Colon carcinoma is one of the most frequently occurring cancers worldwide^[1,2]. The liver is frequently involved in dissemination as up to 30 % of colorectal cancers develop liver metastases^[3]. Presently, therapeutic options for liver metastases are diverse^[4], but liver surgery is the sole potentially curative treatment. Owing to improved perioperative management, increasing surgical expertise and new surgical approaches, more patients are able to undergo surgery^[5] and even extensive Colorectal Liver Metastases (CRLM) can be completely excised^[6]. However, this progress does not reduce the recurrence rates of CRLM^[7]. Therefore, surgery is presently part of a multimodal therapeutic concept^[8]. According to present guidelines, neoadjuvant chemotherapy is justified in downsizing for initially irresectable metastases^[9,10], but its use in initially resectable liver metastases remains controversial^[11-13]. The EORTC trial (40983) induced perioperative chemotherapy to become a standard treatment in patients with resectable CRLM, but its efficacy for overall survival remains to be demonstrated^[11]. Long-term follow-up data on recurrence and survival are especially sparse.

For evaluating the impact of perioperative chemotherapy in long-term follow-up, we performed a matched case-control study with patients after the resection of CRLM. We also investigated otherprognostic factors in our real-world cohort.

Methods

Patient cohort and study design

Patients with CRLM who were treated at our centre between September 1997 and May 2012 were included if they fulfilled the following criteria:

• Complete resection of CRLM was deemed feasible before surgery;

• No extrahepatic dissemination or second primary cancer;

• No previous liver surgery;

• Liver surgery or chemotherapy was the first step of treatment;

• No ablation in patient history or simultaneously with surgery;

• Postoperative survival of ≥ 3 months and constitution to undergo chemotherapy starting at least three months after liver resection.

• First staging three months after surgery with no sign of recurrence.

The following patient characteristics were obtained from the department database: gender, age, tumour (T) and nodal (N) status of the primary cancer, the occurrence of CRLM (synchronous or metachronous), size and number of CRLM, tumour spread (uni- or bilobal), carcinoembryonic antigen (CEA) at liver resection, Fong-Score^[14], approach to surgery (minor versus major resection), status of liver resection (R0 or R1), chemotherapy before and / or after surgery, progress, progression-free survival, death and overall survival.

Synchronous liver metastases were defined as liver metastases occurring within 12 months after surgery of the primary colorectal cancer. The size of the liver metastases was defined as the largest diameter of a single metastasis preoperatively assessed using Computed Tomography (CT) or Magnetic Resonance Imaging (MRI) or postoperatively assessed in the pathology report. Follow-up included CT or MRI scans one month after surgery and then every three months in the first year and every six months in the second and third years. After the third year, scans were annually performed. Progress was defined as a sign of tumour at follow-up.

Of the patients who met the inclusion criteria, pairs of patients with or without perioperative chemotherapy were matched according to the following matching criteria (Figure 1): nodal status of the primary (involvement: yes or no); the occurrence of CRLM (synchronous or metachronous); the status of liver resection (R0 or R1). This study was performed in accordance with the Declaration of Helsinki. Patient care and study conduct complied with good clinical practice.

Figure 1: Study design

Surgical approach and chemotherapy

Surgical procedures were classified as minor or major liver resections. Minor resections were defined as the resection of three or fewer liver segments, while major resections were defined as the resection of four or more liver segments. Minor resections included wedge resections as well as segmentectomies, bisegmentectomies, left lateral sectionectomies, left medial sectionectomies, right posterior sectionectomies and right anterior sectionectomies. Major resections included right and left trisectionectomies as well as right and left hepatectomies^[15].

Perioperative chemotherapy was defined as chemotherapy ≤ 3 months pre- and postoperatively, postoperatively only without any sign of new metastases. Patients undergoing chemotherapy because of new metastases ≤ 3 months postoperatively were excluded. For the analysis, all chemotherapeutical agents were pooled because the chemotherapy regimen changed over the 15-year investigation period. Chemotherapy regimens were chosen at the physician’s discretion with respect to the guidelines at the time of treatment. Chemotherapy regimens included inter alia 5-FU and leucovorin in accordance with the Ardalan protocol, 5-FU and calcium folinate in accordance with the MACHOVER or Mayo clinic protocol, the AIO protocol, FOLFIRI + / - bevacizumab, FOLFOX 2,3 or 4 + / - bevacizumab and capecitabine.

Statistical analysis

Statistical analysis was performed in cooperation with the Institute of Biostatistics and Mathematical Modeling, Department of Medicine, Goethe University, Frankfurt, using IBM SPSS Statistics 21 (NY, USA). For avoiding selection bias for patients with and without chemotherapy, we matched patients according to factors known to influence outcomes after the resection of CRLM (nodal status of the primary cancer, the occurrence of CRLM, the R status of liver resection)^[14]. Matching was performed with R (version 3.4.2, Foundation for Statistical Computing, Vienna, Austria) using the package ‘MatchIt’^[16].

