Introduction

Hereditary hemorrhagic telangiectasia (HHT) is an autosomal dominant genetic disorder involving vascular malformations in different organ systems. Vascular malformations may present in the lungs, brain, liver and sometimes other visceral organs (i.e., uterus, spleen) as well as involvement of mucocutaneous membranes including the nasal mucosa, lips and oropharyngeal mucosa, throughout the gastrointestinal tract and elsewhere. ^{Reference Shovlin, Guttmacher and Buscarini1} The diagnosis of HHT is based on the Curaçao Diagnostic Criteria, which includes (i) recurrent epistaxis, (ii) mucocutaneous telangiectasia, (iii) visceral arteriovenous malformations (AVMs) (iv) and a family history of HHT in a first-degree relative. ^{Reference Shovlin, Guttmacher and Buscarini1} Those with three or four of these criteria have definite HHT; two criteria are consistent with possible HHT, and one criterion is unlikely – based on the Curacao criteria. HHT has an estimated world prevalence of 1 in 5000. ^{Reference Shovlin2} However, in Alberta, the reported prevalence is higher at 1 in 3800. ^{Reference Chowdhury, Chandrarathne and Masilamani3} Populations with Dutch genetic backgrounds appear to have an even higher prevalence of HHT at 1 in 1331, and Alberta is home to many Dutch settlements and rural communities with families affected by HHT. ^{Reference Westermann, Rosina, de Vries and de Coteau4,Reference Bletz5} With an estimated 1200 individuals affected and more with suspected HHT, the demand for BVM screening will continue to grow as these patients interact with the healthcare system. ^{Reference Westermann, Rosina, de Vries and de Coteau4,6}

The prevalence of brain vascular malformations (BVMs) in patients with HHT is reported to be 10%. ^{Reference Brinjikji, Iyer, Wood and Lanzino7} BVMs can result in acute hemorrhage with hemorrhagic strokes and/or chronic bleeding. Embolic strokes, transient ischemic attacks and brain abscesses are seen in patients who have coexistent pulmonary arteriovenous malformations (PAVMs) with right-to-left shunts; blood clots (strokes) or infected material bypass the normal capillary bed architecture allowing for this. ^{Reference Brinjikji, Iyer, Wood and Lanzino7,Reference Akers, Ball and Clancy8}

Three subtypes of BVMs exist, including brain arteriovenous fistula, nidus-type AVM and capillary vascular malformation. ^{Reference Matsubara, Mandzia, ter Brugge, Willinsky and Faughnan9,Reference Krings, Ozanne, Chng, Alvarez, Rodesch and Lasjaunias10} Capillary-type vascular malformations previously known as “micro-AVMs” do not demonstrate an enlarged artery or vein and simply represent dilated capillary beds as evidenced by a contrast blush on cerebral angiography and likely have a lower risk of hemorrhage. ^{Reference Krings, Kim and Power11} The rate of hemorrhage in HHT-associated BVMs versus sporadic BVMs in the population remains a topic of debate; however, retrospective data have demonstrated a similar bleeding risk of around 2.0% per patient-year. ^{Reference Easey, Wallace, Hughes, Jackson, Taylor and Shovlin12} Capillary-type BVMs are the most prevalent vascular malformation among patients with HHT ^{Reference Krings, Kim and Power11} and do not typically represent a bleeding risk. ^{Reference Matsubara, Mandzia, ter Brugge, Willinsky and Faughnan9,Reference Krings, Ozanne, Chng, Alvarez, Rodesch and Lasjaunias10} However, it is important to note that although capillary-type malformations with their low bleeding risk represent the most frequent BVM in patients with HHT, patients with HHT do still demonstrate nidus- and fistulous-type BVMs, and as a result, the average bleeding rate for the HHT population may be falsely low when compared to sporadic BVMs overall. ^{Reference Krings, Kim and Power11} More importantly, patients with true nidus- or fistulous-type BVMs may present with a catastrophic bleed or seizure as a revealing event. ^{Reference Söderman, Andersson, Karlsson, Wallace and Edner13,Reference Choi and Mohr14}

