Prognostic factors after ultrasound‐guided percutaneous ablation of colorectal liver metastases: A systematic review

Abstract

Introduction

Liver metastases from colorectal cancer are common and ablation therapy is a favourable treatment option for selected patients not suited for surgical resection. This study aimed to systematically review the literature and present prognostic factors associated with survival and local recurrence after percutaneous ultrasound‐guided ablation treatment.

Materials and methods

This review is reported according to the PRISMA. PubMed, Embase and Scopus were searched and records were independently screened by two authors, initially on title and abstract and subsequently on full‐text basis. The quality of the studies was assessed using the Newcastle‐Ottawa quality assessment scale.

Results

Of 2.882 records screened, 18 studies were included. The median survival was 23 months. One‐year survival was median 95% and 3‐year survival was median 58%. Complete ablation response and adjuvant chemotherapy produce considerably improved survival and low local recurrence rate outcomes.

Conclusion

Percutaneous ultrasound‐guided ablation technique for colorectal liver metastases provides impressive survival rates for patients not suited for surgical resection. However, there are some factors related to poorer prognosis, which may be considered when selecting patients.

Introduction

Up to 50% of patients with colorectal cancer will develop metastases to the liver,1 and untreated the prognosis is poor with a median survival of less than 12 months, and a 5‐year survival of less than 6%.2, 3, 4, 5, 6

The main treatment for hepatic metastases from colorectal cancer is surgical resection with a 5‐year survival rate of 30–58%.2, 7, 8, 9 However, less than 20% of patients are candidates for resection due to anatomic location, large tumour size, or comorbidities.8, 10 Ablation therapy is an alternative to metastasectomy and is widely used to treat both primary liver cancer and liver metastases.11, 12 Several advantages exist regarding ablation therapy including percutaneous ablation being a less invasive, repeatable procedure, which has the potential to be applied on a wider population. To provide optimal patient care predictors for a poor prognosis should be identified and patient selection should be based on an evidence‐based risk factor assessment. However, this has not been systematically reviewed.

Therefore, the aim of this review was to identify studies investigating risk factors for survival and local recurrence after ablation therapy performed percutaneously and ultrasound‐guided for colorectal liver metastases.

Materials and methods

The study was reported according to Preferred Reporting Items of Systematic Reviews and Meta‐Analysis (PRISMA).13 A flow chart of the selection process of studies can be seen in Figure 1. Before initiation of the review, a protocol was registered at PROSPERO (registration number CRD42015027837).14

Figure 1.

Inclusion of Studies ( PRISMA 2009 Flow Diagram).

We included studies evaluating patients with colorectal liver metastases. Included patients were treated with any ablation therapy with ultrasound‐guided and percutaneous techniques. The outcomes were survival and recurrence. We included studies that reported survival as median survival, overall survival and progression‐ and disease‐free survival. Recurrence was defined in all studies as time from ablation treatment to either development of new focal areas of contrast enhancement in/or adjacent to the ablated area (local recurrence or local tumour progression), or as occurrence of new metastases at a site different from the one treated. Studies of interest were observational studies, randomised clinical trials, cohort and case–control studies, and case series with more than four patients. Furthermore, included studies had obtained informed consent from each study participant and/or approval by an ethics committee or institutional review board. Records that met the inclusion criteria were obtained as full‐text.

PubMed and Embase were searched on 26th November 2017 using the following search terms: ((liver OR liver tumour OR liver tumour OR liver mass OR tumour OR tumour OR metastasis OR metastases OR neoplasm metastasis) AND (colorectal cancer OR rectal cancer OR colon cancer OR rectum cancer OR large bowel cancer OR colorectal neoplasms OR colon OR rectum) AND (percutaneous OR cutaneous administration OR dermal administration OR ultrasonography OR ultrasonic imaging OR ultrasonic diagnosis OR ultrasound‐guided OR ultrasound) AND (ablation OR ablative techniques OR radiofrequency OR thermal OR laser OR coagulation OR thermal coagulation OR ablation therapy OR RFA OR microwave ablation)). The search strategy was modified to fit Scopus: (‘liver’ AND ‘metastasis’ AND ‘colorectal cancer’) AND ‘percutaneous’ AND ‘ablation’ AND (‘survival’ OR ‘recurrence’). Two additional studies were found usingsnowball search (Figure 1). Only English language articles available in full‐text were included.

Screening and selection were performed with the use of the Covidence online platform (http://covidence.org). Duplications in title and abstract were removed and two of the authors performed title and abstract screening and subsequently full‐text screening, independently. Any discrepancies were solved by discussion.

