Percutaneous image-guided cryoablation of venous malformation and fibro-adipose vascular anomaly: prognostic factors of clinical efficacy.

The median follow-up period was 13 months (IQR:6.00–29.5 months), and 11 patients (20%) underwent multiple sessions. The clinical effectiveness rate was 72%, with 40 out of 55 patients reporting partial or complete pain relief at the last follow-up. Complete pain relief was achieved in 50% (28 patients). Pre-therapeutic pain scores ranged from 10 to 100, with a mean of 59, while post-therapeutic pain scores ranged from 0 to 100, with a mean of 21.9 and a median of 0 (IQR: 0–37.5). This difference was statistically significant ( p < 0.0001). Technical success was achieved in 100% of cases. Technical efficacy was 69% (38 patients), with 33% ( n = 18) of patients experiencing a complete response and 36% (20 patients) experiencing a partial response, and with a 47% average reduction in lesion size (IQR: −31 to −88%). An unsuccessfully treated lesion was observed in 31% (17 patients) of cases. Thirteen patients (23%) experienced a recurrence of symptoms during the study follow-up without requiring additional intervention. Female patients reported significantly higher levels of post-therapeutic pain compared to male patients ( p = 0.013). Goyal grade 1 VMs were associated with significantly lower post-therapeutic pain compared to other grades ( p = 0.029). No statistically significant differences were observed in the location of the VM, the performance of pre-therapeutic sclerotherapy and/or surgery, or subcutaneous or intramuscular topography. Pre-therapeutic diameter ( p = 0.014), post-therapeutic diameter ( p = 0.013), and post-therapeutic volume ( p = 0.012) significantly correlated with post-therapeutic pain reduction ( Table 2 ). Eight patients experienced worsening post-therapeutic pain at the 6-month assessment. Among these patients, three had Goyal grade 1, two had Goyal grade 2A, and three had Goyal grade 3 VM. In one patient, post-therapeutic periostitis resolved within 12 months, and three patients demonstrated a complete therapeutic response on MRI. Multivariable analysis assessed factors influencing post-therapeutic pain scores. Significant predictors included Goyal grade ( p = 0.035), post-therapeutic volume ( p = 0.048), and completeness of treatment ( p = 0.029). Other variables such as sex ( p = 0.324), pre-therapeutic diameter ( p = 0.346), and pre-therapeutic VAS score ( p = 0.382) were not statistically significant ( Table 3 ). Among the 55 patients in this study, 27% (15 patients) reported adverse events ( Table 4 ), with 9% (5 patients) experiencing severe complications (SIR grade C–E; CIRSE grade 3–5). All five patients with severe complications exhibited neurological impairments, of whom four were transient and one was permanent (sciatic injury). No fatalities were reported.