Screening results in four possible outcomes for the patient, and these are benefit, no benefit, harm and no harm. ^{Reference Raffle and Gray15} The assumption is that a patient benefits from the identification and discussion around the treatment of a BVM. However, whether or not identifying and treating a lesion results in a better outcome for the patient than not detecting a lesion and subsequently no treatment will vary considerably based on the individual and the clinical presentation. ^{Reference Eker, Boccardi and Sure16} Consideration must be given to the potential harm in screening that may be related to treatment, psychological factors, potential lifestyle modification and financial implications related to insurance. ^{Reference Eker, Boccardi and Sure16} There is no conservative management for BVMs, and therefore, treatment options are invasive including surgical resection, arterial embolization and/or stereotactic radiosurgery. ^{Reference Eker, Boccardi and Sure16} Due to the lack of randomized controlled trials showing definitive management for BVMs and the varied nature of the individualized clinical context, the decision to screen should be patient-centered. The North American standard in keeping with the second international HHT guidelines recommends the use of screening MRI in pediatric and adult patients with possible or definite HHT to identify BVMs that may be amenable to treatment. ^{Reference Faughnan, Mager, Hetts, Palda and Ratjen17}

It is recommended in international guidelines that individuals with possible or definite HHT obtain a screen for BVMs. In Alberta, the absence of a neuroimaging standardized operating protocol (SOP) to screen for BVMs resulted in a myriad of screening modalities and protocols, some of which were nondiagnostic and ultimately led to duplicate testing. The reference standard for the identification of BVMs is invasive digital subtraction angiography (DSA); however, any invasive procedure is associated with risk, and therefore, DSA is not used for screening. ^{Reference Tranvinh, Heit, Hacein-Bey, Provenzale and Wintermark18} CT has excellent spatial resolution, and with the addition of IV contrast, CT angiograms can detect BVMs; however, it is still not as sensitive or specific as MRI with the addition of susceptibility-weighted imaging (SWI). ^{Reference Tranvinh, Heit, Hacein-Bey, Provenzale and Wintermark18,Reference Jagadeesan, Delgado Almandoz, Benzinger and Moran19} CT also includes radiation to the patient and the added risk of a contrast reaction, neither of which is relevant to screening using non-contrast MRI with SWI. MRI is widely available and, although costly and time-consuming, is the most sensitive and specific noninvasive diagnostic imaging modality available for the detection of BVMs.

Alberta is a single healthcare jurisdiction, which makes it opportune for screening. In Alberta, neuroimaging may be performed internally at Alberta Health Services (AHS) facilities or externally at private for-profit facilities. AHS is the provincial health authority that represents a province-wide integrated health system that is further subdivided into 5 health regions operating 106 acute care facilities (Figure 1). ²⁰ Any one of the 106 acute care facilities operated by AHS or an external privately run facility may be the first interaction with diagnostic imaging for a patient with suspected or confirmed HHT; therefore, a standardized protocol is critically important to ensure the accuracy of the screening exam. ²⁰ The larger urban hospitals in the province have subspecialty-trained neuroradiologists who are responsible for the appropriate protocolling of screening MRIs for patients with HHT; however, many acute care facilities and private for-profit centers within the province are small and remote without dedicated neuroradiologists or MRI scanners. Subsequently, patients referred to smaller acute care or private facilities would often be incorrectly protocolled and undergo nondiagnostic exams. These exams varied considerably depending on the study location. Inappropriate screening studies performed included CT heads, portable MRI and MRI protocolled without the use of highly sensitive and specific sequences for arteriovenous shunting such as SWI. ^{Reference Jagadeesan, Delgado Almandoz, Benzinger and Moran19,Reference Hodel, Leclerc and Kalsoum21} Inappropriate screening studies resulted in repeat imaging, increasing the workload to radiology departments and subsequently the cost to the healthcare system. Edmonton is an accredited HHT center of excellence, and given the increased prevalence of patients with HHT in the province and the potential for negative health outcomes, it was imperative to develop an SOP that would be both accessible and sensitive for the detection of BVMs. ²²

Figure 1. Alberta Health Services zones. Population statistics retrieved from https://www.albertahealthservices.ca/assets/about/publications/ahs-ar-2021 on December 4, 2022.