If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
In patients with stage IV colorectal cancer (CRC), minimally invasive surgery (MIS) may offer optimal oncologic outcome with low morbidity. However, the relative benefit of MIS compared to open surgery in patients requiring multistage resections has not been evaluated.
Methods
Patients who underwent totally minimally invasive (TMI) or totally open (TO) resections of CRC primary and liver metastases (CLM) in 2009–2016 were analyzed. Inverse probability of weighted adjustment by propensity score was performed before analyzing risk factors for complications and survival.
Results
The study included 43 TMI and 121 TO patients. Before and after adjustment, TMI patients had significantly less cumulated postoperative complications (41% vs. 59%, p = 0.001), blood loss (median 100 vs. 200 ml, p = 0.001) and shorter length of hospital stay (median 4.5 vs. 6.0 days, p < 0.001). Multivariate analysis identified TO approach vs. MIS (OR = 2.4, p < 0.001), major liver resection (OR = 4.4, p < 0.001), and multiple CLM (OR = 2.3, p = 0.001) as independent risk factors for complications. 5-year overall survival was comparable (81% vs 68%, p = 0.59).
Conclusion
In patients with CRC undergoing multistage surgical treatment, MIS resection contributes to optimal perioperative outcomes without compromise in oncologic outcomes.
Introduction
Technical refinements and growing experience in minimally invasive surgery (MIS) for stage IV colorectal cancer (CRC) have led to better patient selection and improved perioperative safety and effectiveness. Newly developed MIS techniques have contributed to significantly improved perioperative outcomes, including reduced blood loss and transfusion rates, shorter hospital length of stay, decreased complication rates, and optimized cost-effectiveness.
Among patients managed by MIS, the therapeutic course of patients with stage IV CRC tends to be prolonged due to multiple surgical resections (primary, one or several liver resections) and the need for coordinating several treatment modalities (colorectal surgery, liver surgery, chemotherapy, radiation therapy or liver-directed treatments).
It has been shown that minimally invasive resection of the CRC primary and/or LM may be associated with lower complication rates than open resection. However, the series to date comparing MIS to open CRC resection have focused on outcomes after a single surgical step (either the primary or the liver metastases), rather than evaluating MIS for the entire treatment sequence that includes management of both primary and metastatic disease.
This study aims at comparing outcomes after totally minimally invasively (TMI) and totally open (TO) surgical management of a cohort of patients with stage IV CRC requiring multistage surgical resection. To reduce the risk for selection bias, propensity score weighting modeling was used.
Methods
Patient selection
The study protocol was approved by the Institutional Review Board of The University of Texas MD Anderson Cancer Center (protocol number PA17-0643). This institutional database was queried to identify a cohort of consecutive patients who underwent MIS resections (sequential and combined) of a primary CRC tumor, LM and/or lung metastasis. Minimally invasive techniques were defined as pure laparoscopic, robotic, and hand-assisted approaches. Patients who underwent repeat abdominal or liver resections and additional lung or transthoracic liver surgery were included.
A total of 43 patients were identified from March 2009 to September 2016. 121 patients underwent TO resection during the same period and were included to serve as the control group. Patients who had staged minimally invasive and open surgery and patients without complete available data on the surgical approach were not included.
Perioperative management and surgical procedures
The indications for surgical resection were discussed and the approach to surgical resection was determined by a multidisciplinary tumor board, which included surgeons, radiologists, oncologists, and pathologists and outlined a precise coordinated treatment plan for each patient. Neoadjuvant chemotherapy was administrated routinely except in few patients with easily resectable LM and with no negative prognostic factors.
All resections were performed by dedicated colorectal, liver, and/or thoracic surgery teams, with multiple teams involved in the case of combined resections. The specific minimally invasive approach was chosen during the multidisciplinary meeting according to the tumor characteristics, the estimated probability of safely achieving a margin-negative resection, and surgeon preference.