This single-institution retrospective study received approval from the institutional review board, and the requirement for informed consent was waived. This study included all patients who presented with symptomatic (pain or discomfort) VMs and FAVAs, and who were treated with CA from January 2011 to January 2023. A total of 55 patients were retrospectively included: 43 females (78.2%) and 12 males (21.8%). The ages of the patients at the time of inclusion ranged from 12 to 64 years, with a mean age of 31.6 years and a median age of 30.0 years (interquartile range (IQR): 22–40 years). Pain scores prior to treatment ranged from 70 to 100 (IQR:40.0–80.0), with an average score of 60.2 (standard deviation: 27.0). Table 1 summarizes the patient characteristics. Prior treatments included sclerotherapy and/or surgery for 58% ( n = 32) of patients, while 42% ( n = 23) of patients had not received any prior treatment. Exclusion criteria included capillary or lymphatic malformations, vascular tumors such as hemangiomas, and prior treatment for high-flow arteriovenous malformations with a residual venous component. The decision to offer CA as either first- or second-line therapy was determined during multidisciplinary team meetings. The diagnosis of VM and FAVA was primarily based on clinical presentations and supported by radiological evidence for most patients; however, only three patients had their diagnoses confirmed histologically. In total, 51 VMs and 4 FAVAs were identified based on Doppler ultra-sound (US) and magnetic resonance imaging (MRI) characteristics. MRI features of VMs typically include T2-hyperintense and T1-hypointense lesions with heterogeneous post-contrast enhancement. In contrast, FAVA is characterized by a predominant solid component with associated phlebectasia, a moderately hyperintense signal on T2-weighted images, and moderate to strong homogeneous post-contrast enhancement [ 25 ]. The maximal diameter of the VMs ranged from 11 to 119 mm, with a mean diameter of 59.2 mm. The volumes of the VMs varied from 0.3 to 1003 cm 3 , with a mean volume of 43.0 cm 3 and a standard deviation of 139. Vascular malformations were assessed using Doppler US and MRI. Initial evaluations used multiple scanners, with some examinations including MRIs performed externally. Imaging was primarily conducted using GE Discovery 3-T MR750 (General Electric) and Diamond Select Achieva 1.5-T (Philips) scanners. The imaging protocol included T1-weighted sequences (repetition time (TR): 409–765 ms; echo time (TE): 7.8–11 ms) before and after the administration of contrast medium (Gadovist; Bayer Healthcare) at a dose of 0.1 mL/kg with fat saturation, and T2-weighted sequences (TR: 4440–6340 ms, TE: 96–100 ms) with fat saturation in three basic planes. The field of view was set at 23 × 23 cm, voxel size ranged from 0.5 × 0.5 × 3.0 mm to 0.8 × 0.8 ×4.0 mm, and the flip angle varied between 90 and 150°. Follow-up monitoring of all patients was performed using MRI with contrast injection. Two radiologists classified the VMs according to Goyal’s 2002 classification [ 9 ] ( Fig. 1 ). The diagnosis of FAVA primarily relied on ultrasonographic and MRI findings, notably the presence of fat content along with venous dysplasia [ 25 ]. Histology was reserved for cases where the vascular mass appeared atypical or nonspecific to confirm the diagnosis. The Puig classification [ 25 ] was noted in patients who underwent sclerotherapy, which included 21 patients experiencing either temporary clinical effects or no improvement, classified as grade 1 (1 patient), grade 2 (6 patients), grade 3 (11 patients), and grade 4 (3 patients). Following an initial consultation with an interventional radiologist, patients attended a pre-anesthesia consultation and were admitted to the hospital for a period of 1–2 days. CA procedures were performed under US guidance and fluoroscopy by four interventional radiologists with 10, 10, 20, and 30 years of experience, respectively. The technique, previously described in references [ 17 , 18 ], is summarized below. CA was performed using cryoprobes (Boston Scientific Medical) connected to a generator and argon/helium gas tanks. For superficially located lesions, cryoprobes were inserted along their longitudinal axis, parallel to the skin. A minimum safety distance of > 5 mm was maintained between the cryoprobes and the skin or any nearby nerves to allow for hydrodissection with saline and to minimize the risk of complications. To fully encompass the lesion lengthwise, the initial CA involved complete insertion of the cryoprobe. A second session of CA was optional and could be performed by either retracting the probe along its axis or repositioning the probes if needed. The procedure involved freezing the tissue, followed by a passive