Critical appraisal of predictive tools to assess the difficulty of laparoscopic liver resection: a systematic review

A machine learning analysis of difficulty scoring systems for laparoscopic liver surgery

INTRODUCTION

In the last decade, several difficulty scoring systems (DSS) have been proposed to predict technical difficulty in laparoscopic liver resections (LLR). The present study aimed to investigate the ability of four DSS for LLR to predict operative, short-term, and textbook outcomes.

METHODS

Patients who underwent LLR at a single tertiary referral center from January 2014 to June 2020 were included in the present study. Four DSS for LLR (Halls, Hasegawa, Kawaguchi, and Iwate) were investigated to test their ability to predict operative and postoperative complications. Machine learning algorithms were used to identify the most important DSS associated with operative and short-term outcomes.

RESULTS

A total of 346 patients were included in the analysis, 28 (8.1%) patients were converted to open surgery. A total of 13 patients (3.7%) had severe (Clavien-Dindo ≥ 3) complications; the incidence of prolonged length of stay (> 5 days) was 39.3% (n = 136). No patients died within 90 days after the surgery. According to Halls, Hasegawa, Kawaguchi, and Iwate scores, 65 (18.8%), 59 (17.1%), 57 (16.5%), and 112 (32.4%) patients underwent high difficulty LLR, respectively. In accordance with a random forest algorithm, the Kawaguchi DSS predicted prolonged length of stay, high blood loss, and conversions and was the best performing DSS in predicting postoperative outcomes. Iwate DSS was the most important variable associated with operative time, while Halls score was the most important DSS predicting textbook outcomes. No one of the DSS investigated was associated with the occurrence of complication.

CONCLUSIONS

According to our results DDS are significantly related to surgical complexity and short-term outcomes, Kawaguchi and Iwate DSS showed the best performance in predicting operative outcomes, while Halls score was the most important variable in predicting textbook outcome. Interestingly, none of the DSS showed any correlation with or importance in predicting overall and severe postoperative complications.

Appraisal of multivariable prognostic models for post-operative liver decompensation following partial hepatectomy: a systematic review

BACKGROUND

Few reports have evaluated prognostic modelling studies of tools used for surgical decision-making. This systematic review aimed to describe and critically appraise studies that have developed or validated multivariable prognostic models for post-operative liver decompensation following partial hepatectomy.

METHODS

This study was designed using the CHARMS checklist. Following a comprehensive literature search, two reviewers independently screened candidate references for inclusion and abstracted relevant study details. Qualitative assessment was performed using the PROBAST tool.

RESULTS

We identified 36 prognostic modelling studies; 25 focused on development only, 3 developed and validated models, and 8 validated pre-existing models. None compared routine use of a prognostic model against standard clinical practice. Most studies used single-institution, retrospective cohort designs, conducted in Eastern populations. In total, 15 different outcome definitions for post-operative liver decompensation events were used. Statistical concerns surrounding model overfitting, performance assessment, and internal validation led to high risk of bias for all studies.

CONCLUSIONS

Current prognostic models for post-operative liver decompensation following partial hepatectomy may not be valid for routine clinical use due to design and methodologic concerns. Landmark resources and reporting guidelines such as the TRIPOD statement may assist researchers, and additionally, model impact assessment studies represent opportunities for future research.

A systematic review and meta-analysis of randomized controlled trials comparing laparoscopic and open liver resection

BACKGROUND/PURPOSE

The dissemination of laparoscopic liver resection (LLR) has been based on non-randomized studies and reviews of these. Aim of this study was to evaluate if the randomized evidence comparing LLR to open liver resection (OLR) supports these findings.

METHODS

A prospectively registered (reviewregistry866) systematic review and meta-analysis following Cochrane and PRISMA guidelines comparing LLR to OLR for benign and malignant diseases was performed via Medline, Web of Science, CENTRAL up to 31.12.2020. The main outcome was postoperative complications. Risk of bias was assessed with the Cochrane Risk of Bias tool 2.0, certainty of evidence was assessed using the GRADE approach.

RESULTS

The search yielded 2080 results. 13 RCTs assessing mostly minor liver resections with 1457 patients were included. There were reduced odds of experiencing any complication (Odds ratio (OR) [95% confidence interval (CI)]: 0·42 [0·30, 0·58]) and severe complications (OR[CI]: 0·51 [0·31, 0·84]) for patients undergoing LLR. LOS was shorter (Mean difference (MD) [CI]: -2·90 [-3·88, -1·92] days), blood loss was lower (MD: [CI]: -115·41 [-146·08, -84·75] ml), and functional recovery was better for LLR. All other outcomes showed no significant differences.