Outcome variables
Data recorded from patients' medical records included baseline demographic and clinical characteristics, details regarding management of the primary CRC tumor and metastases (combined or staged resections, total number of oncological procedures [combined resections were counted as 1 procedure], type of resection, perioperative chemotherapy, length of hospital stay, postoperative complications, and blood loss), and pathologic characteristics (T category, N category, tumor size and number, presence of extrahepatic disease at LM diagnosis, and margin status). The primary outcome was the occurrence of 90-day postoperative complications. The proportions of patients with complications after (i) resection of the primary tumor were calculated, (ii) resection of LM (any resection of LM, for patients with multiple liver resections), and (iii) any surgical resection. Complications were graded according to the Clavien-Dindo classification.
Secondary outcomes were median length of hospital stay, median blood loss, and 5-year overall and disease-free survival rates. For median length of hospital stay, median length of stay was calculated after resection of the primary tumor, median length of stay after resection of LM (for patients with multiple liver resections, length of stay was calculated as the mean of the length of stay after each procedure), and median cumulative number of days in hospital after all resections. Median blood loss is documented after resection of the primary tumor, after resection of LM (for patients with multiple liver resections, blood loss was calculated as the mean of the blood loss from each procedure), and as median cumulative blood loss after all resections.
Statistical analysis
Differences between continuous variables were compared by the Wilcoxon rank-sum tests or t tests, and differences between categorical variables were compared using Fisher's exact tests or Chi-square tests as appropriate. Propensity score matching with inverse probability weighting was implemented to reduce the possibility of selection bias.
Age at diagnosis, T stage, synchronous liver metastases, multiple tumor, major liver resection (>1 vs. 1) and preoperative chemo were used as the matching criteria to estimate the propensity scores. Matching criteria selection was based on the effective differences seen between the groups in Table 1 in one hand, and on relevant factors associated with technical difficulty and oncologic outcome in the other hand, to capture biases as much as possible. The standardized differences were used to evaluate the balance of these preoperative covariates between resection type groups before and after propensity score adjustment. After adjustment, the differences between groups were compared using weighted t tests or weighted Chi-square tests as appropriate. Overall survival (OS) was defined as the time interval between primary resection date and death date, and was censored at the last follow-up date for patients who were alive. The survival curves were estimated using the Kaplan–Meier methods and weighted Kaplan–Meier methods, and the Cox proportional hazards regression models and weighted Cox proportional hazards regression models were applied to assess the association between patient characteristics and OS before and after propensity score adjustment. Due to the weighted Kaplan–Meier method, patient at risk were not integer numbers and were expressed as percentages. Multivariate logistic regression models and weighted logistic regression models were applied to assess association between patient characteristics and the risk of cumulated complications before and after propensity score adjustment. Backward elimination was implemented in the multivariate model until all remaining covariates had p-values less than 0.05. All statistical analysis was performed using SAS 9.3 [SAS Institute, Cary NC, USA] and Stata 13.1 [Stata Corp, College Station, TX].
Table 1Balance of baseline covariates for unweighted and weighted data for patients who underwent TMI resection and TO resection
Baseline characteristics for the TMI (n = 43) and TO (n = 121) groups, before and after inverse probability of weighted adjustment by propensity score, are summarized in Table 1.
Patients were similar between the two groups except for the proportion of synchronous liver metastases, multiple LM and major liver resection. After inverse probability of weighted adjustment by propensity score, the absolute values of standardized differences were reduced for all variables and ranged from 0.3% to 5%, indicating that the two groups were comparable.
Perioperative characteristics
Perioperative characteristics for the TMI and TO groups, before and after inverse probability of weighted adjustment by propensity score, are summarized in Table 2. Patients in the TMI and TO groups had similar rates of combined primary tumor and liver resection, or repeat liver resection. A lung resection was performed in 19% and 20% of the TMI and TO patients, respectively (p = 0.75). Among them, 50% of patients in each group had combined lung and liver resection (p = 0.18). The proportion of patients undergoing at least one major liver resection was higher in the TO group (33% vs. 16%, p = 0.037) before adjustment.
Table 2Perioperative characteristics for patients who underwent TMI resection and TO resection, before and after inverse probability of weighted adjustment by propensity score
The surgical approaches for resection of the primary tumor and LM in the TMI group are summarized in Table 3. Before and after weighted adjustment, median blood loss was significantly lower in the TMI group than in the TO group for primary tumor resection, liver resection, and cumulative blood loss during all surgical procedures (100 ml vs. 200 ml p < 0.001, after adjustment).
