Radical lymphadenectomy in esophageal cancer: from the past to the present.

Lymphadenectomy as an essential part of the surgical treatment has been one of the most controversial aspects in the management of esophageal cancers. The purpose of this article was to review the evolution, the current role, and the optimal extent of lymphadenectomy for the treatment of esophageal cancers. Studies discussing the outcome of esophagectomy with lymph nodes dissection and comparing among different extent of lymphadenectomy were used in the analysis. Several studies including recently published articles reveal that additional radical lymphadenectomy may be beneficial in some patients with non-extreme esophageal cancer undergoing esophagectomy, whereas two-field lymph node dissection is suitable for distal esophageal cancers regardless of the histology of the tumor. Minimally invasive surgery and neoadjuvant therapy combined with radical surgery seem to show more benefit in selected cases, but further studies should be required to clearly demonstrate their efficacy and safety. The expertise and experience of the surgeons should also be taken into account in determining the success of these radical procedures.

Thoracoscopic esophagectomy with extended lymph node dissection in the left lateral position: technical feasibility and oncologic outcomes.

The aim of this study was to estimate the technical and oncologic feasibility of video-assisted thoracoscopic radical esophagectomy (VATS) in the left lateral position. From January 2003 to December 2011, 132 patients with esophageal cancer underwent VATS. The mean duration of the thoracic procedure and the entire procedure was 294 ± 88 and 623 ± 123 minutes, respectively. Mean blood loss during the thoracic procedure and the entire procedure was 313 ± 577 and 657 ± 719 g, respectively. The mean number of dissected thoracic lymph nodes was 32.6 ± 12.9. There were four in-hospital deaths (3.0%); two patients (1.5%) died of acute respiratory distress syndrome and two patients (1.5%) died of tumor progression. Postoperative unilateral or bilateral recurrent laryngeal nerve (RLN) palsy, or pneumonia was found in 33 (25.0%), 21 (15.9%), and 27(20.5%) patients, respectively. The patients were divided into the first 66 patients who underwent VATS (Group 1) and the subsequent 66 patients (Group 2). The numbers of cases who underwent neoadjuvant or induction chemotherapy for T4 tumor and intrathoracic anastomosis were higher in Group 2 than in Group 1. The duration of the procedure, amount of blood loss, and the number of dissected thoracic lymph nodes were not different between the two groups. The total number of dissected lymph nodes was higher in Group 2 than in Group 1 (72.6 ± 27.8 vs. 62.6 ± 21.6, P = 0.023). The rate of bilateral RLN palsy was less in Group 2 than in Group 1 (7.6% vs. 24.2%, P = 0.042). The mean follow-up period was 38.7 months. Primary recurrence consisted of hematogenous, lymphatic, peritoneal dissemination, pleural dissemination, and locoregional in 15 (11.3%), 20 (15.1%), 3 (2.3%), 4 (3.0%), and 5 patients (3.8%), respectively. The rate of regional lymph node recurrence within the dissection field was only 4.5%. The prognosis of patients with lymph node metastasis was significantly poorer than that of patients without lymph node metastasis. However, the prognosis of the 11 cases that had metastasis only around RLNs was similar to that of node-negative cases. Thirteen patients with pathological remnant tumor (R1 or R2) did not survive longer than 5 years at present. The overall 5-year survival rate of stage I, II, and III disease after curative VATS was 82.2%, 77.0%, and 52.3%, respectively. Expansion of VATS criteria for patients after induction chemotherapy for T4 tumor or thoracoscopic anastomosis did not adversely affect the surgical results by experience. Although the VATS procedure is accompanied by a certain degree of morbidity including RLN palsy and pulmonary complications, VATS has an excellent locoregional control effect. In addition, the favorable survival after VATS shows that the procedure is oncologically feasible.

Learning of thoracoscopic radical esophagectomy: how can the learning curve be made short and flat?

