Minimally Invasive Versus open AbdominoThoracic Esophagectomy (MIVATE) - study protocol for a randomized controlled trial DRKS00016773

Background: The only curative treatment for most esophageal cancers is radical esophagectomy. Minimally invasive esophagectomy (MIE) aims to reduce postoperative morbidity, but is not yet widely established. Linear stapled anastomosis is a promising technique for MIE because it is quite feasible even without robotic assistance. Aim of the present study is to compare total MIE with linear stapled anastomosis to open esophagectomy (OE) with circular stapled anastomosis with special regard to postoperative morbidity in an expertise-based randomized controlled trial (RCT). Methods/design: This superiority RCT compares MIE with linear stapled anastomosis (intervention) to OE with circular stapled anastomosis (control) for Ivor-Lewis esophagectomy. It was initiated in February 2019 and recruitment is expected to last for 3 years. For inclusion, patients must be 18 years of age or more with a resectable primary malignancy in the distal esophagus. Participants with tumor localizations above the azygos vein, metastasis or inltration into adjacent tissue will be excluded. In an expertise-based approach the allocated treatment will only be carried out by the single most experienced surgeon of the surgical center for each respective technique. The sample size was calculated with 20 participants per group for the primary endpoint postoperative morbidity according to comprehensive complication index (CCI) within 30 postoperative days. Secondary endpoints include anastomotic insuciency, pulmonary complications, other intra- and postoperative outcome parameters such as estimated blood loss, operative time, length of stay, short-term oncologic endpoints, adherence to a standardized fast-track protocol, postoperative pain, and postoperative recovery (QoR-15). Quality of life (SF-36, CAT EORTC QLQ-C30, CAT EORTC QLQ-OES18) and oncological outcomes are evaluated with 60 months follow-up. Discussion: MIVATE is the rst RCT to compare OE with circular stapled anastomosis to total MIE with linear stapled anastomosis exclusively for intrathoracic

Minimally invasive approaches and fast-track protocols attempt to lower postoperative complication rates (3,4).
Consequently, there have been several studies attempting to lower postoperative morbidity by using minimally invasive approaches for esophagectomy (5)(6)(7)(8)(9)(10). The existing studies evaluating minimally invasive esophagectomy (MIE) have shown technical feasibility and suggest similar oncological outcomes (11). The different surgical approaches relevant for esophagectomy can be differentiated according to surgical access, site of anastomosis and anastomotic technique. In terms of surgical access, the procedures can be performed via open technique, hybrid or total minimally invasive surgery comprising conventional laparoscopic/thoracoscopic as well as robotic approaches. The anastomoses can be performed with manual suturing or as stapled anastomoses with either end-to-side circular stapled anastomosis or side-to-side linear stapled anastomosis, the latter being a commonly used anastomotic technique in other elds such as obesity surgery (12). The combination of these aspects results in a great variety of possible techniques for esophagectomy that need to be carefully evaluated.
MIE is gaining more popularity due to expected lower postoperative morbidity compared to open and hybrid procedures while oncological outcomes have been promising as well (13). It is hypothesized that with the total minimally invasive approach, postoperative morbidity can be reduced while oncological outcomes stay at least equivalent to open surgery (14). Studies that have evaluated MIE in a randomized setting so far mostly used either intrathoracic circular stapled end-to-side anastomosis or cervical anastomosis (7,15). A simple and reliable alternative to presented techniques is the totally minimally invasive approach with side-to-side linear stapled intrathoracic anastomosis that was established in obesity surgery. Especially for gastric bypass surgery, linear stapled side-to-side anastomosis is a well-established technique and has the potential to decrease the number of postoperative complications in esophagectomy (12). Reasons advocating linear stapled anastomosis include technical feasibility of intrathoracic anastomosis, no requirement for robotic assistance, the fact that it is wellestablished and safe in the frame of other surgical disciplines such as gastric bypass surgery for obesity.
However, until now no randomized trials for total MIE with linear stapled side-to-side anastomosis vs. open esophagectomy (OE) with circular stapled end-to-side intrathoracic anastomosis exist. Besides the novel character of the study regarding its comparison of two surgical strategies that have not been compared so far, this trial also addresses a methodical problem that often accompanies surgical clinical trials i.e. the different skillset and experience level of participating surgeons as one of the main sources for heterogeneity and bias (16)(17)(18). Multicenter randomized controlled trials (RCTs) are the centerpieces in evidence-based surgery. These RCTs require that surgeons in different centers with varying training and expertise of a certain intervention are pooled in order to reach patient numbers that are su cient for proper statistical analysis. For optimal outcomes, surgeons should only perform procedures that they are experienced and comfortable with. This is especially momentous when investigating complex procedures such as esophagectomy and can lead to poor recruitment and biased results. A possibility of coping with this issue are clinical trials that are performed in an expertise-based manner (19), in which the surgeons performing surgeries for trials exclusively are the most experienced and peeracknowledged senior surgical members of staff. Since esophagectomy is a treatment prone to this type of bias MIVATE is conducted in an expertise-based manner.
Aim of the present study is thus to compare postoperative morbidity between total MIE with linear stapled side-toside anastomosis to open surgery with circular stapled end-to-side anastomosis in an expertise-based randomized trial.

