A Systematic Review of Minimally Invasive Trans-thoracic Liver Resection to Examine Intervention Description, Governance, and Outcome Reporting of an Innovative Technique

Supplemental Digital Content is available in the text Introduction: The number of laparoscopic liver resections undertaken has increased. However, lesions located postero-superiorly are difficult to access. This may be overcome by the novel use of trans-thoracic port(s). Methods for the safe and transparent introduction of new and modified surgical procedures are limited and a summary of these issues, for minimally invasive trans-thoracic liver resections (MITTLR), is lacking. This study aims to understand and summarize technique description, governance procedures, and reporting of outcomes for MITTLR. Methods: A systematic literature search to identify primary studies of all designs describing MITTLR was undertaken. How patients were selected for the new technique was examined. The technical components of MITTLR were identified and summarized to understand technique development over time. Governance arrangements (eg, Institutional Review Board approval) and steps taken to mitigate harm were recorded. Finally, specific outcomes reported across studies were documented. Results: Of 2067 screened articles, 16 were included reporting data from 145 patients and 6 countries. Selection criteria for patients was explicitly stated in 2 papers. No studies fully described the technique. Five papers reported ethical approval and 3 gave details of patient consent. No study reported on steps taken to mitigate harm. Technical outcomes were commonly reported, for example, blood loss (15/16 studies), operative time (15/16), and margin status (11/16). Information on patient-reported outcomes and costs were lacking. Conclusions: Technical details and governance procedures were poorly described. Outcomes focussed on short term details alone. Transparency is needed for reporting the introduction of new surgical techniques to allow their safe dissemination.

L iver resection for primary and secondary tumors is an effective oncological treatment and is increasingly being performed by laparoscopic techniques. [1][2][3] Although systematic reviews suggest potential advantages of laparoscopic approaches (such as less blood loss and shorter hospital stay) compared to large abdominal wall incisions, the included studies were nonrandomized. [4][5][6][7] Recently, a single-center randomized controlled trial compared open and laparoscopic surgery for patients undergoing parenchyma-sparing liver resection. 8 This is the only reported randomized controlled trial to compare the 2 techniques and although the study included over 200 patients, it was single-center thus limiting generalizability. Despite the paucity of evidence, guidelines recommend that patients are considered for laparoscopic liver resections wherever possible. 9 Factors that can influence the decision to perform an operation laparoscopically include surgeon and multi-disciplinary team expertise, center policy, infrastructure, and approach to innovation. 10,11 Issues such as anatomy also need to be considered. Tumors situated posteriorly and superiorly within the liver are difficult to access using standard laparoscopic techniques because of the surrounding rib cage and diaphragm. 12 In these circumstances, many surgeons undertake open surgery with an upper abdominal incision. To overcome these challenges, modifications to current minimally invasive techniques have been described. These modifications, known as minimally invasive trans-thoracic liver resection (MITTLR), involve placing intercostal transthoracic ports to improve access to the dome of the liver, in addition to the laparoscopic abdominal ports. A totally trans-thoracic approach is where all the ports are placed through the diaphragm or pleura. The technique was first described in 2003. 13 Currently, MITTLR is not in widespread use although early studies often suggest that the technique is promising. 14 It is well recognized that patients offered new techniques are highly selected. 15 In addition, surgeons may under-report adverse events and are overly optimistic about innovation. 16 The complexity of surgery (eg, surgeon and team expertise, variations in pre-and postoperative management, and variations in outcome reporting) poses challenges for rigorously evaluating new procedures. 11 Historically, surgical innovations have often been adopted without adequate evidence of efficacy or safety. Without proper evaluation, unevidenced based innovations have the potential to harm patients. Therefore, the idea, development, exploration, assessment, and long-term follow up (IDEAL) framework and recommendations were developed to facilitate evaluation of the introduction of complex interventional procedures. 17,18 However, there remain concerns around the processes of patient selection, information provision to patients, technique description (specifically, modifications and their rationale), governance procedures (including steps taken to mitigate against harms), and the selection of appropriate outcomes. 19,20 In view of these issues, it is hypothesized that a systematic review of the literature of an innovative technique and detailed analysis of reporting of study design and conduct would be valuable to understand current methods for surgical innovation and identify areas where further research and improvement is needed. To our knowledge, there has not been a previous systematic review of MITTLR and no previous study has assessed the introduction of this new procedure into clinical practice. The aim of this study is; therefore, to report how MITTLR has been introduced into clinical practice by summarizing technique description, governance procedures, and how outcomes for MITTLR have been reported.

