Introduction

Liver resection is the only potentially curative treatment for patients with colorectal rectal liver metastases (CRLM). Unfortunately, due to local irresectability, extra-hepatic disease, and/or co-morbidity, only 15–20 % of patients with CRLM are eligible for liver resection. In an attempt to increase resectability rates, a number of combined surgical and chemotherapy strategies are increasingly applied for patients with CRLM [1–3]. Neo-adjuvant chemotherapy can change the surgical options and downstage patients who initially present with non-resectable disease [4]. In patients with initially resectable disease, secondary tumours can be downsized in order to achieve a radical margin with a smaller resection [5]. The most commonly applied chemotherapy is oxaliplatin or irinotecan combined with capecitabine or folinic acid (leucovorin)/fluorouracil (5-FU) [6, 7]. In some patients, targeted molecular therapy (bevacizumab or cetuximab) is added to this regimen to achieve higher clinical response rates [8, 9].

Unfortunately, there is a clinical paradox, since although preoperative chemotherapy may improve surgical options, chemotherapy also has considerable hepatotoxic effects [10, 11]. Oxaliplatin can cause vascular changes in the liver, such as sinusoidal obstruction syndrome, which increases morbidity after liver resection [11, 12]. 5-FU and irinotecan regimens can lead to steatosis of the liver and consequently increase the risk of impaired liver regeneration and postoperative liver failure [13, 14]. However, the exact effect of preoperative chemotherapy on liver regeneration after major liver resection in humans is unknown. The vast majority of liver regeneration occurs within the first week after major liver resections [15, 16]. The aim of the present study was to investigate whether preoperative chemotherapy impairs liver regeneration in patients undergoing major liver resection. To that purpose, computed tomography (CT)-based liver volumetry was used before and 7 days after straight-forward anatomical right hemihepatectomy.

Method

Patients

All patients undergoing anatomical right hemihepatectomy were identified retrospectively from the prospective database of the FRESCO trial. The FRESCO trial was a multi-center randomized controlled trial (RCT) on the efficacy of fibrin sealant in reducing resection surface-related complications after partial liver resections [17]. Patients were intra-operatively randomized to application of fibrin sealant or no application of fibrin sealant. The primary endpoint of the study was the incidence of resection surface-related complications after partial liver resection. The study included a CT scan at day 7 after liver resection to evaluate resection surface-related complications. Patients were eligible to participate in this trial irrespective of whether they had undergone preoperative chemotherapy, and this was not a factor in the randomization procedure.

To investigate the effect of preoperative chemotherapy on liver regeneration in the present study, all patients who had undergone a right hemihepatectomy only, and had undergone pre- and postoperative scans as per protocol in the FRESCO trial were divided into two groups: patients who were not treated with preoperative chemotherapy before undergoing liver resection (group A, no chemotherapy group) and patients who were treated with chemotherapy, either in an adjuvant or neoadjuvant setting, before undergoing liver resection (group B, chemotherapy group). The following information was recorded: type of chemotherapy regime, number of cycles, and the date of last chemotherapy administration. The influence of chemotherapy–surgery interval and the number of chemotherapy cycles was analyzed in different subanalyses.

Operative procedure

Liver resection was performed as detailed elsewhere [18]. In short, laparotomy was performed by bilateral subcostal incision, followed by intraoperative ultrasonographic assessment of the liver. Once resectability had been confirmed, mobilization of the liver was performed to prepare for hepatic parenchymal transection. In all these right hepatectomies, transection followed Cantlie’s line from the top of the gallbladder, paralleling the middle hepatic vein straight to the suprahepatic inferior caval vein. In all patients, the middle hepatic vein remained in situ with the liver remnant.

CT imaging details

All patients underwent a contrast-enhanced triphasic CT scan in their routine preoperative assessment, and a scheduled CT scan 7 days after liver resection. The CT-scanning protocol was predefined and agreed between centers in the context of the FRESCO trial.

Volume measurements

Two investigators (SD, PK), who were blinded to the different groups of patients, performed all measurements. This was supervized by a specialist consultant abdominal radiologist (EJ vdJ) and a hepatobiliary surgeon (M dB). After transferring imaging data, volume measurements were performed on a Siemens Syngo workstation (version CT 2007A) with the ‘volume calculation’ application. Scans were retrieved from the individual patients in various centres, transferred and then all measured in the University Medical Centre of Groningen.

Regions of interest (ROIs) were manually drawn in the axial view on 2-mm slices in the portal venous phase. ROIs were drawn in every other slice to minimize partial volume effects. For definition of the anatomical segments Couinaud’s classification was used. The volume program automatically interpolated between two pending ROIs. Each interpolation was revised and corrected manually if necessary. Approximate volumes were calculated with automatic multiplication of the circumscribed areas by the CT-slice thickness.

