博 士 学 位 論 文

博 士 学 位 論 文

進 行 食 道 癌 に お け る 術 前 化 学 療 法 に よ る リンパ節微小転移の制御

平 木 洋 子

進行食道癌における術前化学療法によるリンパ節微小転移の制御平木洋子

近 畿 大 学 大 学 院 医 学 研 究 科 医 学 系 専 攻

Doctoral Dissertation

Controlling lymph node micrometastases by neoadjuvant chemotherapy affects the prognosis in advanced esophageal squamous cell carcinoma

Yoko Hiraki August 2020

Department of Surgery, Major in Medical Sciences

Kindai University Graduate School of Medical Sciences

Controlling lymph node micrometastases by neoadjuvant chemotherapy affects the prognosis in advanced esophageal squamous cell carcinoma

Yoko Hiraki¹, Yutaka Kimura¹, Motohiro Imano¹, Hiroaki Kato¹, Mitsuru Iwama¹, Osamu Shiraishi¹, Atsushi Yasuda¹, Masayuki Shinkai¹, Tomoki Makino², Masaaki Motoori³,

Makoto Yamasaki², Hiroshi Miyata⁴, Takao Satou⁵, Taroh Satoh⁶, Hiroshi Furukawa¹, Masahiko Yano⁴, Yuichiro Doki², Takushi Yasuda¹

1Department of Surgery, Kindai University Faculty of Medicine

2Department of Gastroenterological Surgery, Graduate School of Medicine, Osaka University

3Department of Surgery, Osaka General Medical Center

4Department of Surgery, Osaka International Cancer Institute

5Department of Diagnostic Pathology, Kindai University Hospital

6Department of Frontier Science for Cancer and Chemotherapy, Graduate School of Medicine, Osaka University

Abstract

　Purpose. The purpose of this study is to determine the clinical significance of micrometastases after neoadjuvant chemotherapy （NAC） and the difference in controlling micrometastases using different NAC regimens in resectable advanced esophageal squamous cell carcinoma （ESCC）.

　Methods. We analyzed patients with ESCC who underwent esophagectomy with lymph node dissection after NAC with Adriamycin + cisplatin + 5-fluorouracil （ACF） or docetaxel + cisplatin + 5-fluorouracil （DCF）. Micrometastasis was defined as a single isolated cancer cell or cluster of cancer cells on the cervical, recurrent nerve, or abdominal LNs as shown by immunohistochemical staining with anti-cytokeratin antibody （AE1/AE3）. The associations between micrometastases, recurrence, prognosis, and regimen differences were investigated.

　Results. One hundred and one cases （ACF group: 51 cases; DCF group: 50 cases） were analyzed.

Micrometastases occurred in 24 patients （23.8%）: 17/51 （33.3%） in the ACF group and 7/50 （13.5%） in the DCF group （P=0.0403）. The 5-year recurrence-free survival （RFS） rates for patients without （n=77） and with （n=24） micrometastases were 62% and 32%, respectively （hazard ratio, 2.158; 95% confidence interval, 1.170–3.980; stratified log-rank test, P=0.0115）. A multivariate analysis showed that stage pN1 or higher and micrometastases were significant risk factors affecting RFS.

　Conclusion. In resectable advanced ESCC, controlling micrometastases in the LNs after NAC varied by regimen and may be associated with preventing ESCC recurrence.

Keywords： esophageal squamous cell carcinoma, neoadjuvant chemotherapy, micrometastases

INTRODUCTION

　Advanced esophageal squamous cell carcinoma （ESCC） has a poor prognosis, and multidisciplinary treatment combining chemoradiotherapy （CRT） and surgery is regarded as the standard treatment worldwide ^［1−4］. Although the Japanese guidelines on the clinical practice for esophageal cancer recommend neoadjuvant chemotherapy （NAC） with cisplatin/5-fluoruracil （CF） for resectable stage II/III ESCC, this standard treatment has a low response rate of 30–40% and does not improve the prognosis for cStage III cancer ^［5−8］. Comprehensive registry data in Japan show that the 5-year survival rate for cStage II cancer is 50–60%, whereas the 5-year survival rate for cStage III cancer is approximately 40%, and the prognosis worsens as the stage progresses ^［9］. Thus, researchers seek to develop a NAC regimen with an improved antitumor effect. As epirubicin was added to CF in Europe, Adriamycin was added to CF in Japan, with Adriamycin/CF （ACF） therapy yielding a reported response rate of approximately 50%, slightly higher than that of CF alone ^{［10−12］}. The authors added docetaxel to CF to create DCF therapy and conducted a multicenter randomized controlled trial to demonstrate the effectiveness of ACF vs. DCF NAC （OGSG1003）

in treating resectable advanced ESCC. The response rates from our previous study for the main tumor were 48% （ACF） vs. 77% （DCF）, showing a favorable histological result for DCF ^{［13, 14］}. The 2-year recurrence- free survival （RFS） rates were 45.7% vs. 66.3%, respectively, yielding a significantly better result for DCF

