Primary and secondary liver malignancies are among the most common and clinically significant tumors worldwide, representing a major cause of cancer-related morbidity and mortality (1–3). Hepatocellular carcinoma (HCC) and liver metastases, particularly from colorectal cancer, are frequently encountered in clinical practice and often necessitate locoregional treatment strategies. The 2025 European Society for Medical Oncology (ESMO) clinical practice guidelines recommend percutaneous thermal ablation (PTA) along with surgery for very early [Barcelona Clinic Liver Cancer (BCLC)-0] and early (BCLC-A) HCC patients who are not candidates for liver transplantation (4). In a similar setting, the results of the COLLISION trial indicate that PTA can even represent the standard of care for small liver metastases from colorectal cancer (5–11).

Imaging plays a central role in PTA, both for treatment planning and guidance during the procedure itself. Among imaging modalities, contrast-enhanced magnetic resonance imaging (MRI) offers the highest sensitivity for liver tumor detection, followed by multiphase contrast-enhanced computed tomography (CECT) (12, 13). Despite its superior lesion visibility, MRI guidance remains limited in clinical practice due to equipment availability, higher cost, and the need for MR-compatible interventional tools. CT “step-and-shoot” guidance is widely used for percutaneous procedures but has limitations due to the single axial plane view and the lack of real-time feedback during needle progression. Additionally, because liver enhancement after contrast media administration is transient, lesion visibility may be limited to a short time window, making consistent targeting during needle placement more difficult (14, 15).

Ultrasound (US) is the preferred image-guidance modality for many interventional radiologists due to real-time, multiplanar capabilities, wide availability, and low cost. However, its sensitivity is limited, particularly for lesions of small size, located deep in the liver (i.e., liver dome) or obscured by factors such as the presence of background macronodular cirrhosis or heterogeneous liver echostructure. To overcome these limitations, real-time image fusion of US with pre-acquired CT or MRI datasets has emerged as an efficient technique, enabling better lesion localization and improving the feasibility of image-guided interventions for their detection by radiologists (12–14).

The use of tumor landmarking with small fiducial markers has been employed in the management of potentially operable liver metastases at risk of becoming radiologically occult following neoadjuvant chemotherapy (15). When placed percutaneously before initiation of chemotherapy, these markers have proven useful for precise localization of the initial tumor site, facilitating precise surgical resection or thermal ablation—even in cases of complete tumor disappearance (16–18). Fiducial markers are also commonly used to enable accurate lesion tracking during stereotactic body radiation therapy (SBRT) (21). Their visibility across all imaging modalities—including US, contrast-enhanced CT, and MRI—makes them a reliable tool for long-term lesion tracking and interventional planning.

Treating tumors that are occult or poorly visible on ultrasound and/or CT with PTA can be particularly challenging. Haochen et al. (20) already described the use of MR-guided fiducial implantation before PTA. In the present study, we describe a two-step strategy to facilitate the targeting of occult lesions by the percutaneous insertion of a small metal fiducial marker in these tumors under real-time fusion imaging guidance prior to PTA. Shortly after, the fiducial placement is confirmed with CT or MRI. This study aims to evaluate the feasibility and precision of such fiducial marker placement guided by fusion imaging.

Our institutional review board approved this retrospective study and waived written informed consent for the review of images and records.

A total of 30 patients (mean age 64 years, range 40–80 years; 23 males, 7 females) with 38 lesions were included in the study. Twenty-one patients had HCC, and nine patients had liver metastases from colorectal cancer (7), pancreatic cancer (1), and ovarian cancer (1). The mean lesion size was 13 mm (range 4–27 mm). Two lesions were <6 mm, 13 were between 6 and 10 mm, 18 were between 11 and 20 mm, and 5 were >20 mm. Twenty-five lesions were located at the deep superior or posterior parts of the liver, partially hidden by lung, scars from previous treatment, or interposed bowel structure (segments I, VI, VII, and VIII), and 13 lesions were located in regions more accessible to US detection (segments II, III, and V) (Table 1). The mean depth of the lesions was 7 cm (range 3–14 cm).

