Attaining sufficient ablation margins (>5 mm, and, ideally, >10 mm) can result in significant lowering of local tumor progression rates after percutaneous radiofrequency ablation of colorectal cancer liver metastases; in addition, a modified clinical risk score for ablation can be used as a prognostic stratification tool.
Abstract
Purpose
To identify predictors of oncologic outcomes after percutaneous radiofrequency ablation (RFA) of colorectal cancer liver metastases (CLMs) and to describe and evaluate a modified clinical risk score (CRS) adapted for ablation as a patient stratification and prognostic tool.
Materials and Methods
This study consisted of a HIPAA-compliant institutional review board–approved retrospective review of data in 162 patients with 233 CLMs treated with percutaneous RFA between December 2002 and December 2012. Contrast material–enhanced CT was used to assess technique effectiveness 4–8 weeks after RFA. Patients were followed up with contrast-enhanced CT every 2–4 months. Overall survival (OS) and local tumor progression–free survival (LTPFS) were calculated from the time of RFA by using the Kaplan-Meier method. Log-rank tests and Cox regression models were used for univariate and multivariate analysis to identify predictors of outcomes.
Results
Technique effectiveness was 94% (218 of 233). Median LTPFS was 26 months. At univariate analysis, predictors of shorter LTPFS were tumor size greater than 3 cm (P < .001), ablation margin size of 5 mm or less (P < .001), high modified CRS (P = .009), male sex (P = .03), and no history of prior hepatectomy (P = .04) or hepatic arterial infusion chemotherapy (P = .01). At multivariate analysis, only tumor size greater than 3 cm (P = .01) and margin size of 5 mm or less (P < .001) were independent predictors of shorter LTPFS. Median and 5-year OS were 36 months and 31%. At univariate analysis, predictors of shorter OS were tumor size larger than 3 cm (P = .005), carcinoembryonic antigen level greater than 30 ng/mL (P = .003), high modified CRS (P = .02), and extrahepatic disease (EHD) (P < .001). At multivariate analysis, tumor size greater than 3 cm (P = .006) and more than one site of EHD (P < .001) were independent predictors of shorter OS.
Conclusion
Tumor size of less than 3 cm and ablation margins greater than 5 mm are essential for satisfactory local tumor control. Tumor size of more than 3 cm and the presence of more than one site of EHD are associated with shorter OS.
© RSNA, 2015
Introduction
Liver metastases arise in 45% of patients with colorectal cancer (1). At initial presentation, 15%–25% of patients have liver metastases (synchronous), while an additional 15%–20% develop liver metastases after primary tumor resection (metachronous) (1,2). Surgery is considered to be the standard of care for resectable disease, with 5-year overall survival (OS) rates of 28%–58% (3–5). Unfortunately, only 10%–20% of patients have resectable disease at the time of presentation (1,2). Historically, OS for patients with unresectable disease was reported to be 6–12 months (2,6). Recent advances in systemic chemotherapy have extended this to 20–24 months (7,8).
Percutaneous image-guided radiofrequency ablation (RFA) is a generally accepted alternative therapy for patients who are poor surgical candidates and for patients with posthepatectomy recurrences (9–12). This minimally invasive approach has a favorable safety profile and a low major complications rate of 1.3%–2.2% (13,14). Patients treated with RFA for unresectable colorectal cancer liver metastases (CLMs) have 5-year OS rates of 21%–48%, approaching those for surgical resection (13,15–17). Despite these results, relatively high local tumor progression (LTP) rates remain an obstacle for the widespread use of RFA (10,15,18).
Major variations in baseline characteristics and inclusion criteria make it difficult to compare outcomes of RFA series from different institutions, let alone with outcomes of surgical series (18). Use of a stratification prognostic score, such as the clinical risk score (CRS) described in the surgical literature (4,10,19), may facilitate such comparisons.
The aim of our study was to identify predictors of oncologic outcomes after percutaneous RFA of CLMs and to describe and evaluate a modified CRS adapted for ablation as a patient stratification and prognostic tool (10).
Materials and Methods
Study Population
We performed an institutional review board–approved retrospective review of our prospectively maintained Health Insurance Portability and Accountability Act–registered clinical ablation database. All percutaneous ablations performed to treat CLMs between December 2002 and December 2012 were noted. We identified 184 patients treated with RFA (n = 165), microwave ablation (n = 10), irreversible electroporation (n = 8), or cryoablation (n = 1). Only the 165 patients treated with RFA were eligible for inclusion and analysis in this study. We excluded three patients: One patient in whom the primary tumor was in situ and two patients who were lost to follow-up before the first 4–8-week post-RFA contrast material–enhanced computed tomographic (CT) examination. The study cohort consisted of 162 consecutive patients treated in 188 sessions for 233 CLMs with a median largest tumor diameter of 1.8 cm (range, 0.5–5.7 cm). From this group, data in 93 patients have been previously reported in at least one of three previous smaller studies (10,20,21). We undertook this study to reach enough statistical power (by including more patients and a longer follow-up time) to allow identification of prognostic factors that influence oncologic outcomes after RFA of CLMs.
Our study population reflects two patterns of referral: patients with liver recurrences after hepatectomy (n = 116) (72%) and patients who were considered to have unresectable disease or who refused hepatectomy (n = 46) (28%) and were offered RFA instead. The reasons the latter group did not undergo resection included extrahepatic disease (EHD) (n = 17), comorbidities (n = 15), patient choice (n = 8), disease being deemed technically unresectable at the time of laparotomy (n = 4), and the presence of a small solitary deep tumor requiring major resection (n = 2).
