1 Introduction

Lymph node metastasis is one of the most important prognostic factors in patients with gastrointestinal (GI) cancer including esophageal and gastric cancers as well as colorectal cancer (CRC) [1,2,3,4,5,6,7,8]. However, a pathological node-negative (pN0) state at resection confirmed by conventional histological hematoxylin-eosin (HE) staining does not guarantee long-term survival without recurrence. The existence of lymph node micrometastasis (LNM), including isolated tumor cells (ITCs), which are not detected by conventional HE staining, has been attracting attention as a candidate for recurrence and/or as a prognostic factor. The LNM status of GI cancer patients may be evaluated using immunohistochemistry (IHC) and a reverse transcription polymerase chain reaction (RT-PCR). Due to advances in the development of diagnostic tools, the detection rate of LNM is increasing. The significance of LNM in GI cancers has long been investigated and discussed but currently remains unclear. Perioperative treatments are widely used against various cancers. Neoadjuvant therapy potentially results in the downstaging/downsizing and elimination of LNM, which contributes to improved resectability and curability rates. On the other hand, adjuvant therapy is directed toward micrometastases existing at distant sites and outside of the surgical field and contributes to improvements in the prognosis of patients with GI cancers. However, since it is impossible to investigate the therapeutic effects of adjuvant therapy after surgery for LNM, clinical samples of LN and LNM are examined after neoadjuvant therapy in order to evaluate the influences of chemo- and/or radiation therapy on LNM. Limited information is available on the rates and effects of chemo- and/or radiation therapy in clinical samples of LNM, and the clinical aspects of chemo- and/or radiation therapy and LNM have not yet been evaluated or discussed. This chapter focuses on the clinical significance of LNM after chemo- and/or radiation therapy as well as the influences of chemo- and/or radiation therapy on LNM in carcinomas of the GI tract such as esophageal and gastric cancers as well as CRC.

