-center, prospective, stratification, randomized clinical trials are needed to verify the feasibility of IFI in this sub-group population.