Setting
This is a single-center RCT at the Department of General, Visceral and Transplantation Surgery, Heidelberg University Hospital, Germany. It was initiated in February 2019 and recruitment is expected to last for 3 years. The study protocol was accepted by the independent Ethics Committee of the University of Heidelberg (registration number S-317/2017) before start of the study The trial was registered at DRKS under registration number DRKS00016773 on February 18th, 2019 (20). No Secondary Identifying Numbers such as a Universal Trial Number have been assigned. Recommendations of the SPIRIT (Standard protocol items: Recommendations for Interventional Trials) checklist (Additional le 1) were followed (21).

Patient recruitment
Recruitment exclusively takes place at the Department of General, Visceral, and Transplantation Surgery at University Hospital of Heidelberg in Germany. Possible participants will be screened for eligibility. To be eligible for the study, participants must be 18 years or older with a resectable primary malignancy in the distal esophagus including adenocarcinoma of the esophagogastric junction type 1 and 2 with curative intention and no apparent metastases. Participants with tumor localizations above the azygos vein, emergency situations such as major bleeding or perforation as well as metastasis or in ltration into adjacent tissue are excluded.
All eligible participants will be informed about either operation technique, as well as their potential bene ts and side effects. Written informed consent will be obtained. Only patients who sign the informed consent form will be included. Reasons for exclusion from the MIVATE trial will be documented and explained in the screening form. Thereafter, patients will be randomized to the intervention arm (MIE with linear stapled anastomosis) or the control arm (OE with circular stapled anastomosis). For the duration of the treatment until discharge no concomitant interventions besides the in-house standards and study-related interventions are allowed.

Outcome measures
During the MIVATE trial ( Fig. 1), participants will be monitored before surgery, intraoperatively, on postoperative days (POD) 1-7 and on the day of discharge. Follow-up will be conducted on POD 30, as well as 3 months, 12 months, 36 months and 60 months postoperatively. During follow-up, patients will always complete a professionally administered questionnaire. Demographic and baseline clinical data, intraoperative ndings and postoperative results will be recorded. To enhance participant retention and to avoid loss to follow-up, patients will be called during the follow-up period to remind them of scheduled visits and to arrange appointments. When a patient is not able to participate in a follow-up visit in person, questionnaires will be administered by telephone interview. Informed consent will be obtained and trial data will be collected by trained assessors by using CRFs and established questionnaires,

Primary endpoint
The primary endpoint will be postoperative morbidity assessed as the comprehensive complication index (CCI) within 30 days after index operation (22), which enables to compare the severity of postoperative complications (23).
Postoperative morbidity is de ned as any deviation from the normal postoperative course according to the Dindo-Clavien classi cation (24). Speci cally, it includes anastomotic insu ciency or loss of anastomotic integrity veri ed by either CT scan with detection of contrast agent externally from the anastomosis within the abdominal or thoracic cavity, endoscopy or the detection of methylene blue in the drainage after oral application. Also, it includes pneumonia with radiological veri cation of pneumonic in ltrates and a minimum of 3 of 4 possible symptoms including a body core temperature above 37.5 °C, purulent expectoration, leucocyte count above 12 000 or below 4 500 /ml or increased C-reactive protein (CRP) levels. Postoperative complications also include pancreatic stula, which is de ned by the International Study Group for Pancreatic Surgery (ISGPS) as the "drain output of any measurable volume of uid with an amylase level > 3 times the upper limit of institutional normal serum amylase activity, associated with a clinically relevant development/condition related directly to the postoperative pancreatic stula". The former grade A stula is now called a "biochemical leak", as it has no clinical relevance. Fistula grades B and C are de ned more precisely. Grade B requires drains that are either left in place > 3 weeks or repositioned. Grade C requires reoperation or leads to single or multiple organ failure (25,26).
Recommendations for reporting of complications after esophagectomy will be obeyed in order to improve comparability with future trials (27). Other aspects are postoperative bleeding with a hemoglobin relevant decrease beyond 3 g/dl or the necessity for transfusion of erythrocyte concentrates due to bleeding into the abdominal or thoracic cavity. Wound healing disorders with special wound treatment, abscess and lymphatic stula caused by damage to the lymphatic system with leakage of chyle uid into the cavities (de ned as a milkycolored uid from a drain, drain site, or wound on or after POD 3, with a triglyceride content ≥ 110 mg/dL respectively ≥ 1.2 mmol/L) also accounts for postoperative morbidity (28,29). Further aspects are tracheal injuries with stula between esophagus and trachea and loss of tracheal integrity as well as radiologically con rmed deep leg vein thrombosis and pulmonary embolism. Acute kidney failure in direct context to surgery de ned as a doubling of plasma creatinine levels or necessity for hemodialysis as well as stroke and myocardial infarction are further criteria included in postoperative morbidity.