At the time of liver RFA, 51 patients (31%) had radiographic evidence of EHD in the lung only (n = 29), in the lung and lymph nodes (mediastinal, abdominal, both mediastinal and abdominal, or supraclavicular nodes) (n = 14), in the abdominal lymph nodes (n = 5), in the abdominal lymph nodes and a peritoneal nodule (n = 1), in a solitary spine metastasis (n = 1), and in a solitary peritoneal nodule (n = 1). Patients were classified into one of three groups: patients with EHD in the lung only (n = 29), patients with EHD in an other single site (n = 7), and patients with EHD in more than one site (n = 15).
RFA Procedure
All patients signed an informed consent form prior to RFA and were given a prophylactic antibiotic (cefazolin 1 g) intravenously. Ablations were performed with general anesthesia and continuous electrophysiologic monitoring by the anesthesiologist. CT guidance was used for all procedures. CT fluoroscopy and/or ultrasonography were utilized as needed for real-time needle positioning or ablation zone monitoring. In some of the most recent sessions, we used additional positron emission tomography (PET)/CT (n = 8) or split-dose PET/CT (n = 23) guidance, as previously described (22). Briefly, the split-dose technique uses the standard diagnostic PET dose of 12 mCi (444 MBq) of fluorine 18 fluorodeoxyglucose (FDG), divided into 4 mCi (148 MBq) given before the ablation to help guide needle placement and 8 mCi (296 MBq) given immediately after completion of the ablation to help verify adequate tumor coverage and absence of residual activity (22). The ablation electrodes used included LeVeen electrodes (n = 51) (Boston Scientific, Natick, Mass), RITA XL/XLi electrodes (n = 66) (Angiodynamics, Latham, NY), and Valley Laboratory electrodes (n = 116) (Covidien, Mansfield, Mass). All ablations were performed according to the manufacturer’s protocol and with the aim of creating an ablation defect at least 5 mm larger than the largest tumor diameter. Starting in November 2009, an immediate postprocedure triple-phase CT examination was performed for evaluation of the ablation zones and margins. A total of nine interventional radiologists performed the ablations, with the majority of the ablations (70%) being performed by three interventional radiologists (K.T.B., S.B.S., and C.T.S.).
Definitions
We used the guidelines regarding terminology and reporting initially described by Goldberg et al (23) and the recent update by Ahmed et al (24). By referencing the first post-RFA study, technique effectiveness was defined as an ablation defect that completely encompassed the targeted tumor, and failure was defined as any evidence of residual tumor within 1 cm of the ablation defect. The ablation defect at the first 4–8-week post-RFA contrast-enhanced CT examination was considered the new baseline for future comparisons (25). LTP was defined as any new peripheral or nodular enhancement within 1 cm or an enlargement of the baseline ablation defect. LTP-free survival (LTPFS) was defined as the time interval between initial RFA and the first radiographic evidence of LTP. Assisted LTP rate was the rate of LTP after repeated local treatment to control LTP of the same tumor, while the assisted LTPFS was the cumulative time interval from the initial RFA to the latest follow-up, including all additional re-treatments of LTP for the same tumor. OS was defined as the time between initial RFA and the patient’s death of any cause or most recent follow-up.
Imaging Follow-up after RFA
A contrast-enhanced CT examination was performed 4–8 weeks after RFA to evaluate technique effectiveness and to serve as a new baseline for subsequent imaging (25). Imaging follow-up continued at 2–4-month intervals with contrast-enhanced CT examinations. If evaluation was unclear, FDG-PET and/or contrast-enhanced MR imaging was performed. Radiology reports and FDG-PET reports dictated by faculty body imaging radiologists and nuclear medicine radiologists were retrieved from the electronic medical records and were reviewed. Questionable imaging findings were independently reviewed for the purpose of this study by the senior author (C.T.S.) or by a hepatobiliary radiologist. In addition, after November 2009, all imaging studies were prospectively assessed by the operating interventional radiologist (K.T.B., A.M.C., W.A., M.M., L.A.B., S.B.S., and C.T.S.), the senior author (C.T.S.), and a hepatobiliary radiologist. This was performed independently, and any discrepancies were resolved by consensus or by ordering an additional FDG-PET or contrast-enhanced MR imaging examination as needed. By using these assessments, the technique effectiveness and the LTP at subsequent follow-up were recorded.
Technical and Clinical Predictors of Outcomes
Modified CRS for ablation.—Similar to a surgical CRS, the modified CRS for ablation is composed of five clinical factors: nodal status of the primary tumor, disease-free interval from primary tumor resection to CLM detection, number of liver tumors, size of the largest tumor, and carcinoembryonic antigen (CEA) level (3,4). To optimize the surgical CRS for ablation, we made modifications to the cutoff points of tumor size (3 cm instead of 5 cm) and CEA level (30 ng/mL instead of 200 ng/mL) (Table 1). The CEA level modification was to better describe our population, becaues only four patients had levels greater than 200 ng/mL (µg/L) at the time of RFA. The CEA level was available in our records for 153 (94%) of 162 patients, and thus the modified CRS could be calculated only for these patients. Because of the smaller number of patients in our study (162 patients, vs 1001 patients in the surgical article), we grouped each two points of the modified CRS to arrive at three levels of risk: low risk (0 or 1 points), intermediate risk (2 or 3 points), and high risk (4 or 5 points).