CONCLUSIONS

LLR shows significant postoperative benefits. RCTs assessing long-term outcomes and major resections are needed.

The learning curve of laparoscopic liver resection utilising a difficulty score

BACKGROUND

This study aimed to quantitatively evaluate the learning curve of laparoscopic liver resection (LLR) of a single surgeon.

PATIENTS AND METHODS

A retrospective review of a prospectively maintained database of liver resections was conducted. 171 patients undergoing pure LLRs between April 2008 and April 2021 were analysed. The Halls difficulty score (HDS) for theoretical predictions of intraoperative complications (IOC) during LLR was applied. IOC was defined as blood loss over 775 mL, unintentional damage to the surrounding structures, and conversion to an open approach. Theoretical association between HDS and the predicted probability of IOC was utilised to objectify the shape of the learning curve.

RESULTS

The obtained learning curve has resulted from thirteen years of surgical effort of a single surgeon. It consists of an absolute and a relative part in the mathematical description of the additive function described by the logarithmic function (absolute complexity) and fifth-degree regression curve (relative complexity). The obtained learning curve determines the functional dependency of the learning outcome versus time and indicates several local extreme values (peaks and valleys) in the learning process until proficiency is achieved.

CONCLUSIONS

This learning curve indicates an ongoing learning process for LLR. The proposed mathematical model can be applied for any surgical procedure with an existing difficulty score and a known theoretically predicted association between the difficulty score and given outcome (for example, IOC).

Impact of liver cirrhosis on the difficulty of minimally-invasive liver resections: a 1:1 coarsened exact-matched controlled study

INTRODUCTION

The impact of liver cirrhosis on the difficulty of minimal invasive liver resection (MILR) remains controversial and current difficulty scoring systems do not take in to account the presence of cirrhosis as a significant factor in determining the difficulty of MILR. We hypothesized that the difficulty of MILR is affected by the presence of cirrhosis. Hence, we performed a 1:1 matched-controlled study comparing the outcomes between patients undergoing MILR with and without cirrhosis including the Iwate system and Institut Mutualiste Montsouris (IMM) system in the matching process.

METHODS

Between 2006 and 2019, 598 consecutive patients underwent MILR of which 536 met the study inclusion criteria. There were 148 patients with cirrhosis and 388 non-cirrhotics. One-to-one coarsened exact matching identified approximately exact matches between 100 cirrhotic patients and 100 non-cirrhotic patients.

RESULTS

Comparison between MILR patients with cirrhosis and non-cirrhosis in the entire cohort demonstrated that patients with cirrhosis were associated with a significantly increased open conversion rate, transfusion rate, need for Pringles maneuver, postoperative, stay, postoperative morbidity and postoperative 90-day mortality. After 1:1 coarsened exact matching, MILR with cirrhosis were significantly associated with an increased open conversion rate (15% vs 6%, p = 0.03), operation time (261 vs 238 min, p  < 0.001), blood loss (607 vs 314 mls, p  = 0.002), transfusion rate (22% vs 9%, p  = 0.001), need for application of Pringles maneuver (51% vs 34%, p  = 0.010), postoperative stay (6 vs 4.5 days, p  = 0.004) and postoperative morbidity (26% vs 13%, p  = 0.029).

CONCLUSION

The presence of liver cirrhosis affected both the intraoperative technical difficulty and postoperative outcomes of MILR and hence should be considered an important parameter to be included in future difficulty scoring systems for MILR.

Validation and comparison of the Iwate, IMM, Southampton and Hasegawa difficulty scoring systems for primary laparoscopic hepatectomies

BACKGROUND

Various difficulty scoring systems (DSS) have been formulated to grade the complexity of laparoscopic hepatectomies (LH). This study aims to externally validate and compare 4 contemporary DSS including the Iwate, Institut Mutualiste Montsouris (IMM), Southampton and Hasegawa DSS in predicting the intraoperative technical difficulty and postoperative outcomes after LH.

