A real-world study of sirolimus in the treatment of pediatric head and neck lymphatic malformations

Abstract

Objective

Sirolimus has shown promise in treating lymphatic malformations (LMs); however, previous studies were all single-arm clinical trials. Prior studies have reported that some LMs lesions can spontaneously reduce, which poses a potential confounding factor in evaluating therapeutic outcomes. Our study aimed to assess the effectiveness and safety of oral sirolimus objectively for LMs in the head and neck by comparing the outcomes in a control group.

Methods

This real-world study involved 57 children with LMs from 2020 to 2023 at Beijing Children's Hospital. The patients were divided into nonintervention and oral sirolimus groups. The primary outcome was lesion volume change on magnetic resonance imaging, and the secondary outcome was improvement in clinical symptoms. Safety was assessed based on adverse reactions (regular laboratory tests and patient follow-ups) and sirolimus blood concentrations in the oral sirolimus group.

Results

Twenty-one children were enrolled in the nonintervention group (14 male and 7 female), with a median age of 33 months (interquartile range [IQR], 9-53.5 months); 36 children were enrolled in the oral sirolimus group (15 male and 21 female), with a median age of 24 months (IQR, 7.5-51.25 months). The median volume reduction ratio in the nonintervention group was 0.08 (IQR, 0.30-0.40), and eight cases (38.1%) had an effective response. The median volume reduction ratio in the oral sirolimus group was 0.67 (IQR, 0.40-0.92), and 33 cases (91.7%) had an effective response. Clinical symptoms improved in 9 children (42.9%) in the nonintervention group and 35 (97.2%) in the oral sirolimus group. The effective rate and volume changes differed statistically significantly (P < .001). The most common adverse reaction in the oral sirolimus group was an increase in myocardial enzyme levels (n = 23 [63.9%]), followed by oral ulcers (n = 16 [44.4%]). The overall blood concentration of sirolimus was low during the drug administration in the oral sirolimus group, among which 10 children (27.8%) had a mean blood concentration of <5 ng/mL.

Conclusions

Sirolimus significantly decreased lesion volume and improved clinical symptoms in most patients with LMs, especially those with macrocystic components. Conservative observations were only helpful in some patients, with relatively modest volume changes.

Article Highlights.

• •Type of Research: Single-center retrospective cohort study
• •Key Findings: Compared with the nonintervention group, the oral sirolimus group showed a significant volume reduction (0.67 vs 0.08) and a higher objective response rate, defined as a complete or partial radiological response with associated clinical improvement (91.7% vs 38.1%). Differences were considered statistically significant (P < .001). The most common adverse reaction in the oral sirolimus group was an increase in myocardial enzyme levels (n = 23 [63.9%]).
• •Take Home Message: Using sirolimus is safe and feasible for most patients with pediatric head and neck lymphatic malformations, compared with conservative observation, which is only helpful in a minority of patients.

Lymphatic malformations (LMs) are congenital diseases caused by abnormal development of lymphatic vessels during embryonic stages,¹ which are primarily observed in infants and children under two years old and are most common in the head and neck.² The International Society for the Study of Vascular Anomalies classifies LMs into common (cystic) LM, generalized lymphatic anomaly, LM in Gorham-Stout disease, channel type LM, acquired progressive lymphatic anomaly, primary lymphedema, and others.³ Among these, the most common type is common (cystic) LM, which presents as solitary lesions of different sizes and can be confirmed using magnetic resonance imaging (MRI).^4,5 The typical clinical manifestation of LMs in the head and neck is a facial or cervical mass or swelling. Because these patients usually have aesthetic problems, some may only undergo observation without intervention.^6,7

The phosphoinositide 3-kinase catalytic subunit alpha/mammalian target of rapamycin (PI3K/AKT/mTOR) pathway plays vital roles in cell proliferation, growth, angiogenesis, survival, and metabolism in normal cells.⁸ Mutations in PIK3CA in lymphatic endothelial cells within LMs can lead to sustained activation of the PI3K/AKT/mTOR pathway, which is a significant contributing factor to the occurrence and development of LMs.^1,9, 10, 11, 12