（p=0.003） ^［15］. Although DCF provided more effective local control compared with that of ACF, the number of pathologically metastatic lymph nodes （LNs）, which is strongly correlated with postoperative recurrence, did not differ significantly between the groups. The goal of NAC is to improve the resection rate via its antitumor effect, prevent intraoperative dissemination, and control systemic micrometastases. The recurrence of distant metastases after NAC and surgery may be particularly associated with micrometastases, and the type of NAC regimen may alter the control of micrometastases. Distant LN recurrences are occasionally seen in patients diagnosed with no LN metastases via conventional histopathological examinations （i.e., N0）. Thus, micrometastases that cannot be detected by conventional methods are reported to affect the prognosis ^{［16, 17］}. Nevertheless, no studies have so far reported the differences between NAC regimens or the relationship between NAC regimen and prognosis and recurrence relative to micrometastasis occurrence after NAC for ESCC.

　This study determined the clinical significance of micrometastases after NAC and the difference in controlling micrometastases via NAC regimens in resectable advanced ESCC.

PATIENTS AND METHODS Patients

　This study accompanied the prior randomized phase II trial （OGSG1003）. In this study, we examined patients with resectable advanced ESCC registered in the original trial, who underwent esophagectomy after NAC with ACF or DCF. Of the 162 patients enrolled in the OGSG1003 study between November 2010

─ 2 ─

and October 2012, this study examined patients from 4 hospitals that had a sufficient number of registered cases and could provide excised specimens. Written informed consent was obtained from all patients after ethics committee approval. This study was conducted in consecutive patients recruited into the original trial, which was registered in the University Hospital Medical Information Network Clinical Trials Registry of Japan （identification number UMIN000004555/000004616）. Details of the eligibility criteria and methods of the OGSG1003 study have been previously reported ^{［14, 15］}. The eligibility criteria were as follows: （i） histologically proven squamous cell carcinoma of the thoracic esophagus; （ii） clinical confirmation according to the seventh edition of the Tumor Node Metastasis Classification of the International Union Against Cancer （UICC-TNM）

as T1–T4a, any N category ^［18］; （iii） M0 or M1LYM metastasis （confined to the supraclavicular LNs）; （iv）

age ≥20 years; （v） Eastern Cooperative Oncology Group performance status of 0–1; （vi） no serious vital organ dysfunction （heart, pulmonary, liver, renal, or hematologic）; and （vii） provision of written informed consent before randomization.

Procedures

　A preoperative NAC regimen of either 35 mg/m² Adriamycin, 70 mg/m² cisplatin 1-hour intravenous infusion on day 1, and 700 mg/m² 5-fluorouracil continuous infusion for 7 days （ACF） every 4 weeks or 70 mg/m² docetaxel, 70 mg/m² cisplatin 1-hour intravenous infusion on day 1, and 700 mg/m² 5-fluorouracil continuous infusion for 5 days （DCF） every 3 weeks was randomly allocated to patients after confirming eligibility. After 2 courses of NAC, radical esophagectomy was performed using the Ivor-Lewis or McKeown method with two or three-field lymphadenectomy through a right thoracotomy or video-assisted thoracoscopic surgery. Regional lymphadenectomy included the mediastinal, perigastric, and celiac nodes and distant lymphadenectomy included the cervical nodes. Neither group received adjuvant chemotherapy.

No treatment was prescribed for the presence of residual cancer after resection or for recurrence. Patients were examined every 3 months for the first 2 years after the date of random assignment, every 6 months for the next 3 years, and annually after that. The follow-up and diagnosis of recurrence were performed via computed tomography （CT）, integrated fluorodeoxyglucose positron emission tomography/CT, ultrasound, endoscopy, blood tests, and physical examinations. The RFS and overall survival （OS） rates were calculated from the date of random assignment. The histopathological tumor response was evaluated according to the histological criteria of the Japanese Society for Esophageal Disease ^［19］. Briefly, evaluations were classified into five categories according to the proportion of viable residual tumor cells affected by degeneration or necrosis.

