Background
The liver is the most common site of colorectal cancer metastasis, and approximately 50% of patients develop liver metastasis at some point during their disease course [
1]. Hepatic resection is a potentially curative treatment for colorectal cancer liver metastasis (CRLM) patients with 5-year survival rates of 30–50% [
2]. Unfortunately, more than half of CRLM patients recur after hepatic resection, and the majority of such recurrences occur within 2 years [
3].
Recent advances in surgical approaches have dramatically changed the treatment strategy for recurrent CRLM patients, which has allowed the description of risk factors for survival after a second round of local treatment. Previous studies have provided evidence that repeat hepatic resection is feasible in selected patients with intrahepatic recurrence [
4,
5]. However, which individuals would yield maximum benefit from aggressive treatment following first recurrence is not clear, as there is a paucity of data on the risk factors for survival in this patient population. There has been no consensus introduced on how to select candidates.
The present study aimed to investigate prognostic factors for survival and select candidates for the local treatment of CRLM after first recurrence.
Methods
Study population
From January 2000 to September 2019, a total of 1027 patients underwent curative hepatic resection for CRLM at the Hepatopancreatobiliary Surgery Department I of Peking University Cancer Hospital. The demographic and clinical data were retrospectively obtained from a prospectively collected patient database. Patients were enrolled in the study based on the following criteria: (1) first recurrence developed after initial radical resection of both the liver metastasis and primary tumor; (2) the initial recurrence was considered resectable and liver-limited by a multidisciplinary team (MDT); (3) no extrahepatic metastasis had occurred since the first hepatic resection; and (4) there were no other simultaneous malignancies. The technical criteria of resectability related to the liver remnant after resection were as follows: (a) anticipated ability to preserve two contiguous segments; (b) anticipated ability to preserve adequate vascular inflow, outflow and biliary drainage; and (c) anticipated ability to preserve an adequate future liver remnant volume (30% in normal livers and 40% in livers pretreated with chemotherapy) [
6]. The present study was approved by an Institutional Review Board.
Preoperative management for first recurrence
At first recurrence, all patients underwent routine laboratory tests, including serum carcinoembryonic antigen (CEA) and carbohydrate antigen (CA) 19–9 levels and liver function tests. Routine imaging modalities, including enhanced computed tomography (CT) scans of the abdominal, thoracic and pelvic regions and hepatic magnetic resonance imaging (MRI), were used to determine the disease stage. Strict criteria were used to detect evidence of cancer recurrence (local, metastatic or secondary colorectal cancer). Recurrence was diagnosed by CT scans, MRI, PET‐CT, or pathology (biopsy or specimen). All imaging reports that suggested recurrence were reviewed carefully and compared to baseline imaging reports to confirm that disease was not present at the time of liver resection. Preoperative chemotherapy was recommended for patients who presented a heavy tumor burden (CRS > 2) and/or recurred within a short period (disease-free survival time of less than 6 months) after the first hepatic resection. The response to chemotherapy was evaluated by MRI according to the Response Evaluation Criteria in Solid Tumors (RECIST 1.1) [
7].
Indications for local treatment
The treatment strategy at first recurrence after hepatic resection was basically the same as that at first hepatic resection; that is, recurrent disease was treated locally only when the overall strategy was considered curative, and all detectable lesions had a tumor-free margin. The resectability was discussed, and clinical treatment decisions were made by an MDT. Systemic chemotherapy was recommended if the patient refused secondary local treatment. Resection of three or more segments was considered major hepatic resection. Percutaneous radiofrequency ablation (RFA) was performed using a CELON system (Teltow, Germany). The bipolar electrode needles were 16G, and guidance ultrasonography was performed using Aloka α-10 (Tokyo, Japan) and GE Logiq E9 (Connecticut, USA) devices. In general, RFA was recommended for deeply located tumors that would require extended resection of the normal parenchyma. RFA was contraindicated when (1) the diameter of the largest tumor exceeded 3 cm and (2) the tumor was adjacent to major bile ducts or large blood vessels or the colon/gallbladder were strictly restricted.
Postoperative work-up
After completing treatment, all patients underwent regular follow-up examinations with hepatic MRI and CT scans of the abdominal/thoracic/pelvic region, and levels of tumor markers were measured every three months. Adjuvant chemotherapy was usually recommended [
8].
Statistical analysis
Continuous variables are summarized as the means, and categorical variables are summarized as frequencies and percentages. The means of variables were compared with chi square analysis or Fischer’s exact test (depending on the sample size) or with the independent Student’s t test or Mann–Whitney U test, as appropriate. The optimal cutoff points for the definition of early recurrence were determined using the minimum p value approach, which was calculated using the log-rank test for OS after first recurrence. Survival analyses were carried out using the Kaplan–Meier method with the log-rank test. OS was calculated from the date of local treatment or the date of systemic chemotherapy after first recurrence until death or the last follow-up. OS was the primary endpoint for studies with repeat hepatic resection [
3,
9]. Cutoff values for continuous variables were determined based on the C-statistic, with a Cox regression model for survival data including censored patients. The estimated cutoff values for each variable were tumor number, 1–5; tumor size, 20–550 mm; CEA level, 5 (upper limit of the normal range), 10, 20, 50, 100, 200 ng/ml; and CA19-9, 37 (upper limit of the normal range), 50, 100 units/ml
7. All statistical analyses were performed using SPSS 26.0 and R version 3.2.6 (
http://www.r-project.org), and p values < 0.05 were considered statistically significant.
