Management of Complex Pulmonary Vein Stenosis at Altitude Combining Comprehensive Percutaneous Interventional Treatment with Sirolimus, Pulmonary Hypertension Medications and Intraluminal Imaging with Optical Coherence Tomography

Pulmonary vein stenosis (PVS) is a growing problem for the pediatric congenital heart population. Sirolimus has previously been shown to improve survival and slow down the progression of in-stent stenosis in patients with PVS. We evaluated patients before and after initiation of sirolimus to evaluate its effects on re-intervention and vessel patency utilizing Optical Coherence Tomography (OCT). We performed a retrospective study, reviewing the charts of patients with PVS, who had been prescribed sirolimus between October 2020 and December 2021. OCT was performed in the pulmonary vein of interest as per our published protocol. Angiographic and OCT imaging was retrospectively reviewed. Statistical analysis was performed using Chi square and Wilcoxon signed-rank test to compare pre-and post-sirolimus data. Ten patients had been started and followed on sirolimus. Median age at sirolimus initiation was 25 months with median weight of 10.6 kg and average follow-up of 1 year. Median total catheterizations were 7 for patients prior to starting sirolimus and 2 after starting treatment (p = 0.014). Comparing pre- and post-sirolimus, patients were catheterized every 3 months vs every 11 months (p = 0.011), median procedure time was 203 min vs 145 min (p = 0.036) and fluoroscopy time, 80 min vs 57.2 min (p = 0.036). 23 veins had severe in-stent tissue ingrowth prior to SST (luminal diameter < 30% of stent diameter). Post-sirolimus, 23 pulmonary veins had moderate to severe in-stent tissue ingrowth that responded to non-compliant balloon inflation only with stent luminal improvement of > 75%. Our study suggests that the addition of sirolimus in patients with moderate-severe PVS helps to decrease disease progression with decrease frequency of interventions. Reaching therapeutic levels for sirolimus is critical and medication interactions and side-effects need careful consideration. OCT continues to be important for evaluation and treatment guidance in this patient population.


Introduction
Pulmonary vein stenosis (PVS) is a growing problem for pediatric patients with congenital heart disease. Approximately 0.4% of congenital heart disease diagnoses involves PVS, however, specific centers have noted higher rates [1,2]. PVS represents a growing area of research and innovation for pediatric interventional cardiology and has proven to be a difficult disease to treat in the pediatric population both surgically and percutaneously with a high rate of recurrence [3][4][5]. Histopathological studies 1 3 in PVS have revealed that severe intimal hyperplasia and the development of hyper-cellular vessel walls lead to venous obstruction [6]. This understanding has catalyzed the development of various medical treatments to combat these cellular level changes. Vinblastine and methotrexate were first trialed [7] but had significant adverse side effects. Some centers have demonstrated a survival benefit when imatinib and bevacizumab are combined to help treat PVS [8].
More recently there has been an interest in the use of systemic sirolimus therapy (SST). Animal studies of the effects of sirolimus, a mammalian target of rapamycin (mTOR) inhibitor, on the vascular cell proliferation seen in PVS have been encouraging [9][10][11][12]. Patel et al. recently published their single center retrospective study demonstrating a survival benefit in patients treated with SST [11] and Callahan et al. demonstrated slower progression of in-stent stenosis in patients with SST [13]. These positive clinical and animal studies have inspired more institutions to start SST in patients with severe pulmonary vein stenosis. Our group has been using SST in high-risk patients with moderate to severe recurrent pulmonary vein stenosis.
Given the fact that there is a significant prevalence of pulmonary hypertension (PH) in our patient population, due to being a large referral center for the disease based in a city that is at an altitude of 5280ft, many of our patients with pulmonary vein stenosis are also aggressively treated with PH medications. In our institution, severe PH is comprehensively treated with multiple agents used in the treatment algorithm. The regimen typically includes an endothelin receptor antagonist (ERA), bosentan or ambrisentan, and a phosphodiesterase 5 (PDE5) inhibitor, sildenafil or tadalafil [14]. Sirolimus is metabolized hepatically by cytochrome P450 3A4 [15]. Bosentan is known to induce cytochrome P450 3A4 activity [16] indicating higher doses of sirolimus may be required to achieve therapeutic serum drug concentrations.
