Introduction

Infantile hemangioma (IH) is the most common pediatric vascular tumor, with a reported incidence between 2 and 10%.1,2 The liver is the most common extracutaneous location for IH, followed by the gastrointestinal tract, brain, mediastinum, and lungs.3 A subgroup of patients with infantile hemangiomas exhibits associated structural anomalies of the brain, cerebral vasculature, eyes, aorta, and chest wall in the neurocutaneous disorder named PHACE syndrome (posterior fossa anomalies, hemangioma, arterial lesions, cardiac abnormalities/coarctation of the aorta, eye anomalies).4,5

According to the revised diagnostic criteria proposed in 2016,6 the diagnosis of definite-PHACE requires either the presence of a facial hemangioma >5 cm in diameter, plus one major or two minor criteria or a hemangioma of the neck, upper trunk or trunk and proximal extremity plus two major criteria. Alternatively, the diagnosis of possible-PHACE can be made without meeting all the aforementioned criteria. If a patient lacks the presence of a cutaneous hemangioma, a diagnosis of possible-PHACE can be made if two major criteria are met (Supplement A). Characterization of clinical features of PHACE has been made by frequency across the following organ systems: arterial, structural brain, cardiovascular, ocular, midline, neurologic signs and symptoms, endocrine, hemangioma-related complications, and miscellaneous. Hemangioma-related complications including impairment of the visual axis, stridor, and ulceration are considered “less common,” whereas gastrointestinal (GI) bleeding from hemangiomas is considered “rare” (Supplement B).6

We hypothesize that, despite the lack of cutaneous involvement, the presentation of our patient is more consistent with findings seen in patients diagnosed with definite- or possible-PHACE when compared to patients not diagnosed with PHACE. In this systematic review, we aim to determine: (1) how often patients with infantile intestinal hemangioma (IIH) exhibit diagnostic and/or clinical features of PHACE and (2) the current treatments and outcomes for IIH.

Case report

A 3-month-old female with a prenatal diagnosis of right aortic arch, aberrant left subclavian artery (ALSA), ventricular septal defect (VSD), and patent foramen ovale (PFO) presented to the emergency department with 3 weeks of pallor, restlessness around feeding, and failure to gain weight. Weight on admission was 5.22 kg (25th percentile). Initial work-up was notable for hemoglobin 7.1 g/dL, which dropped to 6.0 g/dL during admission, prompting blood transfusion. Labs were significant for a reticulocyte count of 144 k/μL and a haptoglobin of 406 mg/dL, consistent with active bleeding and hypo-proliferation.

Abdominal ultrasound was performed after the patient exhibited continued restlessness around eating, demonstrating impressive fullness in lower left abdomen, medial to left kidney, in which a 7–8 cm collection of blood vessels passed with no clear boundaries. Computed tomography (CT) of the chest and abdomen showed strong centralized destruction at the level of the adrenal gland, left pancreas, and other retroperitoneal processes, with some free fluid and no evidence of hepatosplenomegaly. Abdominal magnetic resonance imaging (MRI) demonstrated high signal intensity on T2-weighted imaging (Fig. 1), demonstrating extensive involvement of the small bowel and mesentery. Full-body MRI excluded additional involvement in the brain, chest, abdomen, and pelvis. A small core biopsy of the mass retrieved under ultrasound visualization revealed a vascular neoplasm composed of delicate capillaries with closely packed blood vessels, lined by cytologically bland endothelium. GLUT-1, CD34, and CD31 staining were positive, and podoplanin (D2–40) was negative, suggestive of an infantile hemangioma (IH). No cutaneous hemangiomas were noted on physical exam. Ophthalmology exam was normal.

Fig. 1: Diagnostic MRI of our 3-month-old patient's intestinal hemangioma prior to treatment.
figure 1

Abdominal T2-weighted MRI demonstrating extensive hemangioma involvement of the small bowel and mesentery with enlargement of the superior mesenteric artery and vein (a: transverse view; b: anteroposterior view).

