Introduction
The adrenal glands are a common site of tumor metastasis, and such adrenal metastasis (AM) is often associated with poor patient outcomes [1]. While adrenalectomy can improve the survival of patients with isolated AMs [2–5], patients with some comorbidities are ineligible for this procedure [1]. Imaging-guided percutaneous ablation (PA) is often implemented as an alternative to adrenalectomy [6, 7], with several studies having reported these two techniques to exhibit comparable levels of clinical efficacy for benign adrenal tumors, with PA additionally being associated with reduced intraoperative blood loss, a shorter operative duration, and more rapid postoperative recovery [6, 7].
Imaging-guided PA is commonly utilized as a treatment for AM, and reported technical success, local recurrence, and complication rates associated with this procedural approach are in the range of 96–97%, 8.8–25%, and 8.6–18%, respectively [8, 9]. Moreover, the 3-year rates of local recurrence-free survival and overall survival (OS) associated with such PA-based treatment are reported to be in the range of 52–69% and 34–52%, respectively [8, 9]. However, the guidance approaches, PA methods, and primary tumor types included in these studies have the potential to influence the conclusions of associated studies. Hence, comprehensive analyses of the safety and clinical efficacy of imaging-guided PA for the treatment of AMs are warranted, as is confirmation of the factors that influence such safety and efficacy.
Aim
Here, a meta-analysis was performed with the goal of assessing the clinical efficacy, safety, and long-term outcomes associated with imaging-guided PA-based treatment of AMs.
Material and methods
Study selection
Study selection was conducted based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklists [10].
Relevant studies in the PubMed, Embase, and Wanfang databases published as of June 2022 were identified with the following strategy: ((ablation) AND (adrenal)) AND ((metastasis) OR (metastatic)). This meta-analysis was registered at INPLASY.COM (No. INPLASY202270032).
Studies eligible for meta-analysis were:
studies focused on computed tomography (CT)- or ultrasound (US)-guided PA treatment of AM;
studies with > 20 patients;
studies on inoperable patients or patients who refused surgery;
studies reporting a minimum of one of the following: PA technical success, local hemorrhage, pneumothorax, hypertensive crisis, local recurrence, 1-year OS, and/or 3-year OS rates;
no language limitations were imposed.
Studies were not eligible for inclusion if they:
Data extraction
Two authors were responsible for the independent extraction of relevant data from eligible studies, and any discrepant data were resolved through discussion with a third investigator. Extracted baseline data included first author, publication year, country, patient number, study design, number of AMs, patient age, imaging guidance approach, tumor diameter, and PA methodology. Outcome data extracted from these studies included rates of technical success, local hemorrhage, hypertensive crisis, local recurrence, pneumothorax, 1-year OS, and 3-year OS.
Quality assessment
The Newcastle-Ottawa scale [11] was used to evaluate observational study quality. Studies were assigned points according to selection (4 points), comparability (2 points), and outcome (3 points) criteria, with studies exhibiting a NOS score ≥ 7 being considered of high quality.
Endpoints
Technical success was defined by the completion of PA-based AM treatment as per the planned treatment protocol without any visible evidence of tumor enhancement on contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI) performed 2–5 days after treatment, given that the targeted ablation zone in the surrounding fat tissue is difficult to detect on CT or MRI [8, 9]. Hypertensive crisis was defined by systolic blood pressure ≥ 180 mm Hg or diastolic blood pressure ≥ 110 mm Hg [6, 7].
Statistical analysis
Random-effects models were used for all pooled analyses owing to the presumption of heterogeneity, with weighting being performed in accordance with the inverse variance of these studies. Heterogeneity was assessed with the Q test and the I2 statistic, with I2 > 50% corresponding to high levels of heterogeneity. Sources of heterogeneity were investigated with meta-regression and subgroup analyses. Egger’s test was used to assess publication bias, with p < 0.05 as the threshold of significance. Stata 12.0 was used for all pooled analyses.
Results
Study selection
The study selection process is detailed in Figure 1. In total, 15 articles were included in this meta-analysis [12–26], all of which were retrospective (Table I). These studies included 538 patients with 562 AMs who underwent treatment via imaging-guided PA. Of these 15 articles, 8 employed radiofrequency ablation (RFA) [14–16, 18, 20, 22, 25, 26], 5 employed cryoablation (CA) [12, 13, 17, 23, 24], and 2 employed microwave ablation (MWA) [19, 21]. In addition, 10 studies employed CT guidance [12, 13, 15, 19, 20, 22–26], while 5 employed US guidance [14, 16–18, 21]. Of these studies, 9 included patients with multiple primary cancers [12, 14, 15, 17, 18, 21, 23, 25, 26], while 3 included patients with AMs secondary to lung cancer (LC) [13, 19, 24], and 3 studies enrolled individuals with AMs secondary to hepatocellular carcinoma (HCC) [16, 20, 22]. All articles reported NOS scores in the range 7–8.