As an additional part of our centers' protocol, patients undergoing pulmonary vein intervention in the cardiac catheterization laboratory have Optical Coherence Tomography (OCT) assessment of any diseased veins with lumen or stent lumen < 5 mm. OCT allows us to identify abnormal luminal wall, degree of stent apposition, intraluminal clot and severity of in-stent tissue ingrowth [17].
In this study we evaluated patients before and after the initiation of SST as an adjunct to standard interventional therapy to assess the effects of the medication on intervention outcomes, recurrent pulmonary vein stenosis and vessel patency utilizing OCT. As secondary objective we determined to describe the medication and dosing adjustments required when SST is used in conjunction with standard PH medications.

Methods
This was a single-center retrospective study approved by the Colorado Multiple Institutional Review Board. Patients with pulmonary vein stenosis undergoing cardiac catheterization at Children's Hospital Colorado and started on sirolimus between October 2020 and December 2021 were included in the study. A retrospective chart review was performed. OCT was performed using the Dragonfly Optis Imaging catheter (Abbott Vascular, Santa Clara, CA, USA) in the pulmonary vein of interest as per our center protocol. Images from OCT were retrospectively analyzed quantitatively and qualitatively.
After referred for catheterization due to pulmonary vein stenosis, each patient would be seen for a pre-catheterization visit at our institution or as a clinic patient if the cardiologist was part of our institutional group. If percutaneous intervention was performed, the patient was then followed with an echocardiogram the morning following catheterization and then in clinic by their primary cardiologist in 2-4 weeks. These visits were tracked by our groups advanced practice provider (APP). If the patient only had angioplasty of a stenotic vein they were followed with serial echocardiography, and a plan made for re-intervention only if there was a change on imaging or clinical indications. Patients that had a stent placed that was less than 7 mm in diameter were scheduled for a follow-up catheterization in two to six months, and patients with stents larger than 7 mm in diameter were scheduled for follow up catheterization at 6 months to 1 year. Any patient that had severe stenosis of a pulmonary vein, regardless of stenting or angioplasty were scheduled for a repeat catheterization in 2-3 months. Follow-up catheterizations were adjusted per clinical and echocardiographic changes if necessary. Patients did not routinely have lung perfusion scans as per our standard follow up care. Cross sectional imaging was performed when deemed necessary due to disease progression but not routinely after initial diagnosis Patients who underwent angioplasty were put on aspirin only. Patients who had stent implantation were placed on aspirin and clopidogrel and patients with severe stenosis were put on aspirin and enoxaparin or a Direct-Acting Oral Anticoagulant (DOAC) when history of severe stent thrombosis had been reported (Fig. 1).
We adapted the categorical grading system used in our center for CTA vascular stenosis, applied for degree of pulmonary vein stenosis by angiography and OCT with severity ranging from 1 to 7; with 1 representing no stenosis to 7 representing diffuse long segment atresia (Table 1). SST was started if there was worsening distal PVS in interval cardiac catheterizations, persistent severe intrastent intimal growth or, if the assessment included OCT, there was persistent severe pulmonary vein stenosis with distal pulmonary vein intimal wall thickening of > 0.4 mm. A standardized algorithm for patients being treated with SST was created with a multi-disciplinary approach that incorporated surgical, interventional, pharmacy and pulmonary hypertension specialists in its making (Fig. 2).