Immunohistochemistry

　In the present study, cervical paraesophageal LN specimens （Nos. 101R and 101L）, recurrent nerve LNs

（Nos. 106R and 106L）, and the paracardial and lesser curvature LNs （Nos. 1, 2 and 3）, which are frequently involved in LN metastasis in surgically dissected LNs and demonstrate high efficacy for LN dissection, were

examined.²⁰ Consecutive 4-μm histological sections were created from paraffin blocks of lymph nodes that had been fixed in 10% formalin and embedded. These sections were stained with hematoxylin-eosin （HE） and immunostained with cytokeratin （CK） AE1/AE3.

　Immunohistochemical staining was performed using the avidin-biotin-peroxidase complex technique.

After deparaffinization, the antigens were activated by autoclaving, and immunostaining was performed using anti-CK antibody, AE1/AE3 （Nichirei Biosciences, Tokyo, Japan） ^{［17, 21］}. The sections were incubated overnight at 4℃ with mouse anti-CK antibody at the recommended dilution of 1:100. Negative control staining was performed using normal mouse IgG instead of anti-CK antibody. Esophageal epithelium and cancer specimens were used as positive controls for CK.

Micrometastasis

　Micrometastasis was defined as a single isolated cancer cell or small cluster of cancer cells not more than 0.2 mm in greatest extent when stained with the anti-CK antibody according to the UICC TNM classification, even if the LNs were judged to be uninvolved via standard histological examination with HE staining

（Figure 1） ^［18］. Micrometastasis was examined in one section per node, similar to standard histological examinations. If fibrosis or xanthogranulomatous changes believed to be caused by chemotherapy were found near CK-positive cells, micrometastasis was not diagnosed. Two pathologists separately determined the presence of micrometastasis without any clinical information. Where there was disagreement, final agreement was achieved using a two-headed microscope. The whole area of each LN within a section was examined.

When the micrometastasis was observed via immunohistochemical staining using anti-CK antibody in the LN without LN metastasis, the patient was classified as MM（+）. Conversely, when no micrometastasis was observed, the patient was classified as MM（−）.

Figure 1． Identification of lymph node micrometastasis （MM）. MM with a single isolated cancer cell （a） or cluster of cancer cells （b） detected by cytokeratin immunostaining. Each scale bar represents 20 µm.

a b

─ 4 ─

Statistical analysis

　A statistical analysis was performed using JMP Pro, version 13.0 software program （SAS Institute, Tokyo, Japan）.

The relationships between NAC regimen, micrometastasis, and other factors were analyzed using chi-squared tests, Fisher’s exact test, t-tests and the Mann-Whitney U test. Survival （RFS and OS） was analyzed using the Kaplan-Meier method, and groups were compared via the log-rank test. A proportional hazards model was used to perform Cox regression multivariate analyses. Clinicopathological factors considered to be associated with prognosis on the basis of a previous report were selected as variables ^［16］. P values less than 0.05 were considered to be statistically significant.

RESULTS

Patient demographics and factors associated with micrometastases

　This study examined 101 patients （51 in the ACF group and 50 in the DCF group） from 4 hospitals, and 1157 LNs were judged metastasis-free based on histopathological examination via HE staining （Figure 2）. Forty-three micrometastases were found in 29 LNs in 24 （23.8%） of the 101 patients. The median number of LNs per patient in which micrometastases were identified was one （range 1–3） and the median number of micrometastases per patient was also one （range 1–4）. Table 1 shows the relationship between the patient characteristics and the clinicopathological factors by micrometastases. Micrometastases were found in 17/51 patients （33.3%） in the ACF group and in significantly fewer patients （7/50, 14.0%） in the DCF group （P=0.0403）.

Figure 2． CONSORT diagram. ACF, Adriamycin + cisplatin + 5-fluorouracil; DCF, docetaxel + cisplatin + 5-fluorouracil.