METHODS
The methods are based on a published protocol for the analysis of innovative invasive procedures 21 and the review was performed in accordance with the preferred reporting items for systematic reviews and meta-analyses guidelines. 22

Inclusion Criteria
All primary research study designs (eg, case reports, case series, comparative series) describing ''transthoracic'' or ''transdiaphragmatic'' minimally invasive liver resections for benign or malignant conditions in adults or children were eligible for inclusion. Although review articles were not included per se, where identified the included studies were cross-referenced to ensure all sources of evidence were accessed. Preclinical and cadaveric studies were excluded as were editorial/expert opinion articles. 23 Presentations and conference abstracts were excluded. Studies reporting outcomes on patients undergoing purely laparoscopic (without a trans-thoracic component) resections of postero-superior liver segments were excluded as were all duplicate studies.

Search Strategy and Study Selection
Searches were undertaken in Medline, Embase, PubMed, Web of Science, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, and Google Scholar, from inception until the April 1st, 2018. Search terms for trans-thoracic liver resection were combined using the Boolean ''AND'' operator (see Appendix 1, http://links.lww.com/SLA/B888). Bibliographies of relevant studies and the ''related articles'' link in PubMed were used to identify additional studies. Search results were imported into EndNote reference management software and de-duplicated.

Data Extraction and Management
A customized inclusion/exclusion proforma was developed to screen titles and abstracts. Screening was performed independently by at least 2 of 4 authors (SP, BGM, NSB, HFR). Differences between the authors were resolved by consensus with the study group. When no abstract was available, or the details were inadequate to enable a decision about eligibility, the full article was reviewed. The full-text versions of papers included after abstract screening were accessed for further assessment of their eligibility. Reference lists were manually searched for additional relevant articles. Data extraction was undertaken independently by at least 2 authors (SP, BGM, NSB, HFR) using a standardized proforma, and any disagreement resolved by consensus and/or discussion with the senior author (JMB). The following categories of data were extracted from the included papers:

Characteristics of Included Studies
The type of study design, year and journal of publication, country of origin and number of participating centers and patients were recorded. If studies reported patient selection criteria these were documented to provide clarity about patients not suitable for the technique (to prevent harm to future patients). Funding details or other potential conflicts of interest were documented.

Technique Description and Co-interventions
The surgical description of the MITTLR technique reported was assessed using a typology which allows systematic deconstruction of an intervention to understand the individual surgical components and steps. 24 This was undertaken because existing guidance (SPIRIT and TIDierR) for intervention description in study protocols does not request sufficient detail to understand how to reproduce a surgical procedure. 25,26 MITTLR was deconstructed according to its component parts (pre skin incision considerations, incision and access, dissection, resection, hemostasis, reconstruction, after skin closure, and use of adjuncts). Then each paper was studied to understand which components had been reported. We also examined the authors' given rationale for changing components or adding or removing them based on their experience. This was done for each article chronologically. It allowed us to understand how the technique was undertaken and how it evolved over time Co-interventions were documented. They were defined as interventions that naturally accompany or are associated with the intervention itself and can occur before, during or afterward. 27 For example, 1 lung ventilation using a dual lumen endo-tracheal tube may be considered a co-intervention during MITTLR.

Analysis of Studies Against IDEAL Recommendations
Studies were retrospectively classified into IDEAL stages. 17 Each study was then assessed to determine whether stage-specific IDEAL guidelines had been followed.

Surgeon Expertise
Details of the type of center undertaking the intervention, including previous case volumes were recorded. The number and level of seniority of surgeon(s), including their experience with minimally invasive liver surgery and MITTLR, were extracted. Documentation of whether any surgical learning curves were described was also recorded.

Governance Arrangements
Reporting of information describing governance approvals [ethics committee, hospital clinical governance departments, Institutional Review Board (IRB), clinical effectiveness committee, and the National Institute for Health and Care Excellence] was documented because of the importance of transparent introduction of innovation.
Details regarding whether patient consent for the novel technique was obtained was recorded, specifically examining if patients were informed about the innovative nature of MITTLR. If reported, the number of patients declining the intervention was recorded as a measure of patient acceptability. Details of any steps taken to mitigate harm were recorded (eg, whether an independent oversight committee reviewed the new procedure and its outcomes).

Outcome Reporting
Information related to outcome reporting was collected in the following categories: adverse events (unintended injury or complication resulting in prolonged hospital stay, disability at the time of discharge or death), 28 clinical (eg, operative time and histological margin status), process (eg, length of stay), patient-reported (report of the status of the patient's health condition that comes directly from the patient, without interpretation of the patient's response by a clinician), 29 and economic outcomes. It is expected that the types of outcomes reported in studies evaluating innovative procedures focus on safety and efficacy -and acceptability to surgeon and patient.

