Prospective Observational Study of Pain Severity and Pain Interference Outcomes Following Percutaneous MRI-guided Laser Ablation or Cryoablation for Focal Painful Peripheral, Soft Tissue Vascular Malformations: 12-month Outcomes

Abstract

Background: 

Percutaneous image-guided thermal ablation has emerged as a treatment option for vascular malformations (VM). However, prospective data on intermediate to long-term effectiveness with uniform follow-up are lacking.

Purpose: 

To prospectively evaluate the 12-month pain severity and pain interference outcomes following percutaneous MRI-guided laser ablation and cryoablation of focal painful, peripheral soft tissue VM.

Materials and Methods: 

Participants undergoing clinically indicated MRI-guided laser ablation and cryoablation for focal painful (worst pain ≥4 out of 10 on Visual Analog Scale), peripheral soft tissue vascular malformations (VM) were enrolled in an IRB-approved prospective, observational study after giving informed consent. Participants completed the Brief Pain Inventory questionnaire prior to ablation and 1, 3, 6, and 12-month postablation. Differences in pain severity and interference outcomes from baseline on a 0 to 10 scale were compared using a paired t-test.

Results: 

Seventeen participants (15 female; mean age 28.3 ± 9.3 years) underwent 24 MRI-guided and monitored ablation sessions including laser (n = 21), cryoablation (n = 2), or both (n = 1) for treatment of painful slow-flow (n = 16) or high-flow (n = 1) VM. Median maximal VM diameter was 5.7 cm (range 1.1 to 31.0 cm). Mean (±SD) preablation worst pain score was 7.9 ± 1.4. There was a significant decrease in worst pain at 1-month postablation (–3.5 ± 2.9; P = .0007) that was sustained at 3 (–3.4 ± 3.2; P = .0007), 6 (–4.2 ± 3.6; P = .0003), and 12 months (–3.6 ± 3.9; P = .002), with similar results for average, current, and least pain scores. There was a significant improvement in pain interference outcomes at 12-month postablation: general activity (P = .018), walking ability (P = .008), work (P = .003), sleep (P = .007), and enjoyment of life (P = .033).

Conclusions: 

These data suggest that MRI-guided and monitored thermal ablation is safe and provides early and sustained significant improvements in pain severity and pain interference outcomes during the first year of follow-up in participants with focal painful peripheral soft tissue VM.

Introduction

Vascular anomalies (VA), including both vascular malformations (VM) and vascular tumors (VT), may cause a variety of symptoms such as pain, swelling, coagulopathy, or thrombophilia, high output heart failure, gait disturbance, other musculoskeletal issues, and cosmetic disfigurement.¹ Treatment of VA is often multidisciplinary and aimed at symptom improvement. First-line therapies include percutaneous image-guided sclerotherapy (slow-flow or low-flow VA) or transarterial or transvenous embolization (high-flow VA) depending on VA size, location, tissue composition, and flow characteristics with surgical resection ± reconstruction generally reserved as a second-line therapy.^2–7 Nonetheless, treatment efficacy is variable, often not sustained and may require multiple planned treatments or subsequent repeat sessions for progressive or recurrent symptoms.^2,3 Consequently, for patients with persistent or recurrent symptoms following optimal first-line therapy for focal VA, few additional treatment options have existed.

Over the last decade, percutaneous image-guided and monitored thermal ablative therapies have emerged as a safe and effective treatment option for symptomatic, focal VA—primarily slow-flow venous malformations but also VT, most commonly as an alternative to surgery following failed sclerotherapy or after failed surgery.^8–15 However, existing studies are limited by small sample size, retrospective design, lack of long-term uniform follow-up, heterogenous patient populations, and diverse symptomatology. There is also limited insight into how VA-associated pain may interfere with patient function. A major underlying limitation is the lack of specific standardized tools to quantify VA signs and symptoms.

The Brief Pain Inventory (BPI) is one of the most widely utilized measurement tools for assessing clinical pain and its impact on function.¹⁶ With the BPI, participants rate pain severity and degree to which their pain interferes with common dimensions of feeling and function on a 0–10 point scale. While the BPI has most commonly been studied in the setting of cancer-associated pain, it has also been validated in noncancer associated pain and palliative thermal ablation applications.^17,18

The aim of the present study was to prospectively evaluate the 12-month pain severity and pain interference outcomes following percutaneous MRI-guided and monitored thermal ablation of focal painful, peripheral soft tissue VM.