On the preoperative scan, total liver volume (TLV), the volume of the resection specimen (V resection), and tumour volume (V tumour) were measured. V resection was outlined according to Couinaud’s classification and on the basis of hepatic (vascular) anatomy. Remnant liver volume after 7 days (V RLV7days) was measured on the postoperative scan (Fig. 1).

Fig. 1
figure 1

An outline of the remnant liver volume after 7 days (V RLV7days) measured on the postoperative scan. Arrow shows the middle hepatic vein

Full size image

Calculations of Volumes

Functional TLV (TLV func) was calculated as TLV − V tumour. Future remnant liver volume (FRLV) was calculated as TLV − V resection. The early regeneration index, 7 days after surgery (RI early), was calculated as [(V RLV7days − FRLV)/FRLV] × 100. The functional resection percentage (Funct Resection%) was calculated as [(V resection − V tumour)/TLV func] × 100.

Volumetric analysis Syngo®

For assessment of the accuracy of volumetric measurements of the liver with Syngo®, volumes of resection specimens were compared with actual weights of the resection specimens. The resection weights were measured in only two of the seven participating medical centers as this was not part of the FRESCO trial protocol. This was done immediately after liver resection: weights of resection specimens were recorded in the operating theatre or at arrival on the pathology department. The actual weights of the resection specimens remained blinded to the investigators conducting CT volumetry.

Postoperative complications

Postoperative complications were registered according to the Clavien–Dindo score in the prospective database [19].

Liver function and liver cell damage markers

Patients were admitted to the hospital 1 day preoperatively, and routine blood tests were performed by the clinical chemistry department in the individual hospital. Routine blood tests for liver function and liver cell damage were performed on preoperative day 1, on postoperative day 0, 1, 3, 5, and 7, and on the day of discharge. The levels of alanine-aminotransferase (ALT), aspartate-aminotransferase (AST), gamma-glutamyl transpeptidase (γGT), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), bilirubin total, and prothrombin time (PT) were assessed.

Ethics

The study was approved by the medical ethical committee of the University Medical Center of Groningen and conducted according to the Declaration of Helsinki. All patients gave written informed consent.

Statistics

All data are expressed as median (interquartile range). To compare different subgroups, the nonparametric Mann–Whitney U test was applied. Dichotomous data were compared using Fisher’s exact test. Multiple group comparisons for continuous data were conducted via the Kruskal–Wallis test, with Dunn’s post hoc test. A p-value <0.05 was considered to indicate statistical significance. Statistical analysis was performed using Prism 4.0 for Windows (Graphpad software, Inc, San Diego, CA, USA).

Results

Patients

A total of 310 patients were included in the FRESCO trial, 102 of whom underwent a right hepatectomy. In 30 patients it was not possible to perform a complete CT-volumetric analysis (i.e. no adequate preoperative or postoperative CT scan). This left 72 patients (35 male; 37 female) who underwent a right hepatectomy for benign (n = 7) or malignant liver tumours (n = 65) for the final analysis. Of 72 patients, 27 (37.5 %) had been treated preoperatively with chemotherapy and 45 (62.5%) did not receive preoperative chemotherapy (group A) (Fig. 2).

Fig. 2
figure 2

Flowchart of the study. CT computed tomography, MRI magnetic resonance imaging, RFA radiofrequency ablation

Full size image

There were no significant differences between groups in baseline characteristics (Table 1). The vast majority of patients in group B (59.3 %) received preoperative chemotherapy consisting of oxaliplatin, capecitabine, and bevacizumab. The other patients in group B (40.7 %) received a variety of other chemotherapy regimens (Table 2). Of the patients in group B, 18.5 % received adjuvant chemotherapy as part of treatment of the primary colorectal tumour and 81.5 % received neoadjuvant chemotherapy to treat the liver metastases. There was no significant difference in intraoperative blood loss between group A (750 mL [875 mL]) and group B (800 mL [900 mL]) (p = 0.57). There was also no significant difference in operation time between group A and B: 310 min (185 min) and 320 min (202 min), respectively (p = 0.49).

Table 1 Preoperative patient characteristics
Full size table
Table 2 Specification of the chemotherapeutic strategies
Full size table

CT-measured liver volumes

The weight of the resection specimen as measured in 47 patients was 882 g (383 g). A strong significant correlation was found between the resection weight and resection volume measured with Syngo® (p < 0.0001; r = 0.92) (Fig. 3). In the total series (group A + B), the preoperatively calculated TLV, V resection, and V tumour were 1,591 mL (368 mL), 995 mL (322 mL), and 34 mL (80 mL), respectively. The TLV func and FRLV were 1,527 mL [390 mL] and 594 mL [233 mL], respectively. V tumour was significantly larger in group A than in group B: 38 mL (215 mL) versus 19 mL (46 mL) (p < 0.05) (Table 3).