thaw and an optional brief active thaw. Two cycles of freezing were standard, with a third cycle left to the discretion of the operator based on factors such as the lesion’s location, morphological characteristics, and the volume of the ice ball formed. Ice growth was monitored in real time using US, and adjustments to the duration of freezing or system power were made accordingly to ensure comprehensive treatment of the target lesion while protecting surrounding skin and critical structures such as nerves. Following the completion of CA, the probes were removed either during a passive or an active thaw phase. Patients were monitored in the recovery area for 2–4 h before being discharged the following day. Clinical efficacy was defined by the reduction of post-therapeutic pain, assessed either at the 6-month follow-up or during subsequent appointments. Pain levels were measured using a visual analog scale (VAS) and were systematically recorded at baseline (M0) and at 6 months (M6), as well as during optional follow-up visits. Patients experiencing pain recurrence were advised to seek consultation. Technical success was achieved when the lesion was treated according to the protocol and fully covered. Technical efficacy was determined by an objective response, evaluated using the RECIST criteria, with partial or complete response indicated on MRI ( Fig. 2 ). A lesion was considered unsuccessfully treated if it remained stable or showed progression on MRI ( Fig. 3 ). The initial MRI assessment occurred 4–6 months after ablation but prior to the 6-month visit, focusing on evaluating primary effectiveness. Subsequent MRI scans, based on clinical follow-up results, aimed to assess local effectiveness and detect any signs of local tissue progression. Lesion measurements were taken in three planes using T2-weighted sequences both before and after CA. Pre- and post-therapeutic volumes were calculated using the ellipse method for ovoid lesions and manually on a slice-by-slice basis for flat lesions. Complications were recorded following the criteria set by the Society of Interventional Radiology (SIR) [ 26 ] and the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) [ 27 ]. Collection of imaging and clinical outcome data continued until the end of May 2023. Univariable analyses were performed to assess variables such as pre- and post-therapeutic pain, technical success, radiological response, sex, Goyal grade, lesion location, pre-therapeutic treatment (sclerotherapy and/or surgery), lesion topography (subcutaneous and intramuscular), and lesion type using the Mann–Whitney test. Spearman coefficients were calculated to correlate post-therapeutic pain with pre- and post-therapeutic diameters, volumes, and pre-therapeutic pain levels. Multivariable linear regression was used to examine the relationship between post-therapeutic VAS scores and factors including pre-therapeutic diameter, prior treatments, completeness of treatment, and pre- and post-therapeutic volumes. Data was checked for multi-collinearity using the Belsley–Kuh–Welsch technique [ 28 ]. The Breusch-Pagan test and the Shapiro-Wilk test were used to assess heteroskedasticity and the normality of residuals, respectively. A p -value of ≤ 0.05 was considered statistically significant. Patients with missing data were excluded from the analysis.

This study demonstrated a 72% clinical efficacy (pain relief) rate for VM and FAVA through CA, with half of the patients achieving complete pain relief. These results are consistent with previous studies that examined the short- to mid-term effects of CA on VM and FAVA, either as a first-line treatment or after unsuccessful sclerotherapy [ 17 – 23 ]. A systematic review [ 29 ] reported a 92% reduction in lesion size, a 77% reduction in pain, and a 63.6% complete response rate, noting that patients who underwent prior sclerotherapy had lower pre- and post-procedural pain scores. Multivariable analysis identified residual tissue volume as a key indicator of clinical failure CA, highlighting the importance of achieving complete lesion ablation to maximize therapeutic success, consistent with findings from sclerotherapy. Achieving this may require close monitoring and multiple treatment sessions. Precision in treatment may be enhanced by selecting appropriate needles or clusters of needles [ 30 – 32 ]. Additionally, combining sclerotherapy with CA may be a viable approach for treating larger lesions. Univariable analysis suggests that the severity of the malformation, as indicated by its Goyal grade, could influence treatment outcomes. However, factors such as the location of the VM, prior treatments (sclerotherapy and/or surgery), and lesion topography did not significantly impact post-therapeutic pain levels, nor did the pre- or post-therapeutic