Sirolimus, an mTOR inhibitor, binds to mTOR and disrupts its ability to signal to its downstream effectors, inhibiting downstream protein synthesis, cell proliferation, and subsequent vascular formation.13, 14, 15 It has been widely studied in vascular malformations, with evidence supporting its efficacy in treating venous malformations and LMs.¹⁶ In particular, it has been demonstrated to reduce the volume of LMs on imaging and improve clinical symptoms in patients with LMs.17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28 Furthermore, LMs tend to reduce spontaneously.29, 30, 31, 32 Some studies have shown that after 4.5 months (interquartile range [IQR], 4-7 months) of observation in patients with LMs, 12.5% and 50.3% showed lesion regression.²⁹ Bonilla-Velez et al³¹ found that spontaneous regression occurred in 36 of the 65 malformations (58%) that were observed initially. Therefore, with a deeper understanding of LMs, some researchers have proposed a treatment strategy for observation and follow-up that could be adopted for certain patients, particularly those without functional impairment or severe aesthetic damage.30, 31, 32

Previous studies on sirolimus for LMs have primarily been case reports or open-label single-arm clinical trials rather than case-control studies or randomized controlled clinical trials with control groups.33, 34, 35 Considering the potential for spontaneous regression in LMs, it remains unclear whether the decrease in lesion volume or symptom improvement is due to medication or spontaneous regression, limiting the certainty of sirolimus efficacy in treating LMs and its broader clinical adoption. Therefore, this study aimed to evaluate oral sirolimus's therapeutic efficacy and safety in LMs objectively by establishing a control (nonintervention) group.

Methods

Study design

This study used a retrospective design. The medical records of patients with LMs who visited Beijing Children's Hospital Affiliated with Capital Medical University between January 2020 and December 2023 and satisfied the enrollment criteria were collected, including diagnostic evaluation, treatment modalities, imaging data, laboratory examination, and follow-up.

The inclusion and exclusion criteria for this study were as follows. The inclusion criteria were diagnosis of common LMs in the head and neck confirmed by MRI, age <18 years, receipt of either oral sirolimus or conservative observation during the study period, availability of at least two MRI scans, and no other treatment modality received in the 12 weeks before enrollment in this study (except for life-saving treatments). The exclusion criteria were lack of primary efficacy evaluation data (such as MRI before and after observation or treatment), follow-up data, and receipt of other treatment modalities during the study. Patients were categorized into control (nonintervention) and treatment (oral sirolimus) groups based on the different treatment approaches. The control group underwent conservative observation without intervention, and the treatment group received oral sirolimus alone. This study was approved by the Ethics Committee of the Beijing Children's Hospital (BCH-IEC-2022-0129).

Treatment protocols

Both groups of patients underwent routine blood and urine tests, coagulation function tests, biochemical examinations, electrocardiography, and other routine tests before enrollment. In the sirolimus group, the initial dose was 0.8 mg/m² (body surface area), administered orally every 12 hours. Sirolimus blood concentrations were monitored on day 7 after dosing, and routine blood, urine, and biochemical examinations were reviewed. The drug dose was adjusted to maintain the blood trough concentration within the 5 to 15 ng/mL range.

Outcome evaluation

Changes in lesion volume were the primary outcome measures for evaluating the efficacy of observation or treatment. The volume of the LMs was quantified using MRI volume segmentation. The cross-sectional area of each lesion slice was traced manually on T2-weighted MRIs, and the lesion volume was calculated by multiplying the sum of all cross-sectional areas by the slice thickness. The MRI studies were interpreted independently by two radiologists in a blinded manner, with the final lesion volume calculated as the average of their measurements. The volume reduction ratio describes the degree of change in lesion volume.