Categorical data were analysed using cross tabulation and Chi-quadrate test. Correlation between observed data was calculated using Pearson’s product-moment coefficient (r) for normally distributed variables and the Spearman’s rank correlation coefficient (r) for metric but non-normally distributed variables. Time-dependent data were estimated using Kaplan-Meier method. Progression-free and overall survivals were calculated from the day of liver surgery. For progression-free survival, progress and death were counted as events. Significant prognostic factors in the univariate analysis were also analysed using multivariate Cox proportional hazards model. Statistical significance was defined as p  0.05.

Results

Of all patients undergoing liver surgery for CRLM between September 1997 and May 2012, 45 patients received perioperative chemotherapy and met the inclusion criteria. Of these, 29 received neoadjuvant chemotherapy, 13 received adjuvant chemotherapy and three received neoadjuvant and adjuvant chemotherapies. They were matched with 45 corresponding patients with CRLM treated using only surgery. Table 1 presents the patient and tumour characteristics of both groups.

Median follow-up was 1252 days (range: 352–6057days) from liver surgery for patients with perioperative chemotherapy and 1705 days (range: 460–5405 days) for patients without chemotherapy. Progression-free survival was available for all patients and the median was 498 days (range: 112–6057 days) for patients with perioperative chemotherapy and 812 days (range: 187–5405 days) for patients without chemotherapy. Besides the matching criteria, other patient and tumour characteristics were equally distributed in both groups (Table 1). The following surgical procedures were performed in the 90 patients: three wedge resections; 41 segmentectomies; one bi segmentectomy; one left lateral sectionectomy; three right posterior section ectomies; 14 right and six left trisectionectomies; 16 right and five left hepatectomies. This resulted in 49 minor and 41 major resections with no significant differences between the groups (major resections in patients with chemotherapy 42% (19 / 45) vs. 49% (22 / 45) in patients without chemotherapy; p = 0.672). Surgery resulted in R0 resections in 11 patients and R1 resections in 79 patients (88%),which were equally distributed in patients with (13%, 6 / 45) and without chemotherapy (11%, 5 / 45; p = 1.000). Five-year overall survival was 39% for patients with chemotherapy vs. 59% for patients without chemotherapy; ten-year overall survival was 36% for patients with chemotherapy vs. 38% for patients without chemotherapy (Figure 2B).

Figure 2: Kaplan-Meier-curves for progression-free survival (A) and overall survival (B) in patients receiving perioperative chemotherapy and patients without chemotherapy.

Table 1: Patient demographics, tumour characteristics and surgical issues.

*matching criterion; **CEA: Carcino Embryonic Antigen

Overall survival did not significantly differ in terms of gender, location of primary tumour (colon vs. rectum), T status of primary tumour, occurrence of liver metastases (synchronous vs. metachronous), bilobal tumour spread, approach of liver resection (minor vs. major), R status of liver resection or perioperative chemotherapy (Table 2). There was no correlation for survival and age (r = –0.196, p = 0.064), size of metastases (r=0.053, p=0.619), number of metastases (r = –0.010, p = 0.926), CEA (r = –0.138, p = 0.210) or Fong’s score (r = –0.171, p = 0.122). Overall survival was significantly worse in patients with disease progress (p < 0.001) and a positive N status of the primary tumour (p = 0.032).

Table 2: Univariate analysis of predictors of progression-free and overall survivals after liver surgery.

*Statistically significant

Progression-free survival was not influenced by gender, location of primary tumour (colon vs. rectum), T status of primary tumour, occurrence of liver metastases (synchronous vs. metachronous), approach of liver resection (minor vs. major) or perioperative chemotherapy (Figure 2A), whereas positive N status of primary tumour (p = 0.002), tumour spread in both liver lobes (p = 0.016) and R1 liver resection (p < 0.001) significantly reduced progression-free survival (Table 2). There was no correlation between progression-free survival and age (r = 0.059, p = 0.583), size of metastases (r = 0.115, p = 0.282), number of metastases (r = –0.199, p =0.061), CEA (r = –0.206, p = 0.061) or Fong’s Score (r = –0.198, p = 0.072).

In the Cox proportional hazards model, the N status of the primary tumour and R status of resection were significantly correlated with progression-free survival; N status of the primary tumour was the sole predictor of overall survival (Table 3).

Table 3: Multivariate Cox proportional hazards model of predictors of progression-free and overall survivals.