Secondary endpoints
Secondary endpoints can be separated into short-term endpoints and long-term endpoints. Short-term endpoints Long-term endpoints are QoL and oncological outcomes such as disease-free-survival, rate of local recurrence and overall-survival. QoL will be assessed with different questionnaires. SF-36 and CAT EORTC QLQ-C30 measure general aspects of health with scores ranging from 0 to 100 and with higher scores representing better well-being.
CAT EORTC QLQ-OES18 assesses several aspects of esophageal function, ranging from 0 to 100 with lower scores indicating better function (31).
Extended details of the secondary endpoints can be found in Table 1 ASEPSIS score for wound infection (33) and the standardization of data collection for complications associated with esophagectomy from the Esophagectomy Complications Consensus Group (ECCG) (27).

Standardized therapy and trial interventions
OE, hybrid esophagectomy and totally MIE with circular stapled anastomosis are established in single highvolume centers and described to have similar oncological outcomes (14,34,35). In the present study, the singularizing aspect is the linear side-to-side stapled anastomosis derived from bariatric surgery where it is established as a technique with excellent risk-bene t ratio and low anastomotic stricture rate compared to circular stapled and hand-sewn anastomosis (34,35).
In the present study, for both interventions intubation is done with a double lumen tube and patients receive antibiotic prophylaxis perioperatively with Ampicillin-Sulbactam (3 g single-shot) or other in case of allergies. Surgery starts with the abdominal part.
In case of open surgery, the patient is placed in "Crawford"-position. After median laparotomy, the surgeon performs inspection of the abdominal cavity to ensure the absence of metastases and peritoneal carcinomatosis. For standardization purposes surgeons performing this type of surgery for following trials must have surpassed the learning curve which is described to last until at least 50 esophagectomies with the speci c technique (37,38). This is clearly relevant in order to reduce surgeon-related in uences as it could be shown by Nimptsch (2). Therefore, an expertise-based design was chosen for the MIVATE trial and only the single most experienced surgeon of the center for the respective technique is performing trial surgeries.

Modi cation of the protocol
The current protocol version from March 2020 is the protocol the trial was initiated with (protocol version 1.0). In case of protocol amendments, these will be submitted to the ethics committee for approval and no further recruitments will take place until the modi cations are accepted.

Assessment of safety and termination criteria
All adverse events will be documented and analyzed because complications form the study's primary endpoint in the form of the CCI. Participants will be excluded from the study if they withdraw their consent to participate in the trial. A participant may withdraw consent at any time without explanation and without affecting further medical care. The principal investigator may terminate the trial at any time in consultation with the key research associates and the biostatistician. Possible reasons for termination include high morbidity or mortality rates and any indication of potential health hazards caused by either the study treatment or external factors.

Randomization and blinding
After informed consent, patients will be randomized on the day before operation to the intervention or control group. The random sequence generation was performed by an employee of the clinical trial center not otherwise involved in the trial. Block sizes of 4, 6 and 8 were used in a variable order. Allocation was performed by opening envelopes. Therefore, cards displaying "Endoscopic" or "Open" were put in sequentially numbered, sealed, opaque envelopes according to the generated random sequence.
Blinding of study contributors (39): No attempt will be done to blind patients and the access sites will be covered with standard wound dressings until discharge. The blinding of the operating surgeon is not possible. The severity of pain and rescue analgesic will be evaluated by anesthesiologists otherwise not involved in the study (data collectors). All cases are reviewed regarding the primary endpoint by a neutral outcome assessor. The statistical analysis will be performed according to the outlined protocol; no additional attempts are made to blind the statistician as this will have no in uence on the prede ned statistical analysis of previously recorded and saved data.