$$Volumereductionratio=\frac{pre-treatmentvolume\left({cm}^3\right)-post-treatmentvolume\left({cm}^3\right)}{pre-treatmentvolume\left({cm}^3\right).}$$

The secondary efficacy index was the improvement in the clinical symptoms of patients with LMs. The classification of LMs was determined based on the diameter of individual cysts on pre-treatment MRI (macrocystic ≥2 cm; microcystic <2 cm; mixed type; classification in this study only means these three types). Clinical staging was performed by using the De Serres staging system: stage I, unilateral subhyoid; stage II, unilateral suprahyoid; stage III, unilateral suprahyoid and subhyoid; stage IV, bilateral suprahyoid; and stage V, bilateral suprahyoid and subhyoid).³⁶

Referring to previous prospective clinical studies related to vascular anomalies, treatment outcomes were categorized as follows. A complete response (CR) was defined as complete disappearance of the lesion (radiologically) and normalization of quality of life (QoL). A partial response (PR) was defined as a decrease of ≥20% in the size of the vascular lesion and improvement of symptoms or QoL. A stable response (stable disease [SD]) showed no evidence of a response or disease progression. Progressive disease (PD) showed an increase of ≥20% in lesion size and/or decreased QoL assessment.^37,38 CR and PR were categorized as effective, whereas SD and PD were categorized as ineffective.

Safety evaluation

We identified and recorded the clinical symptoms and drug-related adverse events at each outpatient follow-up visit. In the oral sirolimus group, we collected data on changes in relevant laboratory indicators, including routine blood tests, liver and kidney function tests, and coagulation function tests during the medication period. Relevant adverse drug reactions were graded according to the Common Terminology Criteria for Adverse Events.

Statistical analysis

Statistical analyses were performed using SPSS 29.0 (IBM Corp., Armonk, NY). Numerical values are expressed as the median (IQR). Multiple linear regression and binary logistic regression analyses were performed to assess the relationships among the volume reduction ratio, efficacy (effective or ineffective), and factors such as age, sex, baseline volume, treatment method (nonintervention or oral sirolimus), laterality (unilateral or bilateral), and follow-up time (interval between first observation/treatment and follow-up MRI). The Mann-Whitney U test was used to determine whether there was a statistical difference in age, baseline volume, or change in volume between the two groups. The χ² test was used to determine whether there were statistically differences in sex, staging, and previous treatment between the two groups. Interobserver variability was assessed using intraclass correlation coefficients for agreement, with 95% confidence intervals (CIs) calculated via Bland-Altman analysis. Statistical significance was set at P < .05 as significant.

Results

Basic clinical characteristics

Twenty-one patients in the nonintervention group and 36 in the oral sirolimus group were enrolled. An inclusion flowchart is shown in Supplementary Fig 1 (online only). The baseline demographics and clinical characteristics of the patients in the two groups are shown in Table I. The two groups had no statistically significant differences in the baseline data. The interobserver agreement for MRI volume changes was excellent, with an intraclass correlation coefficient of 0.992 (95% CI, 0.987-0.995). The mean difference between the two radiologists was 1.33 cm³ (95% limits of agreement: −3.463 to 6.123 cm³) (Supplementary Table I, online only). The nonintervention group included 14 boys and 7 girls with a median age of 33 months (IQR, 9-53.5 months). The median interval between the first observation and follow-up MRI was 7.37 months (IQR, 5.63-10.93 months). The clinical presentations mainly consisted of painless masses (n = 9), upper airway obstruction (n = 4), recurrent infections (n = 3), pain (n = 2), exophthalmos (n = 2), and facial deformities (n = 1). In the oral sirolimus group, there were 15 boys and 21 girls, with a median age of 24 months (IQR, 7.5-51.25 months). The median interval between the first sirolimus treatment and follow-up MRI was 6.34 months (IQR, 4.80-8.13 months). The clinical presentations were upper airway obstruction (n = 10), facial deformity (n = 8), recurrent infections (n = 5), painless mass (n = 6), dysphagia (n = 3), recurrent infections (n = 2), pain (n = 2), and others (n = 5).

