Discussion
This preSINO study is a diagnostic trial, investigating the combination of diagnostic modalities that has been tested previously in the preSANO trial. It will provide evidence concerning the optimal diagnostic strategy for detecting residual locoregional disease after nCRT and, if successful, will be applied in a future SINO trial comparing active surveillance with standard esophagectomy in patients with esophageal SCC.
Esophageal SCC is a major public health issue for Central, East- and South-East Asian countries. Highest incidences are found in these areas with rates up to 13.6 per 100,000 in men and 4.3 in women (compared to a global average of 7.7 in men and 2.8 in women). In China, esophageal cancer is the fourth most common cancer, making it a major (financial) burden for Chinese national healthcare [
20]. Hence, a potential reduction in the number of esophagectomies might not only have a positive impact on patients’ wellbeing, but also on the burden of esophageal cancer on Chinese healthcare systems.
Active surveillance after nCRT has been proposed for esophageal cancer since high pCR rates are being achieved by current nCRT regimens [
5]. The Dutch CROSS trial showed that 5-year overall survival significantly increased from 33 to 47% by applying nCRT consisting of carboplatin and paclitaxel with concurrent 41.4 Gy radiotherapy prior to surgical resection [
3]. Following this nCRT regimen, 49% of patients with SCC and 23% of patients with AC had pCR. However, only a small part of the included patients had SCC (23%) [
4]. The Chinese NEOCRTEC 5010 trial showed that 3-year overall survival increased from 58.9 to 69.1% by giving nCRT consisting of vinorelbine and cisplatin with concurrent 40.0 Gy radiotherapy prior to surgical resection [
21]. In this Chinese trial including only esophageal SCC, 43.2% of patients achieved pCR. Although pCR rates in patients with SCC are comparable between both regimens (49% vs. 43.2%), the CROSS regimen has been proposed as optimal nCRT regimen for the preSINO trial and future SINO trial. The major issue with nCRT according to the NEOCRTEC 5010 regimen is the relatively high toxicity rate. In the CROSS trial fewer grade 3–4 adverse hematological events (7.6%) occurred compared to 54.3% in the NEOCRTEC 5010 trial [
4,
21].
The significance of the current preSINO study is that after nCRT according to CROSS, pCR rates tend to be higher in patients with SCC compared to patients with AC (49% vs 23%) [
4]. In these patients surgical resection can potentially be postponed or even omitted. Hence, almost half of patients with SCC could benefit from an active surveillance strategy. However, the results of the preSANO trial cannot be simply extrapolated to the planned SINO trial since the preSANO trial included only 21 patients (17%) with SCC who underwent bite-on-bite biopsies and 43 patients with SCC overall (21%) [
8].
Compared to the preSANO trial, the current study differs on several aspects. In the preSANO and subsequent SANO trial, operable patients who underwent or who were planned to undergo nCRT according to the CROSS regimen are/were considered eligible [
7,
8]. However, patients with AJCC cN3 disease have a high risk of an incomplete locoregional response and a low chance of remaining tumor-free during follow-up in the planned future SINO trial. Also, these patients often have suspected lymph nodes outside the maximum tolerated radiation field and can therefore not receive CROSS chemoradiotherapy. Since inclusion of cN3 patients in the preSINO and SINO trial could reduce the proportion of cCR and pCR, patients with cN3 disease will not be included in the preSINO trial. Moreover, in the preSANO trial PET-CT was used to identify distant metastases at baseline and preoperatively. In the preSINO trial, patients who have a positive CRE-I (i.e. histologically proven locoregional residual disease or a non-traversable tumor) are also allowed to receive a high-dose CT scan instead of a PET-CT scan in order to identify distant metastases. Although high-dose CT scan is inferior to PET-CT scan in detecting distant metastases for esophageal cancer [
22], Chinese insurance does not yet cover a preoperative PET-CT scan. However, both in the present preSINO trial and in the planned subsequent SINO trial the group of patients with a positive CRE-I (and CRE-II) will be excluded. Therefore, performing a high dose CT scan instead of PET-CT scan after a positive CRE-I will not affect the preSINO and SINO trials.
Furthermore, in the preSANO trial the reference standard that has been used for calculating the accuracy of the CREs was TRG in the resection specimen [
8]. This grading system evaluates the response of the primary tumor site to neoadjuvant treatment only. However, the modalities used in the CREs also evaluate residual nodal disease. After nCRT, 4–9% of patients with SCC have a pathologically complete response at the primary tumor site with residual disease in the regional lymph nodes (ypT0N+), compared to 3–5% of patients with AC [
23‐
25]. For these patients, higher ypN stage is correlated with worse overall survival [
25]. However, these studies did not clearly specify the nCRT regimens that patients received. Moreover, the majority of patients probably did not receive nCRT according to the CROSS regimen since they were included before the results of the CROSS trial were published [
23,
24]. Accordingly, data of a large Dutch CROSS cohort comprising patients that were included in the CROSS-I and CROSS-II trials, post-CROSS cohort and the preSANO trial were analyzed. In this large CROSS cohort, 8 of 122 patients with SCC (7%) and 11 of 415 patients with AC (3%) who underwent nCRT followed by a surgical resection had ypT0N+ stage. The percentage of patients with ypT0 who had ypN+ was comparable between SCC and AC (14% vs. 13%, resp.), suggesting that the higher rate of ypT0N+ patients is mainly caused by the better response of SCC to nCRT according to CROSS. In an active surveillance strategy that focusses on the primary tumor site only, patients who have ypT0N+ stage run a high risk of developing distant metastases. Therefore, in contrast to the preSANO trial, the preSINO trial will aim to identify patients with residual disease at the primary tumor site as well as in regional lymph nodes.