Table 1.
Clinical Factors Included in the Surgical and Modified Ablation CRS
Note.—DFI = disease-free interval from primary resection to the diagnosis of liver metastasis.
Minimal ablation margin.—The minimal ablation margin was measured at the 4–8-week post-RFA contrast-enhanced CT examination, as previously described (20). Briefly, by using a picture archiving and communication system workstation, both the pre- and the post-RFA contrast-enhanced CT studies were displayed side by side. Anatomic landmarks around the tumor were chosen, and the distances from the edge of the tumor to the landmarks were measured. Then, the corresponding distances from the edges of the ablation defect to the same landmarks were measured in the post-RFA contrast-enhanced CT study. For each landmark, the pre-RFA distance was subtracted from the post-RFA distance to yield the margin at that site; the smallest value was considered to be the minimal margin and was recorded (Fig 1). This was feasible for 174 (75%) of 233 tumors. The 59 tumors with unreported margins either had margins that were impossible to accurately determine (fused defects) (n = 37) or did not have a baseline or first post-RFA contrast-enhanced CT study (n = 22). The minimal margin size was classified into four groups: 0 mm, 1–5 mm, greater than 5 mm to 10 mm, and more than 10 mm. This assessment was performed by the following physicians: a hepatobiliary radiologist; C.T.S, an interventional radiologist; a research-fellow diagnostic body radiologist; and W.S., a research-fellow physician.
Figure 1a:
Example of ablative margin measurement in a 74-year-old man with a 2.7-cm tumor. (a) Axial portal venous phase CT image shows the tumor before RFA; distances 1, 2, and 3 were measured from the edges of the tumor to the chosen landmarks. (b) Axial portal venous phase CT image obtained 5 weeks after RFA shows the ablation defect; the same distances 1, 2, and 3 were measured from the edge of the ablation defect to the same landmarks. The corresponding distances were subtracted, and the smallest value was chosen as the minimal margin—that is, distance 1 (16.3 mm – 12.5 mm = 3.8 mm). (c) Axial portal venous phase CT image obtained 4.5 months after RFA shows LTP (arrow) at the same site of the insufficient margin (distance 1).
Figure 1b:
Example of ablative margin measurement in a 74-year-old man with a 2.7-cm tumor. (a) Axial portal venous phase CT image shows the tumor before RFA; distances 1, 2, and 3 were measured from the edges of the tumor to the chosen landmarks. (b) Axial portal venous phase CT image obtained 5 weeks after RFA shows the ablation defect; the same distances 1, 2, and 3 were measured from the edge of the ablation defect to the same landmarks. The corresponding distances were subtracted, and the smallest value was chosen as the minimal margin—that is, distance 1 (16.3 mm – 12.5 mm = 3.8 mm). (c) Axial portal venous phase CT image obtained 4.5 months after RFA shows LTP (arrow) at the same site of the insufficient margin (distance 1).
Figure 1c:
Example of ablative margin measurement in a 74-year-old man with a 2.7-cm tumor. (a) Axial portal venous phase CT image shows the tumor before RFA; distances 1, 2, and 3 were measured from the edges of the tumor to the chosen landmarks. (b) Axial portal venous phase CT image obtained 5 weeks after RFA shows the ablation defect; the same distances 1, 2, and 3 were measured from the edge of the ablation defect to the same landmarks. The corresponding distances were subtracted, and the smallest value was chosen as the minimal margin—that is, distance 1 (16.3 mm – 12.5 mm = 3.8 mm). (c) Axial portal venous phase CT image obtained 4.5 months after RFA shows LTP (arrow) at the same site of the insufficient margin (distance 1).
Other clinical and technical predictors.—Additional clinical factors were assessed as predictors of OS and LTPFS (Tables 2 and 3). Tumors were classified according to location into either subcapsular (<10 mm from the liver edge) or nonsubcapsular (≥10 mm from the liver edge); the LTPFS and LTP rates were compared for the two locations. Also, LTP rates for ablations performed before November 2009 were compared with those for ablations performed after November 2009, when immediate assessment of margins with triple-phase CT was initiated.
Table 2.
Results of Univariate Analysis for Predictors of LTPFS
Note.—DFI = disease-free interval from primary resection to the diagnosis of liver metastasis, HAIC = hepatic arterial infusion chemotherapy, NR = not reached.
Table 3.
Results of Univariate Analysis for Predictors of OS
Note.—DFI = disease-free interval from primary resection to the diagnosis of liver metastasis, HAIC = hepatic arterial infusion chemotherapy.
Re-treatment of Liver Recurrences
Patients with multifocal liver and/or extrahepatic progression were not considered candidates for local liver re-treatment, while patients with limited liver progression at the ablation site (LTP) and/or at a new liver focus were considered for re-treatment. Re-treatment was mostly performed percutaneously, although a few patients were treated surgically with resection or RFA. At the end of the study, patients were classified into one of three groups: those with no liver progression after initial ablation, those who were re-treated percutaneously with ablation for their liver progression, and those who were not candidates for liver ablative re-treatment. OS was compared among the groups.
