Background
Medical and surgical treatment of advanced metastatic colorectal cancer (CRC) has undergone enormous improvement during the last years and is still evolving. Accounting for 12.2% of all cancer-related deaths in Europe, the majority of patients diagnosed with metastatic CRC present with unresectable metastatic disease [
1]. In contrast, surgery remains the only curative approach to colorectal liver metastases (CLM) and liver resection is indicated in all “resectable” patients [
2]. Definition of resectabiliy however – including “manageable” margin-negative resection, “adequate” residual perfusion and biliary drainage as well as a minimum of 20% residual healthy liver, remains vague and necessitates interdisciplinary evaluation by expert oncologists, surgeons and radiologists [
1,
3]. Furthermore, perioperative chemotherapy has been shown to facilitate surgery in extended metastatic disease, to increase resectability of initially unresectable CLM by 7-40%, and to increase progression-free survival in resectable CLM [
1]. Single centre publications have shown CLM resection rates of 20-45% and 5-year survival rates after CLM resection of up to 64% [
1‐
6]. However, single, high-volume academic centres represent highly selected patients and the results do not always reflect the clinical reality. In a first population-based study of CLM patients in Southern Germany, we had analysed 884 patients diagnosed with CRC in 2002. An overall CLM resection rate of 19.1% was described; a higher resection rate (28.3%) was observed in a subgroup of patients treated in centres specialized in liver surgery. However, a relevant undertreatment of CLM patients was seen [
7]. We therefore decided to analyse the evolution of CLM management over a period of ten years. The present study is the first long-term analysis of 5772 patients diagnosed with colorectal adenocarcinoma between 2002 and 2007 in Germany. Observation was continued until 2012, resulting in a minimum follow-up of 5 years (maximum 10 years).
Discussion
Surgical resection represents the only potentially curative approach to colorectal liver metastases [
1,
2]. During the past years, enormous improvements in both, surgical technique and perioperative chemotherapeutic treatment options, have increased CLM resection rates to 20 - 45% and five-year survival after hepatic resection to up to 64% in selected, single-centre analyses [
2]. However, these results are obtained from highly selected and ideally monitored patients and do not reflect clinical reality. Therefore, we decided to analyse the surgical management and long-term clinical outcome of CLM patients using data from a German tumour registry over a period of ten years.
In the present study, a total of 5772 patients diagnosed with CRC were analysed. In 2002 and 2003, respectively 799 and 856 patients with complete records were analysed, representing a completeness of 80%. Increased and constant numbers of patients analysed in later years do represent a completeness rate of >90%. CLM were diagnosed in 24.7% of CRC patients. This rate is at the lower range as compared to previously published population-based analyses as well as single-centre publications [
3,
5,
6,
8‐
13]. Furthermore, our rate of synchronous metastases is at the higher range as compared to other population-based studies using the same definition [
5,
8‐
13]. We therefore assume a slight deficit in completeness of documented metachronous CLM in our tumour registry. 94% of CLM are diagnosed within 24 months of primary diagnosis, explaining the constant CLM rates until 2009, i.e. 24 months after study entry of the last patients.
Although definition of resectability does no longer include number of hepatic metastases, we assume that the majority of patients diagnosed with 3 or less CLM would be considered to be resectable. Surprisingly, only 52.2% of these patients underwent liver resection, indicating a high potential for further improvement in surgical therapy in this subgroup. However, a significant increase in resection rate in this subgroup from 46.6% in 2002/2003 to 61.8% in 2006 and later years was observed. Overall resection rate of CLM was 26.2%, which compares favourably to other population-based as well as single-centre analyses [
5,
8‐
13]. Over time, a significant increase in resection rates from 16.6% in the 2002 cohort to >21% in the 2003–2005 cohorts to >30% in cohorts after 2005 was observed. In 2006, certification of comprehensive colorectal cancer centres was initialized by the German Cancer Society. During the study period, 8 of all 30 hospitals being part of the present analysis have been certified as colorectal cancer centres. The increase in CLM resection rates may in part reflect the implication of standard operating procedures during interdisciplinary management of CRC patients treated in certified centres. These standards include mandatory discussion of the surgical options of each individual patient by comprehensive tumour-board meetings.
A subgroup of 423 CLM patients (29.7%) in the present study were treated in one of two academic centres specialized in liver surgery, which is a lower rate than in comparable population-based analyses [
5,
7‐
13]. Of those 423 patients treated in academic centres, 197 (46.6%) underwent liver resection, compared to a resection rate of 22% in regional hospitals. The higher resection rate of patients treated in centres specialized for liver surgery is consistent with data published from large academic single-institution studies and may reflect selection and referral bias, but may also indicate that more resectable patients really underwent surgery in these two centres [
1,
3]. Nevertheless, a resection rate of 46.6% is at the higher range as compared to selected, single-centre analyses [
1,
2]. Together with the increase in hepatic resection rates over time in the present study as well as the number of still unresected CLM patients, especially in those patients diagnosed with three or less CLM, we here show for the first time a pronounced improvement in surgical management of CLM over time in Germany, but also a still striking potential for further increases in hepatic resection rates.
