Learn more about the benefits of registering on the new BJS website

Propensity score‐matched outcomes analysis of the liver‐first approach for synchronous colorectal liver metastases. BJS 2016; 103: 600-606.

Published: 10th February 2016

Authors: F. K. S. Welsh, K. Chandrakumaran, T. G. John, A. B. Cresswell, M. Rees


Liver resection before primary cancer resection is a novel strategy advocated for selected patients with synchronous colorectal liver metastases (sCRLM). This study measured outcomes in patients with sCRLM following a liver‐first or classical approach, and used a validated propensity score.


Clinical, pathological and follow‐up data were collected prospectively from consecutive patients undergoing hepatic resection for sCRLM at a single centre (2004–2014). Cumulative disease‐free survival (DFS), cancer‐specific survival (CSS) and overall survival (OS) were calculated by means of Kaplan–Meier analysis. Survival differences were analysed in the whole cohort and in subgroups matched according to Basingstoke Predictive Index (BPI).


Of 582 patients, 98 had a liver‐first and 467 a classical approach to treatment; 17 patients undergoing simultaneous bowel and liver resection were excluded. The median (i.q.r.) BPI was significantly higher in the liver‐first compared with the classical group: 8·5 (5–10) versus 8 (4–9) (P = 0·030). Median follow‐up was 34 months. The 5‐year DFS rate was lower in the liver‐first group than in the classical group (23 versus 45·6 per cent; P = 0·001), but there was no difference in 5‐year CSS (51 versus 53·8 per cent; P = 0·379) or OS (44 versus 49·6 per cent; P = 0·305). After matching for preoperative BPI, there was no difference in 5‐year DFS (37 versus 41·2 per cent for liver‐first versus classical approach; P = 0·083), CSS (51 versus 53·2 per cent; P = 0·616) or OS (47 versus 49·1 per cent; P = 0·846) rates.


Patients with sCRLM selected for a liver‐first approach had more oncologically advanced disease and a poorer prognosis. They had inferior cumulative DFS than those undergoing a classical approach, a difference negated by matching preoperative BPI.

Full text

Your comments


Kirsty McFarlane

3 years ago

We read with interest the study by Welsh et al. analysing the dilemma of the timing of primary tumour (PT) and liver metastases (LM) resection for stage IVA patients (1). The level of evidence is weak, and available retrospective studies (2, 3) and systematic reviews (4, 5) comparing the classical (PT resection first) and the reverse strategy (LM resection first) have yielded conflicting results.

The present retrospective study is of value because it has performed over a recent 10-years period, and included the largest single-centre population LM resection first ever reported. In addition it provided a sufficient follow-up to capture almost all recurrences.

However, the comparison suffers from the main caveat of the intrinsic non-comparability of the two groups, partly controlled by the propensity score matching. For example, data about extrahepatic metastatic disease is missing, and baseline PET-CT was not routinely performed.

While the authors reported an intent to treat survival rate of 44% at 5 years, the survival rate of 82 patients following completed liver-first approach is lacking. Mortality should optimally be measured at 90 days, and excluded upon long-term outcome assessment. Furthermore, timing and pattern of recurrence is not reported.

In the patient’s perspective, many questions remain unanswered such as am I resectable with a curative intent? If yes, am I operable? In case of recurrence, when and where will this occur? What will be the survival after recurrence, and with which quality of life? We agree with the authors that these patients should be managed “à la carte” rather than according to binding criteria.

Chetana Lim
Alexandre Doussot
Daniel Azoulay

Department of Hepatobiliary and Pancreatic Surgery and Liver Transplantation
Henri Mondor Hospital
51 Avenue du Maréchal de Lattre de Tassigny
94010 Créteil, France

1. Welsh FKS, Chandrakumaran K., John TG, Cresswell AB, Rees M. Propensity score-matched outcomes analysis of the liver-first approach for synchronous colorectal liver metastases. Br J Surg 2016 Feb 10. doi: 10.1002/bjs.10099. [Epub ahead of print].
2. Brouquet A, Mortenson MM, Vauthey JN, Rodriguez-Bigas MA, Overman MJ, Chang GJ, et al. Surgical strategies for synchronous colorectal liver metastases in 156 consecutive patients: classic, combined or reverse strategy? J Am Coll Surg 2010; 210(6): 934-41.
3. Andres A, Toso C, Adam R, Barroso E, Hubert C, Capussotti L , et al. A survival analysis of the liver-first reversed management of advanced simultaneous colorectal liver metastases: a LiverMetSurvey-based study. Ann Surg 2012; 256(5): 772-8; discussion 778-9.
4. Mayo SC, Pulitano C, Marques H, Lamelas J, Wolfgang CL, de Saussure W, et al. Surgical management of patients with synchronous colorectal liver metastasis: a multicenter international analysis. J Am Coll Surg 2013; 216(4): 707-16; discussion 716-8.
5. Jegatheeswaran S, Mason JM, Hancock HC, Siriwardena AK. The liver-first approach to the management of colorectal cancer with synchronous hepatic metastases: a systematic review. JAMA Surg 2013; 148(4): 385-91.

Kirsty McFarlane

3 years ago

Dear Sirs,

We thank Azoulay and co-workers for their interest and comments. With regard to some specific points raised, the groups are well-matched, using a recognised and validated prognostic scoring system, the Basingstoke Predictive Index (BPI). Data on extra-hepatic disease is included as part of the BPI. It is not our protocol to perform a baseline PET-CT in all cases.

The survival data is presented on an intention to treat basis, for the entire cohort (Figures 1a and 2a). The post-operative deaths were not included in the matched groups (Figures 1b and 2b). We could of course improve our survival data in the liver-first group by only presenting that of the 82 patients who completed the liver-first strategy or by excluding the 3 post-operative deaths, but did not think this appropriate.

Data on the timing and pattern of recurrence is not presented in this paper as it is not the main focus. However, one of the strengths of this paper is that no patient was lost to follow up and the percentage of patients in either the liver-first or classical groups who underwent subsequent re-intervention for either intra- or extra-hepatic metastatic disease is presented in the Results Section. Whilst it failed to reach statistical significance, more patients underwent reintervention (resection, ablation or cyberknife) for recurrent intrahepatic or extrahepatic disease in the liver-first group, reflecting these patients’ higher BPI and our proactive approach to surveillance and treatment of recurrent disease.

FKS Welsh
K Chandrakumaran
TG John
AB Cresswell
M Rees

Hepatobiliary Unit
Hampshire Hospitals NHS Foundation Trust
Aldermaston Road
Basingstoke, RG24 9NA