Materials and methods

The authors report no biomedical financial interests or potential conflicts of interest related to the content of this manuscript. In this Institutional Review Board (IRB)-approved, Health Insurance Portability and Accountability Act (HIPAA)-compliant study, a prospective, single-arm, observational study with participants serving as their own controls was performed to systematically investigate the 1-year pain severity and pain interference outcomes following MRI-guided laser ablation and cryolablation for treatment of symptomatic peripheral soft tissue VM. Participants were enrolled between November 1, 2016, and November 30, 2020. Inclusion criteria included (1) adult participants ≥18 years of age diagnosed with a focal peripheral soft tissue vascular malformation limited to a single anatomic location, (2) persistent or recurrent moderate to severe pain defined as a pain score of ≥4 on a scale of 0 to 10, (3) declined repeat treatment following prior first-line therapies including surgery, percutaneous sclerotherapy, and embolization or refused or were declined first-line surgery, sclerotherapy, or embolization based on VM anatomic considerations or conspicuity on ultrasound (i.e., proportion of stromal component versus vascular channels); (4) had clinical focal pain that corresponded with VM anatomic location on MRI; (5) had a VM that was in a location that was safe to treat with thermal ablation based on neuroanatomic considerations; (6) referred for MRI-guided ablation after review by the institutional multidisciplinary VM clinic; and (7) consent to study participation utilizing Brief Pain Inventory (BPI). Exclusion criteria included (1) mild pain (<4 on a scale of 0–10), (2) nonfocal or widespread pain, (3) clinical pain localization that was discordant from diagnostic MRI findings, or (4) a VM that was not safe to treat with ablation based on neuroatomic considerations. Participants provided written informed consent. Demographics, clinical, imaging, and follow-up data were collected from the comprehensive electronic medical record.

Preprocedure evaluation

Participants were evaluated in the multidisciplinary VA clinic and considered for focal MRI-guided ablation if they met the inclusion criteria noted above. All participants underwent preablation diagnostic MRI at either 1.5T or 3T to assess VM size, location, and flow characteristics before MRI-guided ablation.

Brief Pain Inventory

Participants filled out the BPI at baseline and then at 1-, 3-, 6-, and 12-month postabation. A pain severity score of 0 equals “no pain,” whereas a score of 10 equals “pain as bad as you can imagine.” A pain interference score of 0 equals “does not interfere whereas a score of 10 equals “completely interferes.”¹⁶ BPI was approved for use in this study (study ID 16-004662, nonfunded academic research) for evaluation of painful VA by the Department of Symptom Research Assessment Tools at MD Anderson Cancer Center.

Ablation procedure

All ablation procedures were performed under general anesthesia with participant positioning to facilitate needle placement as previously described.^8,9 The decision to use laser ablation or cryoablation was determined on a case by case basis. Laser ablation is the default procedure for thermal ablation of symptomatic focal VA at our institution.^8,9 It was noted early in our VA ablation experience that some participants experienced significant short-term periprocedural pain in the first 1–2 weeks following cryoablation, particularly for cryoablation of intramuscular VA. This periprocedural pain exacerbation was not observed with laser ablation and therefore laser ablation became our preferred ablation modality when safe to perform. Cryoablation is the preferred ablation modality if the targeted lesion is located near critical adjacent structures (e.g., nerves, tendons, skin) and precise visualization of the ice ball delineating the ablation zone is required for safe and effective treatment.⁸ Needle placement was performed under either MRI guidance at 1.5T (Siemens, MAGNETOM Espree, Erlangen, Germany or Philips Healthcare, Ingenia, Best, Netherlands) or a combination of US guidance (Philips Healthcare, EPIC, Best, Netherlands) and MR guidance. Multiple cryoprobes (Boston Scientific, Marlborough, MA) or a single laser fiber in several positions (Visualase/Medtronic, Minneapolis, MN, or CLS, Copenhagen, Denmark) were used to create overlapping ablation zones to encompass the VM while minimizing ablation of adjacent nontarget tissue.

Cryoablation was performed under continuous MRI monitoring, and the freezing portion of the procedure was stopped when the lesion was encompassed by the ice ball. Laser ablation was performed with proton-resonance frequency shift MR thermometry monitoring every few seconds until the lesion was encompassed by the calculated thermal damage map and there was resultant decreased T2 signal within the lesion.¹⁹

Participants underwent same-day discharge, 24-hour observation, or admission following the procedure depending on degree of postablation pain and recovery from general anesthesia.

Follow-up

Participants were followed in the Interventional Radiology (IR) clinic beginning at 6–12 months after the procedure. Participant self-reported subjective symptomatic improvement was reviewed. Routine follow-up MR imaging was not performed but obtained on an individual participant basis depending on participant self-reported degree of clinical symptom improvement.

Outcome measures

Primary outcome measures were change in (1) worst pain over past 24 hours and (2) average pain from baseline to 1-year follow-up on the visual-analogue scale (score of 0–10) from the BPI. Secondary outcome measures were (1) other pain severity outcomes—change in least pain in the last 24 hours, pain right now; (2) pain interference outcomes—change in degree to which their pain interferes with general activity, mood, walking ability, normal work including housework, relations with other people, sleep, and enjoyment of life); (3) overall percent (%) pain relief; (4) baseline BPI information (i.e., VA treatment history, pain medication usage and pain characterization) as predictors of 12-month change in worst pain score; and (5) complications.

Pre- and postablation MR images, if available, were independently reviewed by a fellowship trained musculoskeletal (MSK) radiologist for maximum VM diameter (cm) of the T2 hyperintense portion of the VM in the T2-weighted (T2W) images and semiquantitatively % enhancement of total VM in the gadolinium-enhanced T2-weighted (T1W) images at baseline (0–25%, 26–50%, 51–75%, or 76–100%) and % decrease in total VM T2 signal in the T2W images and in total VM enhancement in the T1W images at follow-up (0%, 1–25%, 26–50%, 51–75%, 76–99%, 100%) as previously described.⁹ Minor and major complications were defined based on the Society of Interventional Radiology (SIR) Classification System for Complications by Outcome.²⁰

Power/sample size determination

With baseline worst pain ranging from 4 to 10 (mean 6) on a scale from 0 to 10, the goal was to detect a clinically relevant reduction of 2 points (minimum 20% reduction if baseline pain 10 of 10). By conservatively estimating that follow-up pain scores would range from 0 to 10 with an approximate standard deviation (SD) of 2.5 (range/4), this translated into an effect size of 0.8. Thus, with a two-sided alpha of 0.05 and a power of 0.8, 17 participants were required.