Table 1. Micrometastases and the clinicopathological findings

MM（−） （n=77） MM（＋） （n=24） P-value Age, years

Median （range） 67（46-78） 67.5（51-76） 0.935

Sex Male：Female 68：9 19：5 0.314

Performance Status （ECOG）

0：1 63：14 20：4 >0.9999

Location

Ut：Mt：Lt 8：41：28 3：11：10 0.7799

Clinical T

1：2：3 2：14：61 0：6：18 0.7134

Clinical N

0：1：2：3 17：34：25：1 2：16：3：3 0.6220

Clinical M

0：1 68：9 24：0 0.1100

Clinical stage

IB 6 0 0.8153

IIA 11 2

IIB 8 6

IIIA 24 10

IIIB 18 3

IIIC 1 3

IV 9 0

Regimen

ACF：DCF 34：43 17：7 0.0403

Operative time, min

Median （range） 484 （328-780） 506 （344-960） 0.1996 Blood loss, g

Median （range） 570 （109-3460） 527 （40-31740） 0.4050 LN dissection

2-field：3-field 28：49 6：18 >0.9999

No. of LNs dissected

Median （range） 56 （19-108） 77 （16-139） 0.0512 Pathological T

0 5 0 0.2967

1a 6 1

1b 14 4

2 13 6

3 35 10

4a 2 2

4b 2 1

Pathological N

0：1：2：3 28：26：19：4 11：10：1：2 0.2151

Pathological M

0：1 75：2 22：2 0.2386

Pathological stage

0 3 0 0.6447

IA 10 3

IB 8 3

IIA 8 3

IIB 7 3

IIIA 22 5

IIIB 9 1

IIIC 8 4

IV 2 2

Residual tumor

R0：R1/2 72：5 22：2 0.6692

Pathological response （Grade）

0 4 3 0.0346

1a 29 13

1b 21 4

2 18 4

3 5 0

MM, micrometastasis; ECOG, Eastern Cooperative Oncology Group; LN, lymph node; ACF, Adriamycin + cisplatin + 5-fluorouracil; DCF, docetaxel + cisplatin + 5-fluorouracil.

─ 6 ─

Relationship between micrometastases and survival or recurrence

　The 5-year RFS rates in the MM（+） and MM（−） patients were 33.3% and 62.3%, respectively （HR, 2.158; 95% CI, 1.170–3.980; stratified log-rank test, P=0.0115; Figure 3a）. The 5-year OS rates were 50.0% and 66.2%, respectively （HR, 1.725; 95% CI, 0.869–3.423; stratified log-rank test, P=0.1145; Figure 3b）. The number of micrometastases did not significantly affect RFS or OS （data not shown）. An analysis of MM（+） and MM（−） patients showed that recurrence occurred in a significantly higher proportion of MM（+） patients

（P=0.0014）. Locoregional recurrence did not differ between MM（+） and MM（−） patients （P=0.1474）, but metastasis recurrence, especially distant LN recurrence, occurred significantly more frequently in MM（+）

patients （P=0.0359; Table 2）.

Table 2. The site of recurrence according to micrometastasis

recurrence MM（−） MM（＋） p

（n=77） （n=24）

total − 50 6 0.001

＋ 27 18

locoregional total 13 8 0.147

metastatic total 19 12 0.036

distant LN 7 8 0.010

lung 4 2 0.653

liver 4 4 0.089

pleura 5 3 0.391

other 3 1 >0.9999

MM, micrometastasis; LN, lymph node.

Figure 3． Kaplan-Meier estimates of the 5-year recurrence-free survival （a） and overall survival （b） in all cases. MM, micrometastasis.

Analysis by pathological N

　Among 39 patients with pathological （p） N0, micrometastases were found in 11 patients （28.2%）. In the patients with pN0, the 5-year RFS rates in the MM（+） and MM（−） patients were 54.5% and 89.3%, respectively （HR, 4.748; 95% CI, 1.133–19.898; stratified log-rank test, P=0.0187; Figure 4a）. The 5-year OS rates were 81.8% and 92.9%, respectively （HR, 2.634; 95% CI, 0.371–18.709; stratified log-rank test, P=0.3143;

Figure 4b）.

　In the patients with pN1-3, the 5-year RFS rates in the MM（+） （n=13） and MM（−） （n=49） patients were 15.4% and 46.9%, respectively （HR, 2.370; 95% CI, 1.162–4.833; stratified log-rank test, P=0.0176; Figure 4c）. The 5-year OS rates were 23.1% and 51.0%, respectively （HR, 2.270; 95% CI, 1.075–4.794; stratified log- rank test, P=0.0316; Figure 4d）. In the patients with pN1, HR of RFS in the MM（+） （n=10） compared with the MM（−） （n=26） patients was 7.684 （95% CI, 2.520–23.434, stratified log-rank test, P=0.0003）, but in the patients with pN2-3, HR of RFS in the MM（+） （n=3） was 0.805 （95% CI, 0.186–3.479, stratified log-rank test, P=0.7717）.

Figure 4． Kaplan-Meier estimates of the 5-year recurrence-free survival in pathological N0 （a）, overall survival in pathological N0 （b）, the 5-year recurrence-free survival in pathological N1–3 （c） and overall survival in pathological N1–3 （d）.

MM, micrometastasis.

