External Validation of the Saved Score for Venous Thromboembolism Risk Stratification in Patients with Multiple Myeloma Receiving Immunomodulatory Drugs

Background: High thrombogenicity is a well-established phenomenon in multiple myeloma (MM) and can lead to high morbidity and mortality. The risk of venous thromboembolism (VTE) is particularly elevated in the subset of MM patients treated with immunomodulatory drug (IMiD) containing regimens. The National Comprehensive Cancer Network (NCCN) and International Myeloma Working Group (IMWG) have implemented VTE risk assessment models, and recommend prophylactic anticoagulation based on individual patient, disease, or treatment-related factors specifically for MM patients receiving IMiD. With the recent incorporation of the SAVED (PMID 31319391) scoring system in the NCCN guidelines for VTE risk stratification, we sought to externally validate this risk assessment model and evaluate its predictive ability in our MM patient cohort.

Methods: We retrospectively evaluated 501 adult patients with either newly diagnosed or relapsed/refractory MM who were newly initiated on an IMiD-based regimen between 2010 and 2019. Patients already on therapeutic or prophylactic anticoagulation were excluded from the study (aspirin was allowed). The SAVED score was calculated at IMiD initiation according to the following five parameters: surgery within last 90 days (+2), Asian race (-3), prior VTE history (+3), age ≥80 years (+1), dexamethasone use (+2 for high dose [>160 mg/30-day], or +1 for standard dose [120-160 mg/30-day]). Patients with SAVED scores of <2 and ≥2 were considered of low and high risk, respectively, for developing VTE after IMiD initiation. Time to VTE was estimated by Kaplan-Meier failure method. Logistic regression model was used to test components of the SAVED score with 1-year VTE status. Model discrimination was assessed via c statistic by calculating the area under the receiver operating characteristic curve (AUC) with bootstrapped 95% confidence interval (CI).

Results: The median age of our cohort at the IMiD initiation was 64 years (range 34-89); 264 (53%) patients were male and 378 (75%) were white. IgG and IgA heavy chain subtypes were present in 280 (56%), and 105 (21%) patients respectively, whereas 110 (22%) patients had light chain disease only. The International Staging System (ISS) was I, II, III, and unknown for 161 (32%), 149 (30%), 136 (27%), and 55 (11%) of patients, respectively. Most patients (94%) received lenalidomide, whereas 14 (3%) and 15 (3%) patients were treated with thalidomide and pomalidomide, respectively. The distribution of SAVED covariates is shown in Table 1. The cumulative incidence of VTE at 12-month after IMiD initiation for the entire cohort was 16% (95% CI: 13-20%). Among the 84 patients who developed new VTE after IMiD initiation, 60 (71%) had lower extremity deep venous thrombosis (DVT), 10 (12%) had pulmonary embolism, 9 (11%) had both DVT and PE, and 5 (6%) had upper extremity DVT. The SAVED score stratified 112 patients (22%) as high-risk and 389 patients (78%) as low-risk (similar proportion as the original derivation study). Cumulative incidence of VTE at 6 and 12 months was 31% and 41% in the high-risk group, versus 6% and 9% in the low-risk group, respectively (Figure 1). Each point of SAVED score increase was associated with significantly higher chance of VTE within 1-year post IMiD initiation with excellent discrimination (odds ratio = 2.10 [95% CI 1.71-2.56], c statistic (AUC) = 0.76 [95% CI 0.70-0.81]).

Conclusion: Our results successfully externally validated the SAVED model to predict IMiD-associated VTE in a real-life cohort of patients with MM. Patients with SAVED score of ≥2 had significantly higher incidence for VTE. This indicates that more aggressive pharmacologic thromboprophylaxis is needed in this specific patient population. The SAVED score outperformed several other similar VTE prediction models that were also validated in the current cohort (c statistic=0.68 for IMPEDE [PMID: 33997961] and 0.62 for PRISM [PMID: 35772005]). In contrast to other risk models that focused on all patients with MM, the SAVED score was derived and now validated specifically for patients starting IMiD as this it is the only cancer population with a strong guideline recommendation for ambulatory thromboprophylaxis. Future prospective validation would be useful to determine the predictive nature of SAVED while evaluating safety and efficacy of thromboprophylaxis with direct oral anticoagulant in high-risk patients.

Anwer:Janssen: Consultancy; Allogene Therapeutics: Research Funding; BMS: Consultancy, Research Funding, Speakers Bureau. Valent:Alexion, AstraZeneca Rare Disease: Research Funding.