Angiography-Assisted Cone-Beam CT-Guided Radiofrequency Ablation for Hepatocellular Carcinoma: Single-Center Workflow and Early Outcomes
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors- On what basis do the authors maintain that nodes #10–11 are “most often” the first site of metastasis, given that alternative drainage routes and ~7% of lymphatic vessels ending outside the “classic” stations have been described?
- How do the authors justify the use of a fixed SUV threshold > 2.5 in PET-CT for node assessment in the absence of standardization and known metabolic variability of adenocarcinoma?
- Can the authors quantitatively compare the prognosis and surgical implications in patients with skip metastases (N2 without N1) and with cross-over metastases, broken down by lobe and tumor side?
- What practical algorithm for choosing between EBUS-NA, EUS-NA, video mediastinoscopy, and TEMLA do the authors recommend for suspected N0–N2 depending on location (#2–#9, #10–#14) and CT/PET results, with a clearly defined end point for diagnosis?
Author Response
Point to point response
First, we would like to thank the reviewers for their in-depth reviews. We've responded to your points and made corrections, which we hope to meet the reviewer's intent.
# Reviewer 1
- On what basis do the authors maintain that nodes #10–11 are “most often” the first site of metastasis, given that alternative drainage routes and ~7% of lymphatic vessels ending outside the “classic” stations have been described?
A: Thanks for the good points. Non-classical drainage pathways are, by definition, alternative routes and therefore uncommon; as reported, lymphatic channels terminating outside the classic stations occur at a relatively low frequency of about 7%. In our angio–CBCT–tracked cohort, the sample size was small and we did not perform formal statistical testing; however, simple descriptive tabulation suggested that stations #10–11 had the highest observed frequency.
- How do the authors justify the use of a fixed SUV threshold > 2.5 in PET-CT for node assessment in the absence of standardization and known metabolic variability of adenocarcinoma?
A: The SUV cutoff of >2.5 is the most commonly used threshold at our institution and is consistent with national (Korean) clinical practice guidelines. If the Editor or the reviewer prefers, we will add appropriate supporting references to the manuscript.
- Can the authors quantitatively compare the prognosis and surgical implications in patients with skip metastases (N2 without N1) and with cross-over metastases, broken down by lobe and tumor side?
A: Patients with skip metastasis and cross-over nodal metastasis are generally not considered candidates for RFA; however, it is known that when complete ablation (technical success) is achieved with RFA, the prognosis can be comparable to that of successful surgical resection. Therefore, our team concludes that the prognosis after RFA can likewise be interpreted in terms of lobe and tumor laterality, as is done in surgical settings.
- What practical algorithm for choosing between EBUS-NA, EUS-NA, video mediastinoscopy, and TEMLA do the authors recommend for suspected N0–N2 depending on location (#2–#9, #10–#14) and CT/PET results, with a clearly defined end point for diagnosis?
A: Good point by the reviewer. e consider video mediastinoscopy the most definitive method; however, we are concerned about invasiveness comparable to surgery. If a minimally invasive test is desired, selective contrast-enhanced imaging via angio-CBCT can also be an option. However, if cell and tissue confirmation must be included, we recommend the sequence of screening CT, selective contrast-enhanced imaging with angio-CBCT, and finally EBUS-NA.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsPage 3, patient selection.
It is stated that all consecutive patients who underwent CBCT RFA was included. The benchmark group treated in the same period shold also be mentioned here, as a comparison is later presented regarding results. Also details regarding procedures and electrode placement for the benchmark group should be presented.
p 4, 2.5 and 2.6:
How was needle posistion and ablation zone coverage evaluated? Visual evaluation or overlay of images? probably no use of some confirmation software? A clearer description would be appropriate. Furtehrmore: Immediate CT-control after RFA often shows hyperemia in the periphery of the ablation sone. Was this observed here, and if yes, how was this differatiated from remnant tumour?
p5, l 186-189: Was the cone-beam CT the only CT to confirm techical success. In that case, this shuold be clearly stated. Also, it should be presented here how technical success was assessed in the benchmark group.
Result.
I table 1 , or antother Table, presentation of tumour size and number in both the CB-CT and benchmark group should be presented. Are they comparable? Demographic data of the benchmar group should also be presented. Are the croups comparable I assume many treated patients treated in the period have been excluded, but a clearer description of this than what is this given in page 5 l202-209 is needed. How many patients was excluded and for what reason could be clearyfied with a flow-chart for the benchmark group.