As we only included observational studies, the quality of the studies was assessed using Newcastle‐Ottawa quality assessment scale (NOS).15 It is a star‐based assessment system developed to evaluate bias in non‐randomised studies, where a maximum of nine stars can be obtained. It consists of eight questions evaluating three topics: selection of participants (maximum of four stars), comparability of study groups (maximum of two stars) and outcome assessment (maximum of three stars). The NOS score was categorised so 1–3 stars were considered as high risk of bias #bib4–6 as moderate risk of bias, and 7–9 as low risk of bias. The compiled NOS score is presented as median and range.

Results

Eighteen studies were included with a total of 883 patients (Figure 1). Of these studies, radiofrequency ablation (RFA), microwave ablation (MWA) and laser ablation were evaluated by ten, four, and four studies, respectively. In one study, six of 22 patients also received percutaneous alcohol injection.16 Another study only included patients treated with RFA while waiting for hepatic metastasectomy.17 The hypothesis behind this was that RFA would decrease the number of patients operated due to complete ablation response after RFA.

Patient demographics are given in Table 1. Between 70–100% of the patients received neoadjuvant and/or adjuvant chemotherapy. One study included only patients with progressive disease undergoing second or third line palliative chemotherapy18 and three studies did not have information on the use of chemotherapy.19, 20, 21 A single study included patients treated with microwave ablation combined with synchronous trans‐catheter arterial chemoembolization.22 Three of the eighteen studies also included patients with extrahepatic disease.22, 23, 24 Five studies had no information regarding extrahepatic disease.17, 18, 19, 25, 26

Table 1.

Patient demographics

Study	No. of patients (n)	Agea (years)	Male sex (n (%))	No. of metastases (n)	Solitary metastases (%)	Tumour sizea (cm)
Amin et al. (1993)16	22	66 (28–77)b	–	76	–	2.0 (0.5–9)b
Chiou et al. (2005)19	63	65 (34–89)	39 (62)	109	49	2.9 (1.2–5.0)
Choi et al. (2004)25	24	63 (29–77)b	13 (54)	45	46	3.4 (1.7–13.0)b , c
Facciorusso et al. (2016)28	143	68 (41–85)b	102 (71)	215	–	2.6 (10–48)b , c
Hänsler et al. (2007)26	17	71 (57–87)	12 (71)	26	59	3.6 (2.2–6.0)
Ierardi et al. (2013)32	17	68 (52–83)	13 (77)	21	82	4.0 (1.7–7)
Knudsen et al. (2009)29	36	67 (46–83)b	20 (56)	158	33	2.1 (0.7–6.5)b , c
Lichun et al. (2017)27	43	66 (38–78)	24 (56)	113	–	–
Livraghi et al. (2003)17	88	65 (39–75)	53 (60)	134	56	2.1 (0.6–4.0)
Nishiwada et al. (2014)23	32	< 65: 13 ≥ 65: 19	19 (59)	47	78	≤ 2.0: 9 > 2.0: 23
Nouso et al. (2010)33	23	67 (43–80)b	13 (57)	–	44	2.3 (1.1–4.5)b
Pacella et al. (2006)30	44	68 (42–82)b	29 (66)	75	–	3.4 (0.5–9.0)b
Seki et al. (1999)31	15	66 (55–75)	10 (67)	15	100	2.1 (1.3–2.6)
Solbiati et al. (2012)24	99	65 ± 12 (37–88)	54 (55)	202	42	2.2 ± 1.1 (0.8–4.0)
Wang et al. (2014)20	115	60 ± 12 (30–86)	78 (68)	165	42	3.1 ± 1.1 (1.3–5.0)
Wietzke‐Braun et al. (2004)18	45	62 (38–79)b	30 (67)	–	–	–
Wu et al. (2016)22	30	62 (44–78)	19 (63)	43	67	4.4 ± 2.6 (1.4–10.0)
Yoon et al. (2011)21	27	–	–	35	–	2.0 (0.7–3.8)

^aMean value (range) of reported values.

^bMedian value (range) of reported values.

^cSize of largest tumour.

Survival

Survival was investigated in fourteen studies with a total of 692 patients (Table 2). Survival was reported as time from first ablation treatment to death or to last day of follow‐up, unless otherwise noted. Median survival was reported in ten studies with a median of 23 months (range 9–53).16, 18, 22, 24, 25, 27, 28, 29, 30, 31 One‐year survival was reported in eight studies with a median of 95% (range 47–100%)20, 21, 22, 24, 25, 26, 28, 30 and three‐year survival was reported in six studies with a median of 58% (range 26–79%).20, 21, 23, 24, 29, 30 Only one study reported 7‐ and 10‐year survival of 25% and 18%, respectively.24 Among the included studies, we did not find a clear relationship between survival rates and ablation technique, use of chemotherapy, or number of metastases.