Data Synthesis and Statistical Analysis
Results are summarized in a narrative synthesis. No metaanalysis was performed because the purpose of this review was to describe how MITTLR is reported in the surgical literature rather than to compare outcomes with either open or laparoscopic liver resections.

RESULTS
Of 2067 articles, 23 were identified for further review and 16 full papers were included (Fig. 1).

Characteristics of Included Studies
The 16 studies were performed in 6 countries and included 145 adult patients. The first was reported by a Japanese group in 2003. 13 There were 6 case reports [30][31][32][33][34][35] and 10 noncomparative single-center studies (Table 1). Two stated patient inclusion criteria. 13,36 Eleven papers stated that there were no conflicts of interest. Eight papers declared financial conflicts of interest, although only 3 studies were funded. All 3 37-39 were from educational grants.

Technique Description and Co-interventions
The index paper was a Japanese series describing 3 cases where tumors were excised using a totally trans-thoracic approach. 13 The index paper did not provide a full deconstruction of the technique, although some components were well described (patient and port positioning, use of intraoperative ultrasound, method of liver transection, and specimen extraction). The subsequent paper (62)   reference the index paper and provided further understanding of patient suitability, anesthetic co-interventions, port type used, technical factors such as hemostasis, and use of drains ( Table 2). The next paper to reference the index paper was 45 the seventh report. One other paper 41 referenced the index paper but neither study provided further insight into technique evolution. Nine studies cited other papers describing MITTLR but did not reference the index paper. 30,31,35-38,41 -43 None of these subsequent papers described each component step in detail and minor technical modifications were not documented or justified. It was therefore not possible to determine how the technique had evolved with time or whether specific components had evolved ( Table 2). An opportunity for incremental learning to build upon the initial single center reports was lost.
Steps of the procedure that were well reported (and continued to be well reported as the procedure evolved) included laparoscopic/thoracoscopic stack settings, patient positioning, port placements, use of intraoperative ultrasound, liver mobilization (if required) and parenchymal transection. Although most studies described port placement, there were variations between studies

Analysis of Studies Against IDEAL Recommendations
All the studies were single-center retrospective case studies and were; therefore, classed as IDEAL stage 2a (''procedure under development, performed by a few innovators and early adopters on a small number of patients''). Whilst the first paper would ordinarily be categorized as a stage 1 IDEAL study (ie, first in human), this paper itself references a ''nonindexed'' description of the first case. No studies made a protocol available for review. Selection criteria and proportion of eligible patients selected should also be reportedno papers provided the proportion of eligible patients recruited. Whilst patient characteristics and short-term clinical outcomes were generally well reported, cases were not reported sequentially and iterative changes to the technique (common in IDEAL stage 2a) were not clearly described. Ethical approval was reported in 5 studies and informed consent was not adequately described in any study (Table 3). Learning curves and methods to mitigate harm (eg, mentoring for first few cases) were also not described in any of the papers.

Surgeon Expertise
None of the papers disclosed the center-specific caseload for either minimally invasive liver surgery or MITTLR. Only 1 study 44 stated the number of surgeons performing MITTLR (single surgeon) and level of seniority of participating surgeon was only provided in 1 paper. 38 Furthermore, no studies reported specific training regarding the use of MITTLR for participating surgeons, the expertise of the team supporting the surgeon or use of a ''preceptorship'' during introduction of MITTLR to their center.

Governance Arrangements
Only 5 studies 37,38,41,44,45 reported that they had sought and been given ethical approval. Two studies documented registration with an IRB or hospital clinical governance department. 41,46 Three studies documented informed patient consent 32,37,38 although specific details about whether patients were informed about the innovative nature of MITTLR, and its associated risks, were not provided. None of the remaining papers mentioned consent at all. None of the included studies reported the number of patients declining the intervention. No studies mentioned any steps taken to mitigate harm and whether independent oversight committees reviewed MITTLR and its outcomes.
Report and explain all exclusions

Outcome Reporting
Outcome reporting is summarized in Table 4. Two studies 38,44 provided a priori definition of adverse events.
Fifteen studies detailed complications without a single mortality reported. In terms of clinical outcomes, most of the studies reported blood loss (15/16) and margin status (11/16) as surrogate markers of surgical quality. Only 3 studies 36-38 mentioned factors likely to cause a failure of the minimally invasive trans-thoracic approach and conversion to an open approach. Reasons for conversion included bleeding, poor views, and concern about tumor margins. Operative time (14/16) and length of stay (14/16) were used as descriptors of process in most of the studies. Follow up was described in 4 studies and ranged from 4 to 61 months. Clinical outcomes are summarised in Table 5.
Patient-reported outcomes and cost-effectiveness were not measured in any of the studies.
Three studies drew no conclusion regarding MITTLR practice in the future, 3 studies felt further evaluation was warranted, 7 studies felt the procedure was safe and feasible whilst 3 studies concluded that MITTLR should be adopted by surgeons.