─ 8 ─

Comparison of survival between ACF and DCF regimen

　Regarding the patient backgrounds by regimen, the LN dissection, residual tumor, and pathological grading did not differ significantly between the two groups, but patients who received the DCF regimen had significantly better histopathological tumor response grades （ESM_1.pdf）. The 5-year RFS rates in the ACF and DCF groups were 45.1% and 66.0%, respectively （hazard ratio ［HR］, 1.879; 95% confidence interval ［CI］, 1.027–3.435; stratified log-rank test, P=0.0373）. The 5-year OS rates were 56.9% and 68.0%, respectively （HR, 1.554; 95% CI, 0.816–2.962; stratified log-rank test, P=0.1761） （ESM_2.pdf）.

Multivariate analysis for survival

　Examining the risk factors of RFS via a multivariate analysis with the pathological factors and regimens considered to be involved in recurrence and tumor progression revealed that pN1 or higher and a MM（+）

status were significant risk factors （Table 3）.

Table 3. A multivariate analysis for the recurrence-free survival

Variable Risk ratio 95%CI p

Regimen ACF （DCF） 1.276 0.614-2.652 0.514

No. of LNs dissected <60 （> 60） 1.474 0.800-2.716 0.214

pT3-4 （pT0-2） 1.404 0.684-2.882 0.355

pN1-3 （pN0） 5.181 2.273-11.765 <0.0001

pM1 （pM0） 1.553 0.441-5.464 0.493

Pathological response Grade 0-1a （Grade 1b-3） 1.611 0.720-3.602 0.246

MM（＋） ［MM（−）］ 2.488 1.291-4.792 0.006

CI, confidence interval; ACF, Adriamycin + cisplatin + 5-fluorouracil; DCF, docetaxel + cisplatin + 5-fluorouracil; LN, lymph node; MM, micrometastasis.

DISCUSSION

　This study, conducted as an adjunct to a multicenter randomized controlled trial （OGSG1003）, was the first to show that controlling lymph node micrometastases via NAC in advanced ESCC was significantly associated with an improved RFS in patients at stage pN0 as well as those at stages pN1–3 and that control by NAC varies by regimen. Patients with ESCC who underwent surgery after receiving NAC with ACF or DCF showed a better RFS in the DCF arm, and DCF controlled micrometastases better than did ACF. Thus, to prevent recurrence via NAC in patients with advanced ESCC, using a NAC regimen that produces an antitumor effect in addition to controlling micrometastases may be beneficial.

　Advanced ESCC often leads to distant metastatic recurrence when only surgery is performed, even in cases of radical excision, and the reported prognosis is poor ^{［3, 7, 22］}. The standard treatment in Japan is to administer NAC before surgery, with the expectation that this will control distant metastatic recurrence and improve the therapeutic outcomes ^［6］. However, distant LN recurrence outside the LN dissection area can

occur even after NAC and R0 surgery, and such distant recurrence is a prognostic factor. The ideal NAC is a regimen that effectively controls both local recurrence by tumor shrinkage and distant recurrence by systemically preventing micrometastases. A triplet combination NAC is currently being developed with a potentially stronger antitumor effect than that of previous NACs but that can also control distant metastatic recurrence. The authors have also undertaken a controlled trial of ACF vs. DCF （OGSG1003） and reported favorable RFS results for DCF ^{［14, 15］}.

　Distant LN recurrence also frequently occurs outside the resection area, even after pathological diagnosis of N0 ［19, 23, 24］. Micrometastases may be associated with distant metastatic recurrence, and the prognosis is reported to be poor when micrometastases are observed in the dissected LNs ［16, 17, 25］. Differences in controlling micrometastases due to differences in the NAC regimens and the relevance to prognosis remain unclear although micrometastases after neoadjuvant treatment are reported to be associated with the prognosis ［17, 25, 26］. In the present study, MM（+） patients had a higher rate of distant LN recurrence, possibly due to residual micrometastases outside the LN dissection range, and the RFS was significantly poorer in MM（+） patients at pN0 as well as those with pathological LN metastases. The prognosis was poor in MM（+） and pN1 patients, but in pN2–3 patients with multiple LN metastases, micrometastases had less effect on survival. A multivariate analysis of the data from all patients, including those with pN0 and pN1–3, also identified MM（+） to be an independent predictor of a poor prognosis. Comparison of the hazard ratios of pathologically confirmed LN metastases and micrometastases showed that this association was stronger in patients with pN1–3. Izbicki et al. reported the same result, and pathological LN metastases were considered to more strongly affect the prognosis than were micrometastases ^［16］. However, given that MM

（+） is considered to strongly impact the prognosis in patients with pN0, these results should be interpreted carefully.