Table 2.

Survival outcomes. Survival was reported as time from first ablation treatment to death or to last day of follow‐up unless otherwise noted

Study	Patients (n)	Ablation technique	Median survival (months)	One‐year survival (%)	Two‐year survival (%)	Three‐year survival (%)	Five‐year survival (%)
Amin et al. (1993)16	22	Laser	27	–	–	–	–
Choi et al. (2004)25	24	RFA	17	81	26	–	–
Facciorusso et al. (2016)28	143	RFA	44	91	–	–	42
Hänsler et al. (2007)26	17	RFA	–	94	86	–	–
Knudsen et al. (2009)29	36	RFA	39a	–	–	26	–
Lichun et al. (2017)27	43	Laser	19	–	–	–	–
Nishiwada et al. (2014)23	32	RFA	–	–	–	47	39
Pacella et al. (2006)30	44	Laser	22	95	–	35b	–
Seki et al. (1999)31	15	MW	24	–	–	–	–
Solbiati et al. (2012)24	99	RFA	53	98		69	48
Wang et al. (2014)20	115	MW	–	98	87	79	–
Wietzke‐Braun (2004)18	45	Laser	9	–	–	–	–
Wu et al. (2016)22	30	MW	11	47	25	–	–
Yoon et al. (2011)21	27	RFA	–	100	82	77	–

RFA, Radiofrequency ablation; MW, Microwave ablation.

^aFrom time of diagnosis.

^bAfter complete ablation response.

Two studies described the effect of complete vs. incomplete ablation response.25, 30 A complete ablation response was defined as an effect on the entire tumour and a margin of at least 0.5 cm adjacent normal liver tissue surrounding the tumour. The first study found a median survival after complete ablation of 18 months compared with 13 months after incomplete ablation (P = 0.039). The rate of complete ablation was reported as significantly higher in patients with tumours less than 4 cm in diameter (P < 0.05).25 The second study found an overall survival after complete ablation of 30 ± 13 months compared with an overall survival of 20 ± 10 months after partial ablation or minor response (P = 0.002).30 The same study reported that complete ablation was achieved in 87% of tumours with a diameter of less than 3 cm and in 23% of tumours with a larger diameter.

Three studies described the effect of tumour size in relation to survival.24, 25, 30 One study found a median survival in patients with tumours larger than 4 cm of 12 months compared with a median survival just exceeding 30 months in patients with tumours less than 4 cm (P = 0.008).25 The same study found no correlation between tumour size and progression‐free survival. The second study found no difference in 3‐year survival after complete ablation in patients with tumours over/under 3.0 cm in diameter.30 The third study reported that overall 10‐year survival was not related to tumour size (P = 0.08).24

Two studies found no significant effect of number of metastases (n = 1–4) on overall survival, progression‐free survival, or overall 10‐year survival.24, 25

Two studies evaluated the effect of adjuvant chemotherapy.25, 27One found a median overall survival in patients who received adjuvant chemotherapy of 21 months compared with 13 months for those who did not (P = 0.017).25 In the second study, patients were assigned to a treatment group (laser ablation and chemotherapy) or a control group (chemotherapy) with a median progression‐free survival of 11.8 months and 6.8 months, respectively, (P < 0.05).27 In addition, the median overall survival for the treatment and control groups were 19.1 months and 14.9 months, respectively (P < 0.05).

Lastly, a study evaluated the prognostic role of 25‐hydroxyvitamin D levels and found the median overall survival was 65 months if 25‐hydroxyvitamin D > 20 ng/mL and 34 months if 25‐hydroxyvitamin D ≤ 20 ng/mL (P < 0.001).28

Recurrence

Recurrence was investigated in fourteen studies as local recurrence (LR, n = 9), local tumour progression (LTP, n = 4), and/or new liver metastases (n = 12) in 737 patients (Table 3). LR and LTP had the same definition in all studies, which was the appearance of tumour in/or adjacent to a previously ablated area. LR/LTP were reported in ten studies with a median of 32% (range 12–50%) with a reported follow‐up in range 12–53 months.17, 20, 23, 24, 25, 26, 28, 30, 32, 33 Three studies reported LR/LTP by number of metastases and found LR in 14% with a mean follow‐up of 22 months19, LR in 44% with a median follow‐up of 27 months22, and the third study found LTP in 40% with a median follow‐up of 27 months21. In addition, recurrence rate was assessed after one, two and three years in one study, which reported 28% #bib48% and 59%, respectively.20 Another study assessed the rate of new intrahepatic metastases at similar time intervals and found rates of 29% #bib41%, and 56%, respectively.33 We did not find a clear relationship between LR/LTP and ablation technique, use of neoadjuvant and/or adjuvant chemotherapy, nor between LR/LTP and tumour size or number of metastases in the included studies.