Among the 38 lesions, 28 (74%) were undetectable using ultrasonography alone, and 10 (26%) were doubtful. With the help of the fusion, 26 (68%) lesions showed enhanced visibility, although with unclear boundaries in six cases (Figure 1), while 12 (32%) remained undetectable. The major reasons behind no or poor detection were liver heterogeneity and deep localization (region not accessible by ultrasound). Technically, fusion was primarily useful to distinguish between tumors and neighboring parenchymal heterogeneity or to find an angle of the US beam allowing for a better visualization in cases of deep localization or a hidden part of the liver. All 10 lesions of doubtful detection because of the heterogeneity of the hepatic parenchyma were more distinctly detected. Six lesions located at the very high or posterior part of the right lobe of the liver became visible by adapting the angle of the ultrasound beam according to the fused images.

Overall, the mean distance between the tumor and the fiducial marker was 4 mm (range 0–45 mm). For 30 (79%) lesions, the fiducial marker was successfully placed inside or in contact (27/30) or within <5 mm from the target lesion (3/30). The criterion for success for fiducial marker placement was not satisfied for 8 (21%) lesions. For those failures, the distances between the fiducial marker and the lesion border were 7, 8, 10, 15, 15, 20, 22, and 45 mm on control imaging. These cases were related to the lesion depth and insufficient length of the applicator to place the marker adequately (two cases) or to fusion imprecision (six cases). This inaccuracy was observed in 5 (42%) out of the 12 lesions invisible on ultrasound despite fusion. Interestingly, among these 12 lesions totally invisible to US even after fusion, 7 (58%) had a successful fiducial marker placement (Figure 2).

No complication directly or indirectly related to the fiducial marker placement was recorded (Figure 3).

Potentially curative treatment of primary or metastatic liver tumors requires a comprehensive treatment of all liver lesions, combining liver resection and/or PTA (RFA or microwave ablation (MWA)) (3, 7, 8, 16, 20). With thermal ablation, the aim is to obtain complete tumor destruction and to avoid complications related to damage to the heat-vulnerable organs in the vicinity, such as the bile ducts for central tumors, bowel, or even lung for subcapsular tumors. To overcome these difficulties, the tumor must be precisely targeted by the ablation needle, and the ablation parameters must be carefully adjusted according to the size and location of the tumor in relation to surrounding structures. For these reasons, precise localization of liver lesions is crucial for accurate targeting.

Ultrasound is the modality of choice for interventional procedures in the liver due to several advantages, mainly real-time and multiplanar capabilities. In contrast, it is operator dependent, and its sensitivity for detecting malignancies is lower than that of CT or MRI (21, 22). Bae et al. (23) reported that percutaneous RFA was infeasible in approximately one-fourth of patients with metastatic colorectal cancer. The reason for this infeasibility was mainly an unfavorable tumor location and its invisibility on planning US. Lesion visualization can be improved using US contrast (SonoVue, Bracco, Italy), but the enhanced visibility does not last long after the injection and fails to provide guidance during the whole procedure, particularly for hypervascular tumors such as HCC.

Owing to these limitations, PTA under US fusion imaging guidance has been proposed (24, 25). Image fusion software can be used to superimpose the CT or MRI datasets, clearly displaying the lesions onto real-time ultrasound images in order to make the identification of sonographically occult tumors possible (26). Fusion devices often use electromagnetic fields around the patient to locate the US transducer and reformat the corresponding CT or MR images. These images are reformatted through multiplanar algorithms and are displayed in real time alongside the US image, thus enhancing targeting. The incidence of mistargeting after fusion imaging-guided radiofrequency ablation (RFA) appears very low in the literature, as low as 1.3%, according to Lim et al. (27), but up to 4.4% in ultrasound invisible lesions according to Mauri et al. (28). Recently, in a series of 40 liver metastases to be treated by PTA, Hakime et al. (29) showed a significant increase of the tumor conspicuity after localization on fusion imaging, offering a technical success rate of 83% (33/40) in the intention-to-treat analysis and 100% in an analysis of treated tumors. Such papers will certainly, and rightly, contribute to generalizing image fusion for PTA imaging guidance. But with no mention of the effective reliability of the fusion of images in terms of distance from the targeted lesion, specifically when the lesions remain totally invisible to US despite the fusion, they lack an objective measurement of the anatomical precision obtained.