Attainment of proficiency in video-assisted thoracoscopic radical esophagectomy (VATS) for thoracic esophageal cancer requires much experience. We have mastered this procedure safely under the direction of an experienced surgeon. After adoption of the procedure, the educated surgeon directed induction of this surgical procedure at another institution. We evaluated the efficacy of instruction during the induction period by comparing the results at the two institutions in which VATS had been newly induced. We defined the induction period as the time from the beginning of VATS to the time when the last instruction was carried out. From January 2003 to December 2007, 53 patients were candidates for VATS at Kanazawa University (institution 1). Of these, 46 patients underwent curative VATS by a single operator. We divided this period into three parts: the induction period of VATS, post-induction period, and proficient period when the educated surgeon of institution 1 directed the procedure at Maebashi Red Cross Hospital (institution 2). At institution 1, 12 VATS were scheduled, and nine procedures (75%) (group A) including eight instructions were completed during the induction period (from January 2003 to August 2004). Thereafter, VATS was performed without instruction. In the post-induction period, nine VATS were scheduled, and eight procedures (88.8%) (group B) were completed from September 2004 to August 2005. Subsequently, 32 VATS were scheduled, and 29 procedures (90.6%) (group C) were completed during the proficient period (from September 2005 to December 2007). The surgeon at Maebashi Red Cross Hospital (institution 2) started to perform VATS under the direction of the surgeon who had been educated at institution 1 from September 2005. VATS was completed in 13 (76.4%) (group D) of 17 cases by a single surgeon including seven instructions during the induction period at institution 2 from September 2005 to December 2007. No lethal complication occurred during the induction period at both institutions. We compared the results of VATS among four groups from the two institutions. There were no differences in the background and clinicopathological features among the four groups. The number of dissected lymph nodes and amount of thoracic blood loss were similar in the four groups (35 [22-52] vs 41 [26-53] vs 32 [17-69] vs 29 [17-42] nodes, P = 0.139, and 170 [90-380] vs 275 [130-550] vs 220 [10-660] vs 210 [75-543] g, P = 0.373, respectively). There was no difference in the duration of the thoracic procedure during the induction period at the two institutions. However, the duration of the procedure was significantly shorter in the proficient period of institution 1 (group C: 266 [195-555] minutes) than in the induction period of both institutions (group A: 350 [280-448] minutes [P = 0.005] and group D: 345 [270-420] mL [P = 0.002]). There were no surgery-related deaths in any of the groups. The incidence of postoperative complications did not differ among the four groups. Thoracoscopic radical esophagectomy can be mastered quickly and safely with a flat learning curve under the direction of an experienced surgeon. The educated surgeon can instruct surgeons at another institution on how to perform thoracoscopic esophagectomy. The operation time of thoracoscopic surgery is shortened by experience.

Laparoscopic gastrectomy for gastric cancer: experience with more than 600 cases.

BACKGROUND: Among the less invasive operations noted in recent years, laparoscopic gastrectomy for gastric cancer has become popular because of advances in surgical techniques. The authors performed laparoscopic gastrectomy with regional lymph node dissection for 612 cases of gastric malignancies between March 1998 and August 2006. The technique and results of laparoscopic gastrectomy for gastric cancer are presented. METHODS: Of the 612 gastric malignancy cases, distal gastrectomy was performed in 485 cases, proximal gastrectomy in 42 cases, and total gastrectomy in 85 cases. In all the cases, D1 or D2 lymph node dissection was performed according to the general rule of the Japanese Gastric Cancer Association. RESULTS: Quicker recovery was observed in the laparoscopic gastrectomy cases than in the open cases. The postoperative complications with this technique were within a permissible range. No statistical difference was seen in the survival curve after surgery between the laparoscopic group of advanced cases preoperatively diagnosed as surgical T2N1 or lower and the open group. CONCLUSION: The laparoscopic technique is not only less invasive, but also similarly safe and curative compared with open gastrectomy.

Laparoscopic mesorectal excision with preservation of the pelvic autonomic nerves for rectal cancer.