Complications
Complications were classified according to the Society of Interventional Radiology guidelines as minor (having no consequences and requiring no therapy) or major (requiring therapy and hospitalization) (23,24,26).
Statistical Analysis
Survival times were calculated from the time of RFA by using the Kaplan-Meier method. For OS, the log-rank test was used for the univariate analysis, and a Cox regression model was used for the multivariate analysis. The Cox regression model was used to estimate robust standard errors and for univariate and multivariate analysis of LTPFS to account for multiple tumors per patient. Generalized estimating equations were used to compare the mean tumor size and the proportion of tumors with an ablative margin greater than 5 mm between the time periods occurring before and after November 2009. P < .05 was considered to indicate a significant difference. Analysis was performed by M.G. and W.S. using Stata, version 12.0 (Stata, College Station, Tex).
Results
Technique effectiveness was 94% (218 of 233). All failures were related to poor tumor coverage by the ablation zone. These were attributed to suboptimal overlapping ablations (n = 10), poor tumor visualization (n = 4), and proximity to a vessel (the heat-sink phenomenon) (n = 1).
LTP and LTPFS
Over a median follow-up period of 55 months (range, 4–123 months), LTP occurred in 113 (48%) of 233 tumors. Sixty-seven percent (76 of 113) of LTPs were noted within the 1st year after RFA, and 86% (97 of 113) were noted by the end of the 2nd year after RFA. The median LTPFS was 26 months. At univariate analysis, statistically significant predictors of shorter LTPFS were tumor size larger than 3 cm (P < .001) (hazard ratio [HR]: 2.7; 95% confidence interval [CI]: 1.7, 4.3), ablative margin size 5 mm or greater (P < .001) (HR: 10.2; 95% CI: 5.1, 20.1), high modified CRS (P = .009) (HR: 2.9; 95% CI: 1.3, 6.5), no hepatic arterial infusion chemotherapy before RFA (P = .01) (HR: 1.7; 95% CI: 1.1, 2.5), no prior hepatectomy (P = .04) (HR: 1.6; 95% CI: 1.0, 2.4), and male sex (P = .03) (HR: 1.6; 95% CI: 1.0, 3.6) (Table 2) (Fig 2a–2c). The recurrence rates for minimal margin sizes of 0 mm, 1–5 mm, greater than 5 mm to 10 mm, and more than 10 mm were 90% (36 of 40), 60% (39 of 65), 15% (seven of 48), and 5% (one of 21), respectively (P < .001). Ablations performed after November 2009 had a significantly lower recurrence rate of 28% (26 of 92), compared with 62% (87 of 141) for those performed prior to November 2009 (P < .001) (HR: 2.6; 95% CI: 1.6, 4.2) (Fig 2d). Between the two time periods there was no difference in mean tumor size (1.9 vs 2.1 cm) (P = .8) or the percentage of tumors larger than 3 cm (12 [14%] of 80 vs 20 [13%] of 121; P = .9). However, the percentages of ablative margins larger than 5 mm achieved in the group treated after November 2009 were significantly higher (P < .001), at 61% (45 of 74) versus 24% (24 of 100). LTPFS and LTP rates were similar for subcapsular and nonsubcapsular tumors (26 vs 27 months [P = .9] and 48% [64 of 134] vs 49% [49 of 99], respectively). At multivariate analysis, only tumor size larger than 3 cm (P = .01) (HR: 2.0; 95% CI: 1.2, 3.3) and margin size of 5 mm or smaller (P < .001) (HR: 9.9; 95% CI: 4.9, 20) were independent predictors of shorter LTPFS (Table 4).
Figure 2a:
Graphs show Kaplan-Meier curves for LTPFS classified by (a) tumor size, (b) ablation margin size, (c) modified CRS, and (d) time period (before or after November 2009).
Figure 2c:
Graphs show Kaplan-Meier curves for LTPFS classified by (a) tumor size, (b) ablation margin size, (c) modified CRS, and (d) time period (before or after November 2009).
Figure 2d:
Graphs show Kaplan-Meier curves for LTPFS classified by (a) tumor size, (b) ablation margin size, (c) modified CRS, and (d) time period (before or after November 2009).
Table 4.
Results of Multivariate Analysis for Predictors of OS and LTPFS
Figure 2b:
Graphs show Kaplan-Meier curves for LTPFS classified by (a) tumor size, (b) ablation margin size, (c) modified CRS, and (d) time period (before or after November 2009).
Management of LTP
Treatment of LTP was possible for 47 of 113 tumors in 40 of 87 patients with percutaneous RFA (n = 26), percutaneous microwave ablation (n = 9), surgical resection (n = 8), surgical RFA (n = 3), or percutaneous laser ablation (n = 1). This resulted in local control of 24 of 47 tumors, while 23 tumors were not sufficiently controlled because of treatment failure (n = 5) or subsequent LTP (n = 18). Additional re-treatment was performed for 14 of these 23 tumors with percutaneous RFA (n = 8), surgical resection (n = 5), or percutaneous microwave ablation (n = 1). One tumor was re-treated for a third time with open RFA. Accounting for all subsequent re-treatments of LTP, the assisted LTP rate was 34% (80 of 233), and the median assisted LTPFS was 54 months.
During the study period, 46 new liver tumors (away from the ablation site) developed in 25 patients who were candidates for local ablative therapy; these tumors were ablated percutaneously with RFA (n = 31), microwave ablation (n = 9), irreversible electroporation (n = 4), or laser ablation (n = 2).