Establishment of nomogram
Univariable and multivariable analyses of various clinicopathological factors by Cox’s proportional hazard model were used to identify independent risk factors for OS in 300 patients. The results of multivariable analysis were used to develop an OS prediction nomogram with 1-, 3-, and 5-year OS as the endpoints. The C-index was calculated to assess the degree of discrimination, and calibration plots were generated to visualize the agreement between the predicted and actual 1-, 3- and 5-year OS with bootstrapped samples. Using all categorical risk factors, we found the best separation in terms of survival by permutation and created two risk groups, a high‐risk and a low‐risk group.
Discussion
Curative resection for CRLM improves survival outcomes and curative opportunities; however, the majority of patients will develop recurrence [
13]. It has been suggested that secondary hepatic resection is a safe and feasible procedure for recurrent CRLM patients. However, repeat hepatic resection may not be possible due to anatomical or functional restraints [
9,
10]. RFA is a useful alternative in this situation. The indications for repeat hepatic resection were believed to be the same as those for initial resection [
4,
14]. However, the disease characteristics were not exactly compared between the initial and recurrent resection time points.
The indication and concerning prognostic factors for the local treatment of first recurrence have not been well studied or defined. Previous analyses have been limited by small and single cohorts of patients [
4,
14]. In the present study, 173 patients who underwent local treatment at first recurrence for CRLM and had significantly better survival outcomes than those who only received systemic chemotherapy. Rumor size > 3 cm, node-positive primary, early recurrence, and RAS mutation have been identified to be risk factors for survival outcomes CRLM patients at first recurrence. Furthermore, in the high-risk group (total score ≥ 13), local treatment did not significantly improve OS (median OS: 21 vs. 15 months, p = 0.126). The present model might facilitate the identification of patients who may benefit most from repeat local treatment before an individualized treatment decision is made.
Recurrence occurs in 60–84% of CRLM patients after initial hepatic resection [
15,
16]. The recent expansion of indications for surgery has led to an increase in the number of patients with potentially resectable disease, while these factors can also lead to an increased risk of early recurrence. Early recurrence has been reported to adversely influence survival after hepatic resection and therefore remains a concern for repeat hepatic resection because of worse patient prognosis [
17,
18]. The present study investigated whether the optimal cutoff point for early recurrence was 6 months. Although it remained an independent prognostic risk factor, aggressive treatment should therefore be proactively considered even for patients with early recurrence. If the patient was determined to have unresectable early recurrence, systemic chemotherapy should be recommended. For patients with resectable disease, radical resection should be performed as often as possible. Two cycles of neoadjuvant chemotherapy should be recommended, followed by local treatment. If there is a risk of removing liver metastases, the patient should undergo upfront resection directly. RFA should be performed if the tumor is located deep in the liver.
The liver and lungs are the predominant sites of secondary recurrence, of which liver metastases are especially common [
3,
19]. Therefore, the present study enrolled only patients with resectable, liver-limited disease at first recurrence. Repeat hepatic resection for recurrence has been reported to be associated with an equivalent long-term outcome to first hepatic resection, with a similarly low surgical risk, and the 5-year OS rate ranged from 27 to 45% [
20‐
23]. RFA has emerged as an alternative radical treatment with lower invasiveness and a lower complication rate and is effective for patients with comorbidities and recurrent liver disease [
24,
25]. Previous studies of percutaneous RFA for recurrent CRLM yield similar results [
26,
27]. It has been suggested that tumor diameter and number are the most important factors that influence the efficacy [
28,
29]. In our study, RFA was recommended in deeply located tumors that would require extensive resection of the normal parenchyma.
Given that the survival of patients who undergo liver resection for colorectal cancer metastases is long, surgeons and medical oncologists are now dealing with a “chronic disease” that should be treated differently depending on its presentation upon recurrence [
30]. In the modern era, in addition to the current disease state of CRLM patients, recurrence should be considered when determining the treatment strategy. Preoperative chemotherapy might shrink tumors and increase resectability. In particular, it also likely selected individuals with recurrence who would benefit from resection. Therefore, this treatment might be recommended for patients who develop early recurrence or have a heavy disease burden in the liver.
Limitations
The limitations of the present study included its retrospective design and the gradual change in indications for resection and RFA over the course of the study. First, although the enrolled patients were determined to have resectable disease, bias could still have existed from patient selection in the three groups. Therefore, these results should be validated with further better quality studies. Second, it is possible that other unknown factors could affect the accuracy of the nomogram model and that a higher number of cases might be required to be detect them. This type of study is difficult to perform prospectively, and any trials must be very carefully performed. Moreover, there was no external validation cohort in the present study. Finally, all data were collected from a limited number of institutions.
Acknowledgements
No preregistration was performed for the reported studies reported in this article. We acknowledge QB, who contributed to the study by making substantial contributions to the acquisition of the data, and KMJ, who made substantial contributions to the analysis and interpretation of the data. They were involved in drafting the manuscript but did not meet the criteria for authorship.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.