Data analysis

Study data were collected and managed using REDCap electronic data capture tools.^20,21 Data were analyzed using JMP 16.0 (Raleigh, NC) and Prism 9.0 (GraphPad Software, Inc., La Jolla, CA). Continuous variables were presented as mean ± SD or median (range) and categorical variables counts (proportions; %). Analysis of the primary end point (change in worst pain in last 24 hours and average pain from baseline to 1, 3, 6, and 12-month year follow-up) was performed by calculating difference in pain score from the baseline and comparing the difference with a one sample (paired) t-test (null hypothesis equal to zero) to test the primary hypothesis that pain scores were reduced from baseline. Other data variables from BPI are reported as descriptive statistics. Association between change in worst pain in last 24 hours at 12-month postablation and baseline prior VM therapy, pain medication history, and pain descriptors were compared using an unpaired t-test. Association between change in worst pain in last 24 hours at 12-month postablation and T2 signal and enhancement decrease of the treated area at follow-up MRI (if available) were compared using an one-way analysis of variance (ANOVA). Change in VM size from baseline to follow-up MRI was compared using Wilcoxon signed rank test. Two-sided P value <.05 was considered statistically significant.

Results

Demographics

Seventeen participants (median age 28.3 years, range 18–45) were enrolled including 15 females (88%) and 2 males (12%) with seventeen VM located in the extracranial facial soft tissues (n = 1), upper arm (n = 3), forearm (n = 1), hand (n = 1), hip/thigh (n = 5), or lower leg excluding the foot (n = 6) (Table 1).

Table 1. - Demographics and Vascular Malformation Characteristics of 17 Participants With 17 Unique Vascular Malformations Undergoing MRI-guided Laser Ablation or Cryoablation

Demographics
Gender	n (%)
Male	2 (12)
Female	15 (88)
Age (median, range)	24.6 (18.3 to 44.6)
BMI (median, range)	26.5 (16.5 to 42.4)
VM characteristics	17 unique VM
Location: n (%)
Face	1 (6)
Upper Arm	3 (18)
Forearm	1 (6)
Hand	1 (6)
Thigh/hip	5 (29)
Lower leg (excluding foot)	6 (35)
Classification n (%)
Slow-flow	16 (94)
Venous (primarily)	15 (88)
Venolymphatic	1 (6)
High-flow AVM	1 (6)
Prior therapy—number of treatments	12 (71)
Sclerotherapy: n (%); (median, range)	10 (83)	3 (1 to 23)
Surgery n (%); (median, range)	6 (50)	1 (1 to 3)
Transarterial embolization n (%)	1 (8)	4 total sessions

AVM, arteriovenous malformation; VM, vascular malformation.

Vascular malformation classification

The VM were classified by a combination of flow characteristics and radiologic appearance (Table 1). Sixteen of the 17 malformations were characterized as slow-flow VM which included venous (n = 15) and venolymphatic (n = 1). One malformation was characterized as a high-flow arteriovenous malformation (AVM) with residual malformation after prior embolization.

Prior VM therapy

Twelve of 17 VM (71%) had previously been treated with sclerotherapy, surgery, and embolization. Ten of 12 VM (83%) had undergone previous sclerotherapy, with a median number of 3 prior sessions (range 1–23 sessions). Six out of 12 (50%) had undergone previous surgical resection, with a median number of 1 prior operation (range 1–3). One of 12 participants underwent prior transarterial embolization (TAE) at an outside hospital for an AVM, with 4 prior TAE sessions (Table 1). No participants had previously received molecular targeted therapies such as sirolimus, apelisib, or trametinib.

Baseline pain characterization

Detailed baseline pain characterization data from the BPI are summarized in supplemental table 1https://links.lww.com/JV9/A36. The median (range) time from diagnosis of the VM was 48 months (2–288 months) with 16 participants (94%) reporting pain at the time of original diagnosis and 17 participants (100%) reporting their current pain due to the VM. Eleven participants (65%) had taken pain medication in the last 7 days and 10 participants (59%) reported requiring some form of pain medication each day including nonsteroidal anti-inflammatory drugs (n = 3; NSAIDs), acetaminophen (n = 2), opioid analgesics (n = 1), gabapentin (n = 3), Lyrica (n = 1), cyclobenzaprine (n = 1), and cannabidiol (CBD) oil (n = 1). Four participants reported having side effects from pain medication including constipation (n = 1), foggy feeling (n = 1), or fatique/tiredness (n = 2). Nonmedication treatments for pain included warm compression (n = 8), cold compression (n = 8), distraction (n = 9), and relaxation (n = 5) techniques. Participants reported mean % overall pain relief from pain treatments and medications at 36% ± 31%. Participants described the character of their pain as follows: aching (n = 15, 88%), throbbing (n = 14; 82%), tender (n = 13; 77%), shooting (n = 12; 71%), stabbing (n = 12; 71%), sharp (n = 10; 59%), exhausting (n = 9, 53%), nagging (n = 9, 53%), miserable (n = 9; 53%), tiring (n = 9; 53%), numb (n = 5; 29%), burning (n = 4; 24%), penetrating (n = 4; 24%), unbearable (n = 4; 24%), and gnawing (n = 2; 12%).