BACKGROUND/AIMS: Laparoscopic mesorectal excision with preservation of the autonomic pelvic nerves for rectal cancer including selected advanced lower rectal cancer is now challenging. The aims of the study were to assess the surgical results and short-term outcomes of this procedure prospectively. METHODOLOGY: Seventy-four of 281 rectal cancer patients, since the introduction of laparoscopic colorectal surgery in our hospital, have undergone laparoscopic rectal surgery. The location of the tumor distributed in upper rectum; 33, middle; 22, and lower 19. The mesorectal excision with preservation of the autonomic pelvic nerves was performed for all the patients. The laparoscopic mesorectal excision was performed under 8 to 10 cmH2O CO2 pneumoperitoneum and lymph node dissection was performed along the feeding artery depend on individuals. Ipsilateral lateral lymph node dissection was added for 5 cases of advanced lower rectal cancer. RESULTS: Open conversion occurred in 4 cases, 2 of those were due to locally advanced tumors and 2 technical difficulties in transaction of the distal rectum. There were 15 postoperative complications, 7 anastomotic leakage (10.6%), 3 transient urinary retention (4.1%), 4 wound infection (5.3%), and 1 small bowel obstruction (1.4%). No mortality was recorded in this series. Time of operation was 203 +/- 54 min in mesorectal excision cases and 270 +/- 42 min mesorectal excision with lateral lymph node dissection cases. Blood loss was 92 +/- 90g and 276 +/- 66 g respectively. The hospital length-of-stay was 11.7 days in average. CONCLUSIONS: Laparoscopic mesorectal excision with preservation of autonomic pelvic nerves for rectal cancer patients including selected advanced lower rectal cancer is favorable.

Laparoscopic gastrectomy with regional lymph node dissection for upper gastric cancer.

BACKGROUND: The technique and results of laparoscopic gastrectomy in 110 patients with gastric cancer located in the upper third of the stomach are presented. METHODS: Proximal gastrectomy was performed for lesions in the upper third of the stomach, and total gastrectomy for those that spread over both the upper and middle third. D1 and D2 lymph node dissection was undertaken in patients with T1 or T2 lesions. Anastomosis of the oesophagus was performed intracorporeally using a conventional circular stapling device or a laparoscopic linear stapler. RESULTS: Median operating time was 247 min for proximal gastrectomy and 285 min for total gastrectomy; median blood loss was 207 and 334 ml respectively. A median of 23 lymph nodes was harvested from patients in the proximal gastrectomy group and 34 from those having a total gastrectomy. There was minimal morbidity and fast recovery after surgery. Postoperative recurrence occurred in only one patient, giving a recurrence rate of 0.9 per cent. CONCLUSION: Laparoscopic gastrectomy for upper gastric cancer appears to be a safe and curative procedure.

Laparoscopic distal gastrectomy with regional lymph node dissection for gastric cancer.

BACKGROUND: Recent advances in surgical techniques have led to widespread acceptance of laparoscopic gastrectomy for gastric cancer. We performed distal gastrectomy with regional lymph node dissection in 235 patients with gastric cancer located in the middle and lower third of the stomach. METHODS: In 171 cases, reconstruction was done using the Billroth I method intracorporeally and the aid of laparoscopic linear stapling devices. The Billroth II and Roux-en-Y methods were used in the remaining 56 and eight patients, respectively, RESULTS: Patients who underwent laparoscopic distal gastrectomy had a more rapid postoperative recovery than those treated via the open approach. Postoperative complications with this technique were within a permissible range. In terms of the survival curve, there was no statistical difference between the laparoscopic group diagnosed as clinical T2N0 (c T2N0) Preoperatively and the open group. CONCLUSION: The laparoscopic technique is not only less invasive, but is also similarly safe and curative compared to open gastrectomy.

The necessity of chest physical therapy for thoracoscopic oesophagectomy.

Radical surgery for thoracic oesophageal cancer is highly invasive and often leads to respiratory complications; thoracoscopic surgery is a less-invasive alternative. We examined the need for chest physical therapy (CPT) after thoracoscopic oesophagectomy. Thirty-six consecutive patients, randomly selected for either thoracotomy or thoracoscopic surgery, were included in a randomized clinical trial and received CPT under the same protocol. During short-term post-operative follow-up, both groups showed a marked reduction in respiratory function and responded to CPT to the same extent, although 2 weeks after surgery some parameters of respiratory function were significantly higher in the thoracoscopy group. Thoracoscopic surgery has been reported to be less invasive than standard thoracotomy, but our results suggest that the procedure is also invasive with respect to respiratory function and that CPT should be performed before and after thoracoscopic surgery.

Reconstructive procedure after distal gastrectomy to prevent remnant gastritis.