Similar to the preSANO trial, this trial will use bite-on-bite biopsies during endoscopic response evaluation. After nCRT for esophageal cancer, residual disease is found in deeper layers than the mucosa without residue in the mucosa itself in 24–40% of patients [
26,
27]. In case of bite-on-bite biopsies, the second biopsy is taken at the exact same location as the first biopsy (Fig.
2). It has been suggested that this biopsy technique has the potential to reach into the submucosa, thereby increasing the rate of detecting residual tumor. The preSANO trial was started with regular endoscopic biopsies. However, sensitivity for detecting TRG3–4 residual tumor improved from 69 to 90% after the protocol had been changed from regular biopsies to bite-on-bite biopsies. Moreover, the preSANO trial demonstrated that performing FNA of suspected lymph nodes substantially improved detection of residual disease. It was shown that 29% of residual disease detected during CRE-2 alone was based on results of FNA only. Based on bite-on-bite biopsies only, sensitivity of both CRE-1 and CRE-2 combined for detecting TRG3–4 residual tumor was 83%. With the addition of FNA, this increased to 90% [
8].
As demonstrated by the preSANO trial, the additional value of qualitative and quantitative PET-CT analysis in detecting early residual locoregional disease (up to 12 weeks after completion of nCRT) remains disputable. Qualitative PET-CT analysis alone yielded a reasonable sensitivity of 85% for detecting TRG3–4 residual tumor and a sensitivity of 80% for detecting TRG2–4 residual tumor. However, corresponding specificity was only 37%, caused by a high number of false-positives [
8,
11]. After nCRT, radiation-induced esophagitis apparently increases 18F-FDG uptake throughout the irradiated esophagus including the primary tumor site, thereby increasing false-positive response evaluations. PET-CT detected novel distant metastases (i.e. not detected at baseline) in 9% of patients during CRE-2, saving them from unnecessary surgical resection [
8,
11]. For this reason, in the current study PET-CT will only be used to detect distant metastases and will not be used for evaluating the primary tumor site. In the planned future SINO trial investigating an actual active surveillance strategy, PET-CT might become valuable as esophagitis will probably diminish over time, allowing detection of disease recurrence by qualitative and semi-quantitative analysis of small increments in 18F-FDG uptake.
Moreover, PET-CT will be used in follow-up to compare the distant dissemination rate between both arms of the future SINO trial. For patients undergoing active surveillance in the future SINO trial there is a hypothetical risk of increased distant dissemination since residual vital tumor might be missed during response evaluations. Current assumption holds that this process of spreading and seeding of tumor cells from the primary lesion is an early event. Hence, distant dissemination might have already happened at the time of diagnosis or during locoregional treatment [
28]. This assumption is reflected by the substantial number of patients who develop hematogenous metastases after nCRT plus esophagectomy within 2 years after surgery [
3,
4]. For this reason, PET-CTs will be performed until 2 years after surgical resection in the control arm (immediate surgery) of the planned future SINO trial. Since patients with a clinically complete response in the preSINO trial can be included in this arm of the SINO trial, these patients will also undergo follow-up PET-CT scans.
Although the preSANO trial showed that the combination of endoscopic bite-on-bite biopsies, EUS, and PET-CT was most accurate for detecting residual disease after nCRT, other measurements and modalities for response evaluation have been considered. To our knowledge, however, no other single modality or set of modalities has yet acquired an accuracy that equals the set from the preSANO trial. Moreover, because PET-CT, endoscopic biopsies and EUS are widely used for pretreatment clinical staging, physicians have experience with these techniques and they are available in most Chinese centers. One of the other measurements considered for the preSINO trial was the maximum tumor thickness on EUS as proposed by Jost et al. [
29] In a cohort of 40 patients they found a sensitivity of 86% and specificity of 64% for detecting TRG 2–4 residual tumor. Based on these results, the preSANO trial investigated this method in 123 patients. In this larger cohort, a sensitivity of only 59% and specificity of 58% was found. Since these measurements did not add to the accuracy of the set of diagnostic modalities in the preSANO trial, it was decided that they are not included in the preSINO trial protocol [
8]. MRI is another promising modality, showing sensitivities up to 97% for detecting residual tumor after nCRT [
30,
31]. However, studies that find these high sensitivities all have poor corresponding specificities of at most 58%. In a future SINO trial, this overstaging of patients with a complete response to nCRT would result in unnecessary surgery. Blood-based biomarkers, including circulating tumor DNA, could also become useful for less invasive detection of residual disease. Methods have improved over the last few years, especially for predicting distant recurrences [
32,
33]. To date, however, no biomarker has been identified that can accurately discriminate patients with incomplete locoregional response from patients with complete locoregional response.
If the preSINO trial shows that PET-CT, endoscopic bite-on-bite biopsies and EUS with FNA can detect TRG3–4 residual tumor in patients with esophageal SCC with a sensitivity of at least 81.5%, this combination of diagnostic tests will be used in a subsequent prospective trial comparing active surveillance with standard esophagectomy in patients with SCC and a clinically complete response after nCRT (SINO trial). Similar to the preSANO and SANO trials, it will be possible to include all patients of the preSINO trial having cCR into the control arm of the future SINO trial [
7]. Of patients with SCC receiving nCRT according to CROSS, almost 50% achieve pCR and more than 50% is expected to have cCR. Consequently, the number of study subjects needed for the SINO trial can be substantially decreased compared to the SANO trial. Most probably, all patients needed for the control arm of the SINO trial will be derived from the preSINO trial.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.