Tabel 3.
Local recurrence
What is the follow-up time? For both groups. Is it not to short?
Discussion
l 330 -334
It is stated that a CT for needle verification is essential. This is debatable, as an approch with CT-control after ablation and re-ablation if indicated may also achieve sufficient margins.
Use of confirmation software is also not mentioned. In the recent Cover-all study the superiority of soft-ware based evaluation compared to visual compairson is highlighted and use of such software are by many regarded as mandatory. If this was not used, discuss why. Could it also be implemeted in the described work-flow, or is it not feasible or necessary with the method presented??
Author Response
Point to point response
First, we would like to thank the reviewers for their in-depth reviews. We've responded to your points and made corrections, which we hope to meet the reviewer's intent.
# Reviewer 2
- Page 3, patient selection.
It is stated that all consecutive patients who underwent CBCT RFA was included. The benchmark group treated in the same period shold also be mentioned here, as a comparison is later presented regarding results. Also details regarding procedures and electrode placement for the benchmark group should be presented.
A: Thank you for your valuable comments. We have incorporated your suggestions by adding the benchmark group in Section 2.1 and providing a brief description of the electrode placement procedure for the benchmark group in Section 2.4. The added details aim to provide important information regarding electrode placement and procedural steps, which will enhance the clarity of the study.
“In addition, a benchmark group (conventional CT-guided RFA, ultrasound (US)-guided RFA, and intraoperative RFA) treated during the same period was also included for comparison.”
“The benchmark group included patients treated between July 1, 2022, and August 31, 2025, who received conventional RFA techniques, specifically US-guided, CT-guided, or intraoperative RFA. The procedures for this group were performed according to standard institutional protocols, which involved the use of either conventional CT or ultrasound for guidance.”
“For the benchmark group, conventional methods were used. CT-guided RFA relied on standard CT imaging for electrode placement, while US-guided RFA utilized real-time ul-trasound to direct the electrode. In intraoperative RFA, needle guidance was performed based on real-time ultrasound and direct visualization of mass.”
- p 4, 2.5 and 2.6:
How was needle posistion and ablation zone coverage evaluated? Visual evaluation or overlay of images? probably no use of some confirmation software? A clearer description would be appropriate. Furtehrmore: Immediate CT-control after RFA often shows hyperemia in the periphery of the ablation sone. Was this observed here, and if yes, how was this differatiated from remnant tumour?
A: Thank you for your valuable comments. No confirmation software was used. Needle position and ablation zone coverage were visually assessed using pre- and immediate post-ablation CECT/angio-CBCT. This evaluation method allows for differentiation between residual tumor and hyperemia. By comparing the locations, we could observe differences in enhancement patterns and morphology between the two areas, which allows for clear differentiation between hyperemia and residual tumor. And we will add this point in material and method section. “This visual assessment allowed for differentiation between residual tumor and hyperemia by comparing the enhancement patterns and morphology between the two areas.” Additionally, I will attach new images (figure 6.) of cases where residual tumors were present in result section.
- p5, l 186-189:
Was the cone-beam CT the only CT to confirm techical success. In that case, this shuold be clearly stated. Also, it should be presented here how technical success was assessed in the benchmark group.
A: We thank the reviewer for this valuable comment. In the angio-CBCT cohort, intraprocedural cone-beam CT was the sole imaging modality used to confirm technical success. And the assessment of technical success was made by the interventional radiologist alone during the procedure. we will state this explicitly in the revised manuscript.
For the benchmark cohort, technical success was defined as follows: In the US-guided cohort, the ablation margin was determined by the high-echoic bubbles generated during RFA; in the CT-guided cohort, the margin was determined by the peripheral hyperemia enhancement visible on contrast-enhanced CT. In the intraoperative cohort, as ultrasound guidance was always used, the margin determination was similar to that of the US-guided cohort. These margins were considered successful if they completely covered the pre-ablation tumor bed with a circumferential margin ≥1 cm, and the original tumor's contour and enhancement were no longer visible. We will specify this assessment method in the revised manuscript, along with the imaging modality used for assessment in the benchmark cohort and the time point of evaluation to ensure reproducibility.
