This study aimed to evaluate local tumor progression-free survival (LTPFS) and overall survival (OS) after percutaneous radiofrequency ablation (RFA) for solitary colorectal liver metastases (CLM) <3 cm and to identify the risk factors associated with poor LTPFS and OS after percutaneous RFA.
This study screened 219 patients who underwent percutaneous RFA for CLM between January 2013 and November 2020. Of these, 92 patients with a single CLM <3 cm were included. LTPFS and OS were calculated using the Kaplan-Meier method, and the differences between curves were compared using the log-rank test. Risk factors for LTPFS and OS were assessed using Cox proportional-hazard regression models.
Technical efficacy was achieved in the first (n=91) or second (n=1) RFA sessions. During the follow-up (median, 20.0 months), cumulative LTPFS rates at 1, 3, and 5 years were 92.4%, 83.4%, and 76.5%, respectively. During the follow-up (median, 27.8 months), the corresponding OS rates were 97.5%, 81.3%, and 74.8%, respectively. In multivariable Cox regression analyses, the group with both tumor-puncturing RFA and a T4 stage primary tumor (hazard ratio, 3.3; 95% confidence interval, 1.1 to 10.2; P=0.037) had poor LTPFS. In the univariable analysis, no factors were significantly associated with poor OS.
Both LTPFS and OS were promising after percutaneous RFA for a single CLM <3 cm. The group with both tumor-puncturing RFA and a T4 stage primary tumor showed poor LTPFS. No risk factors were identified for poor OS.
The group with both tumor-puncturing radiofrequency ablation (RFA) and T4 stage primary tumor was associated with poor local tumor progression-free survival. No local tumor progression was found in our study even in patients with T4 staging when no-touch RFA was performed.
Radiofrequency ablation (RFA) has been used to manage metastatic colorectal cancer (CRC) in the liver of patients ineligible for hepatic resection. However, one of the drawbacks of RFA is frequent local tumor progression (LTP) compared to hepatic resection. Tumor size, ablative margin, and perivascular or subcapsular tumor location have been associated with LTP after RFA [
Meanwhile, tumors up to 3 cm in diameter are generally considered good candidates for RFA of colorectal liver metastases (CLM) [
Furthermore, it is expected that even perivascular tumors, which are prone to the so-called "heat-sink effect," may be overcome by these advanced RFA techniques as multiple electrodes can be placed outside the tumor near the peritumoral vessels. In addition, more accurate guidance methods, such as fusion imaging of real-time ultrasonography (US) and pre-acquired computed tomography (CT), magnetic resonance imaging (MRI), or contrast-enhanced US, are now available [
Therefore, this study aimed to evaluate LTP-free survival (LTPFS) and overall survival (OS) after percutaneous RFA for solitary CLM <3 cm at a tertiary cancer center. The risk factors associated with poor LTPFS and OS after percutaneous RFA were also assessed.
The Institutional Review Board of Samsung Medical Center (2021-11-100) of the authors’ medical center approved this retrospective study, and the requirement for written informed consent from patients was waived.
In total, 219 patients who underwent RFA for hepatic metastasis between January 2013 and November 2020 were screened retrospectively. Among them, 127 patients were excluded for the following reasons: (1) multiple metastases (n=119), (2) metastasis from cancers other than CRC (n=5), (3) tumor size >3 cm (n=1), (4) patients who underwent laparoscopic RFA (n=1), and (5) absence of electronic medical records (n=1). Finally, 92 patients who underwent US-guided RFA for a single CLM <3 cm were included. Hepatic metastasis was diagnosed based on percutaneous biopsy (n=2) or typical imaging findings of CT or MRI (n=90) (
All RFA procedures were performed percutaneously by five radiologists who had at least 3 years of experience with RFA. The RFA procedures were performed under conscious sedation or monitored anesthesia, which was conducted under the care of anesthesiologists. Fusion imaging (volume navigation; LOGIQ E9, GE Healthcare, Chicago, IL, USA) of real-time US and CT/MRI scans were used to guide and monitor the RFA procedures [
The type and number of radiofrequency electrodes were chosen based on the tumor size and shape. In the earlier period of this study, conventional tumor-puncturing RFA was performed using a single electrode, especially for small tumors <2 cm. When using multiple electrodes (Octopus electrode, STARmed; Twins electrodes, RF Medical), centripetal RFA was performed after placing electrodes at the peripheral portion of the tumor. Gradually, no-touch RFA also started to be used as it can provide a sufficient ablative margin, leading to effective local tumor control [