Preablation MRI

Before ablation, the mean maximal VM diameter was 10.0 ± 9.4 cm (median, 5.7 cm; range, 1.1–31.0 cm) on preablation diagnostic MRI. All 17 (100%) VM were predominantly hyperintense on T2-weighted MRI and iso or hypointense on T1-weighted MRI relative to adjacent skeletal muscle. All 17 VM underwent gadolinium-enhanced MRI, % enhancement of the total VM was as follows: 0–25%—1 (5.9%), 26–50%—2 (11.8%), 51–75%—4 (23.5%), and 76–100%—10 (58.8%).

Ablation procedure

Twenty-four total ablation sessions (n = 21 laser, n = 2 cryoablation, n = 1 both laser and cryoablation) were performed for the 17 VM among 17 unique participants (Figures 1 and 2). The median number of ablation sessions was 1 session per VM with a range of 1–3 sessions per VM. Multiple repeat ablation sessions were planned for 6 participants due to large VM size and proximity to critical structures. Details of the ablation sessions are summarized in Table 2.

Figure 1.: MR imaging-guided laser ablation in a 42-year-old female participant with recurrent pain (worst pain 9/10, average pain 6/10) at location of a left lateral thigh intramuscular slow-flow venous malformation after two prior percutaneous sclerotherapy treatments. (A) The T2-weighted preablation MR image shows a hyperintense intramuscular slow-flow venous malformation in the left lateral thigh (yellow arrow). The venous malformation was ablated with multiple pullback activations of the laser fiber. (B) Immediate postablation T2-weighted MR images shows expected postablation edema (yellow arrow). (C) The T2-weighted MR image at 3-year follow-up shows minimal residual hyperintense malformation (yellow arrow). Participant reported no residual or recurrent pain at 1-year postablation (worst pain 0/10, average pain 0/10). MR, magnetic resonance.

Figure 2.: MR imaging-guided laser ablation in a 23-year-old female participant with persistent focal pain (worst pain 6/10, average pain 4/10) at location of a left distal forearm intramuscular slow-flow venous malformation after one prior surgery and three prior percutaneous sclerotherapy treatments. (A) The T2-weighted preablation MR image shows a hyperintense intramuscular slow-flow venous malformation in the distal forearm(yellow arrow). The venous malformation was ablated with multiple pullback activations of the laser fiber. (B) Immediate postablation T2-weighted MR images shows expected postablation edema (yellow arrow). (C) The T2-weighted MR image at 6-month follow-up shows mild residual edema and malformation (yellow arrow). Participant reported intermittent mild residual pain at 1-year postablation (worst pain 4/10, average pain 1/10). MR, magnetic resonance.

Table 2. - Details of 22 Laser Ablation and 3 Cryoablation Sessions for Treatment of 17 Unique Vascular Malformations in 17 Participants

Laser Ablation (n = 22 sessions)
Number of laser fibers (median, range)	1 (1)
Number of ablation cycles (median, range)	4 (1-24)
Total ablation time (minutes) (median, range)	12 (2-48)
Average laser ablation time per cycle (median, range)	2 (1 to 3.7)
Average laser power (watts) (median, range)	12 (6 to 25)
MR thermometry utilized	22 (100)
Cryoablation (n = 3 sessions)
Number of cryoprobes (median, range)* ice seed	1 (1)
Number of freeze thaw cycles (median, range)	3 (2-3)

24 total ablation sessions, numbers add to 25 because one participant had both laser ablation and cryoablation in the same session.

Six treatment sessions (25%) had a same-day discharge and 18 sessions resulted in an overnight observation (<24 hours). No participant required inpatient hospitalization beyond 23 hour observation.

Complications

Two participants experienced minor complications with transient paresthesias, which required no or nominal therapy—one occurred after cryoablation of a hand vascular malformation and the other after laser ablation of an upper extremity vascular malformation. Both participants had complete return of full sensation by 6-month postablation. There were no major complications (SIR Grade C–E).

Pain severity and pain interference outcomes

Baseline and follow-up pain severity and pain interference outcomes data are summarized in Table 3. Mean (±SD) preablation worst pain score was 7.9 ± 1.4 (range 6–10). There was a significant decrease in worst pain at 1-month postablation (–3.5 ± 2.9; P = .0007) that was sustained at 3 (–3.4 ± 3.2; P = .007), 6 (–4.2 ± 3.6; P = .0003), and 12 month (–3.5 ± 3.9; P = .002) (Figure 3A), with similar results for least pain in last 24 hours (Figure 3B), pain on average (Figure 3C), and pain right now (Figure 3D) scores. There was a significant improvement in pain interference outcomes at 12-month postablation: general activity (P = .018; Figure 4A), walking ability (P = .008; Figure 4b), work (P = .003; Figure 4c), sleep (P = .007; Figure 4d), and enjoyment of life (P = .033; Figure 4e). There was a significant improvement in mood at 6-month postablation (P = .003) that was not sustained at 12 months (P = .096). Overall percent pain relief increased significantly 12-month postablation—67 ± 35% (P = .005; Figure 4F).