* Statistically significant

Discussion

Node-positivity of the primary tumour, synchronous occurrence of liver metastases, number of lesions, size of the largest lesion, high CEA level and additional extra hepatic dissemination have been repeatedly reported to act as prognostic tools in patients with CRLM^[14,17,18]. The data in the present study confirmed that node-positivity of the primary tumour is a predictor of progression-free and overall survivals in long-term follow-up. R0-resection was also connected with improved progression-free survival, which is in line with previous data^[19].

Given these findings, it seems reasonable to administer chemotherapy for attenuating the growth of occult disseminated tumours at the time of liver surgery^[20]. However, the efficacy of perioperative chemotherapy in resectable CRLM remains controversial^[21]. Randomised controlled trials (RCTs) are ongoing^[22,23]. For adjuvant chemotherapy, data from RCTs show a longer progression-free survival but no improvement in overall survival^[11]. In a recent review, more than 3330 studies were identified, but only four met the criteria for assessing the benefits of chemotherapy, regardless of the regimen, in patients undergoing surgery for CRLM^[24]. While all four studies used 5-fluorouracil, one study also used cetuximab. The results were sobering: in all four studies, progression-free survival was marginally, but significantly, prolonged, whereas overall survival was comparable between patients with and without chemotherapy. Hence, perioperative chemotherapy is not recommended in recently published guidelines^[10]. However, long-term follow-up data concerning survival and the use of chemotherapy are sparse for patients with CRLM^[21]. Our analysis showed no improvement of progression-free or overall survival for patients in long-term follow-up. The presented five-year overall survival is similar to other real-life cohorts^[3], which supports the findings from studies of short-term follow-up.

This study is subject to various limitations including the retrospective design and sample size, which was limited owing to the less number of patients undergoing perioperative chemotherapy. During the investigation period, perioperative chemotherapy was not generally recommended; hence the number of treated patients was less. Perioperative chemotherapy was administered on an individual basis at the physician’s discretion, mostly in patients assumed to have a poor prognosis. This contributed to the high rate of patients with synchronous metastases, bilobal tumour spread and higher Fong’s score included in this matched case-control study^[11]. All chemotherapeutic regimens were combined to perioperative chemotherapy. Most patients received either preoperative or postoperative treatment; only a few patients underwent truly perioperative chemotherapy with treatment before and after surgery. We did not perform any analyses considering individual drugs because of the limited number of patients. Thus, it remains possible that one specific regimen is beneficial. Nevertheless, even targeted agents have not shown efficacy in the studies performed to date^[25].

Conclusion

In conclusion, we performed a matched case-control study for evaluating the impact of perioperative chemotherapy on the long-term outcomes of patients with CRLM. Our results confirmed that perioperative chemotherapy does not improve progression-free or overall survival. Node-positivity of the primary was the sole predictor of overall survival after surgery for CRLM in long-term follow-up.

References

1. 1. Wolf, U., Barnes, B., Bertz, J., et al. The (German) Center for Cancer Registry Data (ZfKD) at the Robert Koch Institute (RKI) in Berlin. (2011) Bundesgesundheitsblatt Gesundheits forschung Gesundheitsschutz 54(11): 1229-1234.

   Pubmed| Crossref| Others

2. 2. Ferlay, J., Soerjomataram, I., Dikshit, R., et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. (2015) Int J Cancer 136(5): 359-386.

   Pubmed| Crossref| Others

3. 3. Manfredi, S., Lepage, C., Hatem, C., et al. Epidemiology and management of liver metastases from colorectal cancer. (2006) Ann Surg 244(2): 254-259.

   Pubmed| Crossref| Others

4. 4. Vogl, T.J., Farshid, P., Naguib, N.N., et al. Thermal ablation of liver metastases from colorectal cancer: radiofrequency, microwave and laser ablation therapies. (2014) Radiol Med 119(7): 451-461.

   Pubmed| Crossref| Others

5. 5. Xie, M., Zhu, J., He, X., et al. Liver Metastasis from Colorectal Cancer in the Elderly: Is Surgery Justified?. (2015) Dig Dis Sci 60(12): 3525-3535.

   Pubmed| Crossref| Others

6. 6. Lam, V.W., Laurence, J.M., Johnston, E., et al. A systematic review of two-stage hepatectomy in patients with initially unresectable colorectal liver metastases. (2013) HPB (Oxford) 15(7): 483-491.

   Pubmed| Crossref| Others

7. 7. Nordlinger, B., Quilichini, M.A., Parc, R., et al. Hepatic resection for colorectal liver metastases. Influence on survival of preoperative factors and surgery for recurrences in 80 patients. (1987) Ann Surg 205(3): 256-263.

   Pubmed| Crossref| Others

8. 8. Otto, G., Duber, C., Hoppe-Lotichius, M., et al. Radiofrequency ablation as first-line treatment in patients with early colorectal liver metastases amenable to surgery. (2010) Ann Surg 251(5): 796-803.