The treatment response and complications were evaluated using contrast-enhanced CT within 12 hours after the procedure. The treatment was considered a technical success if the RFA zone entirely covered the index tumor on CT [
A major complication was defined as an event that led to substantial morbidity and disability that increased the level of care, resulting in hospital admission or a substantially lengthened hospital stay. All other complications were considered minor [
A perivascular tumor was defined as a tumor abutting a portal or hepatic vein branch with a lumen caliber ≥3 mm [
Follow-up contrast-enhanced CT/MRI was performed 1 month after the initial treatment and every 3 to 6 months thereafter. Chest CT, serum carcinoembryonic antigen, carbohydrate antigen 19-9, and liver function tests were also performed. LTP was defined as the appearance of tumor foci at the edge of the ablation zone [
The size of the ablation zone was measured on immediate post-RFA CT. The maximum (Dmax) and minimum (Dmin) diameters of the ablation zone were measured from axial CT images of the portal phase and the longest vertical diameter (Dv) from sagittal or coronal images. The ablation volume was calculated by the formula for ellipsoid lesion (ablation volume=π(Dmax×Dmin×Dv)/6) [
Therapeutic outcomes, including LTPFS and OS, were assessed. LTPFS was defined as the time interval between the initial RFA and the first radiographic evidence of LTP [
Patient demographics and clinical characteristics were analyzed using descriptive statistics. They were presented as numbers with percentages for categorical variables and medians with ranges or means with standard deviations for continuous variables. The ratios of the longest diameter of the ablation zone and tumor size and those of the ablation volume and tumor volume were compared between the no-touch RFA and tumor-puncturing RFA groups using the Wilcoxon rank-sum test. Survival curves were generated using the Kaplan-Meier method, and differences between curves were compared using the log-rank test.
Univariable and multivariable analyses were performed to identify risk factors for LTPFS or OS using Cox proportional-hazard regression models. The risk factors for LTPFS were assessed using the variables with patient-, tumor-, and RFA-related factors. The interaction effect between the T stage and the method of electrode placement was assessed through a
The risk factors for OS were also assessed using a Cox regression analysis with the same variables as for LTPFS. The interaction effect between the T stage and the method of electrode placement was also assessed with
The baseline characteristics of the patients are summarized in
Technical success was achieved in 91 of 92 patients (98.9%). A residual unablated tumor was found in one patient on immediate post-RFA CT images, and a second RFA session was performed, which led to successful local tumor control. Technical efficacy was achieved in all 92 patients (100%) in the 1-month follow-up CT images.
Major complications were found in three patients after RFA. These included biliary sepsis (n=1), subcapsular hematoma (n=1), and large biloma (n=1). All three patients recovered fully after conservative treatment. Minor complications were found in four patients: biloma (n=1), subsegmental infarction (n=2), and diaphragmatic thermal injury (n=1).
The median longest diameter of the ablation zone was 4.6 cm (range, 3.1 to 6.3 cm) in the no-touch RFA group and 3.9 cm (range, 1.9 to 7.3 cm) in the tumor-puncturing RFA group. The median volume of the ablation zone was 24.5 cm3 (range, 9.5 to 58.1 cm3) in the no-touch RFA group and 16.3 cm3 (range, 2.7 to 131.7 cm3) in the tumor-puncturing RFA group (P=0.016). The ratio of the longest diameter and the volume between the ablation zone and those of the tumors between the no-touch RFA and tumor-puncturing RFA groups are summarized in
The median follow-up period was 20.0 months (range, 1.8 to 90.3 months). During follow-up, LTP was observed in 12 of 92 patients (13.0%). The cumulative LTPFS rates at 1, 3, and 5 years were 92.4%, 83.4%, and 76.5%, respectively.
The univariable Cox proportional hazard regression model revealed that the T stage of the primary tumor was a risk factor for LTPFS (P=0.048). The method of electrode placement (P=0.071) and patient age (P=0.066) showed borderline significance as risk factors for LTPFS (
The median follow-up period was 27.8 months (range, 2.3 to 90.8 months). During follow-up, 13 of the 92 patients (14.1%) died. The cumulative OS rates at 1, 3, and 5 years were 97.5%, 81.3%, and 74.8%, respectively.