　In the OGSG1003 study, the groups did not differ in the pretreatment diagnosis of clinical T, N, and M stage, but the response rate and histological effect were greater in the DCF group than in the ACF group.

DCF was also associated with the superior regional control of the main tumor, thus resulting in favorable RFS ^［15］. However, after NAC and surgery, the pN diagnosis of LN metastases did not differ between the groups. The proportions of patients in whom micrometastases were identified via immunostaining with anti-CK antibodies in LNs following a negative histopathological diagnosis for metastases by HE staining were significantly lower in patients who received DCF NAC. In addition, pN0 patients in the DCF group had a more favorable RFS, although this result was not significant, and significantly fewer DCF patients were MM

（+）. Thus, the regimens largely differed in their effectiveness in controlling micrometastases.

　These findings revealed that DCF more effectively prevented micrometastases than did ACF, and because DCF exhibited effective regional and micrometastatic control, DCF may yield a more favorable RFS. In addition to effective regional control, micrometastatic control is an important requirement for the NAC regimens, and DCF is effective for both regional and micrometastatic control.

─ 10 ─

　In Western countries, neoadjuvant CRT （NACRT） is recommended as a neoadjuvant treatment for locally advanced esophageal cancer ^{［3, 4］}. Because NACRT has a high locoregional control effect, it has a reportedly high complete response rate and pN0 ratio,²⁴ and it strongly affects micrometastatic control ^［27］. However, the ability of NACRT to suppress distant metastases, such as distant LN metastases, is weak despite it suppressing locoregional recurrence ^{［28, 29］}. Although it is unclear whether NACRT or NAC is more suitable for preoperative treatment, triplet therapy, such as DCF, which may also suppress distant micrometastases not covered by radiation fields, may be a promising candidate for preoperative treatment ^［30］. Moreover, the number of NAC cycles is typically 2; however, 3 cycles may further enhance the effect on controlling micrometastases. Therefore, we are running a randomized controlled trial to determine whether 2 or 3 cycles are better for NAC with DCF.

　This study is associated with some limitations. First, not all patients enrolled in the OGSG1003 study were investigated. Second, only LNs in regions with a high frequency of metastasis and high efficacy for dissection, such as those around the cervical region, superior mediastinum, or cardiac orifice, were investigated to represent regional LNs, and not all LNs removed during surgery were examined. Third, a more sensitive method for identifying MM（+） patients is preferred. In clinical practice, lymph node metastases are pathologically diagnosed in one section per node; therefore, micrometastases were also investigated and identified in one section per one node. Additionally, this investigation did not assess the presence of cytokeratin deposits （CDs）, which are hyalinized denucleated particles considered to be dead carcinoma cells after NAC, because it is unclear whether CDs originate from dead or dying cancer cells, although the appearance of CDs is reported to be associated with the prognosis ^［15］. The methods for identifying micrometastases using RT-PCR with high sensitivity, molecular biological markers to identify micrometastases other than cytokeratin, and the true clinical significance of and identification methods for CDs thus require further study.

　In this study, we showed that controlling LN micrometastases after NAC for advanced ESCC contributes to preventing ESCC recurrence and that ESCC control varies among the regimens. DCF controlled micrometastases and may prevent recurrence in a high proportion of patients. NAC with DCF effectively controlled micrometastases, which is likely one reason why DCF resulted in favorable outcomes in patients with ESCC. In the future, micrometastatic control may indicate a more appropriate NAC regimen for ESCC.

Acknowledgments Funding: none

　We would like to express our deepest appreciation to Ms. Tomoko Hashimoto for preparing the histological sections from the paraffin blocks. We thank Traci Raley, MS, ELS, from Edanz Group （www.edanzediting.

com/ac） for editing a draft of this manuscript.

Disclosure

　Taroh Satoh serves as a consultant to Daiichi Sankyo, Eli Lilly, Chugai Pharmaceutical Co. Ltd., Merck Serono Co. Ltd., Bristol-Myers K.K., Taiho Pharmaceutical Co. Ltd., ONO Pharmaceutical. Co. Ltd. and Takeda Pharmaceutical Co. Ltd.; Taroh Satoh received a research grant from Chugai Pharmaceutical Co. Ltd., ONO Pharmaceutical. Co. Ltd. and Yakult Honsha Co. Ltd.; Taroh Satoh received lecture fees from Eli Lilly, Chugai Pharmaceutical Co. Ltd., Merck Serono Co. Ltd., Bristol-Myers K.K., Taiho Pharmaceutical Co. Ltd., ONO Pharmaceutical. Co. Ltd. and Takeda Pharmaceutical Co. Ltd.. Yoko Hiraki and other co-authors other than Taroh Satoh have no conflicts of interest.

