Key words
microwave ablation - radiofrequency (RF) ablation - liver metastases - image guidance
Introduction
Percutaneous ablation, e. g., microwave (MWA), radiofrequency ablation (RFA), or irreversible
electroporation (IRE), is a well-established option for the treatment of selected
hepatic tumors [1]
[2]
[3]
[4]
[5]
[6]. Successful ablation needs to cover the whole lesion including a sufficient safety
margin of at least 5–10 mm beyond the tumor margin [7]
[8]
[9], while critical anatomic structures need to be spared. High-quality imaging is an
important prerequisite to determine the ablation method of choice [10]. However, some tumors might be challenging to treat because they are associated
with transient enhancement on multiphase CT, or are only visible on MRI, or even only
on specific MRI sequences such as diffusion-weighted imaging, or hepatobiliary-phase
imaging after liver-specific contrast.
On the other hand, image guidance during ablation is typically achieved by means of
unenhanced CT and/or ultrasound. Different methods are used to improve the detectability
of the target lesion during CT-guided ablation in order to overcome this shortcoming
[11]. One approach to directly improve lesion visibility is to inject small amounts of
contrast agent through an intra-arterial catheter, placed in the hepatic artery [12] – a method that increases the invasiveness of the procedure, requires access to
DSA to place the catheter, and works only for lesions with a strong arterial blood
supply. Indirect approaches use image co-registration of diagnostic CECT or MR imaging
with the corresponding intraprocedural imaging [13]
[14]
[15]. However, co-registration of hepatic lesions can be error-prone.
Another, simple method is to use surrounding anatomical landmarks that can usually
be defined on unenhanced CT, like portal or liver vein branches, or distance to liver
capsule or other anatomical structures [16]. However, such procedures can be challenging since not all landmarks are visible
on unenhanced CT. Image quality is often worse on intra-procedural control scans due
to beam-hardening artifacts. Moreover, gross patient motion, patient breathing, and
bowel motion, as well as deformation of the liver secondary to the penetration of
the probe can all contribute to difficulties when using this approach.
Accordingly, the aim of this study was to investigate whether the detectability of
the target lesions in unenhanced CT would affect the outcome of liver ablation procedures
done on the basis of unenhanced CT, using the anatomical landmark approach to identify
target lesions.
Materials and Methods
This retrospective study was in accordance with the ethical standards of the institutional
research committee and with the 1964 Helsinki declaration and its later amendments.
Patients
We included all patients who underwent CT-guided hepatic ablation procedures (RFA,
MWA, or IRE) in our department between 06/2015 and 06/2018. The decision to perform
percutaneous ablation was established in consensus in a multidisciplinary tumor board,
attended by hepatobiliary surgeons, oncologists, radiotherapists, pathologists, radiologists,
and interventional radiologists. Hepatic tumor ablation was recommended in patients,
who were considered illegible for surgical resection and had no extrahepatic tumor
burden.
All patients had undergone either contrast-enhanced CT or MRI prior to ablation.
Patients, who had less than four weeks of follow-up, were excluded from the study.
Ablation
All procedures were performed by experienced interventional radiologists with more
than 5 years of experience in abdominal interventions.
As per the standard protocol at our institution, all procedures were performed under
general anesthesia. The individual approach and patient positioning during the ablation
were determined by the interventionalist. For planning applicator placement, unenhanced
CT was performed and compared to previous images using the anatomical landmark method
([Fig. 1]). Lesions were defined as visible lesions (DLs) or non-visible lesions (NDLs) if
the lesions could or could not be detected by an experienced interventional radiologist.
The applicator was positioned using intermittent CT scans in expiratory breath hold.
Fig. 1 Using anatomical landmarks for needle placement. A diagnostic, diffusion-weighted MR image. B unenhanced CT image with surface marker grid for planning needle placement (the known
lesion is not visible, but the adjacent portal vein branches are well depictable).
C RFA applicator in the right position.
Abb. 1 Nutzung anatomischer Landmarken für Nadelpositionierung.