No factors were significantly associated with poor OS based on the univariable Cox proportional hazard regression model (
This study analyzed the treatment outcomes of percutaneous RFA of a single CLM <3 cm in size. The results showed an interaction effect of the method of electrode placement and the T stage of primary CRC on LTPFS. The group with both tumor-puncturing RFA and a T4 stage tumor had poorer LTPFS than those with no-touch RFA and T1-T3 stages. The multivariable analysis also showed that the combination of tumor-puncturing RFA and a T4 stage tumor was an independent risk factor for poor LTPFS. However, no interaction effect of the variables mentioned above was found on OS, and there were no significant risk factors associated with poor OS.
A sufficient ablative margin has been reported to be critical for preventing LTP after RFA [
During no-touch RFA, multiple electrodes are inserted outside the tumor boundary, and centripetal ablation of the tumor is feasible. This type of energy deposition can help obtain a sufficient ablation margin and obliterate feeding and draining vessels in the earlier period of the RFA process. Furthermore, as the electrodes are not in direct contact with the tumor, they may prevent tumor spread around the tumor, and there is no risk of tract seeding. In a recent multicenter prospective study [
An interaction effect was found between electrode placement and the T stage of primary CRC on LTPFS. The group with both tumor-puncturing RFA and a T4 stage tumor showed poor LTPFS (hazard ratio, 10.3; 95% CI, 1.2 to 1,345.5; P=0.029) compared to the reference group of no-touch RFA and T1-T3 stages of primary CRC. In addition, in patients undergoing RFA for CLM, the combination of tumor-puncturing RFA and a T4 stage tumor was an independent risk factor for poor LTPFS. Although the association between the T stage of primary CRC and local tumor control of RFA for CLM is not clearly understood, an advanced T stage of primary CRC is known to be associated with poorly differentiated tumors [
Tumor size and a perivascular or subcapsular location of the CLM are well-known factors affecting local tumor control [
It is known that large tumor size (>3 cm) and more than one site of extrahepatic disease are poor prognostic factors for OS [
The present study had several limitations. First, it was a retrospective study performed at a single medical center. Therefore, selection bias may have occurred. Second, although the ablative margin is known to be a factor associated with LTP, it was not assessed because there is no standardized method for evaluating the ablative margin, and an accurate assessment of the ablative margin is sometimes challenging when the index tumor is not demarcated within the ablation zone. However, the effect of a sufficient ablative margin could be indirectly inferred from the difference in ablation zone size between no-touch RFA and tumor-puncturing RFA. This assumption is supported by a randomized controlled trial between no-touch RFA and conventional tumor-puncturing RFA, in which no-touch RFA provided a larger ablative margin than conventional tumor-puncturing RFA [
In conclusion, both LTPFS and OS were promising after percutaneous RFA of a single CLM. LTP was not observed in patients who underwent no-touch RFA. The group with both tumor-puncturing RFA and a T4 stage tumor showed poor LTPFS. No risk factors were identified for poor OS.
Conceptualization: Lee HJ, Lee MW, Kang TW. Data acquisition: Lee MW, Rhim H. Data analysis or interpretation: Lee HJ, Lee MW, Ahn SH, Cha DI, Ko SE, Song KD. Drafting of the manuscript: Lee HJ, Lee MW. Critical revision of the manuscript: Lee HJ, Lee MW, Ahn SH, Cha DI, Ko SE, Kang TW, Song KD, Rhim H. Approval of the final version of the manuscript: all authors.
No potential conflict of interest relevant to this article was reported.
RFA, radiofrequency ablation; US, ultrasonography.
A. Magnetic resonance image of the axial hepatobiliary phase shows a 1.0-cm hepatic metastasis (arrow) in the liver segment 6 subcapsular area. B. Fusion imaging of real-time ultrasonography and pre-acquired magnetic resonance imaging was used to localize the tumor (arrow). C. Ultrasound image shows two parallel radiofrequency electrodes placed outside the tumor (arrow) using the no-touch radiofrequency ablation technique. D. Immediate axial computed tomography image after radiofrequency ablation shows that the tumor (arrow) is completely covered with a sufficient ablative margin. No local tumor progression was noted during the 11.8-month follow-up period.
A. Axial hepatobiliary phase magnetic resonance image shows a 1.3-cm hepatic metastasis (arrow) in the liver segment 4. B. Fusion imaging of real-time ultrasound and pre-acquired magnetic resonance imaging was used to localize the tumor (arrow). C. Ultrasound image shows radiofrequency electrodes placed through the tumor (arrow), using the tumor-puncturing technique. D. Immediate axial computed tomography image after radiofrequency ablation shows that the tumor (arrow) is completely ablated with sufficient margins. E. Axial arterial phase magnetic resonance image acquired 4.7 months after radiofrequency ablation shows a 1.8-cm peripheral rim enhancing lesion (arrow) adjacent to the ablative margin, suggesting local tumor progression.