Approval of the Protocol: Each institution’s review board approved this research.

Author Contributions: Conceived and designed the study: Y. Hiraki, Y. Kimura and T. Yasuda. Participated in data acquisition: Y. Hiraki, H. Kato, M. Iwama, O. Shiraishi, A. Yasuda, M. Shinkai, T. Makino, M. Motoori, H. Miyata and M Yamasaki. Pathological diagnosis: M. Imano and T. Satou. Analysis of survival data: M.

Yamasaki, T. Satoh, and H. Furukawa. Assembly of data and critical revision of the article: M. Yano and Y.

Doki. Statistical analysis and interpretation of data: Y. Kimura and T. Yasuda. Final approval of the version for publication: Y. Kimura and T. Yasuda. All authors proofread and approved the final version of the manuscript. All authors had access to the data and jointly decided to submit the manuscript.

ORCID: Yutaka Kimura ID http://orcid.org/0000-0002-0188-0586

─ 12 ─

References

１． Kelsen DP, Ginsberg R, Pajak TF, Sheahan DG, Gunderson L, Mortimer J, et al. Chemotherapy followed by surgery compared with surgery alone for localized esophageal cancer. N Engl J Med.

1998;339:1979-84.

２．van Hagen P, Hulshof MC, van Lanschot JJ, Steyerberg EW, van Berge Henegouwen MI, Wijnhoven BP, et al. Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med.

2012;366:2074-84.

３．Sjoquist KM, Burmeister BH, Smithers BM, Zalcberg JR, Simes RJ, Barbour A, et al. Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma:an updated meta-analysis. Lancet Oncol. 2011;12:681-92.

４．Mayanagi S, Irino T, Kawakubo H, Kitagawa Y. Neoadjuvant treatment strategy for locally advanced thoracic esophageal cancer. Ann Gastroenterol Surg. 2019;3:269-75.

５．Kitagawa Y, Uno T, Oyama T, Kato K, Kato H, Kawakubo H, et al. Esophageal cancer practice guidelines 2017 edited by the Japan Esophageal Society:part 1. Esophagus. 2019;16:1-24.

６．Ando N, Kato H, Igaki H, Shinoda M, Ozawa S, Shimizu H, et al. A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus （JCOG9907）.

Ann Surg Oncol. 2012;19:68-74.

７．Allum WH, Stenning SP, Bancewicz J, Clark PI, Langley RE. Long-term results of a randomized trial of surgery with or without preoperative chemotherapy in esophageal cancer. J Clin Oncol.

2009;27:5062-7.

８．Yamasaki M, Miyata H, Miyazaki Y, Takahashi T, Kurokawa Y, Nakajima K, et al. Perioperative therapy for esophageal cancer. Gen Thorac Cardiovasc Surg. 2014;62:531-40.

９．Tachimori Y, Ozawa S, Numasaki H, Ishihara R, Matsubara H, Muro K, et al. Comprehensive registry of esophageal cancer in Japan, 2011. Esophagus. 2018;15:127-52.

１０． Motoori M, Yano M, Yasuda T, Miyata H, Peng YF, Yamasaki M, et al. Chemotherapy-induced toxicities and treatment efficacy in advanced esophageal cancer treated with neoadjuvant chemotherapy followed by surgery. Esophagus. 2011;8:81-7.

１１． Shimakawa T, Naritaka Y, Asaka S, Isohata N, Murayama M, Konno S, et al. Neoadjuvant chemotherapy （FAP） for advanced esophageal cancer. Anticancer Res. 2008;28:2321-6.

１２． Tamura S, Takeno A, Taniguchi H, Katsura Y, Ohmura Y, Kagawa Y, et al. Induction chemotherapy using FAP for patients with Stage II/III squamous cell carcinoma of the esophagus. Anticancer Res.

2015;35:2975-82.

１３． Yamasaki M, Miyata H, Tanaka K, Shiraishi O, Motoori M, Peng YF, et al. Multicenter phase I/II study of docetaxel, cisplatin and fluorouracil combination chemotherapy in patients with advanced or recurrent squamous cell carcinoma of the esophagus. Oncology. 2011;80:307-13.

１４． Shiraishi O, Yamasaki M, Makino T, Motoori M, Miyata H, Shinkai M, et al. Feasibility of preoperative chemotherapy with docetaxel, cisplatin, and 5-fluorouracil versus Adriamycin, cisplatin, and 5-fluorouracil for resectable advanced esophageal cancer. Oncology. 2017;92:101-8.