Table 3. - Baseline and Follow-up Pain Severity and Pain Interference Outomes Among 17 Participants Undergoing MRI-guided Ablation for Focal Painful Peripheral Soft Tissue Vascular Malformation

Pain Severity		Baseline	1-month	P	3-month	P	6-month	P	12-month	P
Worst Pain in Last 24 hours; mean±SD	Group mean	7.9 ± 1.4	4.2 ± 3.2	.0007	4.4 ± 3.4	.0007	3.6 ± 3.7	.0003	4.3 ± 3.8	.002
	Mean difference from BL		–3.5 ± 2.9		–3.4 ± 3.2		–4.3 ± 3.6		–3.6 ± 3.9
Least Pain in Last 24 hours		3.0 ± 2.2	1.1 ± 1.8	.025	1.4 ± 1.7	.022	1.4 ± 1.9	.015	1.4 ± 1.9	.005
			–1.6 ± 2.3		–1.5 ± 2.3		–1.4 ± 2.1		–1.6 ± 2.0
Pain on Average		5.1 ± 1.5	2.0 ± 2.2	.0004	2.4 ± 2.0	<.0001	2.0 ± 2.3	.0002	2.1 ± 2.4	.0002
			–2.9 ± 2.3		–2.6 ± 2.0		–3.0 ± 2.4		–3.0 ± 2.6
Pain Right Now		4.5 ± 2.4	1.8 ± 2.4	.011	2.1 ± 2.0	.003	2.0 ± 2.7	.003	2.4 ± 2.7	.020
			–2.3 ± 2.4		–2.2 ± 2.4		–2.4 ± 2.7		–2.0 ± 2.3
Pain Interference
General Activity		5.6 ± 2.5	2.1 ± 2.8	.007	3.0 ± 3.2	.001	1.9 ± 2.6	.0001	3.2 ± 3.6	.018
			–3.1 ± 3.6		–2.6 ± 2.6		–3.7 ± 2.8		–2.4 ± 2.8
Mood		4.8 ± 2.7	2.7 ± 3.0	.059	3.3 ± 3.0	.091	2.3 ± 3.1	.003	3.4 ± 3.9	.096
			–1.6 ± 3.0		–1.4 ± 3.1		–2.4 ± 2.8		–1.5 ± 3.4
Walking Ability		4.8 ± 3.9	2.4 ± 3.5	.012	2.4 ± 3.3	.040	1.6 ± 2.5	.013	1.4 ± 2.7	.008
			–2.6 ± 3.3		–1.9 ± 3.3		–2.9 ± 4.1		–3.1 ± 4.2
Normal Work		5.9 ± 2.9	2.5 ± 3.5	.002	2.7 ± 3.0	.0004	1.8 ± 2.7	<.0001	2.7 ± 3.3	.003
			–3.0 ± 3.0		–3.1 ± 2.7		–4.1 ± 3.0		–3.2 ± 3.7
Relationships with others		2.5 ± 3.2	1.3 ± 2.1	0.183	1.7 ± 2.8	0.402	1.7 ± 2.9	0.352	2.3 ± 3.2	0.733
			–0.9 ± 2.3		–0.5 ± 2.6		–0.6 ± 2.6		–0.2 ± 2.8
Sleep		5.3 ± 2.5	2.3 ± 2.5	.016	3.0 ± 3.3	.025	2.0 ± 2.7	.0006	2.6 ± 3.8	.007
			–2.6 ± 3.5		–2.1 ± 3.4		–3.2 ± 2.9		–2.7 ± 3.6
Enjoyment of life		5.7 ± 3.4	1.6 ± 2.9	.004	2.6 ± 2.9	.007	2.0 ± 2.9	.002	2.9 ± 3.7	.033
			–3.1 ± 3.4		–2.5 ± 3.2		–3.4 ± 3.6		–2.5 ± 4.5
Pain Relief (%)		31 ± 30%	66 ± 40	.012	64 ± 39	.011	66 ± 40	.007	67 ± 35	.005
			+36 ± 46		+34 ± 47		+36 ± 46		+36 ± 46

Pain severity and pain interference scores are on a 0–10 point scale. A pain severity score of 0 equals “no pain” whereas a score of 10 equals “pain as bad as you can imagine.” A pain interference score of 0 equals “does not interfere” whereas a score of 10 equals “completely interferes.” Pain relief score is on a 0-100% scale. Data are reported as mean ± SD of the group mean (top row) and group mean of difference between time-point and baseline (bottom row). Reported P values were calculated using a one sample (paired) t-test after calculating the intraindividual difference in pain severity, pain interference or pain relief score at 1, 3, 6, or 12 months from the baseline score.

BL, baseline.

