Minimally Invasive transCervical oEsophagectomy (MICE) for oesophageal cancer: prospective cohort study (IDEAL stage 2A)

Abstract Background Minimally invasive transcervical oesophagectomy is a surgical technique that offers radical oesophagectomy without the need for transthoracic access. The aim of this study was to evaluate the safety and feasibility of the minimally invasive transcervical oesophagectomy procedure and to report the refinement of this technique in a Western cohort. Methods A single-centre prospective cohort study was designed as an IDEAL stage 2A study. Patients with oesophageal cancer (cT1b–4a N0–3 M0) who were scheduled for oesophagectomy with curative intent were eligible for inclusion in the study. The main outcome parameter was the postoperative pulmonary complication rate and the secondary outcomes were the anastomotic leakage, recurrent laryngeal nerve palsy, and R0 resection rates, as well as the lymph node yield. Results In total, 75 patients underwent minimally invasive transcervical oesophagectomy between January 2021 and November 2023. Several modifications to the surgical technique were registered, evaluated, and implemented in the context of IDEAL stage 2A. A total of 12 patients (16%) had postoperative pulmonary complications, including pneumonia (4 patients) and pleural effusion with drainage or aspiration (8 patients). Recurrent laryngeal nerve palsy was observed in 33 of 75 patients (44%), with recovery in 30 of 33 patients (91%). A total of 5 of 75 patients (7%) had anastomotic leakage. The median number of resected lymph nodes was 29 (interquartile range 22–37) and the R0 resection rate was 96% (72 patients). Conclusion Introducing minimally invasive transcervical oesophagectomy for oesophageal cancer in a Dutch institution is associated with a low rate of postoperative pulmonary complications and a high rate of temporary recurrent laryngeal nerve palsy.


Introduction
Radical oesophagectomy with two-field lymphadenectomy is part of the multimodal curative treatment of oesophageal cancer.Transthoracic minimally invasive oesophagectomy 1 reduces postoperative morbidity compared with open surgery, but the complication rates are still high.Pulmonary and anastomotic complications are seen in up to 35% and 20% of patients respectively 2 .Transhiatal oesophagectomy (open or minimally invasive) is an alternative approach and this technique is associated with a lower pulmonary complication rate [3][4][5] .As only the distal mediastinal lymph nodes can be removed, the oncological safety is under debate, despite studies showing similar survival compared with transthoracic oesophagectomy [6][7][8][9] .
Minimally Invasive transCervical oEsophagectomy (MICE) is a novel procedure, in which the oesophagus is dissected en bloc with thoracic lymph nodes via single-port mediastinoscopy through a neck incision.The lower oesophagus, stomach, and abdominal lymph nodes are dissected by laparoscopy.MICE combines the advantages of transhiatal and transthoracic approaches and, as such, may achieve optimal oncological radicality and a low risk of pulmonary complications.
Several case series on MICE have been published from Japanese and Chinese centres, showing favourable postoperative complication rates and good oncological outcomes [10][11][12][13] .MICE has not been performed within hospitals outside of Asia, given the complexity of the surgical technique, the differences between Asian and Western patients, and the possible lack of prospective data showing its benefit over transthoracic oesophagectomy.Recently, it was shown that MICE can be safely introduced into clinical practice according to the IDEAL framework 14 .The aim of the present study was to report the safety and feasibility of MICE in a Dutch tertiary referral hospital according to the IDEAL 2A recommendations.

Patients
This study was designed as an IDEAL stage 2A single-centre study.Patients with resectable oesophageal cancer (cT1b-4a N0-3 M0) were eligible for MICE.During the study interval, the indication for MICE varied from proximal to mid oesophageal tumours and/or suspected upper mediastinal lymph node metastases to all oesophageal tumours (including distal tumour localization).

Original Article
According to Dutch national guidelines, all patients were offered five weekly cycles of carboplatin and paclitaxel with concurrent radiotherapy (41.4 Gy in 23 fractions, 5 days per week) followed by surgery according to the chemoradiotherapy for oesophageal cancer followed by surgery study (CROSS) 15,16 .After neoadjuvant chemoradiotherapy, patients were restaged using PET-CT to exclude distant metastases, were re-evaluated for surgery, and underwent MICE after oral and written informed consent.

Definitions
Patients were graded according to the ASA risk classification before surgery 17 .Postoperative complications were categorized and described following the Clavien-Dindo Classification of Surgical Complications and the Esophagectomy Complications Consensus Group ('ECCG') classification [18][19][20] .Postoperative pulmonary complications were defined according to the Dutch Upper Gastrointestinal Cancer Audit (DUCA), that is pneumonia, pleural effusion including drainage or aspiration, thorax empyema, acute respiratory distress syndrome ('ARDS'), or reintubation.Pneumonia was defined and classified in accordance with the Uniform Pneumonia Score 21 .When patients displayed signs of postoperative hoarseness, a laryngoscopy was conducted by an ear/nose/throat specialist on postoperative day 7. Recurrent laryngeal nerve palsy was confirmed when the laryngoscopy showed unilateral or bilateral vocal cord palsy.Persistent recurrent laryngeal nerve palsy was defined as a minimum of 1 year of vocal cord palsy, confirmed by laryngoscopy, as described above.Tumour staging was based on the UICC TNM Classification of Malignant Tumours (8th Edition) 22 and lymph node stations were classified according to the Japanese Classification of Esophageal Cancer (11th Edition) 23,24 .