Table I.

Descriptive and demographic data for the study sample

	Nonintervention (n = 21)	Oral sirolimus (n = 36)	P value
Sex			.069
Male	14 (66.7)	15 (41.7)
Female	7 (33.3)	21 (58.3)
Age at presentation, month	33 (9.0, 53.5)	24 (7.5, 51.25)
Interval between first observation/treatment and follow-up MRI, month	7.37 (5.63,10.93)	6.34 (4.80,8.13)
Baseline volume, cm3	17.07 (10.43, 40.54)	38.62 (17.63, 51.57)
Classification			.855
Macrocystic	6 (28.6)	8 (22.2)
Microcystic	5 (23.8)	10 (27.8)
Mixed	10 (47.6)	18 (50.0)
Stages
I	1 (4.8)	0
II	3 (14.3)	4 (11.1)
III	14 (66.5)	12 (33.4)
IV	1 (4.8)	4 (11.1)
V	1 (4.8)	9 (25.0)
Cervical-thoracic	1 (4.8)	7 (19.4)
Clinical symptoms
Upper airway obstruction	4 (19.0)	10 (27.7)
Facial deformities	1 (4.8)	8 (22.2)
Painless mass	9 (42.9)	6 (16.7)
Dysphagia	0	3 (8.3)
Recurrent infection	3 (14.3)	2 (5.6)
Pain	2 (9.5)	2 (5.6)
Othersa	2 (9.5)	5 (13.9)

MRI, Magnetic resonance imaging.

Values are number (%) or median (interquartile range).

^aIn nonintervention group, the others included two cases of exophthalmos. In oral sirolimus group, the others included three cases of abnormal tongue, one case of pleural effusion, and one case of antrum auris discharge.

Analysis of outcomes in the two groups

The situation of the nonintervention group

The median volume reduction ratio in the nonintervention group was 0.08 (IQR, −0.30 to 0.40) (Table II), and eight patients (38.1%) had an effective response (Table III). All eight patients showed PR, with a median volume reduction ratio of 0.53 (range, 0.22-0.79). Among the 13 ineffective patients (61.9%), six patients (28.6%) showed SD and seven patients (33.3%) showed PD, with a median volume reduction ratio of −0.25 (range, −2.31 to 0.14). From the perspective of clinical symptoms, nine patients (42.9%) showed improvement in appearance, snoring, and pain.

Table II.

Comparative statistical analysis of the efficacy (effective or ineffective) between nonintervention and oral sirolimus groups

	Nonintervention		Oral sirolimus		P Value
	Volume reduction ratio	P Value	Volume reduction ratio	P Value
Sex		.799		.170
Male	0.09 (−0.50 to 0.49)		0.74 (0.54 to 0.93)		<.001
Female	0.07 (−0.25 to 0.34)		0.62 (0.34 to 0.85)		.005
Age, months		.414		.226
<24	−0.10 (−0.32 to 0.20)		0.69 (0.52 to 0.94)		<.001
≥24	0.10 (−0.15 to 0.53)		0.59 (0.31 to 0.91)		0.003
Classifications		.779		<.001
Macrocystic	0.03 (−0.11 to 0.70)		0.96 (0.90 to 0.99)		.001
Microcystic	0.30 (−1.31 to 0.40)		0.43 (0.17 to 0.67)		.129
Mixed	0.08 (−0.50 to 0.32)		0.65 (0.42 to 0.82)		<.001
Unilateral/bilateral lesion		.929		.949
Unilateral	0.09 (−0.32 to 0.40)		0.67 (0.40 to 0.88)		<.001
Bilateral	−0.08 (−0.27 to 0.51)		0.67 (0.37 to 0.93)		.023
Stages		–		.488
II	–		0.43 (0.24 to 0.64)
III	–		0.72 (0.48 to 0.97)
IV	–		0.60 (0.21 to 0.94)
V	–		0.67 (0.49 to 0.96)
Cervical-thoracic	–		0.66 (0.23 to 0.89)
Volume reduction ratioa					<.001
Median (IQR)	0.08 (−0.30 to 0.40)		0.67 (0.40 to 0.92)
Range	−2.31 to 0.79		0.06 to 1

Values are median (interquartile range) unless otherwise noted.