OS Results
Median OS was 36 months. One-year, 3-year, and 5-year OS rates were 90%, 48%, and 31%, respectively. By the end of the study, patients without any liver progression had the highest median OS, of 65 months. Patients re-treated with percutaneous ablation for LTP, new tumors, or both had a comparable survival of 51 months (P = .2). However, patients who were not candidates for additional ablation had a significantly lower OS (22 months) versus patients with no liver progression (HR: 4.5; 95% CI: 2.5, 8.0) (P < .001) and patients who underwent reablation for liver progression (HR: 2.9; 95% CI: 1.7, 4.9) (P < .001) (Fig 3a). The minority of patients who underwent surgical re-treatment were excluded from this analysis.
Figure 3a:
Graphs show Kaplan-Meier survival curves for OS classified by (a) liver progression status and re-treatment with percutaneous RFA, (b) tumor size, (c) modified CRS (score of 0 or 1 = low risk, score of 2 or 3 = intermediate risk, and score of 4 or 5 = high risk), and (d) EHD site(s).
Figure 3c:
Graphs show Kaplan-Meier survival curves for OS classified by (a) liver progression status and re-treatment with percutaneous RFA, (b) tumor size, (c) modified CRS (score of 0 or 1 = low risk, score of 2 or 3 = intermediate risk, and score of 4 or 5 = high risk), and (d) EHD site(s).
At univariate analysis, tumor size larger than 3 cm (P = .005) (HR: 2.1; 95% CI: 1.2, 3.4), CEA level greater than 30 ng/mL (µg/L) (P = .003) (HR: 2.2; 95% CI: 1.3, 3.6), high modified CRS (P = .02) (HR: 2.7; 95% CI: 1.2, 6.0), and presence of EHD (P < .001) (HR: 2.3; 95% CI: 1.5, 3.5) were associated with shorter OS (Table 3) (Fig 3b–3d). EHD had a significant effect on OS (25 vs 44 months) (P < .001) (HR: 2.3; 95% CI: 1.5, 3.5). Patients with EHD confined to the lungs had a higher median OS (35 months) than those with EHD in more than one site (14 months) (P = .003) (HR: 1.8; 95% CI: 1.2, 2.7). At multivariate analysis, tumor size larger than 3 cm (P = .006) (HR: 2.1; 95% CI: 1.2, 3.4) and more than one site of EHD (P < .001) (HR: 4.6; 95% CI: 2.4, 8.7) were independent predictors of shorter OS (Table 4).
Figure 3b:
Graphs show Kaplan-Meier survival curves for OS classified by (a) liver progression status and re-treatment with percutaneous RFA, (b) tumor size, (c) modified CRS (score of 0 or 1 = low risk, score of 2 or 3 = intermediate risk, and score of 4 or 5 = high risk), and (d) EHD site(s).
Figure 3d:
Graphs show Kaplan-Meier survival curves for OS classified by (a) liver progression status and re-treatment with percutaneous RFA, (b) tumor size, (c) modified CRS (score of 0 or 1 = low risk, score of 2 or 3 = intermediate risk, and score of 4 or 5 = high risk), and (d) EHD site(s).
Complications
Minor and major complication rates were 16% (31 of 188) and 7% (14 of 188), respectively. Major complications included pneumothorax (n = 8), venous thrombosis and pulmonary embolism (n = 2), pleural effusion (n = 1), biloma (n = 1), intrahepatic hematoma (n = 1), and urinary retention with bacteremia in a patient with benign prostatic hyperplasia (n = 1). These were managed by means of chest tube placement, anticoagulation, drainage, embolization (6 weeks after RFA), and urinary catheterization plus intravenous antibiotics, respectively. There were no deaths within 30 days after ablation.
Discussion
Five-year OS rates of 21%–36% have been reported for RFA of unresectable CLMs (15–17). A study in well-selected patients (13) noted a 5-year OS of 48%, and, for the first time, a 10-year OS of 18% was reported. Percutaneous RFA has also been used as a salvage treatment for recurrent CLMs after hepatectomy (9,10). A prior study (10) treated 56 patients with 71 recurrent CLMs after hepatectomy by using a modified test-of-time approach. These patients had technically re-resectable disease but limited liver reserve or were at high risk of subsequent recurrences (12). A median OS of 31 months and a 3-year OS of 41% were reported (10). The test-of-time approach was originally described by Livraghi et al (12) in a study where 88 patients with resectable CLMs were offered percutaneous RFA. They reported that 52 of 53 patients with resectable disease who were treated completely with ablation were spared unnecessary and potentially morbid surgery because they were either free of disease (n = 23) or they developed extensive liver and/or EHD progression (n = 29) during follow-up, rendering their disease unresectable.
In our study, patients without liver recurrences and those re-treated for LTP and/or new tumors had significantly longer OS than patients with recurrences who were no longer candidates for ablative re-treatment. Similarly, Solbiati et al (13) showed that patients without any LTP had the longest OS (63 months), followed by patients who underwent reablation for LTP (46 months). Patients who were not candidates for reablation of LTP had the lowest OS (31 months) (P < .001) (13). These results indicate that repeated ablation to manage subsequent liver recurrences is safe and effective. In experienced hands, morbidity from liver failure and mortality are low after ablation, and ablation has a favorable safety profile when compared with published surgical data (13,14,27).