Minimally invasive transcervical oesophagectomy
A central left collar incision was made, followed by lateralization of the strap muscles and exposure of the thyroid and trachea.After dissection of the left cervical para-oesophageal lymph nodes (101L), the left recurrent laryngeal nerve was identified, tested with an intraoperative nerve monitoring system (NIM 3.0, Medtronic, Minneapolis, MN, USA), and spared.After mobilizing the cervical oesophagus, a single-port platform (GelPoint Mini, Applied Medical, Rancho Santa Margarita, CA, USA) was placed, and three trocars were inserted.After creating a pneumomediastinum (8 mmHg), sharp dissection opened the plane outside the meso-oesophagus; downward development of this plane was facilitated by pneumodissection in the confined space of the mediastinum, enabling radical en bloc dissection of the upper and mid oesophagus and its lymph nodes, while exposing and preserving, both recurrent laryngeal nerves, the thoracic duct, the azygos vein, both pleurae, and the proximal part of both vagal nerves beyond the inferior borders of both main bronchi.During the transcervical phase, the following lymph node stations were dissected: bilateral paraoesophageal nerves, 105, 108, and 110; bilateral paratracheal nerves, 106, 106tbR, and 106tbL; subcarinal nerves, 107; and bilateral recurrent laryngeal nerves, 106recL and 106recR.Proximal (that is cervical) lymph nodes alongside both recurrent laryngeal nerves (101L, 106tbL, 101R, and 106tbR) could be approached directly through the same left collar incision.Below the level of the thoracic inlet, the lymph node dissection was performed using a single-port platform (GelPoint Mini, Applied Medical, Rancho Santa Margarita, CA, USA); right recurrent laryngeal nerve lymph nodes (106recR) were dissected from the retrotracheal plane.
A rendezvous between the transcervical dissection plane and the mediastinal space was achieved using a laparoscopic transhiatal approach.After mobilization of the greater and lesser curvature of the stomach and abdominal lymph node dissection (including stations 1, 2, 3, 7, 8a, 8p, 9, and 11), a gastric conduit was formed.Under direct laparoscopic view, the gastric conduit was brought up to the cervical incision for a handsewn end-to-side anastomosis.Fujiwara et al. 10,25,26 have reported a more detailed description of the surgical technique.

IDEAL framework
The IDEAL framework describes the stages of surgical innovation: Idea, Development, Exploration, Assessment, and Long-term follow-up.The introduction process of MICE in the Netherlands is in IDEAL stage 2A, in which the development of the procedure is the main goal.Besides the surgical outcomes of this safety and feasibility study, refinement of the surgical technique was registered to reach a state of sufficient stability to allow replication by others.To monitor adjustments to the surgical procedure, an evaluation form was developed to assess each MICE procedure.A total of four different phases were identified: cervical, mediastinal, abdominal, and anastomosis.For each phase, specific steps, such as lymphadenectomy or identification of the recurrent laryngeal nerve, could be assessed with regard to the level of contentment.Comments, details, changes, and areas of improvement for the next procedure were recorded.This form was available online, as well as on a desktop computer and on a mobile device to facilitate accessibility (Microsoft 365, Forms).

Statistics
Data from numerical and categorical variables are summarized using the median (interquartile range (i.q.r.)) and the number of patients (%) respectively.The 2-year overall survival was calculated as the percentage of patients alive at 2-year follow-up.All statistical analyses were performed using SPSS ® (IBM, Armonk, NY, USA; Statistics for Windows, Version 27.0).

Ethics
The introduction of MICE at Radboud University Medical Centre was concordant with the Dutch guideline 'New Interventions into Clinical Practice'.This study does not fall within the scope of the Dutch Medical Research Involving Human Subjects Act.An official exemption from the local Medical Ethics Committee 'Commissie Mensgebonden Onderzoek' Arnhem-Nijmegen and participating hospitals was obtained.This study is reported according to the STROBE guidelines 27 .