   Pubmed| Crossref| Others

9. 9. Gruenberger, T., Bridgewater, J., Chau, I., et al. Bevacizumab plus mFOLFOX-6 or FOLFOXIRI in patients with initially unresectable liver metastases from colorectal cancer: the OLIVIA multinational randomised phase II trial. (2015) Ann Oncol 26(4): 702-708.

   Pubmed| Crossref| Others

10. 11. S3-Leitlinie Kolorektales Karzinom (2017).

    Pubmed| Crossref| Others

11. 12. Nordlinger, B., Sorbye, H., Glimelius, B., et al. Perioperative FOLFOX4 chemotherapy and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC 40983): long-term results of a randomised, controlled, phase 3 trial. (2013) Lancet Oncol 14(12): 1208-1215.

    Pubmed| Crossref| Others

12. 13. Nigri, G., Petrucciani, N., Ferla, F., et al. Neoadjuvant chemotherapy for resectable colorectal liver metastases: What is the evidence? Results of a systematic review of comparative studies. (2015) Surgeon. 13(2): 83-90.

    Pubmed| Crossref| Others

13. 14. Araujo, R.L., Gonen, M., Herman, P. Chemotherapy for patients with colorectal liver metastases who underwent curative resection improves long-term outcomes: systematic review and meta-analysis. (2015) Ann Surg Oncol 22(9): 3070-3080.

    Pubmed| Crossref| Others

14. 15. Fong, Y., Fortner, J., Sun, R.L., et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases. (1999) Ann Surg 230(3): 309-318.

    Pubmed| Crossref| Others

15. 16. Strasberg, S.M. Nomenclature of hepatic anatomy and resections: a review of the Brisbane 2000 system. (2005) J Hepatobiliary Pancreat Surg 12(5): 351-355.

    Pubmed| Crossref| Others

16. 17. Ho, D., Imai, K., King, G., et al. MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. (2011) J Statistical software 42(8): 28.

    Pubmed| Crossref| Others

17. 18. Araujo, R.L., Gonen, M., Allen, P., et al. Positive postoperative CEA is a strong predictor of recurrence for patients after resection for colorectal liver metastases. (2015) Ann Surg Oncol 22(9): 3087-3093.

    Pubmed| Crossref| Others

18. 19. Sadot, E., Groot Koerkamp, B., Leal, J.N., et al. Resection margin and survival in 2368 patients undergoing hepatic resection for metastatic colorectal cancer: surgical technique or biologic surrogate? (2015) Ann Surg 262(3): 476-485.

    Pubmed| Crossref| Others

19. 20. Angelsen, J.H., Viste, A., Loes, I.M., et al. Predictive factors for time to recurrence, treatment and post-recurrence survival in patients with initially resected colorectal liver metastases. (2015) World J Surg Oncol 13: 328.

    Pubmed| Crossref| Others

20. 21. van der Stok, E.P., Grunhagen, D.J., Alberda, W.J., et al. The prognostic value of the primary tumor’s nodal status after surgery for colorectal liver metastases in the era of effective systemic therapy. (2015) Dig Surg 32(3): 208-216.

    Pubmed| Crossref| Others

21. 22. Araujo, R., Gonen, M., Allen, P., et al. Comparison between perioperative and postoperative chemotherapy after potentially curative hepatic resection for metastatic colorectal cancer. (2013) Ann Surg Oncol 20(13): 4312-4321.

    Pubmed| Crossref| Others

22. 23. Ayez, N., van der Stok, E.P., de Wilt, H., et al. Neo-adjuvant chemotherapy followed by surgery versus surgery alone in high-risk patients with resectable colorectal liver metastases: the CHARISMA randomized multicenter clinical trial. (2015) BMC Cancer 15: 180.

    Pubmed| Crossref| Others

23. 24. Kanemitsu, Y., Kato, T., Shimizu, Y., et al. A randomized phase II/III trial comparing hepatectomy followed by mFOLFOX6 with hepatectomy alone as treatment for liver metastasis from colorectal cancer: Japan Clinical Oncology Group Study JCOG0603. (2009) Jpn J Clin Oncol 39(6): 406-409.

    Pubmed| Crossref| Others

24. 25. Khoo, E., O’Neill, S., Brown, E., et al. Systematic review of systemic adjuvant, neoadjuvant and perioperative chemotherapy for resectable colorectal-liver metastases. (2016) HPB (Oxford) 18(6): 485-493.

    Pubmed| Crossref| Others

25. 26. Suenaga, M., Fujimoto, Y., Matsusaka, S., et al. Perioperative FOLFOX4 plus bevacizumab for initially unresectable advanced colorectal cancer (NAVIGATE-CRC-01). (2015) Onco Targets Ther 8: 1111-1118.

    Pubmed| Crossref| Others