Figure 3.: Pain severity scores from the BPI at baseline and 1-, 3-, 6-, and 12-month postablation in 17 participants undergoing MRI-guided ablation for focal painful peripheral vascular malformation. (A) Worst pain in last 24 hours—Mean (±SD) preablation pain score was 7.9 ± 1.4. Significant decrease in worst pain at 1-month (–3.5 ± 2.9; P = .0007), 3-month (–3.4 ± 3.2; P = .007), 6-month (–4.2 ± 3.6; P = .0003) and 12-month (–3.5 ± 3.9; P = .002) postablation. (b) Least pain in last 24 hours—mean (±SD) preablation pain score was 3.0 ± 2.2. Significant decrease in worst pain at 1-month (–1.6 ± 2.3; P = .025), 3-month (–1.5 ± 2.3; P = .022), 6-month (–1.4 ± 2.1; P = 0.015), and 12-month (–1.6 ± 2.0; P = .005) postablation. (C) Pain on average—Mean (±SD) preablation pain score was 5.1 ± 1.5. Significant decrease in worst pain at 1-month (–2.9 ± 2.3; P = .0004), 3-month (–2.6 ± 2.0; P < .0001), 6-month (–3.0 ± 2.4; P = .0002), and 12-months (–3.0 ± 2.6; P = .0002) postablation. (D) Pain right now—Mean (±SD) preablation pain score was 4.5 ± 2.4. Significant decrease in worst pain at 1-month (–2.3 ± 2.9; P = .011), 3-month (–2.2 ± 2.4; P = .003), 6-month (–2.4 ± 2.7; P = .003), and 12-month (–2.0 ± 3.2; P = .020) postablation. Reported P values were calculated using a one sample (paired) t-test after calculating the intraindividual difference in pain severity score at 1, 3, 6, or 12 months from the baseline score. Data are presented as mean ± SEM. ns = P > .05; *P < .05; **P < .01; ***P < .001; ****P < .0001. A pain interference score of 0 equals “does not interfere,” whereas a score of 10 equals “completely interferes.” BPI, Brief Pain Inventory.

Figure 4.: Pain interference and pain relief scores from the BPI at baseline and 1-, 3-, 6-, and 12-month postablation in 17 participants undergoing MRI-guided ablation for focal painful peripheral vascular malformation. Pain interference outcomes included (A) general activity, (B) walking ability, (C) normal work, (D) sleep, and (E) enjoyment of life. A pain interference score of 0 equals “does not interfere” whereas a score of 10 equas “completely interferes.” (F) Overall percent (%) pain relief. Reported P values were calculated using a one sample (paired) t-test after calculating the intraindividual difference in pain severity score at 1, 3, 6, or 12 months from the baseline score. Data are presented as mean ± SEM. ns = P > .05; *P < .05; **P < .01; ***P < .001; ****P < .0001. BPI, Brief Pain Inventory.

There was no significant association between history of prior treatment and change in worst pain in last 24 hours at 12-month postablation (P = .16).

Associations between change in worst pain in last 24 hours at 12-month postablation and baseline pain medication history and pain descriptors are summarized in Table 4. There was a significantly greater reduction in worst pain in last 24 hours at 12-month postablation among participants who at baseline characterized their pain as aching (P = .0002). Conversely, there was a significantly lesser reduction in worst pain in last 24 hours at 12-month postablation among participants who at baseline reported 1) taking pain medication in the last 7 days (P = .029) or (2) requiring pain medication every day (P = .013) or characterized their pain as (1) stabbing (P = .035), (2) burning (P = .010), (3) exhausting (P = .030), (4) tiring (P = .008), (5) numb (P = .005), and (6) unbearable (P = 0.001).

Table 4. - Change in Worst Pain Score in Last 24 Hours at 12-month Postablation by Baseline Pain Medication History and Pain Descriptors

Baseline Pain Medication	Yes	No	P
Did you take pain medications in the last 7 days?	–2.2 ± 3.9	–6.0 ± 2.5	.029
I feel I have some form of pain now that requires medication each and every day.	–1.7 ± 3.6	–6.1 ± 2.9	.013
Baseline Pain Descriptors	Yes	No	P
Aching	–4.2 ± 3.7	1.5 ± 0.7	.0002
Unbearable	0 ± 0.8	–4.6 ± 3.9	.001
Numb	–0.4 ± 1.7	–4.8 ± 3.9	.005
Tiring	–1.3 ± 3.8	–6.0 ± 2.4	.008
Burning	–0.8 ± 0.9	–4.4 ± 4.1	.010
Exhausting	–1.7 ± 4.2	–5.6 ± 2.4	.030
Stabbing	–2.5 ± 3.9	–6.0 ± 2.9	.070
Penetrating	–1.5 ± 2.1	–4.2 ± 4.2	.12
Tender	–3.2 ± 3.9	–4.8 ± 4.3	.54
Shooting	–3.1 ± 2.9	–4.6 ± 6.1	.62
Nagging	–3.0 ± 3.2	–4.0 ± 5.4	.69
Miserable	–3.2 ± 3.7	–3.9 ± 4.4	.74
Throbbing	–3.4 ± 4.1	–4.0 ± 4.0	.84
Sharp	–3.4 ± 4.2	–3.7 ± 3.9	.88
Gnawing	–4.0 ± 4.2	–3.5 ± 4.0	.89

Data are reported as mean ± SD of the intraindividual difference in worst pain score at 12-month postablation from baseline by pain medication history or pain descriptor (yes vs. no). Reported P values were calculated using a two sample t-test.