Before and after weighted adjustment, the rates of postoperative complications were significantly lower in the TMI group compared to the TO group for complications after primary tumor resection and total cumulated complications (Table 2). The rate of Clavien-Dindo grade ≥3 complications did not differ between the TMI and TO groups for complications after primary tumor resection, but was significantly lower in the TMI group after liver resection. No patient had postoperative liver failure in the MIS group compared to one in the open group.
Before and after weighted adjustment, the median length of hospital stay was significantly shorter in the TMI group than in the TO group after primary tumor resection and liver resection; patients in the TMI group also had a lower median number of days in hospital after all surgical procedures (Table 2).
Factors independently associated with postoperative complications
As detailed in Table 4, multivariate logistic regression model before and after adjustment showed that TO approach vs. TMI, major liver resection vs. minor resection (OR = 4.4, p < 0.001, after adjustment) and multiple LM vs. single LM (OR = 2.3, p = 0.001, after adjustment) were independent risk factors for postoperative complications.
Table 4Multivariate analysis (logistic regression) on overall complications occurrence (n = 164) before and after inverse probability of weighted adjustment by propensity score
Before adjustment
After adjustment
Univariate analysis
Multivariate analysis
Univariate analysis
Multivariate analysis
OR (95% CI)
p
OR (95% CI)
p
OR (95% CI)
p
OR (95% CI)
p
Combined resection
1.27 (0.55–2.93)
0.58
1.39 (0.78–2.46)
0.26
Rectal tumor
1.71 (0.72–4.06)
0.23
1.34 (0.73–2.47)
0.35
Major liver resection
2.81 (1.25–6.31)
0.01
2.64 (1.20–5.82)
0.02
4.87 (2.70–8.79)
<0.001
4.38 (2.48–7.74)
<0.001
Multiple LM
2.31 (1.15–4.60)
0.02
2.16 (1.10–4.24)
0.03
2.45 (1.48–4.04)
0.001
2.28 (1.40–3.71)
0.001
Neoadjuvant chemotherapy
1.09 (0.50–2.38)
0.83
1.01 (0.59–1.71)
0.98
Surgical approach (TO vs. TMI)
3.29 (1.50–7.24)
0.003
3.09 (1.42–6.74)
0.004
2.51 (1.54–4.10)
<0.001
2.43 (1.50–3.94)
<0.001
LM, liver metastases; TMI, totally minimally invasive; TO, totally open.
After weighted adjustment, the 5-year overall survival rate was 81% in the TMI group and 68% in the TO group (p = 0.59) (Fig. 1a–b). The median follow-up time from primary resection date among survivors was 32.7 months (range, 3.8–133.1 months).
Figure 1Overall survival of TMI and TO patients (a) before and (b) after inverse probability of weighted adjustment by propensity score
This study analyzes MIS compared to open approach to stage IV CRC and, to our knowledge, is the first to evaluate the MIS approach in the context of the entire multimodality treatment sequence (resection of primary, multiple liver resections and lung resection). The results demonstrate that a completely MIS approach is independently associated with reduced risk of postoperative complications and favorable oncologic outcomes. In non-metastatic colorectal cancer, MIS has been associated with reduced blood loss, postoperative complications, and shorter length of stay in previous studies.
Short-term endpoints of conventional versus laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre, randomised controlled trial.
Colon Cancer Laparoscopic or Open Resection Study Group. Survival after laparoscopic surgery versus open surgery for colon cancer: long-term outcome of a randomised clinical trial.
MIS approach to colorectal liver metastasis has shown similar findings. In this context, a recent randomized controlled trial by Fretland et al. that reported shorter hospital stay, increased cost effectiveness and reduced postoperative complications in the MIS arm.
Short-term endpoints of conventional versus laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre, randomised controlled trial.