^aVolume reduction ratio = [pretreatment volume (cm³) – post-treatment volume (cm³)]/pretreatment volume (cm³).

Table III.

Treatment responses of nonintervention and oral sirolimus group

	Nonintervention (n = 21)				Oral sirolimus (n = 36)
	Macrocystic	Microcystic	Mixed	Total	Macrocystic	Microcystic	Mixed	Total
Effective
CR	0	0	0	0	1 (2.8)	0	0	1 (2.8)
PR	2 (9.5)	3 (14.3)	3 (14.3)	8 (38.1)	7 (19.4)	8 (22.2)	17 (47.2)	32 (88.9)
Total	2 (9.5)	3 (14.3)	3 (14.3)	8 (38.1)	8 (22.2)	8 (22.2)	17 (47.2)	33 (91.7)
Ineffective
SD	3 (14.3)	0	3 (14.3)	6 (28.6)	0	2 (5.6)	1 (2.8)	3 (8.3)
PD	1 (4.8)	2 (9.5)	4 (19.0)	7 (33.3)	0	0	0	0
Total	4 (19.0)	2 (9.5)	7 (33.3)	13 (61.9)	0	2 (5.6)	1 (2.8)	3 (8.3)

CR, Complete response; PD, progressive disease; PR, partial response; SD, stable disease.

Values number (%).

In terms of the classification of LMs, the median lesion volume reduction ratio of macrocystic, microcystic, and mixed LMs was 0.03 (IQR, −0.11 to 0.70), 0.30 (IQR, −1.31 to 0.40), and 0.08 (IQR, −0.50 to 0.32), respectively (Table II). Conservative management was effective in 2 of 6 patients with macrocystic LMs, 3 of 5 patients with microcystic LMs, and 3 of 10 patients with mixed LMs. A comparison of the volume reduction ratios in the three classifications of patients is shown in Fig 1. No statistical difference was observed (P = .799).

Fig 1.

Lesion volume reduction ratio among macrocystic, microcystic, and mixed lymphatic malformations in the nonintervention group. ns, not statistically different.

The situation of the oral sirolimus group

The median volume reduction ratio in the oral sirolimus group was 0.67 (IQR, 0.40-0.92) (Table II). Thirty-three patients (91.7%) had an effective response (Table III), with a median lesion volume reduction ratio of 0.68 (range, 0.2-1.0). Of these effective patients, 32 showed PR and 1 showed CR. Treatment was ineffective in three patients, and all showed SD. In these three cases, one patient showed no change in volume or symptoms during the drug administration period, whereas the other two showed relief in appearance. Notably, one of these two patients showed a clear improvement in symptoms. A 7-month-old girl with large microcystic LMs in the facial region had a hypertrophic, asymmetric, blue-violet tongue (Fig 2, A (i)). Within 2 months of oral sirolimus treatment, although the lesion volume change was not noticeable on MRI (Fig, 2, A (ii), B (ii)), the tongue was smaller and the color was lighter (Fig 2, B (i)). Overall, the clinical symptoms of 35 patients receiving oral sirolimus improved. The treatment was effective in all eight patients with macrocystic LMs, and the median lesion volume reduction ratio was 0.96 (IQR, 0.90-0.99). The effective rate of mixed LMs was second highest (17/18 [94.4%]), and the median lesion volume reduction ratio was 0.65 (IQR, 0.42-0.82). Microcystic LMs had the lowest effective rate (8/10 [80%]), with a median lesion volume reduction ratio of 0.43 (IQR, 0.17-0.67). There were differences in the volume reduction ratios between macrocystic and microcystic LMs (P < .001) and between macrocystic and mixed LMs (P = .01). In contrast, there were no statistically significant differences between microcystic and mixed LMs (P = .355) (Fig 3, A). There were no statistically significant differences in volume changes between the clinical stages (Fig 3, B).

