Predictive Value of Plasma sPECAM-1, t-PA IC, and sCLEC-2 for Early Severe Bleeding in Patients with Acute Promyelocytic Leukemia.

OBJECTIVE: This study evaluated the predictive value of plasma sPECAM-1, t-PAIC, and sCLEC-2 (individually and in combination) for early severe bleeding events in patients with acute promyelocytic leukemia (APL). METHODS: A retrospective review of 319 APL patients was conducted. Patients were divided into Occurrence and Non‑occurrence groups based on severe bleeding events. Plasma markers were measured. Logistic regression, ROC, calibration, decision curve analyses, and internal validation were performed. RESULTS: Plasma sPECAM‑1, sCLEC‑2, and t‑PAIC were significantly elevated in the Occurrence group. Elevated levels of these markers, along with reduced levels of FIB were independently associated with early severe bleeding events in patients with APL. AUCs for sPECAM-1, sCLEC-2, and t-PAIC were 0.759, 0.855, and 0.820, respectively; the three‑marker combination achieved an AUC of 0.929, outperforming any single marker and the traditional model (DIC score, FIB, D-D, and WBC) (AUC = 0.804). Adding the new biomarkers to the traditional model further increased the AUC to 0.959 (P < 0.001). Internal validation yielded a corrected AUC of 0.916, with good calibration (P = 0.069). The combined model provided net clinical benefit across threshold probabilities of 2-98%. CONCLUSION: Plasma sPECAM‑1, t‑PAIC, and sCLEC‑2 are elevated in APL patients with early severe bleeding. Their combined detection demonstrates high predictive efficacy for early severe bleeding in APL. However, the generalizability of this model requires further validation through large‑scale, multicenter, prospective studies.