If needed, overlapping ablation after repositioning of the electrodes was performed
until the interventional radiologist felt that tumor ablation should be complete including
a safety margin of at least 5–10 mm. The safety margin and the size of the ablation
zone were planned according to the size and location of the lesion, independent of
its detectability.
Before removing the electrode, a contrast-enhanced tri-phasic CT scan was obtained
to verify treatment success and assess immediate complications. Track ablation was
performed in all RFA and MWA procedures.
For RFA we used a monopolar system (RF 3000, Boston Scientific Corp, MA, USA) with
umbrella-shaped applicators with a diameter of 2–4 cm and varying shaft lengths (LeVeen,
Boston Scientific Corp, MA, USA). Ablation is performed according to the vendor’s
ablation protocols.
MWA was performed using the Emprint system (Medtronic, MI, USA) with dedicated antennae
with different shaft lengths. We always adhered to the protocols provided by the vendor.
We used the NanoKnife® system (AngioDynamics, Amsterdam, the Netherlands) for IRE with 2 to 5 probes with
a tip exposure of 1.5 cm. IRE was performed with 70–90 pulses per probe pair, a pulse
length of 90 µs, and a maximum voltage of 3000 V.
Evaluation of ablation and follow-up
The size of the ablation zone was measured in the immediate postinterventional contrast-enhanced
CT examination, calculating the average diameter ((short + long axis)/2).
Adverse events were assessed according to the CIRSE classification system [17].
Routine follow-up imaging was performed 1 month after ablation and every 3 months
thereafter and consisted of either contrast-enhanced multiphasic CT or contrast-enhanced
MRI. Technical efficacy was determined on the basis of the 1-month CT/MRI examination
according to the criteria defined by the international working group on image-guided
tumor ablation [8].
Incomplete ablation (IA) was defined as the persistence of enhancing areas within
the ablation zone or peripheral nodular enhancement within a 10-mm margin at the first
follow-up scan four weeks after the procedure.
Local tumor progression (LTP) was defined as recurrent disease within a 10-mm margin
during subsequent follow-up.
Intrahepatic progression-free survival (ihPFS) time was defined as the interval between
ablation procedure and the first detection of local recurrence or new metastases elsewhere
in the liver.
Statistical analysis
All statistical analyses were performed using SPSS software (version 24; IBM). IA
and LTP rates as well as frequency of adverse events (AE) were compared between the
groups using the Chi-square test. The lesion size and the size of the ablation zone
were compared between the groups using the Mann-Whitney U-test. IhPFS was calculated
for all patients using the Kaplan-Meier method and the difference among groups was
determined by the log-rank test.
A value of p < 0.05 was considered statistically significant. All values are expressed
as mean ± standard deviation (SD) or median and interquartile range (IQR).
Results
Patients and tumors
A total of 69 patients with 99 malignant hepatic liver lesions were included in the
study. Patient characteristics are summarized in [Table 1]. The median number of treated tumors per patient was 1 (range 1–3). 59 of all treated
lesions were DLs in the planning unenhanced CT examination and 40 were NDLs.
Table 1
Patients characteristics.
Tab. 1 Patientencharakteristika.
|
|
N = 69
|
|
Age (y)
|
66.6 (± 14.8)
|
|
Gender
|
Male
|
52 (75 %)
|
|
Female
|
17 (25 %)
|
|
Tumor type
|
Hepatocellular cancer
|
20 (29 %)
|
|
Cholangiocarcinoma
|
9 (13 %)
|
|
Colorectal metastases
|
33 (48 %)
|
|
Metastases of melanoma
|
2 (3 %)
|
|
Metastases of ovarian cancer
|
2 (3 %)
|
|
Metastases of gastric carcinoma
|
1 (1 %)
|
|
Metastases of pancreas carcinoma
|
1 (1 %)
|
|
Metastases of bronchial carcinoma
|
1 (1 %)
|
|
Metastases of breast cancer
|
1 (1 %)
|
|
Liver cirrhosis
|
Yes
|
14 (20 %)
|
|
No
|
55 (80 %)
|
|
Number of treated lesions per patient
|
n = 1
|
47 (68 %)
|
|
n = 2
|
15 (22 %)
|
|
n = 3
|
6 (9 %)
|
|
n = 4
|
1 (1 %)
|
The mean lesion size showed no statistically significant difference between lesions
that were visible and those that were not visible, with 14.9 ± 0.9 mm for DLs und
12.9 ± 0.8 mm for NDLs (p = 0.28).