2 Esophageal Cancer

Esophageal cancer is a difficult malignancy to treat, and lymph node metastasis has been identified as one of the most important prognostic factors in patients with esophageal cancer [1,2,3, 8]. A clinical aspect of LNM [9] and the possibility of the sentinel lymph node concept [10, 11] have been reported; however, the significance of LNM has not yet been elucidated in esophageal cancer. Even if complete tumor resection with lymphadenectomy is performed, the disease is already in an advanced stage beyond the scope of curative therapy by the time of surgery. Multiple randomized trials have established neoadjuvant therapy for the management of esophageal cancer patients based on improved survival rates over those achieved with upfront surgery. Neoadjuvant chemotherapy (NACT) or chemoradiotherapy (NACRT) has been shown to extend postoperative survival, and preoperative therapy followed by esophagectomy has become the standard treatment worldwide for patients with resectable locally advanced esophageal or esophagogastric junctional carcinoma [12,13,14,15,16,17,18,19]. The potential benefits of neoadjuvant therapy are considered to be the downstaging/downsizing and elimination of LNM, which contribute to improved resectability and curability rates. Few studies have focused on the relationship between NACT and/or NACRT and LNM in esophageal cancer patients, with only six studies being reported to date [20,21,22,23,24,25]. Table 12.1 summarizes six studies on the relationship between neoadjuvant therapy and LNM in esophageal cancer patients. There were three studies on chemotherapy, two on chemoradiotherapy, and one on chemo- or chemoradiotherapy. The numbers of patients were relatively small, and patients with any disease stage were included. In Eastern countries, squamous cell carcinoma (SCC) was a major histological type, while SCC and adenocarcinoma are both included in Western countries. The CK antibody (AE1/AE3) was commonly used for IHC, except for one study that used CK8/CK18 (CAM 5.2). Single sections were used in two studies and multiple sections in four studies. Estimations of the therapeutic effects on LNM varied according to the authors. Three authors estimated LNM and cytokeratin-positive deposits without nuclei, while another two estimated reductions in LNM, and one author estimated LNM with or without stromal reactions. Cytokeratin-positive particles, named “cytokeratin deposits” (CDs), are frequently observed in lymph nodes by IHC staining for cytokeratin. CDs have been defined as hyalinized denucleated particles and are regarded as the cadavers of carcinoma cells. CDs are round and eosinophilic and do not include a nucleus in serial sections and have been observed more frequently in patients treated with NACT than in those without [25]. Matsuyama et al. [20] evaluated the presence of CD, other than immunohistochemical LNM, in 75 esophageal squamous cell carcinoma (ESCC) patients treated with NACT using cisplatin (CDDP), doxorubicin hydrochloride, and 5-fluorouracil (5-FU) in order to examine the influence of NACT on LNM. The anti-cytokeratin antibody cocktail, AE1/AE3, was used as the primary antibody. They reported that successful NACT converted cancer cells from immunohistochemical LNM to CD and improved the status of ESCC patients from systemic disease to regional disease, and the disappearance of immunohistochemical LNM and emergence of CD suggested the eradication of LNM by NACT in ESCC of the thoracic esophagus. Based on the above findings, they concluded that the clinical benefit of NACT was apparent for immunohistochemical LNM-negative and CD-positive patients. Lee et al. evaluated the histological effects of NACT using 5-FU/CDDP or nedaplatin or NACRT (40 Gy delivered in daily fractions of 2 Gy) on lymph node metastasis in ESCC by performing IHC using the anti-cytokeratin antibody cocktail, AE1/AE3. A total of 3061 lymph nodes were examined from 36 and 26 patients who received NACRT and NACT, respectively, followed by esophagectomy with lymphadenectomy. They also evaluated hyalinized cytokeratin particles (HCP); HCP were defined as cytokeratin deposits without cellular nuclei, similar to CD. Consistent with previous findings for CD, HCP were suggested to reflect a degenerative change in cancer cells in lymph nodes and may predict responses to neoadjuvant therapy [22]. One limitation of these studies is that they did not prove whether cytokeratin-positive particles, CD and HCP, are truly the cadavers of cancer cells. However, this hypothesis is strongly supported by the following findings on cytokeratin-positive particles: their anatomic distribution was identical to LNM and pathological node metastases, associated with the pathological grade of the primary tumor, and reciprocally associated with LNM in its incidence and influence on prognosis [20, 22]. Furthermore, Matsuyama et al. proved the biological origin of cytokeratin-positive deposits using IHC, showing that they were positive for the epithelial marker E-cadherin and negative for the macrophage marker CD68 [25]. The presence of cytokeratin-positive deposits in lymph nodes has not been described in other histological types of GI cancers, except for ESCC. These findings suggest that cytokeratin-positive deposits are immunohistochemically specific to ESCC [25].

Table 12.1 Studies on the relationship between neoadjuvant therapy and LNM in esophageal cancer patients
Full size table