Full size image

The patient began treatment with propranolol 1.5 mg/kg/day and was gradually increased to 3.0 mg/kg/day over 2 weeks with no adverse events. After 1 month of treatment, the patient was readmitted due persistent anemia, suggesting continued bleeding. Rescue therapy was initiated with sirolimus 0.4 mg/m2 twice daily while continuing propranolol treatment and the patient was stabilized and discharged. Overall, sirolimus was well tolerated. The patient experienced episodes of mild fatigue and reduced appetite that may have been associated with treatment. During this time, propranolol was maintained at 3.0 mg/kg/day. After 2 months of both sirolimus and propranolol as an outpatient, the patient successfully reached a weight of 7.7 kg (52nd percentile) and maintained an acceptable hemoglobin level of 11.2 g/dL. Lipid profile remained within normal limits throughout treatment. Repeat abdominal ultrasound showed no evidence of excess fluid in peritoneum, almost complete regression of the intestinal wall thickening, and reduction of the abdominal mass. At 3-month follow-up, the small bowel hemangioma measured 25 × 22 × 32 mL, compared to 40 × 30 × 48 mL at baseline.

Systematic literature review

Methods

A systematic review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement (PRISMA) and was registered at the International Prospective Register of Systematic Reviews (PROSPERO, CRD42021256217).7 The last search was conducted on May 28, 2021. We used the databases PubMed, EMBASE, and Ovid MEDLINE, between the years 1982 and 2021, with the following search terms: (hemangioma OR haemangioma OR haemangiomata) AND (intra-abdominal OR GI OR gastrointestinal OR intestinal OR small bowel). Two reviewers (SG and EK) independently screened titles and abstracts to assess for inclusion in the Rayyan Software and were blinded to each other’s inclusion and exclusion decisions.8 Case reports, case series, randomized control studies, retrospective reviews, and clinical trials were included if they pertained to IIH based on typical clinical dynamics and/or histology and included data points related to patient demographics, clinical characteristics, and treatment management. Patients who presented with intestinal hemangiomas before the age of 5 years were included. We excluded non-English articles, literature reviews, systematic reviews, meta-analyses, and surveys. Cases were excluded if the location of the GI tract was not specified. Non-infantile (e.g., cavernous) or otherwise unspecified subtypes of hemangiomas were excluded.

After unblinding the final decisions, disagreements were resolved with discussion and detailed analysis of the studies in question. The same reviewers extracted data from relevant articles pertaining to demographic data, IIH, PHACE, treatment, and outcome. Extracted data were documented in Microsoft Excel, version 3.04.

Results

The literature search yielded 24 included articles (Fig. 2). Among these, 46 of 111 cases met inclusion criteria for IIHs (Table 1). Data related to the presence and absence of PHACE features, diagnosis, and cutaneous hemangiomas for the included cases of IIH can be found in Table 2. Specific PHACE features and location of cutaneous hemangiomas are further defined in Table 3. Patients were predominantly female (n = 37/45, 82%). Only 1 case (2%) did not present with GI symptoms. Of the symptomatic patients, average age of GI symptom presentation was 5.2 months (range: birth–5 years).

Fig. 2: PRISMA search strategy.
figure 2

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of literature search strategy.

Full size image
Table 1 Included studies and number of cases that met inclusion criteria for treated infantile intestinal hemangioma.
Full size table
Table 2 Data extraction for IIH with the presence/absence of PHACE features, diagnosis, and cutaneous hemangiomas.
Full size table
Table 3 Data extraction for IIH with defined PHACE-related features.
Full size table

A total of 9 cases (19.5%) were diagnosed with definite-PHACE, 5 (11%) possible-PHACE, and 32 (69.5%) no-PHACE. The varying IIH presentations, treatments, and outcomes are organized by PHACE diagnosis in Table 4. Overall, the most common GI presentations were unspecified GI bleed (n = 18/45, 40%), anemia (n = 17/45, 38%), melena (n = 10/45, 22%), FTT (n = 7/45, 15.5%), and vomiting (n = 7/45, 15.5%). The most common treatments were corticosteroids (n = 29/46, 63%), surgical resection (n = 15/46, 32.6%), propranolol (n = 13/46, 28%), vincristine (n = 7/46, 15%), and RBC transfusion (n = 7/46, 15%). No cases in this systematic review underwent treatment with sirolimus for IIH. Outcomes were available for 25 of the symptomatic cases, which were primarily bleeding arrest (n = 21/25, 84%) and infrequently, ongoing bleeding (n = 1/25, 4%) and death (n = 3/25, 12%). Clinical features of PHACE by organ system were primarily hemangioma related (n = 46/46, 100%), followed by arterial (n = 10/46, 22%), cardiovascular (n = 9/46, 19.6%), ocular (n = 8/46, 9%), endocrine (n = 3/46, 6.5%), structural brain (n = 3/46, 6.5%), midline (n = 2/46, 4%), neurological (n = 2/46, 4%), and miscellaneous abnormalities (n = 2/46, 4%).