METHODS

Retrospective review of 548 consecutive patients who underwent LH of which 455 met the study inclusion criteria. Outcomes measures of technical difficulty included operation time, Pringles maneuver, blood loss and blood transfusion rate. Postoperative outcomes measured included morbidity, major morbidity and postoperative hospital stay.

RESULTS

There was a statistically significant progressive increase in blood loss, blood transfusion rate, operation time and postoperative stay associated with all 4 DSS. There was also good calibration with respect to blood loss, operation time, Pringles maneuver, open conversion rate, postoperative morbidity, postoperative major morbidity and postoperative stay for all 4 DSS. The Southampton score demonstrated the poorest calibration in terms of operation time and discrimination in terms of application of Pringles maneuver and major morbidity amongst all 4 systems.

CONCLUSION

All 4 DSS significantly correlated with outcome measures associated with intraoperative technical difficulty and postoperative outcomes. The Southampton DSS was the poorest system in our cohort of patients.

Comparison and validation of three difficulty scoring systems in laparoscopic liver surgery: a retrospective analysis on 300 cases

BACKGROUND

Difficulty scores (DSs) have been proposed to rate laparoscopic liver resection (LLR) technical difficulty increasing surgical safety. The aim of the study was to validate three DSs (Hasegawa, Halls and Kawaguchi) and compare their ability to predict technical difficulty and postoperative outcomes.

MATERIALS AND METHODS

All patients who underwent LLR from January 2006 to January 2019 were analyzed. Exclusion criteria were cyst fenestrations, thermal ablation, missing data for the computation of the DS and a follow-up < 90 days.

RESULTS

The population comprised 300 patients. The DS distribution in the study population was: Halls low 55 (18.3%), moderate 82 (27.3%), high 111 (37%) and extremely high 52 (17.3%); Hasegawa low 130 (43.3%), medium 105 (35%) and high 65 (21.7%); Kawaguchi Grade I 194 (64.7%), Grade II 47 (15.7%) and Grade III 59 (19.7%). Hasegawa and Kawaguchi showed the strongest correlation (r = 0.798, p < 0.001). Technical complexity, evaluated using the Pringle maneuver, Pringle time, blood loss and operative time, increased significantly with Hasegawa and Kawaguchi score classes (p < 0.001 for all comparisons). None of the scores properly stratified postoperative complications. The highest Kawaguchi (23.7% grade III vs. 13.7% grades I and II, p = 0.057) and Hasegawa (24.6% high vs. 13.2% low/medium, p = 0.025) classes had a higher overall morbidity rate than medium-low ones.

CONCLUSIONS

Kawaguchi and Hasegawa scores predicted LLR's technical difficulty. None of the scores discriminated the postoperative complication risk of low classes compared with medium ones.

Laparoscopic liver resection: indications, limitations, and economic aspects

Background

Minimally invasive techniques have increasingly found their way into liver surgery in recent years. A multitude of mostly retrospective analyses suggests several advantages of laparoscopic over open liver surgery. Due to the speed and variety of simultaneous technical and strategic developments, it is difficult to maintain an overview of the current status and perspectives in laparoscopic liver surgery.

Purpose

This review highlights up-to-date aspects in laparoscopic liver surgery. We discuss established indications with regard to their development over time as well as continuing limitations of applied techniques. We give an assessment based on the current literature and according to our own center experiences, not least with regard to a highly topical cost discussion.

Conclusions

While in the beginning mainly benign tumors were laparoscopically operated on, liver metastasis and hepatocellular carcinoma are now among the most frequent indications. Technical limitations remain and should be evaluated with the overall aim not to endanger quality standards in open surgery. Financial aspects cannot be neglected with the necessity of cost-covering reimbursement.

Does the difficulty grade of laparoscopic liver resection for colorectal liver metastases correlate with long-term outcomes?

INTRODUCTION

Prognosis of patients with colorectal liver metastases (CRLM) is strongly correlated with the oncological outcome after liver resection. The aim of this study was to analyze the impact of laparoscopic liver resection (LLR) difficulty score (IMM difficulty score) on the oncological results in patients treated for CRLM.

METHODS

All patients who underwent LLRs for CRLM from 2000 to 2016 in our department, were retrospectively reviewed. Data regarding difficulty classification, -according to the Institute Mutualiste Montsouris score (IMM)-, recurrence rate, recurrence-free survival (RFS), overall survival (OS) and data regarding margin status were analyzed.