A. LTPFS tended to be higher with no-touch radiofrequency ablation (RFA) than with tumor-puncturing RFA. However, this difference was not statistically significant. B. LTPFS was significantly different between the T1-T3 and T4 stages of primary colorectal cancer. C. LTPFS was significantly different between the combinations of the method of electrode placement and the T stage of primary colorectal cancer.
A. OS was not significantly different between the tumor-puncturing and no-touch radiofrequency ablation groups. B. OS tended to be higher in T1-T3 than in T4. However, this difference was not statistically significant. C. OS tended to be different according to the combinations of the method of electrode placement and T staging of primary colorectal cancer. However, this difference was not statistically significant.
Demographic and clinical characteristics of the study patients
Variable | Total (n=92) |
---|---|
Age (year) | 60.2±11.7 |
Sex | |
Male | 63 (68.5) |
Female | 29 (31.5) |
Disease-free interval (month) | |
≤12 | 70 (76.1) |
>12 | 22 (23.9) |
Method of electrode placement | |
No-touch | 21 (22.8) |
Tumor-puncturing | 71 (77.2) |
No. of electrodes used | 2 (1‒3) |
No. of overlapping ablations | 3 (1‒10) |
Ablation time (min) | 12 (3‒26) |
Energy (kcal) | 9.6 (1‒57) |
Tumor size (cm) | 1.3 (0.3‒2.8) |
Tumor location (Couinaud segment) | |
Segment I | 4 (4.3) |
Segments II, III, and IV | 28 (30.4) |
Segments V, VI, VII, and VIII | 60 (65.2) |
Perivascular location | 30 (32.6) |
Subcapsular location | 34 (37.0) |
Chemotherapy history | 44 (47.8) |
Resection history | 11 (12.0) |
Synchronicity | 41 (44.6) |
Presence of extrahepatic metastasis | 3 (3.3) |
T staging | |
Tx | 2 (2.2) |
T1‒T3 | 65 (70.7) |
T4 | 25 (27.2) |
N staging | |
Nx | 3 (3.3) |
N0 | 29 (31.5) |
N1‒N2 | 60 (65.2) |
CEA (ng/mL) | 3.3 (0.5‒594.9) |
CA 19-9 (U/mL) | 14.0 (1.2‒554.74) |
Values are presented as mean±standard deviation, number (%), or median (range).
CEA, carcinoembryonic antigen; CA 19-9, carbohydrate antigen 19-9.
Ratio of longest diameter and volume between the ablation zone and tumor
Variable | No-touch RFA (n=21) | Tumor-puncturing RFA (n=71) | P-value |
---|---|---|---|
Ratio of longest diameter (ablation zone/tumor) | 3.5 (2.3‒5.6) | 3.0 (1.5‒12.5) | 0.035 |
Ratio of volume (ablation zone/tumor) | 32.1 (8.4‒190.5) | 20.6 (4.0‒311.7) | 0.114 |
Values are presented as median (range).
RFA, radiofrequency ablation.
Univariable and multivariable analyses of the risk factors for LTPFS
Variable | Hazard ratio (95% CI) | P-value |
---|---|---|
Univariable analysis | ||
Age | 1.05 (1.00‒1.11) | 0.066 |
Sex [ref: male] | 1.38 (0.44‒4.36) | 0.579 |
Disease-free interval [ref: ≤12 months] | 0.57 (0.12‒2.59) | 0.463 |
Method of electrode placement | 0.15 (0.00‒1.13) | 0.071 |
[ref: tumor-puncturing] | ||
No. of electrodes used | 1.20 (0.54‒2.64) | 0.656 |
No. of overlapping ablations | 1.07 (0.74‒1.54) | 0.736 |
Ablation time | 0.97 (0.85‒1.10) | 0.622 |
Energy | 0.92 (0.82‒1.04) | 0.208 |
CEA | 1.00 (0.98‒1.02) | 0.761 |
CA 19-9 | 0.99 (0.97‒1.02) | 0.567 |
Tumor size [ref: <2 cm] | 0.41 (0.00‒3.11) | 0.471 |
Tumor location (Couinaud segment) | ||
Right lobe [ref: caudate lobe] | 1.26 (0.06‒27.20) | 0.884 |
Left lobe [ref: caudate lobe] | 2.55 (0.12‒55.34) | 0.550 |
Perivascular location | 1.21 (0.38‒3.83) | 0.741 |
Subcapsular location | 0.54 (0.15‒1.99) | 0.352 |
Chemotherapy history | 0.75 (0.22‒2.50) | 0.637 |
Resection history | 2.03 (0.44‒9.34) | 0.366 |
Synchronicity [ref: synchronous] | 1.99 (0.60‒6.62) | 0.264 |
Presence of extrahepatic metastasis | 1.03 (0.01‒7.99) | 0.984 |
T stage [ref: T1‒T3] | 3.15 (1.01‒9.81) | 0.048 |
N staging [ref: N0] | 1.57 (0.43‒5.81) | 0.497 |
Multivariable analysis | ||
Age | 1.05 (1.00‒1.10) | 0.066 |
Tumor-puncturing and T1‒T3 [ref: no-touch and T1‒T3] | 0.33 (0.00‒2.77) | 0.372 |
No-touch and T4 [ref: no-touch and T1‒T3] | 0.54 (0.00‒4.67) | 0.652 |
Tumor-puncturing and T4 [ref: no-touch and T1‒T3] | 3.31 (1.08‒10.17) | 0.037 |
Hazard ratios and P-values were obtained using Cox regression analysis. The reference category for each categorical variable is indicated by square brackets in the first column.