Among patients with EHD, those with lung-only metastases had the highest OS, approaching that of patients without EHD, while patients with more than one site of EHD had the shortest OS. RFA with curative intent should be reserved for selected patients without EHD and probably for those with lung-only EHD, especially when the lung disease can also be managed with ablative or systemic therapy (15,28). Gillams et al (16) reported a lower OS after liver RFA for CLMs for patients with EHD (14 vs 28 months). Similar to our study, in the study by Gillams and colleagues, patients with lung-only disease lived longer than patients with metastases to other sites (26 vs 12 months). The surgical literature shows similar results. In a multi-institutional study of 171 patients who underwent liver and EHD resection, the highest median OS (46 months) was observed in patients with EHD that was confined to the lungs, while the lowest OSs (15 and 13 months) were observed in patients with multiple EHD sites or aortocaval nodal disease (29).
Adequate safety margins are a critical determinant of oncologic outcomes after resection (4,27,30,31). However, unlike with surgery, actual margins cannot be assessed after RFA because of tissue unavailability for pathologic examination. A less precise alternative is radiologic margin evaluation: A prior dedicated study (20) that evaluated margins after RFA of CLMs revealed that a minimal margin of greater than 5 mm was a critical factor for local tumor control and prolonged LTPFS. Although almost all prior studies of RFA for CLMs did not report margins, we strongly believe that high recurrence rates (particularly in small tumors) are a direct consequence of insufficient ablation margins.
In November 2009, we started to evaluate ablation zones with immediate post-RFA triple-phase CT. This resulted in a significantly lower LTP rate when compared with the prior time period. Thus, we attribute the better local tumor control to the higher percentage of tumors ablated with a margin size of more than 5 mm. This reflects the acquired knowledge about the importance of ablative margins and the experience gained in how to achieve those margins. We therefore strongly encourage immediate assessment of the ablation zone with triple-phase or at least dynamic contrast-enhanced CT.
In 2009, the American Society of Clinical Oncology issued an evidence-based review of the role of RFA in the management of CLMs that highlighted the difficulty in objectively comparing various studies (18). This could be explained by the differences among study populations and inclusion criteria. Use of a prognostic risk score to stratify patients could allow more meaningful comparisons (4,19). In our study, we described and assessed the use of the modified CRS as a prognostic tool for ablation of CLMs. The modified four-point CRS (excluding CEA level) has been previously shown to be a significant predictor of oncologic outcomes after salvage RFA treatment of posthepatectomy recurrences (10). In our study, the modified CRS (including CEA level) was a predictor of OS and LTPFS. Specifically, the risk of death or LTP almost triples for patients and tumors in the high risk group. It appears that a modified CRS can be used as a prognostic tool, although further validation in larger studies is necessary. The application of such a CRS may improve patient selection and help identify patients at high risk who might benefit from stricter follow-up schedules or adjuvant therapies.
This study had several limitations. Its retrospective nature and completion in a single institution makes the universal application of the described modified CRS and findings to different institutions and populations questionable. Randomized controlled studies are desirable to define the role of RFA and to validate factors affecting outcomes. However, such randomization is unlikely in the near future (18). Another limitation was the lack of histopathologic evaluation of the target tumors before ablation, as well as the lack of histopathologic evaluation of the ablation zone and margins for viable tumor cells after ablation. Such examinations can offer important information regarding tumor biology and could help identify patients at risk for progression before or immediately after ablation (21,32–34). Although patients who were candidates for repeated local therapy of liver recurrences were treated mostly with percutaneous RFA, a few patients were treated with other modalities or with surgery. These exceptions were not thought to affect the overall results of the study. Finally, an array of therapies, including systemic chemotherapy, hepatic arterial infusion chemotherapy, and, rarely, yttrium 90 radioembolization, was used to treat multifocal progression in the liver and/or extrahepatic sites as encountered. Thus, the OS observed cannot be attributed solely to RFA.
In conclusion, attaining sufficient ablation margins (>5 mm, and, ideally, >10 mm) can significantly lower LTP rates after percutaneous RFA. A modified CRS for ablation can be used as a prognostic stratification tool. In addition, our data agree with what has been previously shown regarding the poor prognostic effects of large tumor size (>3 cm) and more than one site of EHD. Further investigation is necessary to better understand and optimize the role of ablation in the management of CLMs.
Advances in Knowledge
■ Radiofrequency ablation (RFA) of colorectal cancer liver metastases (CLMs) in patients with lung-only extrahepatic disease (EHD) is associated with better overall survival (OS) than in those with more than one site of EHD (P = .003); the median OS for the two groups was 35 versus 14 months, respectively (hazard ratio [HR]: 1.8; 95% confidence interval [CI]: 1.2, 2.7).
■ Patients without any liver progression after RFA had the highest median OS (65 months).
■ Patients with subsequent limited liver progression were treated with repeat ablation, resulting in a significantly longer median OS (51 months) than patients who were not candidates for local ablative therapy because of multifocal liver and/or extrahepatic progression (22 months) (P < .001) (HR: 2.9; 95% CI: 1.7, 4.9).
■ RFA of CLMs with margins of more than 5 mm to 10 mm or more than 10 mm all around the target tumor decreased local tumor progression rates to 15% and 5%, respectively, over a median follow-up period of 55 months.
