The Added Diagnostic Value of Transcatheter CT Hepatic Arteriography for Intraprocedural Detection of Previously Unknown Colorectal Liver Metastases During Percutaneous Ablation and Impact on the Definitive Treatment Plan

Colorectal cancer (CRC) is a frequently diagnosed malignancy and the second leading cause of cancer-related mortality in the world. Approximately half of the patients with CRC develop colorectal liver metastases (CRLM) [1‐5]. Currently, multiple curative-intent local treatment options are available for patients with liver only or liver dominant disease. Over the past decades, thermal ablation has acquired an important role, either as an adjunct or as a less invasive alternative to partial hepatectomy [4‐8].

In conventional computed tomography (CT)-guided liver tumour ablation, a baseline intravenous contrast-enhanced (ce)CT, often including the arterial and portal venous phase, is acquired for treatment planning. Probe placement is usually performed under unenhanced CT fluoroscopy while focusing on unsteady anatomical landmarks or using stereotactic navigation. To timely detect complications and assess technical success, a second ceCT with intravenous contrast is used after the ablation. Confirmation of the ablation zone, if performed at all, is either based on so-called ‘eye-balling’ to subjectively estimate if the tumour-free margin around the initial tumour was achieved or using image registration and confirmation software [9‐12].

The transcatheter CT hepatic arteriography (CTHA) technique [9, 12, 13] is a relatively new technique to assist percutaneous thermal ablation procedures in the treatment of hepatocellular carcinoma (HCC), colorectal and non-colorectal liver metastases. This technique entails the selective placement of a catheter in the common or proper hepatic artery, to enable the repeated admission of small doses of intra-arterial intrahepatic iodine-based contrast agent. Compared to conventional ceCT fluoroscopy guidance, CTHA is hypothesized to optimize (a) pre-procedural planning, by (repeatedly) clearly depicting most liver tumours and surrounding blood vessels, (b) intraprocedural targeting by improving tumour conspicuity and (c) image registration and ablation confirmation [11‐14]. In case of insufficient tumour coverage by the ablation zone, the needle can be directly repositioned to allow for additional overlapping ablations. CTHA is also known for its ability to differentiate between viable residual tumour tissue and ablative scar tissue (‘incomplete ring sign’), thereby improving intraprocedural monitoring [10‐12, 15]. Besides leading to a decrease in the number of patients with local tumour progression (LTP), potentially reducing the number of repeat procedures, CTHA is also thought to visualize metastases at an earlier stage, hence potentially contributing as a diagnostic tool [9‐12]. However, the detection of concomitant, previously unknown liver tumours, does not automatically justify immediate local treatment in the same session, as the rate of lesions representing benign liver tumours or heterogeneous perfusion deficits is unknown.

The primary aim of this study was to determine the added diagnostic value of CTHA for the intraprocedural detection of previously unknown CRLM during percutaneous ablation and the impact on the treatment plan. By defining specific imaging characteristics, we intend to validate criteria that help decide whether to ablate immediately or a wait-and-see.

Between January 2012 and March 2022, 191 patients treated with 273 transcatheter CTHA-guided percutaneous ablation procedures were assessed. Twenty-seven patients were excluded due to intended non-selective placement (n = 15) or luxation of the catheter during transport (n = 12). After reviewing all procedure records, 155 patients treated with 194 CTHA-guided ablations met the inclusion criteria (Fig. 1, Table 1).

In eighteen procedures (in 18 distinct patients) concomitant, previously unknown, CRLM were detected (n = 18/194; 9.3%). Three procedures were discontinued (n = 3/194; 1.5%), because multiple new CRLM were detected alongside rapid tumour growth of the known CRLM (increase in longest diameter of the known CRLM > 20%), excluding the procedure from further analysis in this study. Eventually, 152 patients underwent 191 CTHA-guided percutaneous thermal ablations (Fig. 1, Table 1). Forty-six patients were treated with induction chemotherapy prior to the procedure (30.3%).

Compared to the conventional cross-sectional imaging modalities (ceCT, ceMRI and 18F-FDG PET-CT), adding CTHA was superior for the detection of CRLM. Furthermore, retrospective CTHA image assessment showed a remarkable number of true-positive CRLM (46.3%), unappreciated during the initial treatment, but detected on follow-up imaging. Taking into account the low number of false positives (6.7%), the favourable safety profile of percutaneous thermal ablation and the substantial interobserver agreement (kappa = 0.75), the immediate ablation of typical ring-enhancing supplementary lesions seems justified and sufficiently validated [16]. Although, in this series, only three patients with rapid disease progression, were not treated with thermal ablation because of the detection of additional CRLM, the depiction of multifocal and scattered disease will prevent some patients from receiving futile ablative procedures.

