Application of the Major Predictive Scoring Strategies Assessing KIR-based NK Alloreactivity to Donor/Recipient Couples

The use of haploidentical donors for aHSCT has greatly increased this past decade leading to a major paradigm shift: while finding 10/10 HLA-matched donors represented the prior difficulty for decades, the current problem is about finding the best haploidentical donor among several potential ones. The prediction of NK cells alloreactivity toward leukemic cells provides promising perspectives, although the underlying biological processes remain unclear. To date, many prediction models based on KIR and MHC genotyping have been designed and used across studies, which contribute to blur clinical conclusions.

The investigators hypothesized that the diversity of models used to predict NK alloreactivity in aHSCT could partly be responsible for the current literature discrepancies. The main objective of this work consisted of applying the major KIR-based prediction models in D/R couples undergoing aHSCT in different fashions - with MSD and haploidentical donors - to describe their heterogeneity and potential correlations. As clinical data were available for these two cohorts, the investigators described correlations that could be assessed between the scoring strategies and the clinical outcomes.

As suspected, it was highlighted that the different scoring strategies greatly impact the assessment of alloreactivity within D/R couples. As an example, two broadly used scoring strategies - educational models and missing-ligand models - show clear opposite predictions. Moreover, some scoring strategies seem to be better adapted to genoidentical or haploidentical cohorts, whereas others are robust across the different cohorts. Concerning the clinical-biological correlations, it was highlighted that (i) each scoring strategy is differentially associated to the different outcomes (ii) the different scoring strategies predict one particular outcome with different efficacy (iii) the D/R compatibility greatly impacts the pertinence of the scoring strategy.

This work therefore contributes to unravel the KIR-based alloreactivity prediction of NK cells in aHSCT. This would help to overcome the current literature discrepancies in this field as in making new hypotheses to better understand and predict NK alloreactivity to further develop its use in medical practice.