Fig 2.

Clinical response to oral sirolimus in lymphatic malformations (LMs). (A) Tongue photograph (i) and magnetic resonance imaging (MRI) (ii) of a 7-month-old girl with facial microcystic LMs. (B) Tongue photograph (i) and MRI (ii) 2 months after oral sirolimus treatment.

Fig 3.

(A) Lesion volume reduction among macrocystic, microcystic, and mixed lymphatic malformations (LMs) in the oral sirolimus group. (B) Volume changes across different clinical stages in the oral sirolimus group. ∗P < .05, ∗∗∗P < .001; ns, not statistically different.

Efficacy comparison

Through multiple linear and binary logistic regression analyses, we found that the volume reduction ratio and efficacy were not associated with sex, age, baseline volume, laterality, or follow-up time, but were significantly related to the treatment method (nonintervention or oral sirolimus) (Supplementary Tables II and III, online only). We collected MRI data from patients in the nonintervention and oral sirolimus groups from their initial visit to the end of the study (Supplementary Table IV, online only); the results are presented in Supplementary Fig 2 (online only). Compared with the nonintervention group (Supplementary Fig 2, A-H, online only), the oral sirolimus group (Supplementary Fig 2, I-P, online only) showed a more significant reduction in lesion volume before and after the study. The statistical results indicating the degree of volume reduction (P < .0001) confirmed this finding (Fig 4). Additionally, the efficacy in the oral sirolimus group was significantly better than that in the nonintervention group (P < .001) (Fig 5, A). Lesion volume reduction was compared between the nonintervention and oral sirolimus groups across LM subtypes (Fig 5, B). Significant between-group differences were observed in the macrocystic (P = .001) and mixed (P < .001) subtypes, but not in the microcystic subtype (P = .129).

Fig 4.

Volume reduction ratio between the nonintervention and oral sirolimus groups. ∗∗∗∗P < .0001.

Fig 5.

(A) Treatment efficacy rate between the nonintervention and oral sirolimus groups. (B) Volume reduction ratio across different classifications between the nonintervention and oral sirolimus groups. ∗∗P < .01, ∗∗∗P < .001. ns, not statistically different.

Adverse events

In the nonintervention group, some patients with LMs had a further increase in lesion volume during the follow-up period. However, none had clinical symptoms that required timely intervention. Adverse events were frequent in the oral sirolimus group during the follow-up period, with 33 patients (91.7%) experiencing adverse events. As shown in Table IV, the most common adverse reactions observed in terms of clinical symptoms were oral ulcers (n = 16 [44.4%]), followed by recurrent upper respiratory tract infections (n = 10 [27.8%]) and unexplained fever (n = 1 [2.8%]). Among the adverse reactions detected by laboratory tests, elevated cardiac enzyme levels (n = 23 [63.9%]) was the most common, followed by hepatic dysfunction (n = 12 [33.3%]), hypertriglyceridemia (n = 9 [25.0%]), and hypercholesterolemia (n = 5 [13.9%]). According to the Common Terminology Criteria for Adverse Events 5.0, these adverse reactions were grade 1 or 2, and observation or symptomatic management was sufficient. No hospitalizations or specialized treatments were required.

Table IV.

Adverse events in the oral sirolimus group

Adverse event	Oral sirolimus
	Grade 1	Grade 2	Grade 3	Grade 4	Grade 5
Oral ulcers	11 (30.6)	5 (13.9)	0	0	0
Recurrent upper respiratory tract infections	4 (11.1)	6 (16.7)	0	0	0
Hepatic dysfunction	9 (25.0)	3 (8.3)	0	0	0
Myocardial enzyme increased	22 (61.1)	1 (2.8)	0	0	0
Hypertriglyceridemia	9 (25.0)	0	0	0	0
Hypercholesterolemia	5 (13.9)	0	0	0	0
Unexplained fever	1 (2.8)	0	0	0	0

Values number (%).