volume. This indicates that the nature of the malformation itself, and its response to CA, rather than its size or location, are likely the primary determinants of treatment outcomes. A deeper understanding of the mechanisms driving the effectiveness of CA, such as endothelial injury, thrombus formation, or cryoneurolysis, could enhance treatment optimization. The intralesional injection of various sclerosing agents, such as bleomycin, ethanol, or aetoxysclerol, remains a standard treatment for VM, with repeated procedures often necessary to achieve clinical improvement. Horbach et al [ 33 , 34 ] found no significant predictors of clinical improvement but noted that impairment in work- or study-related activities prior to sclerotherapy negatively affected the physical quality of life at follow-up. Yun et al reported that being female and the absence or delayed visualization of drainage veins, along with a well-defined margin on MRI, were independent predictors of a “good response” to ethanol sclerotherapy [ 35 ]. Lee et al [ 36 ] found that age < 18 years and overlying skin conditions were associated with cutaneous complications post-sclerotherapy. To improve the efficacy of bleomycin, irreversible electroporation Bleomycin ElectroScleroTherapy (BEST) [ 37 ] has been used and demonstrated good clinical efficacy, [ 38 ] achieving an 86% reduction in VM volume. Schmidt et al [ 39 ] reported significantly improved outcomes regarding mobility, aesthetic aspects, and pain in children and adolescents, with an 8.9% incidence of major complications such as skin hyperpigmentation, necrosis, blebs, and local superinfection. Endovenous or percutaneous laser therapy has shown good clinical efficacy [ 40 , 41 ] with 65–97% clinical improvement; however, no prognostic factors were reported, with a 7.89% incidence of paresthesia [ 40 ]. Adverse events associated with CA were observed in 27% of patients, with neurological impairment being the most severe (5 Goyal grade 3). While most neurological symptoms were transient (4/5 patients), one patient experienced persistent post-therapeutic sciatica. Fujimara et al [ 41 ] reported an 11% incidence of neurological lesions. Although careful patient selection and procedural vigilance are necessary, the overall safety of CA in managing VM and FAVA remains promising [ 13 , 22 , 42 ]. This study has several limitations, including its retrospective design, which might introduce bias due to the subjective nature of clinical response assessment. The sample size may also limit the generalizability of the findings, particularly concerning specific FAVA. Only a few biopsies were performed (7%), and histological analysis often yields nonspecific results for vascular malformations [ 43 – 45 ]. In conclusion, this study offers valuable insights into the prognostic factors associated with percutaneous CA of VM and FAVA. Complete lesion treatment was identified as a significant predictor of clinical outcomes, emphasizing the importance of achieving comprehensive iceball coverage of lesions. Further research, including prospective studies with larger cohorts and longer follow-up periods, is essential to validate these findings and optimize patient selection and treatment strategies for VM management.

Percutaneous sclerotherapy is often the first-line treatment for venous malformations (VMs) due to its efficacy and relative simplicity [ 1 – 4 ]. Despite its advantages, the incidence of local complications can reach up to 41%, with an overall morbidity and mortality rate of 0.8% [ 5 – 8 ]. Sclerotherapy also carries systemic risks due to the potential diffusion of the sclerosing agent [ 9 ]. Additionally, sclerotherapy may prove challenging and ineffective for VMs that have minimal vascular components, such as fibro-adipose vascular anomalies (FAVAs) [ 10 ]. Cryoablation (CA) is a minimally invasive procedure that uses extreme cold to ablate soft tissues, including conditions such as desmoid tumors and abdominal wall endometriosis [ 11 – 16 ]. CA is considered an alternative to sclerotherapy, being capable of targeting both the vascular and tissue components of VM [ 17 ]. Since its first documented use in 2013, numerous studies have highlighted the short- to mid-term efficacy of CA for treating VM, whether as a primary or secondary option following unsuccessful sclerotherapy [ 17 – 23 ]. Factors such as small lesion size, low level of infiltration, and poor drainage have been linked to positive outcomes in sclerotherapy [ 9 ]; therefore, identifying similar predictive factors is important to enhance patient selection and improve the success rates of CA [ 24 ]. This study aims to retrospectively assess the prognostic factors for clinical and radiological responses to percutaneous CA in patients with VM and FAVA.