Table I
Baseline data of the included studies
| Author [ref.] | Year | Country | Number of patients | Number of AMs | Mean age [years] | Mean diameter [cm] | Primary tumor | Ablation methods | Guidance | Mean follow-up [m] |
|---|---|---|---|---|---|---|---|---|---|---|
| Aoun [12] | 2021 | USA | 34 | 40 | 63 | 3.2 | Multiple | CA | CT | 21.3 |
| Cheng [13] | 2021 | China | 34 | 34 | 58.4 | 3.0 | LC | CA | CT | 28.8 |
| Gao [14] | 2020 | China | 43 | 43 | 60.9 | Not given | Multiple | RFA | US | 28.6 |
| Hasegawa [15] | 2015 | Japan | 35 | 41 | 64.7 | 3.3 | Multiple | RFA | CT | 30.1 |
| Huang [16] | 2019 | China | 22 | 22 | 53.0 | 4.0 | HCC | RFA | US | 10 |
| Li [17] | 2013 | China | 30 | 30 | 64.2 | 4.2 | Multiple | CA | US | Not given |
| Liu [18] | 2020 | China | 29 | 29 | 52.9 | Not given | Multiple | RFA | US | 24.5 |
| Men [19] | 2016 | China | 31 | 31 | 64.9 | 3.46 | LC | MWA | CT | 11.1 |
| Sui [20] | 2019 | China | 70 | 70 | Not given | Not given | HCC | RFA | CT | Not given |
| Sun [21] | 2010 | China | 23 | 24 | 49.35 | 4.1 | Multiple | MWA | US | Not given |
| Yuan [22] | 2018 | China | 38 | 38 | 54.2 | 3.3 | HCC | RFA | CT | 26.3 |
| Zhang [23] | 2018 | China | 31 | 31 | 57.9 | 3.0 | Multiple | CA | CT | 30.5 |
| Zhang [24] | 2021 | China | 39 | 39 | 58.4 | 3.0 | LC | CA | CT | 33 |
| Zhou [25] | 2018 | China | 33 | 38 | 60 | Not given | Multiple | RFA | CT | 22.8 |
| Zhou K [26] | 2018 | China | 46 | 52 | 64.4 | 3.31 | Multiple | RFA | CT | 11.5 |
Primary technical success
Primary technical success rates were reported in all studies, with a pooled primary technical success rate of 88% (95% CI: 0.85–0.91, Figure 2 A). While these studies exhibited low heterogeneity for this endpoint (I2 = 18.6%, p = 0.246), a high risk of publication bias was detected (Egger’s test: p < 0.01).
Secondary technical success
Secondary technical success rates were reported in four studies [12, 15, 16, 18], with a pooled secondary technical success rate of 93% (95% CI: 0.88–0.97, Figure 2 B). While these studies exhibited low heterogeneity for this endpoint (I2 = 0%, p = 0.508), a high risk of publication bias was detected (Egger’s test: p = 0.005).
Local hemorrhage
Local hemorrhage rates were reported in four studies [15, 19, 25, 26], with a pooled rate of 3% (95% CI: 0.01–0.05, Figure 2 C). These studies exhibited low heterogeneity for this endpoint (I2 = 0%, p = 0.495), and Egger’s test failed to detect any publication bias.
Pneumothorax
Pneumothorax rates were reported in four studies [20, 22, 25, 26], with an overall pooled rate of 6% (95% CI: 0.02–0.09, Figure 2 D). These studies exhibited low heterogeneity for this endpoint (I2 = 0%, p = 0.996), and Egger’s test failed to detect any publication bias.
Hypertensive crisis
Hypertensive crisis rates were reported in 11 studies [12, 14–22, 25, 26], with a pooled rate of 6% (95% CI: 0.03–0.08, Figure 2 E). Low heterogeneity was detected among these studies (I2 = 20.4%, p = 0.249), as well as a low risk of publication bias (Egger’s test: p = 0.935).
Local recurrence
Local recurrence rates were reported in 12 studies [12, 13, 15–19, 22–26], with a pooled local recurrence rate of 19% (95% CI: 0.15–0.23, Figure 2 F). This endpoint was associated with low levels of heterogeneity (I2 = 0%, p = 0.87), as well as a low risk of publication bias (Egger’s test: p = 0.744).
1-year overall survival
Patient 1-year OS was reported in nine studies [13–15, 18–20, 22–24], with an overall pooled 1-year OS of 80% (95% CI: 0.71–0.88, Figure 2 G). This endpoint exhibited high levels of heterogeneity (I2 = 79.6%, p < 0.001), as well as a high risk of publication bias (Egger’s test: p = 0.003).
Meta-regression analyses failed to identify predictors of 1-year OS rates (Table II), and details pertaining to subgroup analyses for this endpoint are provided in Table III. Low heterogeneity was evident in subgroups of patients who had undergone CA (I2 = 0%) and patients with multiple primary cancers (I2 = 0%).
Table II
Meta-regression results
Table III
Subgroup analysis of 1-year OS rates
3-year overall survival
Patient 3-year OS was reported in seven studies [14, 15, 18, 20, 22–24], with an overall pooled 3-year OS of 46% (95% CI: 0.36–0.56, Figure 2 H). This endpoint exhibited high levels of heterogeneity (I2 = 67.1%, p = 0.006), as well as a high risk of publication bias (Egger’s test: p = 0.007).