Blood drug concentration

The dosage in the oral sirolimus group was adjusted based on the sirolimus blood concentration, which was monitored on the day 7 after administration. In this study, the average dose of sirolimus was 0.55 ± 0.25 mg, with the median dose of 0.5 mg (IQR, 0.41-0.63 mg). Among the participants, 15 underwent dose adjustments (2 patients underwent 2 adjustments, while the remaining 13 underwent 1 adjustment). The initial median blood concentration was 7.37 ng/mL (IQR, 4.03-13.20 ng/mL). The median blood concentration during the treatment period was 6.79 ng/mL (IQR, 4.83-9.92 ng/mL). Ten children (27.8%) had mean blood concentrations of <5 ng/mL, but all showed favorable responses to sirolimus.

Discussion

In this study, we evaluated the efficacy and safety of sirolimus in treating LMs using real-world data. This study is the first to establish a nonintervention group. We found that eight patients (38.1%) in the nonintervention group experienced lesion volume reduction and symptom improvement without any treatment measures, which confirms that follow-up observation can be conservative management for patients with LMs. Notably, volume change and symptoms improvement was relatively limited in these patients. In contrast, 33 patients (91.7%) in the oral sirolimus group had an effective response. These patients' symptoms improved significantly, and the adverse reactions were mild, confirming that sirolimus is a safe and effective treatment for LMs. Our findings provide a theoretical reference and basis for understanding the development of LMs and the clinical application of sirolimus in treating LMs.

The effect of sirolimus on volume reduction varies with vascular malformation subtypes.³⁹ An observational-phase randomized clinical trial in low-flow vascular malformations indicated that sirolimus was more effective for LMs than for venous malformations, with simple LM potentially being the optimal indication.⁴⁰ However, there are differences in treatment outcomes among distinct types of LMs. Extensive lesions and higher De Serres stage (predictors of surgical risk and prognosis) were associated with poor responses.^35,41,42 Macrocystic LMs, particularly those located in the lateral neck region, are more likely to undergo spontaneous regression.^{18,22,31,43,44} It has also been found that factors such as age (<2 years), macrocystic LMs, and fewer prior interventions correlate with significant sirolimus efficacy.^18,22,43,44 In our cohort, the within-group analysis revealed that volume reduction differences were dependent on the LM subtype in the sirolimus group; macrocystic LMs exhibited the most pronounced response, followed by mixed LMs, whereas microcystic LMs showed minimal benefit. By contrast, no subtype-dependent differences were observed in the nonintervention group. Between-group comparisons further demonstrated that sirolimus significantly improved outcomes in macrocystic and mixed LMs patients compared with the nonintervention group (P < .01), whereas no difference was observed in microcystic LM patients. However, there were no cases of volume increase in the sirolimus group, implying that sirolimus may have favorable implications for maintaining disease stability in patients with microcystic LMs. Despite the lack of specific guidelines for sirolimus in LMs, our study supports oral sirolimus as an effective treatment for most pediatric LMs, including extensive and older cases, with macrocystic or mixed LMs showing the greatest benefit. For microcystic LMs, combining sirolimus with sclerotherapy and other modalities may be necessary to enhance its efficacy.

Notably, inconsistent classification criteria exist across studies on LMs. In our research, the LM classification (based on single cyst diameter ≥2 cm vs <2 cm) was adopted from prior clinical studies, because this imaging-based classification correlates with significant outcome differences in traditional therapies like surgery and sclerotherapy. Martinez-Corral et al⁴⁵ characterized LMs as macrocystic (large fluid-filled cysts), microcystic (diffuse infiltrative lesions with small lymph/blood-containing vesicles), or mixed (both types). Specific criteria for cyst size were not defined. They used a PIK3CA^H1047R-driven mouse model to confirm that macrocystic and microcystic LMs originated from different periods of mutational activation during the embryonic stage. Therefore, it may be more meaningful to redefine and standardize the classification of LM based on differences in molecular mechanisms and clinical phenotypes in future studies.