Results
A total of 75 patients operated on between January 2021 and November 2023 were included.The patient characteristics are shown in Table 1.A total of 73 of 75 patients (97%) underwent neoadjuvant chemoradiotherapy.CROSS was not indicated in two patients, because of early-stage oesophageal carcinomas in one patient with Barrett's oesophagus and because of a stenotic oesophagus after caustic ingestion in the other patient (more than 40 years ago).
Perioperative outcomes are shown in Table 2. Postoperative pulmonary complications were observed in 12 of 75 patients (16%).This included four patients with pneumoniae (Clavien-Dindo grade II) and eight patients with pleural effusion with the need for drainage or aspiration (Clavien-Dindo grade IIIa).
A total of 9 of 75 MICE procedures (12%) were converted into transthoracic minimally invasive oesophagectomy procedures, due to suspicion of a tumour invading adjacent organs (pericardium or airway), loss of surgical view in the mediastinum because of obesity, or a thoracic inlet too narrow for access to the mediastinum.Anastomotic leakage was observed in 5 of 75 patients (7%), including 2 patients who underwent surgical re-intervention (Clavien-Dindo grade IIIb).
Postoperative hoarseness was present in 36 of 75 patients (48%).In 33 of 75 (44%) patients recurrent laryngeal nerve palsy was confirmed by postoperative laryngoscopy and in 14 patients vocal cord augmentation was performed.None of these 14 patients had relapse of hoarseness at 1-year follow-up.A total of 30 of 33 patients (91%) showed full recovery from recurrent laryngeal nerve palsy 1 year after surgery.
The surgical margin was tumour free (pR0) in 72 patients (96%).The median lymph node yield per patient was 29 (i.q.r.22-37).Fig. S1 shows an intraoperative snapshot after subcarinal lymph node dissection.During the study interval, 30 patients completed 2-year follow-up and 18 were alive at 2 years (giving a 2-year overall survival rate of 60%).
Between February 2022 and September 2023, 18 evaluation forms for MICE were completed.This led to changes in surgical technique, which are shown in detail in the Supplementary Results.Briefly, these changes included: use of recurrent laryngeal nerve monitoring; inclusion of distal oesophageal tumours; preservation of the strap muscles; a different order of procedural steps (mediastinal or abdominal first; dorsal or ventral mobilization of the upper mediastinal oesophagus first); use of different camera systems (5 mm flexible tip, three-dimensional, or 10 mm 30°); exploring different extraction sites (neck or abdomen); a lower threshold for conversion in case of a difficult cervical approach; development of a total laparoscopic pull-through procedure for the gastric tube; and changes in the surgical staff.Most recently, a new approach to the left recurrent laryngeal nerve was implemented, preserving its attachments to the lateral vascular sheet for better vascularization and protection (see Fig. S2).

Discussion
This study shows that MICE after neoadjuvant chemoradiotherapy is associated with low rates of postoperative pulmonary complications and acceptable pathological outcomes.These findings support the careful adoption of MICE, which offers radical oesophagectomy with two-field lymphadenectomy without the need for transthoracic access.The main drawback of the procedure is a high rate of recurrent laryngeal nerve palsy,  although 91% of patients recovered 28 .With progression through the learning curve and the use of tools like intraoperative nerve monitoring, recurrent laryngeal nerve palsy rates might be reduced 29 .Recurrent laryngeal nerve palsy in the present study did not seem to lead to additional morbidity (that is pneumonia) and vocal cord augmentation was of value to reduce hoarseness, as a bridge to spontaneous recovery 30 .
When comparing the outcome of this study with the outcomes of the Dutch Upper GI Cancer Audit data set or Japanese and Chinese studies, the results of the present study are encouraging [10][11][12][13] .An anastomotic leakage rate of 7% is low, even compared with transthoracic oesophagectomy, considering the fact that a cervical anastomosis is usually associated with a higher risk of anastomotic leakage 31 .Optimal exposure in the neck after removal of the cervical lymph nodes, a more distal location of the anastomosis, and increased experience with the cervical anastomosis might have contributed to this result.Moreover, most leakages could be treated by opening of the cervical wound or laparoscopic transhiatal drainage.The less severe sequelae of anastomotic leakage after MICE might be explained by preservation of the pleurae limiting the development of pleural empyema.
MICE can be considered radical oesophagectomy with two-field lymphadenectomy.This was confirmed by an acceptable lymph node yield and R0 resection rate.However, future studies will have to assess and determine the number of resected lymph nodes per nodal station and its anatomical borders.The 2-year overall survival rate of the first 30 patients was 60%, which is in line with data from the National Cancer Registry.Analyses of recurrence rates, recurrence free survival, and overall survival will be performed when the cohort reaches sufficient follow-up.
This study is an IDEAL stage 2A study and was therefore performed during the learning curve.Studies have shown that the learning curve for transthoracic minimally invasive oesophagectomy involves 119 consecutive cases before reaching the plateau phase with regard to postoperative complications 32,33 .Hence, outcomes of MICE may become even better with increasing experience of the surgical team, as only 75 patients have been operated on so far.Another important aspect of the current stage of development within IDEAL stage 2A 34 is the adjustment of the surgical technique.The refinements made during the study are important, as they make MICE more suitable for the patient population, surgical team, and equipment, and will eventually contribute to reaching the plateau phase of the learning curve.
When introducing a new complex surgical procedure, such as MICE, collaboration with peers in the same stage of innovation is key.There are some examples of unnecessary learning-associated morbidity when several pioneers started new procedures on an individual base 35,36 .An international collaborative group on the development of MICE has recently been established to align the surgical techniques and reach consensus on an optimal procedure.Within this group, important topics will be addressed, such as patient selection, handling of recurrent laryngeal nerve palsy, development of training programmes, and data collection for collaborative reporting of outcomes.This collaboration will be the basis for a comparative trial of MICE versus the standard procedure to assess a possible benefit of MICE in patients with oesophageal cancer within an IDEAL stage 3 design.