The type of ablation procedure used is summarized in [Table 2].
Table 2
Ablation characteristics.
Tab. 2 Charakteristika der Ablationen.
|
|
N = 99
|
|
Detectability of tumor in unenhanced CT
|
Detectable (DL)
|
59 (59 %)
|
|
Non-detectable (NDL)
|
40 (40 %)
|
|
Ablation type
|
MWA
|
DL n = 18
|
23 (23 %)
|
|
NDL n = 5
|
|
RFA
|
DL n = 24
|
41 (41 %)
|
|
NDL n = 17
|
|
IRE
|
DL n = 17
|
35 (35 %)
|
|
NDL n = 28
|
Technical success
The size of the ablation zone was not significantly different for both groups: 38.8 ± 1.
3 mm for DLs and 37.1 ± 1.4 mm for NDLs (p = 0.41).
Incomplete ablation was observed in 2 out of 59 cases (3.4 %; 0.4–11.7 %) in the DL
group and in 2 out of 40 cases (5.0 %; 0.6–16.9 %) in the NDL group, which was also
not significantly different (p = 0.69).
Adverse events
AEs were observed in 6 of 69 patients (9 %).
The following AEs were observed: One patient developed pneumothorax, two patients
had perihepatic hematomas, two patients had arterioportal fistulae, and one patient
was found to have segmental liver infarction due to procedure-related portal vein
thrombosis. In two patients, an additional intervention in form of coil embolization
(one of arterial bleeding and one of a symptomatic arterioportal fistula) was required
(grade 3).
All other complications were categorized as grade 1. There was no procedure-related
permanent morbidity or death. All complications were observed in the DL group in 6
out of 59 cases (10 %), with a statistically significant difference (p = 0.037) compared
to 0 out of 40 cases (0 %) in the NDL group.
Follow-up, LTP, and ihPFS
The mean follow-up was 16.2 ± 11.7 months: 14.8 ± 1.5 months for patients after ablation
of DLs and 18. 2 ± 1.8 months for patients after ablation of NDLs.
One patient was lost to follow-up after 4 weeks.
During the follow-up-time, local tumor progression was observed in 10 out of 59 cases
(16.9 %) in the DL group and 6 out of 39 cases (15.4 %) in the NDL group, which was
not significantly different (p = 0.84).
Using Kaplan-Meier analysis, the mean estimated ihPFS time was 15.5 ± 2.2 months for
DLs compared to 14.3 ± 2.4 months for NDLs (p = 0.84). These results are visualized
in [Fig. 2].
Fig. 2 Kaplan-Meier curve for intrahepatic progression-free survival.
Abb. 2 Kaplan-Meier-Kurve für das intrahepatische progressfreie Überleben.
Discussion
This study investigated whether the detectability of target lesions during CT-guided
hepatic ablations would influence the technical and oncological success or adverse
event rates of ablation procedures. Our results demonstrated that there was no significant
difference comparing the groups with detectable and non-detectable lesions in terms
of the completeness of ablation, rates of local recurrence, or intrahepatic recurrence-free
survival. We expected that in patients with non-detectable lesions (NDL), the ablation
zones would be larger than in patients with detectable lesions (DL) – but this was
also not the case. Last, adverse events were even significantly lower in the NDL group,
suggesting that they are not connected with the detectability of the target lesion.
Accordingly, when using the “anatomical landmark approach” to ablate non-detectable
lesions, the outcome was similar compared with that of patients with detectable lesions.
Applying the “anatomical landmark approach” it should be considered that the position
of the target lesion in relation to a landmark might be distorted due to patient positioning
(e. g., supine vs. prone position) or breathing movements. Therefore, we perform every
needle placement and every control CT scan in breath hold under general anesthesia
and we attempt to perform ablation in the same patient position as in the preoperative
imaging.