We evaluated LNM—LNM was defined as tumor cells in lymph nodes with a stromal reaction (granulation tissue or desmoplastic connective tissue) and/or tumor cell micro-involvement (TCM) and tumor cells in lymph nodes without a stromal reaction immunohistochemically using the anti-cytokeratin antibody cocktail, AE1/AE3, in 1052 lymph nodes from 20 ESCC patients treated with NACT using CDDP and 5-FU—and reported that the incidence of LNM and/or TCM in the chemotherapy group was similar to that in the surgical group of patients with ESCC. However, chemotherapy may be effective in patients with TCM alone before cancer cells form clusters with stromal reactions in lymph nodes. Patients with lymph node metastasis including LNM do not benefit from NACT, and LNM, but not TCM, is prognostically equivalent to lymph node metastasis and needs be examined using IHC in order to classify these cases correctly as pathological node-positive (pN1) in ESCC [21]. We also evaluated the usefulness of NACRT for ESCC and reported that fewer metastatic lymph nodes were present in the NACRT group than in the surgery only group [26]. Furthermore, we immunohistochemically evaluated LNM and/or TCM using the anti-cytokeratin antibody cocktail, AE1/AE3, in 663 lymph nodes from 20 ESCC patients treated with NACRT, a total radiation dose of 40 Gy and concurrent intravenous chemotherapy with CDDP and 5-FU. The extent of lymph node metastasis was slightly greater in the surgery group than in the NACRT group. However, the incidence of patients with LNM ± TCM and TCM alone was significantly higher in the surgery group than in the NACRT group. NACT may be effective for patients with TCM alone, whereas NACRT was effective for LNM and TCM. These findings indicate that NACRT is more effective for LNM than NACT in patients with node-positive ESCC [21]. The presence of pathological lymph node metastasis was identified as a prognostic predictor in the surgery alone and NACRT groups. Furthermore, assessments of the simultaneous presence of pathological lymph node metastasis and LNM may facilitate highly accurate predictions of survival in esophageal cancer patients undergoing esophagectomy, regardless of whether they have received NACRT. This finding was consistent with previously reported data from other studies [4]. Prenzel et al. investigated the influence of NACRT on LNM in 52 esophageal cancer patients (21 adenocarcinomas and 31 SCC) who received NACRT by performing IHC using the anti-cytokeratin antibody cocktail, AE1/AE3. Intravenous chemotherapy was performed using CDDP (20 mg/m2/day, short-term infusion) and 5-FU (1000 mg/m2/day, over 24 h) on days 1–5. Radiation was delivered in daily fractions of 1.8 Gy to a total dose of 36 Gy. They categorized the extent of histomorphological regression of the primary tumor into a major (<10%) and minor response (>10% vital residual tumor cells) and evaluated 1186 lymph nodes that were diagnosed as negative for metastases in a routine histopathological analysis. In patients with a major histomorphological response following NACRT, the presence of LNM was significantly less than that in those with a minor response [23]. Wang et al. evaluated the influence of NACRT on LNM in 20 ypN0 esophageal adenocarcinoma (EAC) patients and ypN1 EAC patients who received NACRT followed by surgery by performing IHC for cytokeratin CK8/CK18 (CAM 5.2) and compared the impact of NACRT on LNM with 20 surgery-alone EAC patients. Radiation was delivered in daily fractions of 1.8 Gy to a total dose of 41.4–45 Gy. Concurrent intravenous chemotherapy was performed using paclitaxel (50 mg/m2) and carboplatin (area under the curve = 2). They concluded that a 30% reduction in LNM was achieved with NACRT from that with surgery alone in ypN0 [24]. Based on the six studies that investigated neoadjuvant therapy and LNM in ESCC and EAC, neoadjuvant therapy appeared to reduce LNM and had an equal negative impact on prognosis to that in node-positive patients with esophageal cancer. If the evaluation of LNM detection and effectiveness of neoadjuvant therapy differ between each institution, the findings of different studies will naturally vary. Regarding prognostic impacts, three out of six studies reported that the effectiveness of neoadjuvant therapy against LNM reflects prognosis. The numbers of patients and nodes examined were not high. All studies included early and advanced carcinoma. Four studies included only SCC, one study examined SCC and adenocarcinoma, and one study only investigated adenocarcinoma. Common criteria are needed in order to accurately evaluate the effectiveness of neoadjuvant therapy for LNM.