DISCUSSION
This is the first systematic review examining MITTLR. It identified 16 studies that were case reports and non-comparative case series. Whilst these small studies showed promising results the reporting limits their value to inform practice. Full technical description of the technique was lacking from all publications including the index study published in 2003. 13 Limited data on patient selection criteria and technique description were available. This means that it is challenging for other surgeons (early adopters) to reproduce the technique safely in appropriate patients. Errors may be inadvertently repeated. In addition, the governance arrangements (including patient consent) and regulatory approvals were poorly reported. Overall, current reporting of this innovative and evolving technique was poor and lacks rigor. Better and more transparent methods for evaluating and reporting new procedures are needed so that surgeons adopting the technique can do so ethically and with minimal risk to patients, to optimize safe uptake.
Surgery is a complex intervention making it difficult to evaluate rigorously. The IDEAL framework, developed in 2009, is 1 proposed method for introducing and evaluating new surgical techniques. 18 None of the reports describing MITTLR adhered to all the IDEAL guidelines, despite 15 of the 16 studies being published after IDEAL guidance became available. All the included studies would retrospectively be categorized as IDEAL stage 2a (development stage), suggesting that the technique has not yet stabilized, despite having been first reported over 15 years ago. There has been a lack of incremental learning and stepwise progression over the 15year period -all 16 studies were either case reports or noncomparative single-center, single-surgeon series. An opportunity to build upon this work and perform multi-center studies with short-and long-term outcomes to provide insight into the technique has been missed. Learning curves could have been shortened by collaboration between the ''pioneering'' surgeons to standardize techniques and outcomes to be monitored in future studies. This concept has been demonstrated in laparoscopic liver surgery where ''early adopters'' who received specific training with standardization of techniques had a shorter learning curve than the ''pioneering'' surgeons. 46 The nonadherence to IDEAL in evaluating MITTLR has may have resulted in reduced compliance with governance arrangements (gaining ethical approval and informed consent from patients) and poor study design with inadequate reporting of patient selection and technical modifications. In this review, it was identified that most studies did not report that ethical (IRB) approval was obtained. The lack of formal ethical approval may be because surgeons view MITTLR as a simple variation of laparoscopic surgery and therefore consider ethical approval unnecessary. A new or modified technique; however, could be associated with additional or unknown risks and provision of patients with such information is of importance. There may also be requirements to inform patients if a surgeon is undertaking a new procedure for the first time or if the technique is still evolving as the short and long-term sequalae may not yet be fully defined. 47,48 In addition to not reporting whether patient consent for a new procedure was obtained, it was noted that publications did not report the impact of the surgical learning curve on outcomes and whether steps were taken to mitigate potential harm to the patient as a surgeons' experienced developed. Methods to mitigate the effects of the learning curve include cadaveric training, visiting specialist centers and/or mentoring by visiting surgeons with expertise. 49 No publications clearly reported how the surgeon's own experience and learning curve was relayed to patients.
Whilst this study is the largest and most comprehensive review of the literature reporting MITTLR to date, it has some limitations. Data from the papers were extracted verbatim and it was assumed that if information was not documented it did not happen. In addition, it was not possible to contact all authors individually to obtain further information about technique evolution and local governance arrangements. It is possible that early adopters may have implemented MITTLR in their clinical practice but have not yet published their results. Only full-text publications were included as it was considered that only these would provide the necessary methodological details required to robustly assess technique development, governance procedures, and outcome reporting. This study demonstrates significant deficiencies in the introduction of this innovative procedure with regards to (i) description of patient selection, (ii) information provision to patients, (iii) technique description, and (iv) governance procedures. Surgical innovators should continue to use the IDEAL framework, although descriptions of patient selection, surgical technique, governance procedures, and outcome reporting need to be optimized to ensure new procedures are reproducible.
To enhance the evaluation and adoption of novel procedures, there is a need to (i) identify possible benefits early (and equally importantly, identify unsuccessful or harmful procedures so that they can be stopped in a timely way); (ii) provide good leadership to champion and advocate adoption whilst encouraging continuous assessment of outcomes; (iii) provide accurate descriptions of the technique and careful documentation of when (and why) this evolves; and (iv) consider how impact will be measured. Information provision to patients detailing the innovative nature of the procedure, the surgeon's experience with the proposed procedure, the risks and benefits including uncertainty around risks and alternative treatments, are all crucial. 50 Surgical innovation is required to develop novel procedures that improve patient outcomes. However, innovation it is important that this is conducted safely and transparently to optimize learning and minimize patient harm. Clear guidance and regulatory frameworks that build on IDEAL for surgical innovation are necessary.