The overall rate of local tumor recurrence observed in this study is in line with
previously published data, ranging from 6 % to 33 % [18]
[19]
[20]
[21]
[22]. [Fig. 3] shows an example of an incomplete ablation in one of our patients, most likely due
to the adjacency of the liver vein and thus the heat sink effekt. Several studies
investigated risk factors associated with partial ablation and local recurrence [23]
[24]
[25]
[26]. To our knowledge, there is no previous study investigating the detectability of
the target lesion as a potential risk factor.
Fig. 3 Example of an incomplete ablation. A Pre-procedural T2-weighted MRI-Image of a colorectal metastasis adjacent to the middle
liver vein (arrows). B Post interventional control-CT-scan in hepatic portal phase one day after MWA of
the metastasis. C T2-weighted MRI-Image four weeks after MWA with detectable residual tumor adjacent
to the middle liver vein, in the sense of incomplete ablation, probably due to the
heat sink effect.
Abb. 3 Beispiel einer unvollständigen Ablation. A Preinterventionelles, T2-gewichtetes MRT-Bild einer kolorektalen Metastase angrenzend
an die mittlere Lebervene (Pfeile). B Postinterventionelle Kontroll-CT-Untersuchung in venöser Kontrastmittelphase ein
Tag nach Mikrowellenablation der Metastase. C T2-gewichtetes MRT-Bild vier Wochen nach Ablation mit abgrenzbarem Resttumor angrenzend
an die mittlere Lebervene, im Sinne einer unvollständigen Ablation, am ehesten aufgrund
von heat sink Effekt.
One study that compared the accuracy of computer-assisted versus cognitive registration
for locating liver tumors had shown that non-rigid registration yielded better localization
accuracy than cognitive registration performed by the interventional radiologist [27]. However, no clinical outcome was investigated in this study. In other words, the
clinical benefit of image registration in liver ablation is largely unknown.
Limiting factors of the presented study include the use of unenhanced CT scans as
a criterion for lesion visibility and the lack of additional ultrasound guidance.
We generally use unenhanced CT for ablation planning but use contrast agent after
the presumed completion of the procedure in order to evaluate the ablation result.
We perform contrast-enhanced imaging only in cases requiring visualization of not
only the target lesion but also of critical anatomic structures e. g., the liver hilum
– but usually after placement of the ablation probe near the target lesion. Often
several rounds of contrast-enhanced CT imaging are required, though the reasonably
applied amount of IV contrast is limited. Therefore, we tend to dispense with an initial
contrast-enhanced CT examination for planning purposes in order to safely be able
to perform a contrast-enhanced CT examination in the case of acute complications.
Also, anatomical structures, as well as the target lesion, are visible only for a
short time after application of contrast agent. Nevertheless, their exact position
might change during the procedure as they can be displaced by the puncture needle.
Ultrasound is often used as a stand-alone or assisting technique for tumor ablation.
In our institution we perform tumor ablation solely with CT guidance. The addition
of US could help depict more NDLs and could be used as a real-time tool for accurate
needle placement.
In view of our results, however, there seems to be little need for additional ultrasound
for targeting purposes even in patients with NDLs.
Another often used technique is prior lesion marking by intraarterial lipiodol injection.
In our experience marking was insufficient in many cases of liver metastases due to
poor vascularization. Furthermore, the combination of intraarterial and CT-guided
interventions is often logistically challenging and places an additional strain on
the patient with a further risk of complications. Nevertheless, our experience is
based on very few cases and since there are some good results in the published literature
[28], this technique should be considered in the appropriate cases.
We included a variety of liver malignancies, including primary and secondary liver
tumors, which typically present different features on CT and MRI imaging, such as
arterial hyperenhancement and venous washout. Also, the presence or absence of liver
cirrhosis generates a different image in diagnostic and interventional imaging. However,
the different tumor types and the cases of liver cirrhosis were evenly distributed
in the DL and NDL groups, so that we are confident that this would not confound our
results.
In conclusion, this study demonstrates that successful ablation of liver tumors that
are not detectable on unenhanced CT is possible and safe when using the anatomical
landmark method to guide needle placement.