Our series objectively measures the distance between liver tumors that are poorly or not visible to ultrasound and the position of a needle tip placed under image fusion, thanks to the deposition of a marker whose location was verified by CT or MRI. It demonstrates the feasibility and efficiency of fiducial marker placement under fusion. With an average tumor—fiducial marker distance of 4 mm and successful placement in 8 out of 10 cases, image fusion can be considered reliable for tumor localization before PTA. This is remarkable because five of these lesions remained totally invisible on ultrasound, even with the aid of fusion, and the fiducial marker placement relied solely on the fusion technique.

Fiducial markers are commonly placed by the breast radiologists daily for targeting lesions during ultrasound studies or guided biopsies. This technique has become widely accepted to help the surgeon by facilitating and guiding intraoperative tumor resection procedures (30, 31). In US, fiducial markers appear hyperechoic with posterior acoustic shadowing; in MRI, they produce a small blooming artifact; and in CT images, they are clearly visualized as hyperdense structures.

The second key contribution of our series is to confirm the interest of preoperative tracking of liver tumors potentially difficult to detect during PTA by fiducial markers. Once a fiducial marker is adequately placed inside or in close contact with the liver tumor, any technique (US, CT, or both combined) may be employed to guide the ablation needle in a very simple way, similar to cases of lesions well seen on routine imaging. In our practice, ablations are performed under dual guidance; ultrasound to monitor the needle progression, and CT is used to confirm the final position of the needle and to evaluate the outcome of the thermal ablation. In case of tumors completely invisible or poorly visible by either of the two techniques, the fiducial marker greatly facilitates the placement of the ablation needle regardless of the access route imposed by the tumor location and adjacent structures to be preserved.

There remains, however, the question of whether fiducial marker placement should be performed for all cases of tumors that are difficult to locate and proposed for thermal ablation or only for some of them. If the tumor is finally well identified, thanks to the fusion of images as it was for 2/3 of our lesions, the fiducial marker may not be necessary, and the positioning of the ablation needle could be performed directly under fusion during treatment. If, on the contrary, the tumor remains completely invisible on ultrasound despite image fusion, placing a fiducial marker before ablation appears valuable to verify its position at CT or MRI. In the best cases, the fiducial markers will be perfectly in place; in the other cases, the distance from the fiducial marker to the tumor can be appreciated and considered for the positioning of the ablation needle. We recommend fiducial marker placement in all cases of metastases that are at risk of disappearing after chemotherapy and for which lesion-site excision is being considered following neoadjuvant treatment.

This study has several limitations. First, it was a retrospective, single-center study with interventional radiologists highly trained in image fusion and guided puncture, which may limit generalizability. It may, however, invite other teams to use image fusion and occasionally fiducial markers for the most complex hepatic thermal ablation procedures. Beyond the facilitation of the gesture itself, we were able to note a possible widening of the indications, with virtually no cases remaining ineligible to PTA due to poor tumor visualization. Second, the cohort included a heterogeneous group of tumors (HCC and metastases from different origins), potentially influencing detectability and results. Third, the relatively small sample size introduces the risk of bias. Fourth, the study did not assess oncological outcomes such as local control or survival, as its objective was limited to evaluating the feasibility and accuracy of fiducial marker placement. Finally, in some cases, fiducial placement failed or lesions remained undetectable, underscoring the technical challenges that remain and the need for further prospective studies.

Marking invisible or poorly conspicuous hepatic lesions with a fiducial marker under US guidance with fusion technology is safe and feasible with good precision. Our work suggests that routine fiducial marker marking of those occult liver lesions may make them more suitable for percutaneous ablation, potentially widening the indication of PTA to classically hard-to-treat lesions.