１５． Yamasaki M, Yasuda T, Yano M, Hirao M, Kobayashi K, Fujitani K, et al. Multicenter randomized phase II study of cisplatin and fluorouracil plus docetaxel （DCF） compared with cisplatin and fluorouracil plus Adriamycin （ACF） as preoperative chemotherapy for resectable esophageal squamous cell carcinoma （OGSG1003）. Ann Oncol. 2017;28:116-20.

１６． Izbicki JR, Hosch SB, Pichlmeier U, Rehders A, Busch C, Niendorf A, et al. Prognostic value of immunohistochemically identifiable tumor cells in lymph nodes of patients with completely resected esophageal cancer. N Engl J Med. 1997;337:1188-94.

１７． Doki Y, Ishikawa O, Mano M, Hiratsuka M, Sasaki Y, Kameyama M, et al. Cytokeratin deposits in lymph nodes show distinct clinical significance from lymph node micrometastasis in human esophageal cancers. J Surg Res. 2002;107:75-81.

１８． Brierley JD, Gospodarowicz MK, Wittekind C. TNM Classification of Malignant Tumors, 8th edition.

Oxford:Wiley-Blackwell 2017.

１９． Japanese Society for Esophageal Diseases:Guideline for clinical and pathological studies on carcinoma of the esophagus, ninth edition:part II. Esophagus. 2004;1:107-25.

２０． Tachimori Y, Ozawa S, Numasaki H, Matsubara H, Shinoda M, Toh Y, et al. Efficacy of lymph node dissection by node zones according to tumor location for esophageal squamous cell carcinoma.

Esophagus. 2016;13:138-45.

２１． Komukai S, Nishimaki T, Watanabe H, jioka Y, Suzuki T, Hatakeyama K. Significance of immunohistochemically demonstrated micrometastases to lymph nodes in esophageal cancer with histologically negative nodes. Surgery. 2000;127:40-6.

２２． Ando N, Iizuka T, Ide H, Ishida K, Shinoda M, Nishimaki T, et al. Surgery plus chemotherapy compared with surgery alone for localized squamous cell carcinoma of the thoracic esophagus :a Japan Clinical Oncology Group Study--JCOG9204. J Clin Oncol. 2003;21:4592-6.

２３． MacGuill MJ, Barrett C, Ravi N, MacDonald G, Reynolds JV. Isolated tumour cells in pathological node-negative lymph nodes adversely affect prognosis in cancer of the oesophagus or oesophagogastric junction. J Clin Pathol. 2007;60:1108-11.

２４． Prenzel KL, Hölscher AH, Drebber U, Agavonova M, Gutschow CA, Bollschweiler E. Prognostic impact of nodal micrometastasis in early esophageal cancer. Eur J Surg Oncol. 2012;38:314-8.

─ 14 ─

２５． Prenzel KL, König A, Schneider PM, Schnickmann C, Baldus SE, Schröder W, et al. Reduced incidence of nodal micrometastasis after major response to neoadjuvant chemoradiation in locally advanced esophageal cancer. Ann Surg Oncol. 2007;14:954-9.

２６． Mariette C, Dahan L, Mornex F, Maillard E, Thomas PA, Meunier B, et al. Surgery alone versus chemoradiotherapy followed by surgery for stage I and II esophageal cancer :final analysis of randomized controlled phase III trial FFCD 9901. J Clin Oncol. 2014;32:2416-22.

２７． Yanagi M, Sasaki K, Uchikado Y, Omoto I, Arigami T, Kurahara H, et al. Effect of neoadjuvant chemoradiotherapy on lymph node micrometastases in thoracic esophageal cancer. Anticancer Res.

2018;38:893-900.

２８． Robb WB, Messager M, Dahan L, Mornex F, Maillard E, D’Journo XB, et al. Patterns of recurrence in early-stage oesophageal cancer after chemoradiotherapy and surgery compared with surgery alone.

Br J Surg. 2016;103:117-25.

２９． Oppedijk V, van der Gaast A, van Lanschot JJ, van Hagen P, van Os R, van Rij CM, et al. Patterns of recurrence after surgery alone versus preoperative chemoradiotherapy and surgery in the CROSS trials. J Clin Oncol. 2014;32:385-91.

３０． Klevebro F, Alexandersson von Döbeln G, Wang N, Johnsen G, Jacobsen AB, Friesland S, et al. A randomized clinical trial of neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the oesophagus or gastro-oesophageal junction. Ann Oncol. 2016;27:660-7.