Fig. 3
figure 3

Correlation between volume measured with Syngo® and resection weight measured in the operating theatre

Full size image
Table 3 CT-volumetry of livers and tumours
Full size table

Early liver regeneration

The RI early at day 7 in the whole group of 72 patients was 58 % (39 %). The RI early was not significantly different between group A and B (60 [36 %] and 50 [43 %], respectively; p = 0.47) (Table 3; Fig. 4a). When benign tumours were excluded in a subanalysis, the difference in liver regeneration between the chemotherapy and non-chemotherapy group remained insignificant. In group B, the time interval between chemotherapy and the liver resection was 3 months (3 months) (Table 2). There was no significant difference in liver regeneration between patients who underwent liver resection within 3 months after chemotherapy (n = 17) and patients who underwent liver resection after 3 or more months (n = 10), 49 (40 %) and 58 (52 %), respectively; p = 0.57. There was also no significant difference in liver regeneration between patients who underwent liver resection within 3 months after chemotherapy (n = 17) and patients who did not receive chemotherapy (n = 45), 49 (40 %) and 60 (36 %), respectively; p = 0.38.

Fig. 4
figure 4

a RI early, as a marker of early regeneration in patients without preoperative chemotherapy (group A) and in patients treated with preoperative chemotherapy (group B). Data are medians and interquartile ranges. b RI early, as a marker of early regeneration in patients who were treated with more than six cycles of chemotherapy versus those who were treated with six or fewer cycles of chemotherapy. *p < 0.05. RI early the early regeneration index 7 days after surgery

Full size image

Application of fibrin glue was equally distributed between groups. In the chemo group A, 11 patients (40.7 %) did not receive fibrin glue, while 16 patients (59.3%) did. Equally, in the non-chemotherapy group B, 21 patients (46.7 %) did not receive fibrin glue while 24 patients (53.3 %) did. The application of fibrin glue did not effect liver regeneration. There was no significant difference in liver regeneration between patients without fibrin glue (n = 32; 57.1 [32.4 %]) and with fibrin (n = 40; 58.0 [42.3 %]), p = 0.81.

It can be argued that, if the time interval between the last cycle of preoperative chemotherapy and liver surgery is 6 months or more, the chance is small that chemotherapy still has an effect on liver regeneration. To analyse this, a subanalysis was performed in which patients with a chemotherapy–surgery interval of 6 months or more (n = 5) were included in the group of patients who did not receive chemotherapy (n = 45) and this group was then compared with the group of patients who received surgery within 6 months after chemotherapy (n = 22). Again, there was no significant difference in liver regeneration between these two groups: 59 (37 %) and 51 (42 %), respectively (p = 0.71).

The RI early was significantly lower in patients who had received more than six cycles of chemotherapy (n = 5) than in patients who were treated with six or fewer cycles of chemotherapy (n = 22; 37 [20 %] and 62 [38 %], respectively; p < 0.05) (Fig. 4b). The time interval between chemotherapy and liver resection was not significantly different between patients who had received more than six cycles of chemotherapy and patients who were treated with six or fewer cycles of chemotherapy (5 months [14 months] and 3 months [2 months], respectively; p = 0.13).

Some patients had a postoperative CT scan earlier or later than day 7. This deviation from the original protocol was usually related to early functional recovery of the patient and early discharge from the hospital. There was no significant difference between groups in the number of days between the operation and the first postoperative CT scan (group A: 7 [1] days vs. group B: 7 [2] days; p = 0.24). However, in order to exclude a potential difference in liver regeneration due to difference in time (surgery/CT scan interval), the RI early was also calculated and compared in patients who had a CT scan on exactly postoperative day 7 (26 patients in group A and 15 in group B). In this sub-analysis, the RI early was also not significantly different between groups A and B (58 [55 %] and 73 [47 %], respectively; p = 0.84). As mentioned, RI early in patients who received more than six cycles of chemotherapy was significantly lower than in patients receiving six or fewer cycles, and this remained significantly different when only patients who had a CT scan strictly on postoperative day 7 were analyzed.

In order to investigate the potential negative impact of bevacizumab on liver regeneration, the RI early in patients treated with preoperative bevacizumab was compared with that of patients who had not received bevacizumab. The RI early was not significantly different between patients (n = 20) who had received bevacizumab and those (n = 7) who had not (56 [42 %] and 38 [61 %], respectively; p = 0.18).

Postoperative complications

There was no significant difference between groups in postoperative complications, length of hospital stay, or 30-day mortality (Table 4).

Table 4 Postoperative complications
Full size table

Liver function and liver cell damage markers

There were no significant differences between groups in liver function and liver cell damage markers on the day before surgery. Postoperatively, there were also no significant differences in liver function and liver cell damage markers between groups (Table 1; Fig. 5a–g, supplemental data; Fig. 6a–c, supplemental data).