3 Gastric Cancer

Lymph node metastasis is one of the significant prognostic factors for gastric cancer [5, 27]. Although the prognostic value of LNM remains controversial, its clinical impact is apparently strong in early and advanced gastric cancer [28]. Surgery in combination with adjuvant treatments is the globally accepted standard of care for locally advanced gastric cancer. However, approaches to adjuvant treatment vary among countries. In Asia, the standard adjuvant treatment for stage II or III gastric cancer is postoperative chemotherapy with the oral fluoropyrimidine derivative S-1 for 1 year or capecitabine plus oxaliplatin for 6 months after D2 surgery based on the findings of the ACTS-GC trial or CLASSIC trial [29, 30]. In the USA, surgery followed by NACRT is the standard protocol for T3 or greater and/or positive-node gastric cancer based on findings of the INT-0116 trial [31]. In the UK and some European countries, preoperative and postoperative chemotherapy with epirubicin, CDDP, and 5-FU is used based on evidence from the MAGIC trial [16]. Postoperative adjuvant chemotherapy is directed toward micrometastases that may exist as residual disease after surgery. Chemotherapy in the neoadjuvant setting may also be considered for the downstaging/downsizing and eradication of microscopic disease prior to surgery, which contributes to improved resectability and curability rates [32]. The detection rate of LNM is increasing due to advances in the development of diagnostic tools, such as IHC and RT-PCR; however, the clinical significance of LNM in gastric cancer is still controversial, much less than that for chemo- and/or radiation therapy and LNM. Very few studies have focused on the relationship between chemotherapy and LNM in gastric cancer, and only three studies on LNM and chemotherapy have been reported to date. Becker et al. [33] evaluated 622 lymph nodes that were resected from ypN0 17 locally advanced gastric adenocarcinoma (GAC) patients who received NACT followed by gastrectomy by performing IHC for cytokeratin (AE1/AE3 and Ber-Ep4) and compared the impact of NACT on TCM in ypN0 62 surgery-alone GAC patients. Six patients (35%) and 25 out of 622 lymph nodes (4.0%) had TCM, whereas 93% of patients and 21.8% of lymph nodes had TCM in patients treated with surgery alone. In their previous study on lymph node micro-involvement in 100 GAC patients after primary surgery, the rates of TCM in lymph nodes were 90% and 97% in cases classified by routine histology as pN0 and pN1, respectively [34]. Furthermore, this study indicated that the degree of the pathological responses of primary tumors to NACT correlated with effects on tumor cells in regional lymph nodes [33]. Yokoyama et al. established an in vivo lymph node metastasis model using the green fluorescent protein (GFP)-transfected gastric cancer cell line, GCIY-EGFP, which metastasizes spontaneously to the inguinal lymph nodes when inoculated subcutaneously into the abdominal wall. They also demonstrated that ITCs in lymph nodes regressed spontaneously through natural killer cell-mediated antitumor activity following the resection of primary subcutaneous tumors, whereas micrometastasis in lymph nodes continued to proliferate and may be effectively eliminated by postoperative chemotherapy using this model [35]. Using the same model, Eguchi et al. evaluated the effects of perioperative chemotherapy against micrometastasis in gastric cancer. After the inoculation of GCIY-EGFP into the lower abdominal wall of nude mice, a preoperative treatment with S-1 and docetaxel or postoperative treatment with S-1 was performed in addition to the resection of primary tumors in order to assess the efficacy of chemotherapy on micrometastases, metastatic foci measuring 0.2–2 mm in diameter and ITCs, and metastatic foci measuring less than 0.2 mm, in the lymph nodes. NACT was effective against micrometastases and ITCs in the lymph nodes, despite chemotherapy not being active against primary tumors [36]. These findings indicated that chemotherapy has a marked effect on LNM in regional lymph nodes independent of whether the effects of chemotherapy are active against the primary tumor.

