**6. Follow-Up**

Follow-up is important for patients with HHT to monitor for reperfusion of treated PAVM and growth of existing or previously microscopic PAVMs. As previously mentioned in the screening protocol discussion, patients with negative initial screening or suspected microscopic PAVMs (grade 1 shunts on initial screening TTCE) should have repeat TTCE screening every 5 years.

In patients with initial negative screening CT, or a CT showing a very small PAVM not indicated for treatment, there is conflicting evidence on what constitutes an appropriate surveillance interval. Previous guidelines recommended 3–5 year CT follow-up. However, a 2019 study by Curnes et al. has reported a lack of growth over time for small untreated PAVMs in adults [111]. For each patient in the study, they compared 2 CT exams with the longest interval between them (mean 8.4 years, range 3.1–14.1 years) to assess growth, analyzing a total of 88 PAVMs in 21 patients. They found that untreated PAVMs grew slowly, if at all, and that any demonstrated growth was minimal and clinically inconsequential [111]. Similar findings were reported by Ryan et al. in a 2017 study investigating the natural history of small and microscopic untreated PAVMs in adults [112]. The findings from Curnes et al. and Ryan et al. challenge the guideline of 3–5 year CT follow-up for small untreated PAVMs, suggesting that this interval could be safety extended up to 5–10 years [111,112].

Another study assessed the diagnostic yield of rescreening adult HHT patients with initial negative screening CT [113]. They found that in 172 HHT patients, there is a low rate of newly detected PAVMs, approximately 0.7%/patient-year, most of which are small and not amenable to treatment. No treatable PAVMs were identified at the 5-year mark, and only 1 treatable PAVM was identified after 6 years, further supporting the notion that a longer screening interval of 5–10 years may be warranted [113]. In addition, a survey of providers at HHT Centers of Excellence worldwide has shown that around one fifth of providers already choose to obtain follow-up imaging in 10 years for patients who demonstrate PAVM stability on 2 CT scans in a 5-year period [21]. This is the regimen we follow at our institution.

The existing guidelines state that for patients who have undergone recent embolization of their PAVMs, follow-up CT should be performed within 6–12 months of treatment, then repeated every 3 years [18]. We feel that this interval can be increased in many patients, thereby reducing radiation exposure and reducing costs. At our institution, we recommend initial follow-up with CT within 6 months of embolization, followed by a repeat CT in 3–5 years based on likelihood of persistence, favoring shorter follow-up times for larger and more complex PAVMs. We use the common definition of successful PAVM treatment, that being more than 70% shrinkage of the draining vein or sac [57,61,62]. Either non-contrast or contrast-enhanced CT can be used.

One study has investigated whether graded TTCE can be used post-embolotherapy as a follow-up tool to predict the need for repeat treatment [114]. In 32 patients with prior PAVM embolization, graded TTCE was performed and the results were compared to their most recent chest CT. Two patients had PAVMs requiring repeat embolotherapy (feeding artery diameter ≥3 mm) due to untreated PAVM growth or treated PAVM persistence. All patients with negative TTCE had no visible PAVMs on CT. Both patients who did require repeat embolotherapy had grade 3 shunts on TTCE. The study suggests that post-embolotherapy TTCE can be used to predict the need for repeat embolotherapy and presence of treatable PAVM on CT. These results are promising and may provide an avenue for post-embolotherapy patients to avoid repeated radiation exposure.

Lastly, some authors sugges<sup>t</sup> the use of time-resolved magnetic resonance imaging (MRI) for follow-up, particularly in patients treated with coils as there may be less induced metallic artifact with this modality [3,16,115,116]. A recent pilot study compared the use of ferumoxytol-enhanced MR angiography (MRA) to CT angiography (CTA) for PAVM detection [81]. The two modalities were comparable in detection rate for PAVMs > 2 mm, and ferumoxytol-enhanced MRA was able to detect several persistent PAVMs which were missed by CTA due to beam-hardening artifact from embolization coils. The data are preliminary, but both time-resolved MR and ferumoxytol-enhanced MR may prove to be feasible alternatives to CT for PAVM imaging, especially in the post-embolization setting, while avoiding the use of radiation and nephrotoxic contrast [117].