Kaplan-Meier analyses demonstrated – as expected - that patients without CLM had a significantly better 5-year and 10-year OS than CLM patients. In CLM patients, the number of metastases, irrespective of hepatic resection, was a significant factor for long-term OS. Long-term OS in CLM patients was significantly increased by liver resection, which was most pronounced in patients diagnosed with a singular hepatic lesion, but remained significant also for patients diagnosed with multiple CLM. Survival curves plateaued at 8–9 years after CLM diagnosis, thus being in line with the definition of “cure” from CLM published by Tomlinson and colleagues [
14]. Cure rates from CLM of 17 to 25% have been published and ongoing trials evaluating perioperative triple-chemotherapy combined with targeted therapy for CLM patients report of 5-year survival rates of up to 64% [
2]. Thus, our survival data are in line with previously published population-based and single-centre analyses.
Intraoperative ablative therapies in combination with curative resection have been applied in 24 patients of the present series (10 thermal ablations, 12 radiofrequency ablations, 2 cryotherapies). Previous studies have suggested that ablation of liver metastasis leads to more frequent and quicker liver recurrence [
15]. Since the number of patients treated with a combined resection/ablation approach was very small, no subgroup analyses of recurrence-free or overall survival could be performed in the present series.
The timing of the resection of synchronous liver metastasis is a focus of international discussion. Some experts recommend simultaneously resecting the primary colorectal cancer and the CLM, others recommend staged resection with the primary tumor resected first and finally others recommend a “liver-first” approach with resection of the CLM first, followed by resecting the primary tumor in a following surgery [
16]. In our series, 253 of 374 hepatic resections were performed in patients diagnosed with synchronous liver metastases. Of these 253 patients, 134 received resection of hepatic metastases and primary tumor during a simultaneous approach. In these cases, metastatic tumor burden could be controlled by atypical or segmental hepatic resection. Only 5 patients diagnosed with synchronous hepatic metastases underwent a “liver-first” approach, the remaining 114 patients received the “classic approach”, starting with the resection of the primary tumor. The rationale for this approach, e.g. bowel obstruction or bleeding caused by the primary tumor, remained unknown. Comparing the long-term clinical outcome of patients treated with the liver-first approach versus “classic approach” would be of high interest. Recent publications have discussed the liver-first approach as preferred treatment, since the long-term prognosis is more determined by controlling metastatic disease and hepatic resectability may be jeopardized by prior colorectal surgery [
16]. Furthermore, neoadjuvant chemotherapy prior to liver-first hepatic resection may select patients benefiting most from curative surgery [
16]. Due to the very limited number of liver-first resected patients, we cannot provide a statistically relevant result in this regard.
In the present study, perioperative chemotherapy was given in 180 CLM patients (48%) undergoing hepatic resection. However, chemotherapeutic regimens were manifold, ranging from 5-FU monotherapy to triple chemotherapy plus targeted therapy. Furthermore, timing of perioperative chemotherapy as neoadjuvant, adjuvant, or combined neo-/ and adjuvant regimens in relation to liver resection as well as total duration of chemotherapy was diverse and no significant influence of chemotherapy on resection rate or survival was seen in univariate analysis. Multivariate analysis however showed a significant benefit of perioperative chemotherapy for CLM. Perioperative chemotherapy for CLM is a highly relevant research focus and has been shown to increase resectability of initially unresectable CLM by 7-40% [
1]. An international expert panel recommends chemotherapy for all patients diagnosed with unresectable CLM, followed by surgery as soon as resectability may occur [
3]. The most efficient chemotherapy is presently studied in multiple randomized clinical trials. The CRYSTAL and OPUS trials showed significantly increased resection rates of CLM after treatment with Folfiri, respectively Folfox, combined with the targeted agent Cetuximab [
17]. The CELIM trial, cross-comparing Folfox plus cetuximab versus Folfiri plus cetuximab, showed resection rates of CLM of 40-43% with no significant difference comparing both regimens [
18]. Ongoing clinical trials evaluate the triple chemotherapeutic approach Folfoxiri combined with single or dual targeted agents [
18‐
20]. Expert consensus recommends neoadjuvant chemotherapy in resectable CLM with more than one of the poor prognostic factors multiple metastases, size of CLM >5 cm, synchronous CLM, lymph-node positive primary CRC or high tumor markers [
3].
We interpret our trend of an improved survival of CLM patients diagnosed after 2005 compared to CLM patients diagnosed 2002–2005 to be partly caused by the above mentioned CRC centre certification, partly caused by improved hepatic surgery and last but not least by advances in perioperative chemotherapy.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CH: conception and design of study, data acquisition and analysis, drafting and critical revision of the manuscript. PN: Data acquisition and analysis. MG: Analysis of data, drafting of figures, critical revision of the manuscript. ML and MKS: analysis of data and critical revision of the manuscript. HJS: conception and design, critical revision of the manuscript. All authors read and approved the final manuscript.