BACKGROUND/AIMS: Gastroduodenostomy (Billroth I) or gastrojejunostomy (Billroth II) after distal gastrectomy is associated with duodenogastric reflux and remnant gastritis. This study sought to determine which reconstructive procedure is least likely to cause remnant gastritis and to determine the correlation between duodenogastric reflux and remnant gastritis. METHODOLOGY: Sixty patients who underwent curative distal gastrectomy for gastric cancer were classified into three groups by reconstructive procedure: group A, Roux-Y (n=18); group B, Billroth I (n=25); group C, Billroth II (n=17). Intragastric bile reflux was monitored using the Bilitec 2000 14 days after surgery, and endoscopy was performed and a patient questionnaire was completed 12 weeks after surgery. RESULTS: Bile reflux occurred in 23.9%, 40.4%, and 73.4% of the time (p<0.001), and remnant gastritis developed in 33%, 76%, and 100% of patients (p<0.001), in groups A, B, and C, respectively. Helicobacter pylori infection did not correlate with remnant gastritis (p=0.57). Symptoms following Roux-Y reconstruction were comparable to those following Billroth I and II reconstructions. CONCLUSIONS: Roux-Y reconstruction following distal gastrectomy is superior to Billroth I and II reconstruction in preventing remnant gastritis because it reduces duodenogastric reflux.

A novel laparoscopic technique for stapled colon and rectal anastomosis.

BACKGROUND: We present new techniques of stapling anastomosis at laparoscopic colorectal surgery with retrospective review of data. METHODS: A triangulating stapling technique (T method) was performed in 101 laparoscopic colectomies. Adouble stapling technique (DST) with rectal division by a conventional linear stapler (Abd method) was used in 5 cases of upper/middle rectal cancers and subsequent eversion of the distal rectum from the anus (Ev method) was used for 4 low rectal cancers. Four hundred ninety-six colectomies and 280 rectal surgeries were reviewed. RESULTS: Leakage was lower in the T group (0.5%, n=196) than in the hand-sewn group (3.0%, n=233). Leakage of the DST using a laparoscopic linear stapler (12.1%, n=91) was significantly higher than with conventional DST (2.1%, n=189). There was no leakage with either Abd method or Ev method. The T-method is acceptable after laparoscopic colectomy. CONCLUSION: New methods of rectal division using conventional devices are expected to yield reliable anastomosis at laparoscopic rectal surgery.

Comparison of three-field esophagectomy for esophageal cancer incorporating open or thoracoscopic thoracotomy.

BACKGROUND: Thoracoscopic esophagectomy for esophageal cancer has been performed as an alternative to open surgery to reduce surgical trauma. However, its effect on pulmonary function, exercise tolerability, and quality of life is unknown. METHODS: Fifty-one patients with esophageal cancer underwent thoracic esophagectomy with radical lymphadenectomy by posterolateral thoracotomy (29 cases) or thoracoscopic surgery (22 cases). Patients performed spirometry and exercise tolerance testing and completed a quality-of-life questionnaire before and 3 months after surgery. RESULTS: Pre-to-postoperative change in vital capacity was 74.3 +/- 10.6% in the thoracotomy group and 84.9 +/- 10.4% in the thoracoscopy group (p = 0.021). Maximum oxygen uptake was similar, but dyspnea was the more common factor limiting exercise tolerance postoperatively in the thoracotomy group. Change in pre-to-postoperative performance status was 1.20 +/- 0.62 in the thoracotomy group and 0.55 +/- 0.51 in the thoracoscopy group (p = 0.0003). CONCLUSIONS: Thoracoscopic esophagectomy for esophageal cancer has better preservation of pulmonary function and quality-of-life.

Surgical management of synchronous hepatocellular carcinoma and gastric cancer.

BACKGROUND/AIMS: The improvement of diagnostic modalities and operative techniques has resulted in prolonged survival for cancer patients, but has also led to the diagnosis of an increasing number of patients with synchronous hepatocellular carcinoma (HCC) and extrahepatic primary cancer. It is necessary to determine the optimal surgical strategies for synchronous HCC and gastric cancer. METHODS: In this retrospective study, clinicopathologic findings, diagnostic methods, treatment and outcome were reviewed in 13 patients who underwent curative surgery for synchronous HCC and gastric cancer. RESULTS: Twelve of the 13 patients were men older than 60 years. All patients had chronic hepatic disease, and hepatitis viral infection was detected in 9 patients. Examinations of the esophagus to search for esophageal varices before liver resection for HCC, and imaging studies to rule out liver metastasis before gastrectomy for gastric cancer can lead to the incidental finding of a synchronous carcinoma. The most frequent postoperative complication was massive ascites, which occurred in 4 patients who underwent lymph node dissection, 1 of whom died of perioperative hepatic failure. HCC recurred in 7 patients, 4 of whom died of their disease; only 1 patient died of recurrence of gastric cancer. CONCLUSION: Careful follow-up for recurrence of HCC is necessary because the most common cause of death in patients with synchronous carcinoma is recurrence of HCC.