Data management
All data will be collected and recorded in case report forms (CRFs) by an investigator before transfer to the data management center. Personal information about potential and enrolled participants will be collected, shared, and maintained with third-party only after pseudonymization in order to protect con dentiality. All demographic and baseline clinical data, as well as primary and secondary outcome measures, will be recorded in the CRF. To promote data quality there will be automated checks for double data entry and value ranges. To ensure patient con dentiality, the CRF for each patient will be given an anonymous allocation number. We will obtain permission to continue follow-up and data collection in the event of withdrawal from the study. The responsible investigator must review and sign all completed CRFs.

Statistical analysis
Superiority of the intervention versus the control will be assessed using a two-sided t-test. The primary analysis will be based on the intention-to-treat population. If values do not display normal distribution the Mann-Whitney-U-Test will be used. A per protocol and an as treated set will be evaluated as a sensitivity analysis. Missing data for the primary outcome variable will be replaced by using imputation (40). The primary analysis will test the following hypotheses: whereas the second subintervention is the use of two different anastomotic techniques that are speci ed for each group. The linear stapled anastomotic technique is only used in the MIE group, whereas the circular stapled approach is only used in OE. This is mainly due to practical reasons since the linear stapling technique is facilitated by the minimally invasive access with trocars whereas this technique is more cumbersome in open approach due to angulation and access di culties. On the other hand the circular stapling technique is more practical in open surgery whereas in minimally invasive surgery this proves more di cult. This combination of two different steps of intervention leads to the problem that identi ed differences between the groups can possibly not be ascribed to one speci c interventional aspect. However, while complex interventions are not necessarily suited to explain effects mechanistically -which is not the scope of this trial -they are recognized to be more sensitive to differences between groups and to more effectively re ect on reality. Consequently, the advantage of this complex intervention is its ability to screen for several interventional steps at once resulting in an increased probability of difference detection as well as its representative validity of esophageal surgery representing regularly performed operation methods and thereby increasing the relevance and external validity of this trial. Yet, there will be descriptive subgroup analysis to identify potential major in uencing factors.
Another strength of the MIVATE trial is the extensive surgical standardization. There is no surgical heterogeneity within the groups as every single patient of one group is treated by the same highly experienced surgeon according to the aforementioned description of expertise-based principles. This reduces bias by ensuring that every patient gets the best surgical expertise available and avoids both the "problem of the poor control group" (45) and poor recruitment. Furthermore, patients follow a highly standardized fast track protocol for postoperative recovery, of which the adherence will be documented and reported.
The MIVATE trial speci cally addresses the issue of short-term complications in comparing OE and MIE with two different anastomotic techniques with the intention of making a contribution to an optimized treatment strategy for esophageal cancer. Long-term oncological outcomes will have to be evaluated in further multicenter trials with adequate power.
In summary, this monocenter trial will evaluate the difference between OE and total MIE focusing on postoperative complications. The MIVATE trial will provide further evidence of optimal technique for oncologic esophagectomy comparing the open and total minimally invasive approach. The two singularizing aspects of this trial are the linear stapled anastomosis for total MIE as well as the expertise-based approach. The ndings will serve as a basis for conducting multicenter RTCs to evaluate which procedure is best for patients requiring esophagectomy in order to further optimize outcomes and reduce complications.  Prof. Dr. Müller-Stich is the study sponsor and is involved and has ultimate authority over study design, data collection, interpretation of data and writing of the report. Data management & analysis is done by a third party to ensure objectivity.
The Division of Minimally Invasive Surgery at the University Hospital of Heidelberg is the coordinating centre of this trial and provides the steering committee; statistical planning and analysis is done by the IMBI Heidelberg (Institute of Medical Biometry and Informatics) and randomization as well as assistance with data management and data monitoring is done by the SDGC Heidelberg (Studienzentrum der Deutschen Gesellschaft für Chirurgie), which forms the independent data monitoring committee (DMC). Strategies to improve adherence include a trial routine within the SDGC as well as regular patient visits by the trial physicians or medical students.

Ethics approval and consent to participate
This protocol study received approval from the Ethics Committee of the University of Heidelberg (registration number S-317/2017). All patients receive clari cations regarding the objectives and procedures and written informed consent will be obtained from those who agree to participate.