Table 3.

Rates of recurrence after ablation therapy. Local recurrence and local tumour progression were both defined as time from ablation treatment to development of contrast enhancement in/or adjacent to an ablated area. New metastasis was defined as development of new focal areas at a site different from the one treated

Study	Patients (n)	Follow‐up (median, months)	Ablation technique	LR (%)	LTP (%)	New hepatic metastases
Chiou et al. (2005)19	63	22a	RFA	14b	–	10/63 patients developed new liver metastases
Choi et al. (2004)25	24	12	RFA	50	–	–
Facciorusso et al. (2016)28	143	72	RFA	19	–	76/143 developed new metastases
Hänsler et al. (2007)26	17	852 days	RFA	–	24	11/17 patients developed distant recurrence, four of these also showed LTP
Ierardi et al. (2013)32	17	12	MW	12	–	2/17 patients developed disease progression with new lesions
Livraghi et al. (2003)17	88	33	RFA	40	–	53/88 patients had complete ablation, of these 37 patients developed new lesions. 26/37 patients developed intrahepatic lesions and 7/37 patients developed both intra‐ and extrahepatic lesions. Of the 35/88 patients with partial necrosis, all demonstrated either LR or new lesions.
Nishiwada et al. (2014)23	32	24c	RFA	47	–	15 patients had recurrence in the liver remnant
Nouso et al. (2010)33	23	492 daysa	RFA	17	–	10/23 patients developed liver metastases, of these five had distant intrahepatic recurrence and one patient developed both local and distant intrahepatic recurrence
Pacella et al. (2006)30	44	−	Laser	48	–	13/44 patients developed new liver metastases
Seki et al. (1999)31	15	(9–37)	MW	–	–	One patient demonstrated local recurrence and distant intrahepatic metastases (after incomplete necrosis), and three developed distant intrahepatic recurrence
Solbiati et al. (2012)24	99	53	RFA	–	41	40/99 patients developed new intrahepatic metastases
Wang et al. (2014)20	115	28	MW	–	12	14 developed distant intrahepatic recurrence
Wu et al. (2016)22	30	27	MW	44b	–	–
Yoon et al. (2011)21	27	27	RFA	–	40b	16/27 patients showed new lesions at other liver sites and/or distant areas

RFA, Radiofrequency ablation; MW, Microwave ablation.

^aMean value.

^bReported by number of metastases instead of patients.

^cOverall observation time.

Three studies described the effect of ablation response in relation to recurrence.17, 30, 31 One study found time to progression to be 11 ± 9 months after complete ablation and 6 ± 4 months after partial ablation (P = 0.024).30 Another study reported that 13 of 15 patients showed complete necrosis and none of these developed LR, but of the two with incomplete necrosis one developed LR.31 The third study reported complete ablation in 53 patients and partial ablation in 35 of 88 patients.17 In this study #bib16 of the 53 patients with complete ablation response remained free of disease after the first treatment and 37 developed new intra‐ and extrahepatic lesions. Some intrahepatic lesions were re‐treated, and overall 23 of the 53 patients (43%) remained free of disease. Of the 35 patients with partial necrosis, all lesions demonstrated LR or new lesions on the follow‐up imaging. After re‐treatment #bib20% remained free of disease.

Four studies described the effect of tumour size in relation to recurrence.23, 24, 30, 33 One study reported that local recurrence‐free survival (LRFS) after RFA was the same as after hepatic resection in patients with a tumour size less than 2.0 cm (P = 0.676), but LRFS was significantly lower after RFA in patients with tumours larger than 2.0 cm (P < 0.001).23 The same study showed that tumours above 2.0 cm were a significant risk factor for LR (P = 0.046). A second study stated that none of the single tumours less than 2.0 cm treated with RFA had LR.33 A third study found no difference after complete ablation in time to progression between those having metastases less than 3.0 cm compared to those having metastases larger than 3.0 cm.30 The fourth study found no difference in incidence of new events (neither new metastases nor LTP) in relation to tumour size (P = 0.08).24

Two studies evaluated the effect of number of metastases on recurrence.23, 24 The first study found that multiple liver metastases were a significant risk factor for LR (P = 0.014).23 The second study found no difference in the incidence of new metastases between one, two, three or ≥ four metastases (P = 0.724).24

A single study stated that absence of adjuvant chemotherapy was a significant risk factor for LR (P = 0.019).23

One study evaluated time to recurrence when stratifying by 25‐hydroxyvitamin D level and found median 24 months in the group with 25‐hydroxyvitamin D level ≤20 ng/mL and 50 months in the group >20 ng/mL (P < 0.001).28

Quality assessment

The median overall NOS score for all included studies was 6 (range 4–8). The NOS score was lowered as many studies did not have a non‐exposed cohort and had a short follow‐up.