Postablation MRI

Postablation MRI was performed in 11 of 17 participants, with 10 of 11 performed with gadolinium contrast injection at a mean imaging follow-up of 25.8 ± 10.6 months. On follow-up MRI, the mean maximum dimension of the VM was 7.3 ± 5.7 cm (median 4.8 cm, range 2.2–20.0 cm) with a small but significant reduction in the maximum size from baseline (–1.4 ± 2.1 cm; P = .004). Among 11 participants with follow-up MRI, % decrease total VM T2 signal was as follows: 0%—0 (0%), 1–25%—5 (45.5%), 26–50%—0 (0%), 51–75%—2 (18.2%), 76–99%—5 (36.3.1%), and 100%—0 (0%). Among the 10 participants with follow-up gadolinium-enhanced MRI, % decrease total VM enhancement was as follows: 0%—1 (10.0%), 1–25%—4 (40.0%), 26–50%—0 (0%), 51–75%—2 (20.0%), 76–99%—3 (30.0%), and 100%—0 (0%). There was no significant association between change in worst pain in last 24 hours at 12-month postablation and T2 signal (P = .15) or enhancement (P = .65) decrease of the treated area at follow-up MRI (if available).

Discussion

The present study suggests that MRI-guided and monitored thermal ablative therapies are safe and provide early and sustained significant improvements in both pain severity and multiple pain interference outcomes out to 1 year in participants with focal painful peripheral soft tissue VM. These prospective data with multidimensional evaluation of pain severity and pain interference outcomes using the BPI add to the growing body of literature supporting the safety and effectiveness of image-guided thermal ablation as a treatment option for symptomatic VA.^8–15 Importantly, there was no significant association between change in pain severity at 12-month postablation and prior VM therapy, further suggesting that ablation may be an effective treatment option in both a first or second-line setting. Overall these prospective data are in agreement with prior data reported by Augustine et al. in a retrospective study of MRI-guided laser or cryoablation for symptomatic peripheral VA with non-standardized pain assessment follow-up.⁹ That prior study reported a mean 5.7 ± 1.0 point reduction in pain score on the visual analog scale with an average pain of 1.6 ± 1.8 at 12.2-month follow-up, similar to the mean 2.1 ± 2.4 pain score reported at 12-month follow-up in the current study.⁹”

Predictors of pain response from baseline BPI

Prior studies have demonstrated wide ranging, variable effectiveness of surgical and interventional therapies for symptomatic VA.^2,3,9 However, predictors of pain response to MRI-guided laser and cryoablation have been lacking. In the present study, participant self-reported character of pain at baseline was associated with better (aching) or worse (numb, burning) improvement in pain to MRI-guided laser ablation or cryoablation. Aching pain may be associated with a flow and edema related mechanism whereas paresthesias may be suggestive of a neuropathic component. Additionally, recent need for pain medication and daily need for pain medication at baseline were associated with less improvement in pain following MRI-guided thermal ablation. Furthermore, participants who described their pain as unbearable, tiring, or exhausting at baseline had significantly worse improvement in pain at 12 months, thereby providing insight into the psychological impact of the VM-associated pain in this participant population. Although the mechanism of VM-associated pain as well as the pain relief following thermal ablation are not well understood, pain medication history and pain characterization may be predictors of response to percutaneous image-guided thermal ablative therapies for symptomatic, painful peripheral soft tissue VM, and warrant further investigator. Taken together, incorporation of the BPI into routine clinical practice as a longitudinal VA assessment tool may be useful (1) to better characterize participant pain severity and pain interference, (2) to identify participants who may be more likely to benefit from ablation, and (3) to identify participants who should be evaluated for other surgical, interventional and medical therapies for painful VM.

Ablation candidacy—Mutlidisciplinary participant evaluation

Holistic evaluation of a patient with a painful peripheral soft tissue VM by a multidisciplinary VA team is critical. In particular, participants should be screened for comorbid conditions such as a chronic pain syndrome and major depression, particularly given the often long-standing history of VM-associated pain, and referred for consideration of pain rehabilitation program and/or psychiatric care prior to or concurrently with ablation therapy. In our practice, in order to be a candidate for percutaneous image-guided ablation, participants must (1) have been reviewed by our multidisciplinary VA team and deemed to be a better candidate for ablation than other therapies such as sclerotherapy or surgical resection and (2) have focal pain on exam that corresponds with the anatomic location of the vascular malformation on MRI.^8,9 Participants with diffuse pain and discordant clinical physical exam and imaging findings are not offered focal ablation. Furthermore, setting treatment goal expectations upfront is key. While some participants had complete pain relief, this was not universal. The present data demonstrate an approximate 45% and 60% reduction in worst and average pain from baseline at 1 year following ablation on average, respectively. Furthermore, there was an approximate doubling of reported overall pain relief with an overall ~70% pain relief at 1-year following ablation. These data can be helpful while counseling participants prior to ablation. Of note, thermal ablation is generally not considered for treatment of peripheral high-flow AVMs. In the current study, a single patient had a history of a peripheral tissue infiltrative intramuscular high-flow AVM (Yakes type IV) that had undergone four prior transarterial embolization procedures at an outside hospital.²² This AVM was considered for ablation based on the residual focal tissue infiltrative soft tissue component of the AVM and marked reduction in flow due to prior transarterial embolization. Ablation of a Yakes type IV AVM in the setting of reduced flow from prior therapy is a reasonable consideration if there is residual reduced flow soft tissue component. Nonetheless, ablation should not be considered (1) as a first-line therapy in any high-flow AVM, (2) as a second-line therapy in any AVM with persistent high flow due to heat sink (laser) or flow-mediated warming (cryoablation) that would reduce the efficacy of ablation, or (3) in any high-flow AVM without a dominant infiltrative soft tissue component, namely Yakes I-III AVMs.²²