Compared to other diagnostic modalities, CTHA is considered an effective diagnostic technique with a higher detection rate of both primary and secondary hepatic tumours as is further confirmed by our results [12, 17‐21]. Due to its invasive nature and limited influence on the treatment strategy, CTHA is not widely used as a diagnostic tool. However, CTHA-guided percutaneous ablation has an important and increasing role in today's curative-intent treatment options for CRLM, as several studies have demonstrated that CTHA correlates with a reduced risk of local tumour progression (LTP) and increased odds of progression free survival (PFS) [6, 9‐12, 22]. This study also suggests that adding CTHA to conventional cross-sectional imaging may help visualize vanished tumours. Furthermore, van Tilborg et al. found transcatheter CTHA to increase operator’s confidence by improving distinction between (vital) residual or recurrent tumour tissue and non-enhancing scar tissue (incomplete ring sign) [11].

To our knowledge, no previous report has assessed the validity to immediately ablate additionally detected lesions suspect for metastases versus to opt for a more conservative wait-and-see approach and treat the lesions at a later stage whenever confirmed on conventional cross-sectional imaging. Arguments in favour of immediate ablation would be to reduce the number of repeat procedures and hence improve quality of life and potentially recurrence-free survival, whereas arguments for the conservative approach would be that these potential benefits may not outweigh the added risks of ablating concomitant potentially false-positive benign liver lesions. Given the very low risk of serious adverse events to ablate small-size CRLM and the high positive predictive value of ring-enhancing lesions found in this study, we suggest to immediately treat the concomitant CRLM as long as the location allows for a safe procedure.

This study has some limitations. First, our inclusion criteria consisted in large part of additionally detected CRLM that were immediately ablated within the same session without histopathological or follow-up confirmation that these actually represented CRLM. For this reason, we were unable to identify false-positive lesions, potentially leading to an overestimation of the accuracy of CTHA. Specificity and positive predictive values could not be analysed due to the missing of false-positive lesions. Therefore, we added a retrospective analysis, where follow-up confirmation was available. However, despite the blinded retrospective assessment, there may still have been confirmation bias due to the fact that our observers knew additional CRLM would appear somewhere in the assessed liver. Though it remains unclear how the intra-procedurally detected and ablated CRLM correlate to the retrospectively found CRLM, it seems likely that the more typical and highly suspect lesions were immediately ablated, which would further strengthen our recommendation to immediately ablate. Furthermore, it should be noted that patients did not all receive identical pre-operative imaging (ceCT, ceMRI and/or 18F-FDG PET-CT). Another limitation is the fact that CTHA is in fact an expansion of ceCT, and the operators were aware of the findings on the pre-procedural ceCT. As a result, the McNemar test has to be interpreted with prudence as CTHA was not assessed as a standalone method. Nonetheless, the high number of additionally detected CRLM on CTHA does render the combination of ceCT and CTHA superior for the detection of CRLM compared to ceCT alone.

The CTHA technique also comes with limitations. Variations in hepatic arterial anatomy such as aberrant right or left hepatic arteries frequently occur [23]. In patients with aberrant anatomy, selective catheterization of the segmental vascular supply makes CTHA of other regions not assessable. Additionally, non-selective catheterization within the aorta or catheter luxation also contributes to a decrease in the quality of the CHTA. In this study, the catheter luxated during transport in 12 out of 273 procedures (4.4%), a weakness that can be overcome by the future implementation of combined angio-CT systems. Additionally, CTHA has well-known pitfalls such as non-tumorous perfusion abnormalities and subsequent formation of pseudo-lesions [18, 24, 25]. Our paper demonstrated no catheter-related complications. However, other studies showed that CTHA comes with a non-negligible risk of vascular complications related to catheter placement, occasionally leading to re-interventions, increased direct costs and radiation dose [12, 19]. Though comparative data are not available, CTHA presumably is more time consuming as the estimated time to place the hepatic artery catheter, to transfer patients to the CT suite, to acquire and assess the CTHA scans and to place a femoral artery closure device (estimated additional time per procedure 20–25 min) questionably outweighs time saved by a superior tumour delineation and a theoretical reduction in required treatments. Several of these shortcomings will likely be solved by the advent and rapid spread of combined angio-CT systems.

In conclusion, this study supports that the hypothesis CTHA is able to detect CRLM not visualized on pre-procedural ceCT, ceMRI and 18F-FDG PET-CT, including CRLM that disappeared after systemic chemotherapy. Due to its invasive character, CTHA is unlikely to replace conventional diagnostic modalities soon. We recommend using transcatheter CTHA in percutaneous ablation procedures (a) to improve visualization and detection of (previously unknown) tumours, (b) to improve the outcome of the percutaneous ablation and (c) to allow for targeted follow-up of indeterminate tumours. The latter requires the interventional radiologist to thoroughly review all liver segments before starting the ablative procedure. Taking into account the low number of false positives and the favourable safety profile of percutaneous thermal ablation, we postulate that immediate ablation of typical ring-enhancing supplementary lesions is sufficiently validated and justified and likely to reduce the number of repeat procedures.