Revisiting the PLASMIC score and the future of ADAMTS13 prediction Authors: Mehmet Baysal, Ahmet Emre Eşkazan Published: 2026 TO THE EDITOR: Over the past 15 years, the approach to immune-mediated thrombotic thrombocytopenic purpura has evolved dramatically.1,2 The Bentley score was an early attempt to predict ADAMTS13 deficiency.3 As clinical experience grew, the need for a more robust prediction tool became apparent.4 Ultimately, these early studies led to the development of the PLASMIC score, which demonstrated superior diagnostic accuracy.5 Although prospective comparisons of these systems remain unlikely, retrospective single-center studies demonstrate comparable diagnostic utility, despite smaller sample sizes than the cohorts establishing the PLASMIC score.6 The recent multicenter analysis by Singh et al provides a deeper reevaluation of the PLASMIC score itself, highlighting a necessary shift toward more nuanced, data-driven models.7 Strikingly, their findings reveal a significant diagnostic gap in the older population, with sensitivity dropping to 62% in patients aged ≥60 years.7 Missing cases in such a vulnerable demographic are troublesome for any prediction system, suggesting that the comorbidities common in older adults easily confound traditional PLASMIC parameters.8, 9, 10 Furthermore, we must acknowledge that these systems still lack robust validation in other populations, specifically pediatric immune-mediated thrombotic thrombocytopenic purpura and pregnancy-associated thrombotic microangiopathies. Evaluating individual parameters, the authors provide a clear directive for score simplification. Mean corpuscular volume appears to be the weakest link, lacking independent predictive value, whereas the high odds ratio of lactate dehydrogenase robustly justifies its inclusion in a revised “PLASIC+L” model.7 Although the PLASMIC score remains a reliable framework, medical history is a cycle of continuous refinement. The evidence presented underscores the need to evolve toward the PLASIC+L variant, particularly for older patients.7 As artificial intelligence–driven diagnostic tools emerge, we must question the longevity of static scoring systems. Classical hematology is likely approaching its own “Moneyball” moment,11 replacing traditional intuition with rigorous, machine-learning analytics to transform how we predict and treat thrombotic microangiopathies. However, we must also recognize that artificial intelligence models are only as good as the data we feed them; without massive, high-quality, and structurally diverse training data sets, they will invariably fail. Conflict-of-interest disclosure: The authors declare no competing financial interests.