P725: battle of the giants – comparison of ipss-m and ipss-r in patients with missing molecular data except tp53 mutation status from the düsseldorf mds registry

Background: The IPSS-M was developed to improve prediction of median overall survival (OS) in MDS patients by additionally incorporating molecular data. A known TP53 mutation status, in addition to clinical and cytogenetic data, is the minimum requirement to assign patients to six different risk groups. Incorporation of such an extended molecular panel in the current clinical setting is still scarce, making a precise categorization to one of these groups difficult in the majority of patients. Aims: The aim of our analyses was to assess the applicability of the IPSS-M in cases of missing molecular analyses except TP53-status and to compare it to a cohort of patients with complete molecular genetics using data of the Düsseldorf MDS Registry. Furthermore, a comparison to the IPSS-R and the actual median survival of patients should help to determine which prediction tool is the most useful. Methods: We included 497 patients from our registry whose IPSS-M-requested molecular data were incomplete apart from TP53 mutation status as well as 254 patients with complete molecular genetics and executed the IPSS-M and the IPSS-R. In a second step, we compared the predicted median OS of both tools to the patients’ actual median OS. Finally, the accuracy of prediction regarding the range between the best and worst IPSS-M score of each patient was calculated and compared within the two cohorts. Results: In only 254 of 1648 patients with any known molecular data within our MDS Registry, all IPSS-M-requested molecular data were known. These patients form the group of patients with ‘complete molecular data’. Additionally, another 497 of those 1648 patients had at least a known TP53 mutation status and were included into our analyses comprising the cohort of patients with ‘incomplete molecular data’. In 265 of these 497 patients (53.3%), a minimum of one additional IPSS-M-requested mutation was known which simultaneously meant that in 232 patients (46.7%) within the group of ‘incomplete molecular data’ the TP53 mutation status was the only molecular information available. Median overall survival time of the groups was not different (62 vs 69 months, p = 0.69). Median OS in the six risk groups of the IPSS-M ranged from 192 months in the very-low risk to 11 months in the very high-risk group regarding the cohort of patients with complete molecular data. In contrast, according to the group of patients with incomplete molecular data, median OS ranged from 117 months in the low-risk group (the very-low risk group even had a shorter median OS) to 11 months in the very high-risk group. Comparison of the predicted median OS by IPSS-M and IPSS-R to the actual median OS is shown in table 1. The median delta between the best and worst IPSS-M score in the cohort with complete molecular data ranged from 0.66 to 0.67 while it ranged from 5.16 to 5.97 in the group of patients with incomplete molecular data. Summary/Conclusion: Regarding patients with complete molecular data, the IPSS-M provides the most accurate prediction of median OS, with the highest precision in higher risk groups. In patients with incomplete molecular data, the prediction of median OS became less accurate, in concordance with larger ranges between the best and worst IPSS-M score, thereby showing that missing data lead to imprecision. According to our cohort, the IPSS-R provides accurate prediction of median OS and remains the most powerful tool due to its broad applicability, even in patients with incomplete molecular data.Keywords: TP53, Myelodysplastic syndrome, Molecular markers