Table 4 IIH clinical features, treatment, and outcomes by PHACE diagnosis.
Full size table

Cardiovascular involvement was most commonly seen in cases with definite PHACE (n = 7/9, 78%). Of the available data, death/ongoing bleeding were only reported in cases with possible PHACE (n = 1/2, 50%) or no PHACE (n = 3/18, 17%). The presenting symptoms of our patient was most commonly seen in cases of possible PHACE (anemia 40%, FTT, 40%) and definite PHACE (pallor 11%). Our patient’s outcome of bleeding arrest was most frequently seen in cases with definite PHACE (55.6%). Aside from hemangioma-related complications, our patient’s cardiovascular involvement was most frequently seen in cases with definite PHACE (77.8%).

Discussion

We present a case of an infant with right aortic arch, aberrant left subclavian artery, VSD, and IIH causing anemia and FTT. We hypothesize that our patient’s extreme presentation of a large symptomatic hemangioma in the context of cardiac malformations is not likely due to chance, but rather, due to the association of PHACE syndrome. However, based on the current diagnostic criteria for PHACE, our patient only meets one major (aberrant left subclavian artery) and two minor (aortic arch anomaly, VSD) diagnostic criteria for definite PHACE.7 Thus, we would like to highlight the small bowel as an extra-cutaneous site in which hemangiomas may present in the context of PHACE. By excluding intestinal hemangiomas in the current diagnostic of PHACE, it is possible that diagnoses may be delayed in other similar, yet less extreme cases. With this consideration, cardiovascular or other characteristic clinical PHACE features may guide consideration of PHACE in infants with anemia who lack cutaneous hemangiomas.

If IIH were considered a defining hemangioma in the same way that head/neck hemangiomas are considered in the diagnosis of PHACE, then one case in this review with no PHACE diagnosis would meet criteria for possible PHACE. None of the possible PHACE cases would meet criteria for definite PHACE by assuming this change. This is likely due to the fact that possible PHACE cases tended to already have cutaneous hemangiomas. However, the data in this cohort may underestimate the percentage of PHACE cases if full workups had not been performed.

Results from this review demonstrated that patients with a IIH most commonly present with anemia followed by melena, irrespective of PHACE diagnoses. Possible PHACE cases tended to present with FTT, whereas no PHACE diagnosis tended to present with vomiting. There was a higher frequency of patients with no PHACE diagnosis that underwent surgical resection as a treatment for symptomatic infantile IH. Of note, many of the included articles were published prior to the advent of the beta-blocker era. While these case series and case reports suggest positive outcomes with respect to bleeding cessation for both PHACE and non-PHACE cases using propranolol, our case suggests that sirolimus might represent an additional treatment option for non-responsive cases. Sirolimus, also known as rapamycin, is an mTOR inhibitor, has emerged as a safe and effective treatment modality for slow-flow vascular anomalies and for kaposiform emandoendothelima (KHE).9 Few reports highlight its role in the treatment of IH.10,11 Sirolimus might work by targeting the self-renewal of IH stem cells, diminishing differentiation, and inhibiting vasculogenesis, ultimately leading to regression of hemangioma vasculature.12 Our case suggests that sirolimus may play a role as an adjunct to propranolol in the treatment of IIH.13

This study has limitations. First, as with all systematic reviews, analyzing data in this format is limited by inconsistent reporting among case reports and case series in the literature. Second, the anecdotal evidence from case series and case reports lacks scientific rigor to determine a true association between IIH and PHACE. Third, the cases included in this review may underrepresent the total number cases in the literature given that cases were excluded during the review process if age, hemangioma type, or location in the GI tract were not specified.3,14,15,16,17,18,19,20

The present case is the first reported case in the literature that demonstrates sirolimus’s efficacy in treating propranolol resistant IIH. It is unknown whether sirolimus was effective on its own or due to synergistic effects with propranolol. Other theories to explain our patient’s resolution is that propranolol merely required additional time to take effect. If this is the case, sirolimus might play a role as a “bridge therapy” in actively bleeding, non-stable IIH patients. Future studies are required to validate our findings with a larger sample size.