Head and neck LMs often traverse the fascial planes and surgical modalities may be associated with significant complications.⁶ As shown in Supplementary Fig 2 (online only), G, in a 7-year-old boy with microcystic LM in the neck, the lesion size spontaneously reduced by 79.34% after a 6-month observation period. After this study, the child underwent surgery, which achieved complete resection of the lesion with a shorter surgical incision and lower surgical risk. Therefore, the significance of conservative observation in patients with mild symptoms and SD cannot be ignored.

Adverse reactions associated with sirolimus in treating LM are currently considered frequent but mild and manageable, consistent with most studies.^40,46,47 In this study, oral ulcers were the most common (44.4%), but they could be effectively reduced by mixing sirolimus with porridge and jam and promptly rinsing the mouth after administration. More than half of the children (63.9%) had a history of mild cardiac enzyme elevations. However, none of the patients exhibited clinical symptoms, and this index showed a decreasing trend or returned to normal during the follow-up period. Nonetheless, routine laboratory tests during sirolimus administration are necessary to detect abnormalities promptly, which may benefit children.

Previous studies have reported that adverse reactions are related to blood concentration levels.⁴⁸ This relationship was insignificant in our study, possibly because of the patients' low blood drug concentrations. Consistent with most studies of sirolimus for the treatment of vascular anomalies, the starting dose of our study was set at 0.8 mg/m²/12 h, and the target level of blood concentration was 5 to 15 ng/mL. It has been suggested that the response rate to low target levels of sirolimus (4-10 ng/mL) is comparable with that found in the literature using higher target levels (10-15 ng/mL).⁴⁹ Therefore, our study adjusted the dose to maintain the lowest target concentration of approximately 5 ng/mL. This factor may account for the mild adverse effects observed in this study.

This study had some limitations that should be considered. First, although the regression analysis indicated that the treatment method was the sole independent predictor of volume changes or efficacy (Supplementary Tables I and II, online only), it is important to acknowledge that more balanced data can enhance the reliability and validity of the results. Our research was conducted as a retrospective study with a controlled design based on a real-world setting, which may introduce some data imbalances, such as differences in sex and baseline volume between the groups. These imbalances reflect clinical realities (eg, parental preferences for proactive treatment in female patients or those with larger lesions). A prospective design with stricter inclusion criteria could mitigate baseline heterogeneity while maintaining clinical relevance. Second, the existing nonintervention group cases were insufficient for comprehensive subgroup analyses, highlighting the need for continued case collection and expanded sample size for future research.

Conclusions

Our study confirms that sirolimus is an excellent treatment option for patients who are highly motivated to undergo interventions and for whom clinical evaluations recommend active treatment. Sirolimus is safe and has been shown to significantly reduce the lesion size and improve symptoms in most patients with LMs. Additionally, follow-up observations may serve as a viable conservative approach for certain cases of LMs, allowing careful monitoring without immediate intervention.

Data availability statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Author contributions

Conception and design: JW, SW, XN

Analysis and interpretation: JW, YM, XZ, YZL

Data collection: JW, YM, NS, XC, YHL, TWL, GZ, SW, XN

Writing the article: JW, YM

Critical revision of the article: JW, YM, XZ, YZL, NS, XC, YHL, TWL, GZ, SW, XN

Final approval of the article: JW, YM, XZ, YZL, NS, XC, YHL, TWL, GZ, SW, XN

Statistical analysis: JW, YM, XZ

Obtained funding: Not applicable

Overall responsibility: XN

JW and YM contributed equally to this article and share co-first authorship.

Funding

Funded by the National Natural Science Foundation of China (82271184), the Capital Health Research and Development of Special Fund (2022-1-2092).

Disclosures

None.