Prior to starting SST, patients had baseline labs drawn which included CBC, renal panel, lipids and LFTs (Fig. 2). Our pulmonary vein team pharmacist was involved with dosing recommendations at the initiation of SST. Due to the interactions between sirolimus and bosentan a decision path was constructed (Fig. 2). If the patient was not on bosentan at baseline they were started at a daily sirolimus dose of 1 mg/m 2 . If a patient was on bosentan there was first a discussion with the PH team about switching the patient to ambrisentan which does not interact with sirolimus. If no switch was made and patient remained on bosentan, then the patient was given a loading dose of sirolimus of 2 mg/ m 2 . The patient was then started the following day on a daily dose of 1 mg/m 2 . The goal sirolimus serum concentration for all patients was 6-10 ng/mL. Levels were checked at 4 days after initiation. Following initial level if the sirolimus dose was changed, a follow-up concentration was checked 1 week later. If initial level did not require dose adjustment, a follow-up concentration was checked 2 weeks later. Subsequent dose adjustments and follow-up intervals were determined by APP and pharmacist collaboration. Sirolimus dosing could be adjusted to twice daily in patients unable to achieve therapeutic concentrations despite incremental dose increases.
Long term follow-up for all patients on SST included repeat labs at 1 month post-initiation of medication and then repeat labs at 6-12 months. Patients were discontinued from SST if they had an allergic reaction, recurrent infections, elevated cholesterol or any other side effects that the team felt were a result of starting the SST. Intraluminal pulmonary vein assessment was made when pulmonary vein or stent was < 5 mm. Recorded images were retrospectively reviewed, intimal layer of the pulmonary vein was identified and measured. Intimal growth in the stents was also measured and luminal diameter was also obtained. Statistical analysis was performed using SPSS 28 (IBM software group, Chicago, Illinois). Categorical data were summarized by frequency and percentage. Continuous data were shown as median and range. Chi square and Wilcoxon signed-rank test were used to compare pre-and-post-sirolimus data. For all our analyses, a two-sided p-value < 0.05 was considered significant.

Results
Ten patients had been started and followed on SST. The median age at initiation of SST was 25 months and the median weight was 10.6 kg. Average follow up since initiation of SST was 1 year. The cohort was 50% female. Four of the 10 patients had prior surgical repair and 9 of the patients had stents placed in pulmonary veins prior to starting SST. Prior to starting SST one patient was on sildenafil, two patients on sildenafil and bosentan, three patients on sildenafil and ambrisentan. Other demographic characteristics are demonstrated in Table 2. Patients required an average of 6 (2-10) dose modifications during their treatment period to maintain therapeutic levels with a median time of 22 days for patients to reach therapeutic levels. (Table 3).
Procedural characteristics were analyzed for each patient and compared to corresponding data from before starting SST. The median total number of catheterizations per patient was 7 prior to starting SST and 2 after starting treatment (p = 0.014). Prior to treatment patients were being catheterized an average of every 3 months while after treatment the catheterizations were average every 11 months (p = 0.011). The overall median procedure time decreased from 203 to 145 min (p = 0.036) Fig. 2 The PVS group's Sirolimus management plan from initiation to follow-up  (Table 4).) Forty pulmonary veins total had interventions performed in our patient group during the period of the chart review. The median number of veins intervened upon was 2 veins pre-and 2 veins post-SST treatment. Prior to starting SST 60% of the veins required intervention at catheterization. After starting SST only 40% required intervention (p = 0.04). There was also an improvement in severity of the stenosis. Twenty-three veins had severe in-stent tissue ingrowth prior to SST (luminal diameter < 30% of stent diameter) which underwent cutting balloon angioplasty followed by noncompliant balloon angioplasty with moderate improvement (50-75% increase) of the luminal diameter. After SST these 23 pulmonary veins had moderate to severe in-stent tissue ingrowth that responded to non-compliant balloon inflation only, without the need to use cutting balloons with luminal improvement of > 75%. Ten individual veins were considered atretic and were recanalized before commencing SST. Five were still atretic following SST. There was no progression to atresia on any of the affected pulmonary veins on patients after being treated with SST. Two pulmonary veins that were atretic and unable to be recanalized prior to starting treatment remained completely occluded on SST. No patient needed re-stenting of pulmonary veins during this study period.

Optical Coherence Tomography (OCT) Before and After SST
Seven patients from this cohort had pre and post assessment of the same pulmonary vein using OCT. Native vessel characteristics and/or stent ingrowth was able to be obtained in all cases. All vessels in this group were stented. (Fig. 3  and 4).