4 CRC

Regional lymph node metastasis is a reliable prognostic factor and is used for clinical decision-making [6, 7]. In Western countries, the standard treatment for patients with locally advanced rectal cancer is 5-FU-based NACRT followed by total mesorectal excision (TME) [37]. On the other hand, NACRT is still not a standard treatment in Japan. The Japanese guidelines for the treatment of CRC recommend upfront surgery followed by adjuvant chemotherapy [38]. In either case, the prognostic value of LNM in patients with node-negative CRC has remained unclear because of a lack of evidence from prospective studies. In a meta-analysis by Rahbari et al., the relationship between the molecular detection of occult disease in regional lymph nodes and an increased risk of disease recurrence and poor survival in patients with node-negative CRC was reported, and the necessity for prospective studies was emphasized [39]. In another systematic review and meta-analysis, LNM had a worse prognosis than that for patients without occult tumor cells; however, ITC did not have predictive value in patients with stage I/II CRC [40]. Yamamoto et al. recently reported the findings of a prospective multicenter clinical trial in terms of the usefulness of the micrometastasis volume in lymph nodes in 315 patients with node-negative stage II CRC using the molecular detection of CEA mRNA by quantitative RT-PCR in addition to conventional qualitative RT-PCR [41]. However, few studies have focused on the relationship between NACT and/or NACRT and LNM in CRC; only three studies have been reported to date. Table 12.2 summarizes studies on the relationship between neoadjuvant therapy and LNM in CRC patients [42,43,44]. Kinoshita et al. examined the pathological effects of neoadjuvant radiotherapy (NART) on the intramural spread of tumors and risk factors for local recurrence, including tumor deposits, the budding growth of primary tumors, and LNM by performing IHC using the anti-cytokeratin antibody cocktail, AE1/AE3, for lower rectal cancer. Twenty-five stage-matched patients were enrolled, with 25 patients who received 50 Gy NART and 25 who did not. LNM was significantly smaller in patients who received NART than in those who did not. These findings suggested the beneficial effects of NART for LNM in lower rectal cancer [42]. Perez et al. examined 518 lymph nodes that were resected from 56 distal rectal cancer patients who received NACRT followed by radical surgery by performing IHC using the anti-cytokeratin antibody cocktail, AE1/AE3, and compared the impact of NACRT on LNM. Radiation was delivered in daily fractions of 1.8 Gy to a total dose of 50.4 Gy. Intravenous chemotherapy was performed using 5-FU (20 mg/m2/day) and folinic acid (1000 mg/m2/day, over 24 h) for 3 consecutive days on the first and last 3 days of radiation therapy. The detection rate of LNM was markedly low in patients treated with NACRT (7%), even in high-risk patients (T3 and T4 tumors); however, LNM was not associated with decreased overall or disease-free survival [43]. Sprenger et al. prospectively examined 2412 lymph nodes that were resected from 81 rectal adenocarcinoma patients who received NACRT followed by total mesorectal excision. Radiation was delivered in daily fractions of 1.8 Gy to a total dose of 50.4 Gy. Intravenous chemotherapy was performed using either the continuous infusion of 5-FU (1000 mg/m2 on days 1–5 and 29–33) or a combined regime of 5-FU/oxaliplatin (5-FU, 250 mg/m2 on days 1–14 and 22–35, and oxaliplatin, 50 mg/m2 on days 1, 8, 22, and 29). Conventional HE staining was performed to detect lymph nodes metastases and revealed a markedly high incidence of mesorectal LNM (32.8%) after NACRT. They concluded that residual LNM did not impair disease-free survival or cancer-specific survival [44]. Based on the three studies that investigated neoadjuvant therapy and LNM in CRC, NART or NACRT exhibited the ability to reduce LNM. However, LNM after neoadjuvant therapy had a negative impact on the prognosis of patients with node-negative CRC.

Table 12.2 Studies on the relationship between neoadjuvant therapy and LNM in colorectal cancer patients
Full size table

5 Future Perspectives of NACT and/or NACRT and LNM

According to the findings of our previous review, a high incidence of LNM ≥10% was found in patients with pN0 GI cancer [9]. In our study on gastric cancer, LNM already exhibited proliferative activity, even in ITCs [45]. If LNM is present in patients diagnosed with pN0, these patients need to be categorized as pN1. Prospective randomized controlled studies need to be conducted in order to examine the effectiveness of adjuvant therapies for patients with LNM. In conclusion, further studies on the biological behavior of LNM treated with NACT and/or NACRT are required in order to elucidate the efficacy of NACT and/or NACRT for LNM and may lead to a better understanding of LNM and the development of further treatments for patients with GI cancers.