RESULTS

A total of 520 patients were included. Patients were allocated into 3 groups based on IMM difficulty score of the LLR they underwent: there were 227 (43,6%), 84 (16,2%) and 209 (40,2%) patients in groups I, II and III, respectively. The R₁ resection rate in group I, II and III were 8,8% (20/227), 11,9% (10/84) and 12,4% (26/209) respectively (p = 0.841). Three- and 5-year RFS rates were 77% and 73% in group I, 58% and 51% in group II, 61% and 53% in group III, respectively (p = 0.038). Three and 5-year OS rates were 87% and 80% for group I, 77% and 66% for group II, 80% and 69% for group III respectively (p = 0.022).

CONCLUSION

The higher LLR difficulty score correlates with significant morbidity and worse RFS and OS, although the more technically demanding and difficult cases are not associated with increased rates of positive resection margins and recurrence.

The external validation of a difficulty scoring system for predicting the risk of intraoperative complications during laparoscopic liver resection

BACKGROUND

This study aimed to externally validate and upgrade the recent difficulty scoring system (DSS) proposed by Halls et al. to predict intraoperative complications (IOC) during laparoscopic liver resection (LLR).

METHODS

The DSS was validated in a cohort of 128 consecutive patients undergoing pure LLRs between 2008 and 2019 at a single tertiary referral center. The validated DSS includes four difficulty levels based on five risk factors (neoadjuvant chemotherapy, previous open liver resection, lesion type, lesion size and classification of resection). As established by the validated DSS, IOC was defined as excessive blood loss (> 775 mL), conversion to an open approach and unintentional damage to surrounding structures. Additionally, intra- and postoperative outcomes were compared according to the difficulty levels with usual statistic methods. The same five risk factors were used for validation done by linear and advanced nonlinear (artificial neural network) models. The study was supported by mathematical computations to obtain a mean risk curve predicting the probability of IOC for every difficulty score.

RESULTS

The difficulty level of LLR was rated as low, moderate, high and extremely high in 36 (28.1%), 63 (49.2%), 27 (21.1%) and 2 (1.6%) patients, respectively. IOC was present in 23 (17.9%) patients. Blood loss of >775 mL occurred in 8 (6.2%) patients. Conversion to open approach was required in 18 (14.0%) patients. No patients suffered from unintentional damage to surrounding structures. Rates of IOC (0, 9.5, 55.5 and 100%) increased gradually with statistically significant value among difficulty levels (P < 0.001). The relations between the difficulty level, need for transfusion, operative time, hepatic pedicle clamping, and major postoperative morbidity were statistically significant (P < 0.05). Linear and nonlinear validation models showed a strong correlation (correlation coefficients 0.914 and 0.948, respectively) with the validated DSS. The Weibull cumulative distribution function was used for predicting the mean risk probability curve of IOC.

CONCLUSION

This external validation proved this DSS based on patient's, tumor and surgical factors enables us to estimate the risk of intra- and postoperative complications. A surgeon should be aware of an increased risk of complications before starting with more complex procedures.

Ultrasound liver map technique for laparoscopic liver resections: perioperative outcomes are not impaired by technical complexity

Intraoperative liver ultrasound has a crucial role to guide open liver surgery. A 4-step ultrasound liver map technique for laparoscopic liver resection (LLR) has been standardized in our center. The aim of this study was to evaluate outcomes of our technique according to the hepatectomy technical complexity. A difficulty scale (DS) ranging from 1 to 10 was applied to each LLR. A cumulative sum control-chart analysis identified 3 periods of gradually increasing DS. Perioperative outcomes of the 3 periods were compared. 300 LLRs performed between 2006 and 2018 were analyzed. Median DS was 3 for first 100 cases (P1), 5 for cases 101-200 (P2) and 6 for cases 201-300 (P3). A significantly greater percentage of postero-superior segments resections (P1 11%, P2 36%, P3 46%, p < 0.001) were performed in P3. P3 LLRs had a significantly longer transection time (p < 0.001) and wider cut surface area (p < 0.001), but median blood losses were similar among the 3 periods (P1 100 cc, P2 100 cc, P3 140 cc). There were no differences among periods in overall morbidity (P1 12%, P2 17%, P3 17%), major morbidity (P1 1%, P2 2%, P3 3%) and length of hospital stay (5 days in all the three groups). Despite the increasing surgical complexity of LLR, ultrasound liver map technique allows good perioperative outcomes.