Six patients had adequate visualization of the native pulmonary vein past the stent towards the lung; five of these   before, during and after SST. The diameter was measure at the same site for all measurements patients had improved pulmonary vein wall thickness after SST with a thickness ≤ 0.2 mm which is comparable to normal pulmonary veins based on our experience [17]. Two had OCT of the same vein immediately after discontinuation of SST with evidence of increase thickness of the vessel wall to > 0.3 mm and then repeat evaluation 3 months later with demonstration of vessel wall > 0.3 mm and disease progression on angiography with one patient having a nearly atretic vein. (Fig. 5). Eight pulmonary veins with stents were evaluated in this population with OCT and demonstrated and average improvement of stent lumen diameter of 40% after SST when compared to prior SST initiation. Two veins had stents placed inside another stent with poor apposition with each other due to ingrowth tissue in between them. This was unchanged despite high pressure angioplasty pre SST. On both patients, the stents were completely opposed to each other 3 months after starting SST and confirmed by OCT imaging.

Discussion
Starting in August 2019, Children's Hospital Colorado established a pulmonary vein stenosis program to comprehensively monitor and follow our patient population. This was done to help standardize how this patient population was treated, monitored and followed up. Given our large geographic catchment zone, many patients are not locally followed, instead managed by cardiologists from different groups and geographical locations. Establishing a PVS protocol allowed for our group's pharmacist and APP to communicate more easily with the referring cardiologist and allow smoother follow up management and adherence Fig. 5 Progression of PVS in a patient with a LLPV stent by angiography and OCT while on and off Sirolimus. Panel A demonstrate angiography of a stented LLPV. Panel B demonstrates OCT images of the native pulmonary vein just distal to the stent. Panel C demonstrate OCT images of the stent tissue ingrowth. Each column represents the angiographic and OCT images during the same cardiac catheterization. These images demonstrate the significant change in intimal ingrowth of the native vessel while on sirolimus vs off sirolimus and the angiographic appearance of the same vessel. There is angiographic evidence of stenosis with intimal thickness of 0.3 mm and almost complete occlusion with intimal thickness > 0.5 mm. This vessel's intimal thickness returned to < 0.2 mm when sirolimus was restarted to treatment recommendations. We adopted a multidisciplinary approach to help monitor and follow all patients with PVS. We identified a key APP who, as part of the catheterization team and the PVS group facilitated regular follow up and scheduling of catheterizations providing a single point of contact for the referring cardiologist to navigate the referral process, the assessment and procedures at our institution and then the post catheterization planning. We also identified a key pharmacist on the PVS group who helped to manage and follow sirolimus drug levels and medication interactions. They also collaborated to manage changes in peri-procedural antiplatelet and anticoagulation strategies.
In this retrospective study of the use of SST in patients with moderate to severe PVS and their evaluation using OCT, we found that SST decreased the progression of pulmonary vein disease with resulting decrease in frequency of catheterizations, duration of cases and duration of fluoroscopy. We also noted intravascular imaging evidence of pulmonary vein remodeling determined by intimal vessel wall thickness measurements on OCT. We found objective evidence with a combination of angiography and OCT that in these patients, there is decreased severity of in-stent stenosis, decrease frequency of vessel occlusion and progress of the disease towards the distal pulmonary venous tree. Interestingly, changes in the tissue ingrowth characteristic which responded to balloon angioplasty more readily suggest that it was softer and more responsive without the need for cutting balloons. Our conclusion is that these findings are directly related to the anti-proliferative properties added by SST to our already aggressive catheterization strategy.
The use of SST has increased in recent years in pulmonary vein stenosis. The Emory group found a survival benefit comparing patients managed with SST [18] to those being managed without. It has also been established that intervening initially at a younger age and intervening more frequently provides a survival benefit [19,20]. Our study did not look at survival benefit but did look at frequency of interventions and intravascular changes. We maintain our patients on longer courses of sirolimus that those previously reported, keeping patients on SST for up to and beyond 1 year if well tolerated without evidence of significant adverse events.
