European consensus on essential steps of Minimally Invasive Ivor Lewis and McKeown Esophagectomy through Delphi methodology

Background Minimally invasive esophagectomy (MIE) is a complex and technically demanding procedure with a long learning curve, which is associated with increased morbidity and mortality. To master MIE, training in essential steps is crucial. Yet, no consensus on essential steps of MIE is available. The aim of this study was to achieve expert consensus on essential steps in Ivor Lewis and McKeown MIE through Delphi methodology. Methods Based on expert opinion and peer-reviewed literature, essential steps were defined for Ivor Lewis (IL) and McKeown (McK) MIE. In a round table discussion, experts finalized the lists of steps and an online Delphi questionnaire was sent to an international expert panel (7 European countries) of minimally invasive upper GI surgeons. Based on replies and comments, steps were adjusted and rephrased and sent in iterative fashion until consensus was achieved. Results Two Delphi rounds were conducted and response rates were 74% (23 out of 31 experts) for the first and 81% (27 out of 33 experts) for the second round. Consensus was achieved on 106 essential steps for both the IL and McK approach. Cronbach’s alpha in the first round was 0.78 (IL) and 0.78 (McK) and in the second round 0.92 (IL) and 0.88 (McK). Conclusions Consensus among European experts was achieved on essential surgical steps for both Ivor Lewis and McKeown minimally invasive esophagectomy. Supplementary Information The online version of this article (doi:10.1007/s00464-021-08304-5) contains supplementary material, which is available to authorized users.

and courses aim to improve surgical proficiency and shortening the learning curve. However, surgical techniques are heterogeneous and essential steps of the procedure have not been established, which complicates teaching of a standardized and effective form of MIE. In addition, a consensus on the essential steps of MIE can be a foundation for a widely accepted evidence-based and structured way of training and assessment of surgical technique, which could aid in quality assurance, surgical learning and reducing learning associated morbidity.
Therefore, the primary objective of this study was to achieve international expert consensus on essential steps for both Ivor Lewis (IL) and McKeown (McK) MIE ( Fig. 1) by using the Delphi methodology. Since IL and McK are the most preferred MIE approaches [3], both were incorporated in this study.

Round table meetings
Led by peer-reviewed literature [14], three practicing surgeons from two high-volume hospitals experienced in thoracolaparoscopic esophagectomies (an average of 100 MIE performed so far), accompanied by one expert in surgical education, determined and defined consecutive steps required to complete MIE. The meeting was recorded to transcribe the steps, which were iteratively sent to the three surgeons for verification and refinement until the surgeons approved the version individually. This final list was used in the Delphi rounds.

Delphi methodology
The Delphi methodology was used to achieve consensus on the essential steps of both IL and McK MIE ( Fig. 1) and has been widely used in determining essential steps of other surgical procedures [15][16][17][18]. It is a process in which experts express their ideas using a questionnaire [19,20]. Based on the responses and comments, items are adjusted, removed or added, and then resent for another round. This iterative process is ended when consensus is achieved.

Expert panel
An international panel of practicing MIE surgeons was invited to participate in the Delphi rounds. Experts were selected based on surgical experience and involvement in training and education of surgical residents. A minimum of 100 esophagectomies and at least 3 years of experience in total MIE was required for participation. Based on expertise, involvement in research and education, we invited all members of the European Minimally Invasive Oesophagectomy (MIO) Think Tank as well as the majority of the Dutch high-volume centers. A total of 36 surgeons were invited to participate in the first round of this study. Experts were contacted by a personal invitation email, in which the aim of this study was elaborated. Then the survey was sent, followed by a personal reminder three to four weeks thereafter. Only surgeons that actively declined participation or those that did not meet inclusion criteria were not invited for next rounds.

Ivor Lewis vs. McKeown
Since IL and McK resection are the most common MIE approaches performed [3], an individual list of essential steps was constructed for both procedures. The two lists contained several anastomotic techniques (i.e., hand-sewn end-to-end (E/E), stapled side-to-side (S/S) and stapled endto-side (E/S) anastomosis for the IL approach and hand-sewn E/S, hand-sewn E/E, stapled E/S and stapled S/S for the McK approach). All participants received both lists and were asked to rate the MIE approach(es) and anastomotic technique(s) they regularly practiced. An anastomotic technique was excluded when rated by less than five participants. Additional procedures (i.e., nasogastric tube and jejunostomy placement) were incorporated as well.

Delphi round one
An online questionnaire and database system [www.casto redc.com] was used to send out the Delphi questionnaire and to collect the data and comments. Panel members were asked to rate the importance of each step in MIE on a Likerttype scale; 1-5 (strongly disagree to strongly agree). Furthermore, they were asked to comment on their ratings and on any missing steps.

Delphi round two
Responses and comments on round one were collected and analyzed. Based on the ratings and comments, steps were modified and resent to the same participants. The main modifications in the second round were rephrasing ("redefined") steps and adding new ("new") ones. A redefined step was a step which was changed regarding content, or which was split into multiple steps or vice versa. Modification of the steps was performed in two separate sessions by expert surgeons of two Dutch high-volume centers (> 75 MIE annually). If less than 80% of participants rated a step as 4 ("agree") or 5 ("strongly agree"), the calculated percentages were presented back to the panel members as well. Steps that reached more than 80% agreement could also still be modified based on the comments and resent for another rating to improve agreement.