Immunologic effects of allogeneic versus autologous blood transfusion in patients undergoing radical oesophagectomy.

Blood transfusion is often required in patients undergoing radical oesophagectomy, and is associated with immunosuppression that may worsen postoperative and long-term outcomes. However, the immunologic effects of allogeneic versus autologous transfusion have not been studied in this group of patients. We analyzed 103 patients who underwent radical oesophagectomy for oesophageal cancer, including 45 patients who received allogeneic transfusions (Allo), 16 patients who donated autologous blood but were not transfused (Auto-1) and 42 patients who received autologous transfusions (Auto-2). Peripheral blood lymphocyte subsets and natural killer (NK) cell activity were analyzed for 2 weeks postoperatively. Furthermore, the rate of infectious complications such as pneumonia and wound infection was compared. Patients receiving blood transfusion had decreased CD4+ lymphocyte counts and NK cell activity postoperatively, compared to Auto-1 patients. However, these abnormalities were corrected by day 14 in the Auto-2 group, but not in the Allo group. CD8+ lymphocyte counts were decreased in all groups postoperatively, returning to normal by 14 days in the Auto-1 group only. The rate of infectious complications was significantly higher in the Allo than in the Auto group. Blood transfusion is associated with adverse immunologic effects in patients undergoing radical oesophagectomy. However, autologous blood transfusion is favourable compared to allogeneic transfusion. Autologous transfusion programs should be employed when possible in this group of patients.

Laparoscopic distal gastrectomy with regional lymph node dissection for gastric cancer.

Recently, a minimally invasive operation for gastric malignancies has been advocated, and the laparoscopic operation is noted as a technique that increases the quality of life. We performed distal gastrectomy with regional lymph node dissection on 160 cases of gastric cancer located in the middle or lower third of the stomach. In 123 cases, Billroth I reconstruction was performed intracorporeally using the quadrilateral (square) stapling technique with a laparoscopic linear stapling device to prevent postoperative anastomotic bleeding and stenosis. In the remaining 37 cases, the Billroth II method was performed with a linear stapling device [1]. This technique is not only less invasive but also as safe as open gastrectomy, which was performed on 100 gastric cancer cases of similar staging.

A comparison of video-assisted thoracoscopic oesophagectomy and radical lymph node dissection for squamous cell cancer of the oesophagus with open operation.

BACKGROUND: A direct comparison of open operation and video-assisted thoracoscopic surgery (VATS) for radical oesophagectomy has yet to be published. METHODS: Medical records of 149 patients with oesophageal squamous cell carcinoma who underwent oesophagectomy and three-field lymphadenectomy were reviewed. Seventy-seven patients had the thoracic procedure performed via a 5-cm minithoracotomy and four ports (VATS group); the others were operated on by conventional posterolateral thoracotomy (open group). RESULTS: The mean number of retrieved mediastinal nodes, blood loss and morbidity were similar in the VATS and open groups (33.9 versus 32.8 nodes, 284 versus 310 g, and 32 versus 38 per cent respectively). The thoracic procedure took longer in patients having VATS than in the control group (227 versus 186 min; P = 0.031). Vital capacity reduction was less with VATS than in the open group (15 versus 22 per cent; P = 0.016). The 3- and 5-year survival rates were similar: 70 and 55 per cent respectively for VATS compared with 60 and 57 per cent for the open procedure. CONCLUSION: VATS provides comparable results to open radical oesophagectomy, with less surgical trauma.

Learning curve of video-assisted thoracoscopic esophagectomy and extensive lymphadenectomy for squamous cell cancer of the thoracic esophagus and results.