“In the angio-CBCT cohort, technical success was confirmed by intraprocedural cone-beam CT. For the benchmark cohort, the margin determination varied by imaging modality: in the US-guided cohort, it was determined by the high-echoic bubbles from RFA; in the CT-guided cohort, by the peripheral hyperemia enhancement seen on contrast-enhanced CT. In the intraoperative cohort, as ultrasound was always used. Based on these findings, technical success was defined as meeting both criteria (1) and (2), as in the angio-CBCT RFA cohort.”
- Result.
table 1 , or antother Table, presentation of tumour size and number in both the CB-CT and benchmark group should be presented. Are they comparable? Demographic data of the benchmar group should also be presented. Are the croups comparable I assume many treated patients treated in the period have been excluded, but a clearer description of this than what is this given in page 5 l202-209 is needed. How many patients was excluded and for what reason could be clearyfied with a flow-chart for the benchmark group.
A: Thank you for your valuable comment. We agree that the demographic data of the benchmark group, as well as the tumor size and number for each group, should be included. We have now incorporated these data into Table 1 and revised the manuscript accordingly to explicitly describe these characteristics (see Section 3.1).
- Tabel 3. local recurrence
What is the follow-up time? For both groups. Is it not to short?
A: We appreciate the comment and have clarified follow-up and the endpoint.
-Methods: We renamed the endpoint from “local tumor recurrence” to time to local recurrence (TLR); defined TLR and censoring at last imaging or October 30, 2025 (administrative); and specified Kaplan–Meier summarization with numbers at risk.
-Results: We now report median [IQR] follow-up (460 [205–636] days for angio-CBCT; 967 [801–1076] days for conventional). Using a time-to-event approach, local control at 12 months was 100% in both groups; at 24 months, 100% (95% CI NE; no events; at-risk n=1) for angio-CBCT and 94.1% (95% CI 65.0–99.1) for conventional. Deaths were not modeled as competing events.
-Table 3: We added a row for Imaging follow-up (median [IQR]) and KM local control (12 and 24 months), with a footnote on censoring and the meaning of NE.
We hope these revisions address the concern regarding follow-up duration.
- Discussion l 330 -334
It is stated that a CT for needle verification is essential. This is debatable, as an approch with CT-control after ablation and re-ablation if indicated may also achieve sufficient margins.
A: We sincerely thank the Reviewer for this thoughtful comment. We agree that our previous phrasing suggesting CT for needle verification is “essential” may be debatable. Accordingly, we will revise the Discussion section as follows to soften the language and clarify our rationale. We recognize that other approaches are reasonable.
“Contrast-enhanced cross-sectional imaging (CECT or angio-CBCT) at key steps is recommended when uncertainty must be minimized. In our workflow, we routinely obtain two acquisitions—(i) a pre-ablation scan immediately after needle placement to verify the needle–tumor relationship (to mitigate malposition-related complications and plan adequate margins), and (ii) an immediate post-ablation scan to assess the ablative margin. We acknowledge that some centers rely on post-ablation CT with re-ablation as needed to achieve margins; while reasonable, we favor pre-ablation verification because a baseline scan facilitates discrimination of peri-ablational hyperemia from true residual tumor.”
- Use of confirmation software is also not mentioned. In the recent Cover-all study the superiority of soft-ware based evaluation compared to visual compairson is highlighted and use of such software are by many regarded as mandatory. If this was not used, discuss why. Could it also be implemeted in the described work-flow, or is it not feasible or necessary with the method presented??
A: Thank you for this valuable comment. Software-based confirmation was not used in this cohort because it was not available at our institution during the study period. Needle position and ablative margin were assessed by visual side-by-side comparison on pre-ablation and immediate post-ablation CECT/angio-CBCT, as described. We agree that software-assisted evaluation can improve reproducibility and help distinguish peri-ablational hyperemia from true residual tumor, and it is compatible with our workflow; we plan to evaluate it in future work. We have clarified this point in the manuscript as follows.
-Methods : “Software-based confirmation tools were not available during the study period and were not used; needle position and margins were assessed visually on pre- and immediate post-ablation CECT/angio-CBCT.”
-Discussion: “Software-assisted margin confirmation—reported to outperform visual comparison—could be integrated into the described angio-CBCT workflow and may further enhance standardization; it was not evaluated here.”
-Limitations: “Lack of software-based margin evaluation may affect reproducibility and generalizability.”
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsI find the manuscript and the described approch interesting. As all my inqueries have been adressed properly, the manuscript have been improved, and I have no further comments or suggestions