Role of pre- and postablation MRI

Preablation MRI is necessary for both clinical correlation with physical exam findings and treatment planning. However, follow-up MRI after ablation is not mandatory in our practice.⁹ In the current study, MRI-guided and monitored laser ablation and cryoablation resulted in a small, but statistically significant reduction in the size of the VA at follow-up, and at least half of the cohort demonstrated 50% or greater reduction in T2 signal intensity and enhancement of the treated portion of the lesion on postablation MRI. However, there was no significant association between change in pain score and follow-up T2 signal or enhancement. As noted by the independent diagnostic radiologist reviewing the available follow-up MR images, postprocedural T2 signal hyperintensity and enhancement often took months or years to resolve. Moreover, difference in lesion size can be difficult to measure due to variable signal intensity, enhancement, and the discontinuous and multidirectional nature of the lesions. As such, participant self-reported symptomatic improvement is the primary clinical outcome measure and driver of subsequent clinical decision making, not imaging findings. Nonetheless, there is an inherent risk of bias regarding the follow-up MRI data as patients may be more likely to have a follow-up MRI if persistent or recurrent symptoms, suggestive of therapeutic failure. Postablation MRI may be helpful if there are persistent or recurrent symptoms, planned staged procedure, secondary clinical pathology, or concern for complications as previously described by Augustine et al.⁹ In short, post-procedure MRI should not be obtained unless the patient’s pain is either not improved or worse following ablation.

Limitations

There are limitations to this study. This was a single center, single-arm study with a relatively small number of participants with varying prior therapy histories. While further studies are needed to better understand when ablation may be considered in the first-line setting, ablation should not be considered mutually exclusive with other VA therapies as previously described by Augustine et al.⁹ This multimodal team approach remains key and tailoring different treatments to different aspects of a participants’ VA over time should be considered. Next, this study was limited to adult participants given the need for use of the complex BPI. Nonetheless, ablation for treatment of painful VA has been used in the pediatric population and pain outcomes should be further systematically investigated in the pediatric population.^8,9,14 Furthermore, the study was performed at a large tertiary medical center with a dedicated interventional MRI practice which may not be widely available. Nevertheless, US and CT-guided ablative procedures, which are more commonly available, may be appropriate alternatives in experienced centers. Next, laser ablation is our preferred ablation modality of choice due to our previous experience demonstrating less immediate pain and equal efficacy to cryoablation.⁹ Our long-standing experience with laser ablation has shown that it offers advantages in the ability to tailor ablation power settings and time to the specific vascular malformation, as well its compatibility with use of intraprocedural MR thermometry monitoring. However, numerous ablative modalities have been used successfully for treatment of symptomatic VA including laser ablation, cryoablation, high-intensity focused ultrasound (HIFU) and radiofrequency ablation (RFA).^8–15 Ultimately, the authors recognize that access to an interventional MRI may not be feasible; nonetheless, US or CT-guided cryoablation are feasible alternatives.¹⁴ Moreover, the transient increase in periprocedural pain observed with cryoablation over laser ablation is easily managed with nonopioid and if needed opioid analgesics ±a short course of steroid therapy. While comparative effectiveness data may be helpful, ultimately whichever imaging-guidance, monitoring technology (US, CT, MRI), and ablative device the proceduralist is comfortable with may be sufficient for undertaking treatment of focal painful VA with ablation. Furthermore, while the present study identified pain medication history as a potential predictor of pain response to ablation, only a portion of the cohort reported baseline pain medication usage, the type of analgesic was variable and the quantitative amount of pain medication used was not evaluated. Future studies should seek to assess quantitative changes in pain medication usage, particularly among participants using opioid analgesics given the risk of long-term dependence.²³ Additionally, this was not a double-blind placebo controlled trial and improvement in pain could be related in part to placebo effect as evidenced in other interventional studies such as vertebroplasty for painful osteoporotic spinal fractures.²⁴ Finally, the follow-up period for the BPI was limited to 12 months in this study. Nonetheless, ongoing serial follow-up will help establish longer term 3 and 5 year outcomes in this participant population.

Conclusion

In conclusion, these prospective data suggest that MRI-guided and monitored thermal ablation is safe and provides early and sustained significant improvement in pain severity and pain interference outcomes during the first year of follow-up in adult participants with focal painful peripheral soft tissue VM in both the first- and second-line therapy setting. Baseline pain medication history and participant self-reported vascular malformation-associated pain characterization may be predictors of response to ablation and warrant further investigation. Finally, follow-up MRI should not be routinely obtained unless the patient’s pain is either not improved or worse following ablation.

Acknowledgments

We acknowledge Rickey E. Carter, PhD, Professor of Biostatistics and Consultant, Division of Clinical Trials and Biostatistics, Department of Quantitative Health Sciences, Mayo Clinic Jacksonville, Florida, for his consultation regarding study design and statistical analysis.