Implications for Patient Care
■ In the presence of controllable lung-only EHD, RFA for treatment of CLMs should be considered; alternatively, in patients with more than one site of EHD the median OS can be limited, thus eligibility should be assessed individually, guided by extent of EHD and expected patient survival.
■ Whenever safe, RFA should be performed with the intent of creating a uniform margin of more than 10 mm all around the target tumor.
Acknowledgments
Acknowledgments
We gratefully acknowledge Alessandra Garcia, BA (research study assistant at Memorial Sloan-Kettering Cancer Center), Richard K. Do, MD, PhD (hepatobiliary radiologist at Memorial Sloan-Kettering Cancer Center), and Xiaodong Wang, MD (interventional radiologist at Peking University Cancer Hospital and Institute, Beijing, China) for their contribution to this work.
Received October 29, 2014; revision requested January 5, 2015; revision received April 16; accepted May 14; final version accepted May 21.
Funding: This research was supported by the National Institutes of Health (grant R21 CA131763-01A1).
Disclosures of Conflicts of Interest: W.S. disclosed no relevant relationships. E.N.P. disclosed no relevant relationships. M.G. disclosed no relevant relationships. J.P.E. disclosed no relevant relationships. K.T.B. disclosed no relevant relationships. A.M.C. disclosed no relevant relationships. W.A. disclosed no relevant relationships. J.C.D. disclosed no relevant relationships. M.M. disclosed no relevant relationships. L.A.B. disclosed no relevant relationships. R.H.S. disclosed no relevant relationships. M.I.D. disclosed no relevant relationships. W.R.J. disclosed no relevant relationships. S.B.S. Activities related to the present article: institution has received grants from GE Healthcare and AngioDynamics; is a consultant for GE Healthcare. Activities not related to the present article: none to disclose. Other relationships: none to disclose. N.E.K. Activities related to the present article: none to disclose. Activities not related to the present article: has received a grant from Amgen. Other relationships: none to disclose. C.T.S. disclosed no relevant relationships.
Abbreviations:
- CEA
- carcinoembryonic antigen
- CI
- confidence interval
- CLM
- colorectal cancer liver metastasis
- CRS
- clinical risk score
- EHD
- extrahepatic disease
- FDG
- fluorine 18 fluorodeoxyglucose
- HR
- hazard ratio
- LTP
- local tumor progression
- LTPFS
- LTP-free survival
- OS
- overall survival
- RFA
- radiofrequency ablation
References
- 1.Ruers T, Bleichrodt RP. Treatment of liver metastases, an update on the possibilities and results. Eur J Cancer 2002;38(7):1023–1033. [DOI] [PubMed] [Google Scholar]
- 2.Norstein J, Silen W. Natural history of liver metastases from colorectal carcinoma. J Gastrointest Surg 1997;1(5):398–407. [DOI] [PubMed] [Google Scholar]
- 3.Nordlinger B, Guiguet M, Vaillant JC, et al. Surgical resection of colorectal carcinoma metastases to the liver: a prognostic scoring system to improve case selection, based on 1568 patients. Association Française de Chirurgie. Cancer 1996;77(7):1254–1262. [PubMed] [Google Scholar]
- 4.Fong Y, Fortner J, Sun RL, Brennan MF, Blumgart LH. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases. Ann Surg 1999;230(3):309–318; discussion 318–321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Abdalla EK, Vauthey JN, Ellis LM, et al. Recurrence and outcomes following hepatic resection, radiofrequency ablation, and combined resection/ablation for colorectal liver metastases. Ann Surg 2004;239(6):818–825; discussion 825–827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bengmark S, Hafström L. The natural history of primary and secondary malignant tumors of the liver. I. The prognosis for patients with hepatic metastases from colonic and rectal carcinoma by laparotomy. Cancer 1969;23(1):198–202. [DOI] [PubMed] [Google Scholar]
- 7.Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350(23):2335–2342. [DOI] [PubMed] [Google Scholar]
- 8.Van Cutsem E, Köhne CH, Hitre E, et al. Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. N Engl J Med 2009;360(14):1408–1417. [DOI] [PubMed] [Google Scholar]
- 9.Elias D, De Baere T, Smayra T, Ouellet JF, Roche A, Lasser P. Percutaneous radiofrequency thermoablation as an alternative to surgery for treatment of liver tumour recurrence after hepatectomy. Br J Surg 2002;89(6):752–756. [DOI] [PubMed] [Google Scholar]
- 10.Sofocleous CT, Petre EN, Gonen M, et al. CT-guided radiofrequency ablation as a salvage treatment of colorectal cancer hepatic metastases developing after hepatectomy. J Vasc Interv Radiol 2011;22(6):755–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Benson AB, 3rd, Bekaii-Saab T, Chan E, et al. Metastatic colon cancer, version 3.2013: featured updates to the NCCN Guidelines. J Natl Compr Canc Netw 2013;11(2):141–152; quiz 152. [DOI] [PubMed] [Google Scholar]
- 12.Livraghi T, Solbiati L, Meloni F, Ierace T, Goldberg SN, Gazelle GS. Percutaneous radiofrequency ablation of liver metastases in potential candidates for resection: the “test-of-time approach”. Cancer 2003;97(12):3027–3035. [DOI] [PubMed] [Google Scholar]