Meta-regression analyses failed to identify predictors of 3-year OS rates (Table II) and details pertaining to subgroup analyses for this endpoint are provided in Table IV. Low heterogeneity was evident in subgroups of patient who had undergone CA (I2 = 0%), patients who had undergone US-guided PA (I2 = 0%), patients with AMs secondary to HCC (I2 = 0%), and patients with multiple primary cancers (I2 = 0%).
Table IV
Subgroup analysis of 3-year OS rates
Subgroup analyses based on primary LC
Three studies focused on patients with primary LC [13, 19, 24]. The pooled rates of primary technical success, local hemorrhage, hypertensive crisis, local recurrence, 1-year OS, and 3-year OS were 93% (95% CI: 0.88–0.98), 3% (95% CI: –0.03–0.09), 6% (95% CI: –0.02–0.14), 22% (95% CI: 0.14–0.30), 79% (95% CI: 0.71–0.88), and 53% (95% CI: 0.38–0.69), respectively (Table V).
Subgroup analyses based on primary HCC
Three studies focused on the patients with primary HCC [16, 20, 22]. The pooled rates of primary technical success, secondary technical success, pneumothorax, hypertensive crisis, local recurrence, 1-year OS, and 3-year OS were 88% (95% CI: 0.82–0.93), 86% (95% CI: 0.72–1.01), 6% (95% CI: 0.01–0.10), 10% (95% CI: 0.05–0.15), 21% (95% CI: 0.10–0.31), 89% (95% CI: 0.84–0.94), and 60% (95% CI: 0.51–0.70), respectively (Table VI).
Table VI
Meta-analysis based on the patients with primary hepatocellular carcinoma
Discussion
Here, the safety, short-term efficacy, and long-term outcomes associated with imaging-guided PA-based treatment of AMs were assessed. Overall these analyses yielded positive results, with respective pooled primary and secondary technical success rates of 88% and 93%, suggesting that PA can readily achieve instant efficacy when used to treat AMs. While some patients exhibit residual tumors following an initial PA procedure, this approach can be repeated as an efficacious supplementary treatment.
In these analyses, the primary and secondary technical success rates exhibited low heterogeneity, suggesting that they are unlikely to be impacted by the different PA methods, imaging guidance approaches, or primary tumor types in different patients and studies. The most common PA methods at present are RFA, MWA, and CA. RFA and MWA have previously been reported to exhibit comparable complete ablation rates in several tumor types [27–29]. CA can also achieve efficacy similar to that of RFA and MWA in many malignancies [30–32]. CT and US guidance were not associated with any significant differences in PA-associated therapeutic efficacy [33].
Overall, the imaging-guided PA approach was found to be a safe approach to AM patient treatment, as evidenced by the low pooled local hemorrhage, pneumothorax, and hypertensive crisis rates (3%, 6%, and 6%, respectively). Hypertensive crisis is the most critical adverse event to consider when performing the PA-based treatment of adrenal disease, with incidence rates in the range of 15.2–67.5% for patients with primary adrenal tumors undergoing PA, particularly among pheochromocytoma patients [6, 7, 34]. In contrast, PA is much safer when used to treat AMs.
Pooled long-term outcome analyses of patient local recurrence, 1-year OS, and 3-year OS were also conducted. As the pooled local recurrence rate was just 19%, this may suggest that PA can achieve good local control of AMs, and this local control was stable across analyses (I2 = 0%). Frenk et al. [9] have previously reported local AM recurrence to be unrelated to primary tumor histology, tumor size, or PA modality, while Liu et al. [18] found PA to be associated with significantly higher rates of local recurrence relative to adrenalectomy (24% vs. 6.5%, p = 0.048). However, owing to the minimally invasive nature of the PA procedure, it could be repeated following local AM recurrence.
In this meta-analysis, the respective pooled rates of 1- and 3-year OS were 80% and 46%, suggesting that imaging-guided PA represents an effective disease control strategy in patients harboring AMs. Prior work has suggested primary LC to be a risk factor linked to shorter patient OS [9, 15], but meta-regression analyses conducted herein failed to detect any OS-related risk factors in the present meta-analysis. Subgroup analyses revealed lower levels of 1- and 3-year OS endpoint heterogeneity when evaluating patients who underwent CA or had multiple primary cancer types. Over half of the analyzed studies included multiple primary cancer types, highlighting the need for more studies focused on specific cancers.
There are certain limitations to this meta-analysis. For one, all studies were retrospective. In addition, the 1- and 3-year OS endpoints were subject to high levels of heterogeneity. While some subgroup analyses were conducted for these endpoints, other factors with the potential to impact patient OS such as age, gender, tumor stage, tumor size, the extent of disease, and systematic treatment use were not evaluated as the included studies did not provide sufficient information to stratify patient data based on these parameters. Third, many of the included endpoints exhibited evidence of publication bias. Fourth, this study did not include any control group. Consequently, it was not possible to gauge the relative clinical efficacy of PA and other treatments or to compare different PA modalities.