However, a completely MIS approach to the entire treatment sequence to render stage IV CRC patients disease free has not been evaluated to date. Additionally, comparing MIS and open approaches in LM surgical management is challenging due to concerns for selection bias: the cohort of MIS hepatectomy patients may confounded with technically less challenging cases being considered for MIS. To avoid this bias, our series not only used weighted matched paired analysis to reduce the impact of selection bias, but also included major hepatectomies, combined resections, and re-resections. This is in contrast to previous reports which evaluated resectable patients according to either a staged (sequential) or combined approach respectively.
Impact of totally laparoscopic combined management of colorectal cancer with synchronous hepatic metastases on severity of complications: a propensity-score-based analysis.
Therefore, the inclusion of all treatment sequences and various magnitudes of liver resection suggests reduced risk for selection bias and applicability of the data to a general hepatobiliary practice. Further, in resectable stage IV CRC patients, the need for multiple surgical interventions and the importance of treatment pathway completion suggest that an approach that minimizes morbidity and may shorten inter-treatment intervals is optimal. In this context, our group has recently reported on the importance of timely return to intended oncological treatment.
The reduced morbidity and the favorable outcome in the MIS approach cohort suggest a contribution of MIS to optimal treatment sequencing.
Regarding the technical aspect of a MIS approach, previous reports have shown that laparoscopic colon resection might facilitate subsequent liver resection.
This may have been a barrier to reporting on totally MIS managed series of stage IV colorectal cancer in the past. As the number of reports on learning curve and appropriate outcomes of MIS hepatectomy is increasing, the data shown here on not only technical feasibility and improved postoperative outcomes but oncologic safety is paramount.
In addition, certain clinical scenarios when employing MIS to hepatectomy bear specific technical challenges, such as redo-hepatectomy or resection of lesions in difficult to access locations (e.g. posterosuperior liver).
Operative and short-term oncologic outcomes of laparoscopic versus open liver resection for colorectal liver metastases located in the posterosuperior liver: a propensity score matching analysis.
To this end specific techniques have been developed to achieve safe resection. These include transthoracic laparoscopic surgery for resection of lesions in the postero-superior liver, which was included in this series.
Operative and short-term oncologic outcomes of laparoscopic versus open liver resection for colorectal liver metastases located in the posterosuperior liver: a propensity score matching analysis.
Our study has limitations related to its retrospective nature and design. Some information, in particular the timing for return to intended treatment, was not available in a substantial number of patients as patients received adjuvant chemotherapy at an outside facility. While patient selection, institution, performing surgeon and other biases are likely present, we attempted to reduce these biases by performing a propensity score weighting as well as regression analysis to reduce confounding variables. While our cohort displayed heterogeneity prior to adjustment, adjustment by propensity score weighting allowed us to create comparable cohorts.
Long-term survival in laparoscopic vs open resection for colorectal liver metastases: inverse probability of treatment weighting using propensity scores.
Moreover, unlike propensity score matching, adjustment by propensity score weighting allows for the entire cohort to be included in the analysis, therefore increasing power.
In conclusion, complete TMI surgical management of stage IV CRC is associated with reduced postoperative morbidity, less blood loss, shorter length of hospital stay, and favorable overall survival. The data suggests, if careful patient selection is employed optimal postoperative and oncological outcomes may be achieved when using a MIS surgical approach to stage IV CRC.
Conflict of interest
None of the authors have any conflict of interest to declare.
References
Nguyen K.T.
Gamblin T.C.
Geller D.A.
World review of laparoscopic liver resection-2,804 patients.
Short-term endpoints of conventional versus laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre, randomised controlled trial.
Impact of totally laparoscopic combined management of colorectal cancer with synchronous hepatic metastases on severity of complications: a propensity-score-based analysis.
Operative and short-term oncologic outcomes of laparoscopic versus open liver resection for colorectal liver metastases located in the posterosuperior liver: a propensity score matching analysis.
Long-term survival in laparoscopic vs open resection for colorectal liver metastases: inverse probability of treatment weighting using propensity scores.
Colon Cancer Laparoscopic or Open Resection Study Group. Survival after laparoscopic surgery versus open surgery for colon cancer: long-term outcome of a randomised clinical trial.
30834-7&id=gr1.jpg){kind=link}