- 13.Solbiati L, Ahmed M, Cova L, Ierace T, Brioschi M, Goldberg SN. Small liver colorectal metastases treated with percutaneous radiofrequency ablation: local response rate and long-term survival with up to 10-year follow-up. Radiology 2012;265(3):958–968. [DOI] [PubMed] [Google Scholar]
- 14.Livraghi T, Solbiati L, Meloni MF, Gazelle GS, Halpern EF, Goldberg SN. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications encountered in a multicenter study. Radiology 2003;226(2):441–451. [DOI] [PubMed] [Google Scholar]
- 15.Hamada A, Yamakado K, Nakatsuka A, et al. Radiofrequency ablation for colorectal liver metastases: prognostic factors in non-surgical candidates. Jpn J Radiol 2012;30(7):567–574. [DOI] [PubMed] [Google Scholar]
- 16.Gillams AR, Lees WR. Five-year survival in 309 patients with colorectal liver metastases treated with radiofrequency ablation. Eur Radiol 2009;19(5):1206–1213. [DOI] [PubMed] [Google Scholar]
- 17.Van Tilborg AA, Meijerink MR, Sietses C, et al. Long-term results of radiofrequency ablation for unresectable colorectal liver metastases: a potentially curative intervention. Br J Radiol 2011;84(1002):556–565. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wong SL, Mangu PB, Choti MA, et al. American Society of Clinical Oncology 2009 clinical evidence review on radiofrequency ablation of hepatic metastases from colorectal cancer. J Clin Oncol 2010;28(3):493–508. [DOI] [PubMed] [Google Scholar]
- 19.Guenette JP, Dupuy DE. Radiofrequency ablation of colorectal hepatic metastases. J Surg Oncol 2010;102(8):978–987. [DOI] [PubMed] [Google Scholar]
- 20.Wang X, Sofocleous CT, Erinjeri JP, et al. Margin size is an independent predictor of local tumor progression after ablation of colon cancer liver metastases. Cardiovasc Intervent Radiol 2013;36(1):166–175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sofocleous CT, Garg S, Petrovic LM, et al. Ki-67 is a prognostic biomarker of survival after radiofrequency ablation of liver malignancies. Ann Surg Oncol 2012;19(13):4262–4269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ryan ER, Sofocleous CT, Schöder H, et al. Split-dose technique for FDG PET/CT-guided percutaneous ablation: a method to facilitate lesion targeting and to provide immediate assessment of treatment effectiveness. Radiology 2013;268(1):288–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Goldberg SN, Grassi CJ, Cardella JF, et al. Image-guided tumor ablation: standardization of terminology and reporting criteria. Radiology 2005;235(3):728–739. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ahmed M, Solbiati L, Brace CL, et al. Image-guided tumor ablation: standardization of terminology and reporting criteria—a 10-year update. Radiology 2014;273(1):241–260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Pua BB, Sofocleous CT. Imaging to optimize liver tumor ablation. Imaging Med 2010;2(4):433–443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Burke DR, Lewis CA, Cardella JF, et al. Quality improvement guidelines for percutaneous transhepatic cholangiography and biliary drainage. J Vasc Interv Radiol 2003;14(9 Pt 2):S243–S246. [PubMed] [Google Scholar]
- 27.Ito H, Are C, Gonen M, et al. Effect of postoperative morbidity on long-term survival after hepatic resection for metastatic colorectal cancer. Ann Surg 2008;247(6):994–1002. [DOI] [PubMed] [Google Scholar]
- 28.Petre EN, Jia X, Thornton RH, et al. Treatment of pulmonary colorectal metastases by radiofrequency ablation. Clin Colorectal Cancer 2013;12(1):37–44. [DOI] [PubMed] [Google Scholar]
- 29.Pulitanò C, Bodingbauer M, Aldrighetti L, et al. Liver resection for colorectal metastases in presence of extrahepatic disease: results from an international multi-institutional analysis. Ann Surg Oncol 2011;18(5):1380–1388. [DOI] [PubMed] [Google Scholar]
- 30.Nuzzo G, Giuliante F, Ardito F, et al. Influence of surgical margin on type of recurrence after liver resection for colorectal metastases: a single-center experience. Surgery 2008;143(3):384–393. [DOI] [PubMed] [Google Scholar]
- 31.Tomlinson JS, Jarnagin WR, DeMatteo RP, et al. Actual 10-year survival after resection of colorectal liver metastases defines cure. J Clin Oncol 2007;25(29):4575–4580. [DOI] [PubMed] [Google Scholar]
- 32.Sofocleous CT, Klein KM, Hubbi B, et al. Histopathologic evaluation of tissue extracted on the radiofrequency probe after ablation of liver tumors: preliminary findings. AJR Am J Roentgenol 2004;183(1):209–213. [DOI] [PubMed] [Google Scholar]
- 33.Snoeren N, Huiskens J, Rijken AM, et al. Viable tumor tissue adherent to needle applicators after local ablation: a risk factor for local tumor progression. Ann Surg Oncol 2011;18(13):3702–3710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sofocleous CT, Nascimento RG, Petrovic LM, et al. Histopathologic and immunohistochemical features of tissue adherent to multitined electrodes after RF ablation of liver malignancies can help predict local tumor progression: initial results. Radiology 2008;249(1):364–374. [DOI] [PubMed] [Google Scholar]