Discussion

The basic findings of this systematic review were that complete ablation response and adjuvant chemotherapy seemed to be correlated with improved survival and lower LR rates. A number of studies reported that complete ablation response was associated with tumours having a maximal diameter of 2–4 cm. One study evaluated the prognostic role of 25‐hydroxyvitamin D level and found a better median overall survival and a longer time to recurrence when 25‐hydroxyvitamin D level was lower than 20 ng/mL. The importance of removing or destroying the whole tumour is of paramount importance in invasive procedures whether it is surgery or ablation therapy. These results emphasise the importance of combination therapy with chemotherapy and an invasive procedure to improve the outcome for the patients by making sure that all tumour tissue is removed or destroyed by ablative treatment. The results concerning the prognostic value of 25‐hydroxyvitamin D level is indeed interesting and should be a subject for further research as it could be a valuable and easily assessable prognostic factor in the future. Unfortunately, number of metastases and size were not consistently reported to have a significant effect. This could be explained by the low number of patients included in some of the studies. Furthermore, the reported three‐year survival rates of the included studies with median of 58% (range 26–79%) compares excellently to the reported three‐year survival of 43–67% and five‐year survival of 30–58% after surgical resection.2, 7, 8, 9, 34 In many of the included studies, some of the inclusion criteria for ablation were unresectable hepatic metastases or patients not eligible for surgery. Ablation is presently regarded as an inferior treatment option to hepatic resection, however, with these well comparable survival rates, the time might have come to consider a paradigm shift and the consideration of percutaneous ablation as an equal treatment option to surgery, at least for solitary metastases <3 cm where metanalysis has proven equal treatment results in terms of overall survival and disease‐free survival.35, 36

A few limitations in this study should be mentioned. First, being a literature review the data reported are limited to the search strategy and the restriction of only including English language articles. Secondly, the majority of data reviewed outcomes from single‐centres and were often retrospective studies with inherent risk of bias, primarily selection bias. Furthermore, except for one study27, all the included studies were non‐randomised and without a non‐exposed cohort. Therefore, a causal link between ablation technique and the primary outcomes cannot be concluded.

A number of different factors have previously been proposed as prognostic factors in the literature. Among these are liver volume replaced by tumour, metachronous vs. synchronous metastases, laboratory tests (including carcinoembryonic antigen), bilateral vs. unilateral liver involvement, extrahepatic disease, number of metastases and size of the largest lesion. However, most of these factors were not analysed in the included studies. In contrast to similar studies that included intraoperative and laparoscopic ablation interventions35, we only included studies focusing on a percutaneous method to clarify factors specific for this approach, with the intention to reveal novel information concerning this technique. However, as this comprehensive literature search failed to identify more studies it clearly highlights the need for further research to be performed in the coming years to reveal important and clinically relevant prognostic factors. Ablation performed percutaneously under ultrasound‐guidance is the least invasive method with no risk of radiation and for the sake of the patients, prognostic and risk factors should be identified for this intervention. Currently, no clear guidelines exist concerning the role of percutaneous ablation in the treatment of colorectal liver metastases.

Conclusion

In conclusion, percutaneous ultrasound‐guided ablation technique for colorectal liver metastases provides impressive survival rates and is a good treatment alternative for patients not suited for curative surgical resection. In contrast to surgery, there are no technical limitations to the number of treatments with percutaneous ultrasound‐guided ablation therapy in one patient as long as the lesions are visualised with ultrasound. Survival rates after ablation therapy and resection are well comparable and perhaps these treatment options should be considered equivalent for patients with solitary metastasis <3 cm where data, including metanalysis, indicate equal treatment results. Comparison between interventions such as chemotherapy and ablation therapy would provide valuable information but was not possible to obtain from the available studies and is difficult to evaluate as modern cancer therapy increasingly implies a multimodality treatment regime. However, it may be a subject for further studies.

Disclosure statement

The authors received no financial support for this study.

Authorship declaration

The authorship conforms with the journal's authorship policy, and all authors have read and approved the final version of the manuscript.