The Impact of Liver Iron Concentration Measurements By Machine-Learned R2 Magnetic Resonance Imaging on Clinical Decision Making in a Cohort of Episodically Transfused Lebanese Patients with Sickle Cell Disease

Many patients with sickle cell disease (SCD) require episodic transfusion putting them at risk of iron overload and iron related morbidities. A previous study has indicated that total transfusion volumes may not be a reliable measure of total body iron burden in SCD patients and that transfusion rate may be a better indicator [1]. The objectives of this study were to use a new MRI method (FerriSmart®) based on a machine-learned algorithm to prospectively assess body iron burden by measuring liver iron concentration (LIC) in a cohort of episodically transfused SCD patients in Lebanon, compare results with transfused volumes and transfusion rates, and determine the impact of the MRI findings on modifying the clinical management of the studied patients.

Seventy two episodically transfused patients with sickle cell disease (61 SS; 7 SB0; 4 SB+) were prospectively recruited from a cohort of 450 patients. Inclusion criteria were patients 8 years and above having received ≥10 total life transfusions or having 2 steady-state serum ferritin more than 1000 ng/mL and with no other condition causing iron overload. The median age (range) of the patients was 24 (8 to 54) years with 39 being female and 54/72 patients being treated with hydroxyurea (HU) and 10/72 splenectomised. Lifetime transfusion volumes received by the patients ranged from 3 L to 120 L (median 9.45 L) with the time since last transfusion ranging from zero to 18 years (median 1 year). Four out of 72 patients had received iron chelation therapy. Median (range) serum ferritin was 513 (56 to 10,374) ng/mL.

LIC was prospectively measured for 71/72 patients using the MRI method between February and July 2024. LIC values ranged from 0.6 to 26.2 (median 1.4) mg Fe/g dry tissue with the proportions of patients with LIC <3, 3-7, and >7 mg Fe/g dry tissue being 52/71 (73%), 10/71 (14%), and 9/71 (13%) respectively. There was a highly significant correlation between serum ferritin and LIC with Pearsons R² = 0.87 (95% CI 0.80 - 0.92). There was an extremely weak but statistically significant correlation between total transfused volume and LIC in the 67 chelation naïve patients (R² = 0.18; 95%CI 0.04 - 0.36). Of 50 chelation naïve patients who had received 15 L or less of transfusion volume, 8/50 (16%) had LIC > 3 mg Fe/g dry tissue. Of 17 chelation naïve patients who had received >15 L of transfusion volume, 7/17 (41%) had LIC ≤3 mg Fe/g dry tissue indicating that transfusion volume is a very unreliable predictor of chelation requirements in these patients.

There was no significant correlation between LIC and transfusion rate in the 67 chelation naïve patients (both Pearson and Spearman correlations tested), contrary to the findings of a previous study [1], suggesting that transfusion rate is not a reliable indicator of iron chelation needs in these patients.

A weak but significant negative rank order correlation (r = -0.34; 95% CI -0.54 to -0.10) was found between the ratio of LIC to transfusion volume and the time since last transfusion for the 67 chelation naïve patients suggesting that SCD patients can lose iron from the body during non-transfused periods, possibly via urine, as suggested previously [1, 2].

In light of the LIC measurements, iron chelation therapy was initiated for 19 patients, discontinued for 2 patients, and resumed for 1 patient.

In conclusion, neither transfusion volume nor transfusion rate is a reliable indicator of chelation needs in SCD patients. Loss of iron during extended periods since the last transfusion may confound these measures. A new cost-effective machine-learned algorithm for measuring LIC from MR images has enabled more confident decision making on chelation needs for this cohort of SCD patients.

[1] Inati, et al. (2009) European Journal of Haematology, 83: 565.

[2] Porter and Garbowski (2013) Hematology, 2013: 447.

Inati:Novartis: Consultancy, Research Funding; GBT/Pfizer: Consultancy, Research Funding; Roche: Consultancy, Research Funding; Novo Nordisk: Consultancy, Research Funding; Bausch Health: Research Funding; Vifor: Consultancy, Research Funding; Agios: Research Funding; Pharmacosmos: Research Funding. St Pierre:Resonance Health Ltd: Current Employment.