Our practice has evolved along with our indications to start sirolimus since 2020, influenced by the addition of OCT as a regular imaging modality. Interpreting OCT for pulmonary veins in pediatric patients has been previously published by our group [17]. We started the use of sirolimus in patients that required frequent cardiac catheterizations with evidence of severe progression of the disease in between interventions. Slowly we identified the intimal changes that can predict fast progression in individual veins. In particular, recently when we noted intimal thickening to ≥ 0.3 mm and prominent collateral vasculature, we started SST. We have yet to determine any safety concerns for patients staying on sirolimus for longer than 6 months nor the impact of discontinuing the drug in patients in whom we started this therapy at an early age or disease stage.
One thing that may be unique in our PVS group is the frequent concomitant use of PH medications. Many of our patients live at altitude and have PH diagnoses which are compounded by the presence of PVS. Sixty percent of the patients in our cohort were on sildenafil. Five patients had been on dual therapy with bosentan and three of those patients were converted to ambrisentan due to it having fewer reported interactions with sirolimus. We found no adverse events related to PH in the setting of concurrent treatment of PVS. Most of our patients were able to reach sirolimus therapeutic levels quickly, with the shortest time being 3 days. Our data did include 2 patients who took significantly longer to reach therapeutic levels. The first patient was on bosentan when starting SST and was unable to achieve therapeutic sirolimus levels despite dose increases. This patient was subsequently switched to ambrisentan and quickly reached therapeutic sirolimus levels. The second patient took significantly longer to reach therapeutic levels thought to in part to be due to poor compliance. Due to the demonstrated interaction between sirolimus and bosentan, we have preferentially used ambrisentan instead of bosentan when on SST.
Overall, SST was well tolerated in this patient cohort in keeping with a published report where 2 patients had pauses in SST therapy related to side-effects which resolved and did not recur after recommencement [18]. Minor complications in our cohort included 1 patient with stomatitis that resolved with a 5 day course of antiseptic mouthwash. Two patients had constipation that improved with oral laxatives and 1 patient had a community acquired pneumonia that resulted in a 2 day admission and oral antibiotics. All the patients recovered after their treatments and continued SST. To date, 3 patients have had their SST therapy discontinued. Two patients stopped treatment at 6 months, both stopped due to recurrent respiratory infections related to a known immunodeficiency diagnosis. One was discontinued after stable PVS with no progression after 1 year on sirolimus. With proper management and surveillance, SST has been shown to be well tolerated.

Limitations
Our study is limited in that it was a retrospective study at a single center with a small patient cohort size. Given the rarity of PVS and that not all patients have moderate to severe stenosis, very few patients have been treated with SST in any institution. These patients are also the ones that are catheterized more frequently because they have worse stenosis.
Patients that are older and larger are able to have larger balloon dilations and larger stents placed. These larger stents have been shown to have less instances of recurrent in-stent stenosis and our group likely has an element of this mixed into the cohort. Some of the patients had surgical intervention and others did not, which also confounds results. The follow-up time for patients after starting SST is short. A multi-center RCT with longer follow up and more patient stenosis variability is needed to truly help determine the benefit of SST as an adjunct to aggressive catheter based pulmonary vein interventions.

Conclusion
Our study demonstrates that the addition of SST in patients with moderate to severe pulmonary vein stenosis may help to decrease PVS progression. The likely mechanism is a decrease in abnormal intimal remodeling, resulting in decreased frequency of required interventions and improved results after interventions. Reaching the therapeutic window is relatively fast when interactions with other medications are considered, especially bosentan. Additionally, OCT continues to prove a useful tool in this patient population evaluation and treatment guidance and has demonstrated particular benefit when assessing the effects of SST on the vascular wall.
Author Contributions JEZ conceptualized the idea of this investigation, created figures, wrote and reviewed the main manuscript. MJS collected data and wrote the main manuscript. EM data collection. DI evaluated pulmonary hypertension perspective of manuscript PS and MG contributed in creation of protocols described in manuscript. GJM manuscript editing and figures design. All authors reviewed manuscript.
Funding Abbott Vascular.