Discussion

This study aimed to investigate whether preoperative chemotherapy impairs liver regeneration in patients undergoing major liver resection. To that purpose CT-based liver volumetry was used before, and 7 days after, right anatomical hepatectomy. The results of the present study suggest that chemotherapy prior to major liver resection may have no negative impact on early liver regeneration. It could be that the number of chemotherapy cycles is pivotal, since patients who had received more than six cycles of chemotherapy had significantly less early liver regeneration than patients treated with six or fewer cycles of chemotherapy. However, this should be studied in future larger studies.

Only few studies have focused on early liver regeneration. Most of these studies measure the early liver regeneration index in living donors for liver transplantation, and the results in these studies vary, presenting different regeneration indices after 1 week [15, 16, 20, 21]. The data from the present study showed that ~58 % of liver regeneration takes place in the first week after resection of the right lobe. These results are in accordance with results from Zappa et al. [20] who reported a regeneration index of 56 % in the first week in patients undergoing right hepatectomy. Although several studies warn of the hepatotoxic effects of preoperative chemotherapy on liver regeneration in liver resection patients [10–12, 22], data on the exact effect of preoperative chemotherapy on liver regeneration in humans are scarce. To our knowledge, the present study is the first that analyses the effect of preoperative chemotherapy on liver regeneration in a large group of patients after major liver resection. All patients underwent an anatomical right hemihepatectomy and there were no major differences in mobilization and transection of the liver. Therefore, these patients were comparable and the protocol of this RCT provided the opportunity to investigate early liver regeneration in an adequate and valid human study model. Since there is no reliable biochemical marker to assess liver regeneration in humans, CT volumetry is an attractive approach. The CT-volumetric method with Syngo®, a professional radiological software program used in the present study, was shown to be valid and reliable as evidenced by the excellent correlation between resected weight and volume. This method was comparable with other recently validated CT-volumetry methods [23, 24]. In order to conduct a better validation of the CT-volumetric method with Syngo® in future studies, it would be worthwhile to compare different liver volumetry measurements (TLV, FRLV, etc.) with different CT software programs or even with different modalities, e.g. magnetic resonance imaging (MRI) or ultrasonography.

The type of chemotherapy and the number of chemotherapy cycles a patient receives often varies. One would expect that patients have worse liver regeneration with an increasing number of cycles of chemotherapy because their liver is exposed to hepatotoxic compounds for longer. In line with this, Karoui et al. [25] found that preoperative chemotherapy was an independent predictive factor of postoperative morbidity after liver surgery, especially when an increasing number of cycles was given to patients. In the present study, the hypothesis was confirmed that, in patients who were treated with more than six cycles of chemotherapy a smaller RI early was measured than in those who were treated with six or fewer cycles. Impaired early liver regeneration in these patients is probably also associated with an increased risk of complications after liver resection. However, only a small subgroup in this study had more than six cycles of chemotherapy. Future studies with larger groups of patients on the exact effect of chemotherapy and the length of chemotherapy on liver regeneration are necessary.

Contradictory results have been reported by several researchers regarding the potentially negative impact of bevacizumab on liver regeneration. Aussilhou et al. [26] have shown that bevacizumab impairs liver regeneration after preoperative portal vein embolization. On the other hand, Gruenberger et al. [27] showed that neoadjuvant bevacizumab does not affect liver regeneration 3 months after resection. The same group of authors also showed in another study that, when bevacizumab is combined with oxaliplatin, bevacizumab protects the liver against sinusoidal obstruction syndrome [28]. Data from the present study showed that the preoperative administration of bevacizumab did not impair early liver regeneration after major liver resection.

The present study has several limitations. The most important limitation was that no power analysis was performed and that the sample size was relatively small. However, an important advantage was that the patients enrolled in this study were included in a multicenter prospective RCT on a different subject, for which chemotherapy was not part of the selection criteria and therefore no selection bias took place. It was assumed that percentage early liver volume increase assessed with CT volumetry reflects liver regeneration and also liver function; however, part of the liver volume increase may be caused by edema or hepatic swelling due to high portal pressure [29]. The future liver remnant volume and postoperative true liver remnant (V RLV7days) were measured differently; for future studies it would probably be better to conduct a direct measure of the left liver lobe pre- and postoperatively. There were no significant differences between group A and B in traditional liver function and liver cell damage markers preoperatively and postoperatively in the first week. Yet, it has to be taken into account that these are relatively crude markers [30].

In conclusion, this study shows that preoperative chemotherapy does not seem to have a negative impact on early liver regeneration after partial liver resection. However, there was a significantly lower regeneration capacity in patients who were treated with more than six cycles of chemotherapy. This study shows that strategies combining preoperative chemotherapy with liver surgery are safe, based on liver regeneration in the first week.