BACKGROUND: The efficacy of thoracoscopic radical esophagectomy for cancer of the thoracic esophagus and the learning curve required have yet to be clearly established. METHODS: Eighty treatment-naive patients with esophageal cancer without contiguous spread underwent esophageal mobilization and extensive mediastinal lymphadenectomy through a 5-cm minithoracotomy and four trocar ports. The outcomes in the first 34 patients (group 1) and the last 46 patients (group 2) were compared. RESULTS: There were no differences in background or clinicopathologic factors between the two groups. The duration of the thoracoscopic procedure and blood loss were less (p <0.0001), the incidence of postoperative pulmonary infection was less (p = 0.0127), and the number of mediastinal nodes retrieved was greater (p = 0.0076) in group 2. Multivariate analysis demonstrated that surgical experience (number of cases performed) predicted the risk of pulmonary infection (p = 0.0331). CONCLUSION: Video-assisted thoracoscopic radical esophagectomy can be performed with safety and efficacy comparable to those of open esophagectomy. Morbidity decreases with the surgeon's experience.

Usefulness of self-expandable metallic stent with an antireflux mechanism as a palliation for malignant strictures at the gastroesophageal junction.

BACKGROUND: Patients with unresectable malignant gastroesophageal strictures often are troubled with reflux esophagitis after stent placement. METHODS: A self-expandable metallic stent (SEMS) without an antireflux mechanism was placed in seven patients with unresectable malignant gastroesophageal strictures (group A), and SEMS with an antireflux mechanism was placed in five patients (group B). After we obtained monitoring systems, two patients in group A and all the patients in group B underwent measurement of bilirubin and pH in the esophagus using a 24-h bilirubin and pH monitor. RESULTS: The mean percentage of total time less than 0.14 for use of the bilirubin absorbance unit was 12.4% in group B and 64.0% in group A. The mean percentage of total time for a pH less than 4 was 2.9% in group B and 37.8% in group A. CONCLUSION: The placement of SEMS with the antireflux mechanism can be effective not only for palliation of gastroesophageal stricture, but also for prevention of reflux.

Video-assisted thoracoscopic esophagectomy and radical lymph node dissection for esophageal cancer. A series of 75 cases.

BACKGROUND: The efficacy of thoracoscopic radical esophagectomy for cancer has yet to be established, mainly because previous reports have not included a sufficient number of cases. METHODS: Seventy-five treatment-naive patients with esophageal cancer without contiguous spread underwent esophageal mobilization and extensive mediastinal lymphadenectomy through a 5-cm mini-thoracotomy and four trocar ports. RESULTS: Video-assisted thoracoscopic surgery was performed without major intraoperative complications or emergency conversion to open surgery. We retrieved 34.1+/-13.0 mediastinal nodes, including 11.5+/-3.8 tracheobronchial nodes and 6.2+/-3.0 recurrent laryngeal nodes. Mean time of operation and blood loss were less in the last 39 patients than the first 36 (186.7+/-25.3 min and 165.4+/-101.8 g vs 270. 2+/-96.0 min and 421.5+/-31.2 g, respectively: p <0.0001 and p <0.001). Pulmonary morbidity was 5% in the later 39 patients. Survival was 90%, 80%, and 57% at 1, 2, and 5 years after surgery. CONCLUSION: Thoracoscopic radical esophagectomy has less morbidity and comparable survival to conventional surgery, after a moderate amount of experience. Mini-thoracotomy is essential to perform the procedure safely and effectively.

Prognostic factors after oesophagectomy and extended lymphadenectomy for squamous oesophageal cancer.

BACKGROUND: The prognosis of patients without nodal metastasis of oesophageal cancer is generally good, but recurrence develops in some cases. METHODS: Data on 88 consecutive patients with squamous oesophageal cancer who underwent three-field lymph node dissection from 1986 to September 1998 and who had no evidence of nodal disease were reviewed retrospectively. Disease status was based on histological examination of the section of each node with the largest surface area, stained with haematoxylin and eosin. RESULTS: The 3- and 5-year survival rates of patients without lymph node metastasis were 85 and 81 per cent respectively, better than in patients with metastasis. Twelve patients died from recurrence. Recurrence was haematogenous in nine patients and locoregional in three. Survival was worse in men, for patients with lesions located in the upper thoracic oesophagus, and in those with lymphatic or blood vessel invasion. Only the presence of lymphatic invasion correlated with survival on multivariate analysis (P = 0.04). CONCLUSION: Although survival was generally good in patients without nodal metastasis from oesophageal cancer following three-field lymph node dissection, patients with lymphatic invasion remained at risk for haematogenous dissemination.