Determination of consensus
Consensus among panel members was determined by using Cronbach's alpha [21], which is a measure for how closely related the responses of the experts are. Missing datapoints were replaced by mean values. For scales used in research tools and for comparing groups, an Alpha of 0.7-0.8 is considered satisfactory [21]. A Cronbach's alpha of > 0.7 was deemed satisfactory for the educational and research purpose of this study. Steps were included as an essential step when they were rated as 4 (agree) or 5 (strongly agree) by more than 80% of panel members. A new round was conducted when new steps were proposed by panel members, even when Alpha exceeded 0.7.

Results
Three out of 36 surgeons did not perform total MIE and two surgeons of the remaining 33 did not meet the inclusion criteria at the time of the first invitation and were therefore excluded. In round one, 23

Ivor Lewis steps
Due to the low incidence of the use of the IL hand-sewn E/E technique (one expert), the anastomotic steps of this technique were excluded. After the first round, 68 of the remaining 126 steps were directly included based on both the results and comments of the respondents. Forty-five steps were redefined to be resent for another rating, seven were resent without redefinition, five were excluded and 34 new steps were added. Some steps were redefined into multiple smaller steps or vice versa. For each step, the percentage of agreement and the action after round one ("redefined", "resent", "included" or "excluded") are shown in detail in Online Appendix 1. In the second round, 81 steps (40 "redefined", seven "resent" and 34 "new" steps) were sent to the participants, of which 43 were excluded and 38 were included, resulting in a total of 106 included (Table 2) and 48 excluded steps (Table 3). For all steps in the second round, the origin ("redefined", "resent" or "new"), percentage of agreement and action after round two ("included" or "excluded") are shown in detail in Online Appendix 2.

McKeown steps
Due to the low incidence of the use of the stapled E/S (one expert) and stapled S/S technique (two experts), the steps of these techniques were excluded. After the first round, 64 of the remaining 116 steps were directly included in the final list of essential steps. Forty-one steps were redefined to be resent for another rating, ten were resent without redefinition, one was excluded and 37 new steps were added. The details of round one are shown in Online Appendix 3. In the second round, 87 steps (40 "redefined", ten "resent" and 37 "new" steps) were sent to the participants. In this round, 45 steps were excluded and 42 were included, resulting in a total of 106 included (Table 4) and 46 excluded steps ( Table 5). The details of round two are shown in Online Appendix 4.

Discussion
This is the first study describing consensus-based essential steps of minimally invasive esophagectomy for cancer. Consensus among European MIE experts was achieved on essential surgical steps for both Ivor Lewis and McKeown. This resulted in a distinct list of essential steps with 106 steps for each approach, describing both procedures in detail.

Strengths and limitations
One of the strengths of this study is that the adjustments after the first round were made at two separate occasions with local experts from two high-volume hospitals. A significant increase in consensus was reached after the second round, which demonstrated a high consensus rate compared to similar studies [15][16][17]. Despite a percentage of agreement ≥ 80 being the main perquisite for inclusion, comments have been used to refine or rephrase steps to improve consensus, even when this percentage was reached. Another strength of the study is that compared to the literature, and despite the length of the questionnaires (81-126 items each), high response rates of 74% and 81% were obtained for both the first and second round [16-18, 22, 23]. The international expert panel, greatly involved in education, with a vast experience in MIE and the high response rates make these lists likely to be internationally widely supported. To ensure the widely use and     support of the future assessment tools, we incorporated multiple anastomotic techniques into the questionnaires. A limitation might be that participants were asked to rate the techniques they used "on a regular basis" which could have been interpreted differently by the participants. Due to the lack of expert input on the excluded techniques (IL hand-sewn E/E, McK stapled E/S and stapled S/S), we were not able to construct a consensus-based list of these steps. Another limitation is that in both rounds datapoints were missing (nine and seven percent for first and second round, respectively). Missing datapoints were replaced by respondents' mean values to calculate Cronbach's alpha. This method has been previously described in the literature [15]. Since sufficient consensus rates were already achieved, we believe, like in other studies, inclusion and exclusion of steps in round one was justified [20]. Finally, all experts in this study were European. Despite seven different countries were represented, the results of this study may not be easily translated to other countries and continents. The lists of essential steps that were created present us with a detailed format that can be used to standardize MIE. In addition, it provides a starting point for developing procedure-specific assessment tools for both the entire as well as certain parts of the operation. Since final mastery of the procedure comes literally step-by-step, a validated assessment tool for specific parts of the procedure would facilitate specific and structured feedback for residents, fellows and surgeons. This will help to objectively evaluate and assure a surgeons' proficiency and might potentially shorten the learning curve and, more importantly, diminish the learning associated morbidity and mortality. In bariatric surgery, patients operated by surgeons in the top quartile of skills seem less likely to develop overall complications as compared to the bottom quartile [24]. Moreover, in complex oncologic procedures technical performance among credentialed surgeons varies substantially, which is significantly associated with clinical and pathological outcomes [25]. This emphasizes the need to improve one's surgical skills as fast and efficient as possible, especially in complex oncologic procedures like MIE. Procedure-specific assessment tools differentiate well between different skills levels and they seem to be more suitable for summative assessment than global rating scales [26]. For example, the procedure-specific assessment tool in laparoscopic cholecystectomy seemed to better differentiate between novice, intermediate and almost competent trainees than the OSATS and GOALS. Since complex procedures, like MIE, are taught to surgeons that already have surgical experience in other procedures, technical differences between novice and expert surgeons might be subtler. Therefore, a specific assessment tool may be more suitable to allow for a unique insight in the dependence between different levels of skills and outcome of surgery in minimally invasive esophagectomy.

Conclusion
In this study, we described consensus-based essential steps of minimally invasive esophagectomy for cancer. Future perspectives include the development and validation of an assessment tool targeting essential steps associated with clinically relevant outcome parameters.     Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.