LTPFS, local tumor progression-free survival; CI, confidence interval; CEA, carcinoembryonic antigen; CA 19-9, carbohydrate antigen 19-9.
Interaction effect between the no-touch versus tumor-puncturing technique and the T stage
Group | Hazard ratio (95% CI) | P-value |
---|---|---|
LTPFS | ||
No-touch and T1‒T3 | 1 [reference] | |
Tumor-puncturing and T1‒T3 | 2.90 (0.34‒377.89) | 0.398 |
No-touch and T4 | 2.03 (0.01‒374.76) | 0.727 |
Tumor-puncturing and T4 | 10.31 (1.22‒1,345.54) | 0.029 |
OS | ||
No-touch and T1‒T3 | 1 [reference] | |
Tumor-puncturing and T1‒T3 | 2.07 (0.01‒383.63) | 0.719 |
No-touch and T4 | 2.58 (0.3‒335.63) | 0.455 |
Tumor-puncturing and T4 | 7.30 (0.86‒952.36) | 0.073 |
CI, confidence interval; LTPFS, local tumor progression-free survival; OS, overall survival.
Univariable analyses of risk factors for OS
Variable | Hazard ratio (95% CI) | P-value |
---|---|---|
Age | 1.02 (0.97‒1.07) | 0.490 |
Sex [ref: male] | 1.61 (0.54‒4.79) | 0.395 |
Disease-free interval [ref: ≤12 months] | 0.51 (0.11‒2.29) | 0.376 |
Method of electrode placement [ref: tumor-puncturing] | 0.19 (0.00‒1.40) | 0.123 |
No. of electrodes used | 1.52 (0.72‒3.20) | 0.271 |
No. of overlapping ablations | 1.19 (0.85‒1.68) | 0.315 |
Ablation time | 1.09 (0.97‒1.22) | 0.135 |
Energy | 1.04 (0.97‒1.13) | 0.251 |
CEA | 0.98 (0.90‒1.06) | 0.587 |
CA 19-9 | 1.00 (0.99‒1.01) | 0.899 |
Tumor size [ref: <2 cm] | 0.94 (0.12‒7.29) | 0.951 |
Tumor location (Couinaud segment) | ||
Right lobe [ref: caudate lobe] | 0.48 (0.10‒4.62) | 0.458 |
Left lobe [ref: caudate lobe] | 0.70 (0.14‒6.89) | 0.711 |
Perivascular location | 1.24 (0.40‒3.80) | 0.708 |
Subcapsular location | 1.00 (0.33‒3.05) | 0.997 |
Chemotherapy history | 1.39 (0.47‒4.17) | 0.551 |
Resection history | 1.66 (0.37‒7.50) | 0.511 |
Synchronicity [ref: synchronous] | 1.47 (0.48‒4.49) | 0.502 |
Presence of extrahepatic metastasis | 1.18 (0.01‒8.93) | 0.910 |
T staging [ref: T1‒T3] | 2.62 (0.87‒7.85) | 0.086 |
N staging [ref: N0] | 2.80 (0.62‒12.66) | 0.180 |
Hazard ratios and P-values were obtained using Cox regression analysis. The reference category for each categorical variable is indicated by square brackets in the first column.
OS, overall survival; CI, confidence interval; CEA, carcinoembryonic antigen; CA 19-9, carbohydrate antigen 19-9.