EndoSysScore (ESS)
Synthetic data-augmented ML model · 30-day mortality after surgery for infective endocarditis
AUC
0.785
95% CI 0.74–0.83
DeLong wins
6/6
All p<0.05 vs comparators
O/E ratio
1.05
Good calibration
Validation n
930
109 deaths · 11.7% mortality
Head-to-head — held-out internal validation cohort (n=930)
ScoreAUC (95% CI)ΔAUC vs ESSDeLong zp-valueResult
EndoSysScore (ESS)0.785 (0.74–0.83)reference
EuroSCORE II0.744+0.0412.230.026WIN
EndoSCORE0.749+0.0352.170.030WIN
RISK-E0.722+0.0632.860.004WIN
AEPEI (5-var)0.724+0.0602.680.007WIN
APORTEI0.721+0.0633.090.002WIN
STS-IE0.695+0.0903.600.0003WIN
High-risk phenotype analysis — observed vs ESS predicted
PhenotypeDefinitionnObservedESS predictedGap ESSBest comparator gap
A · Rare high-lethalityMultivalve + S.aureus + CKD≥3/dialysis/PAPs>501154.5%46.2%−8.3 pp−12.5 pp
B · Rare high-lethalityAge≥80 + NVE + (COPD or LVEF<50%)1553.3%46.3%−7.1 pp−22.2 pp
C · Rare high-lethalityAbscess/periannular + isolated mitral + age≥651050.0%50.0%0.0 pp−4.1 pp
D · Rare high-lethalityPVE + S.aureus + mitral + aortic1050.0%46.5%−3.5 pp0.0 pp
E · Rare high-lethalityNVE + age≥80 + LVEF<50%1154.5%52.8%−1.7 pp−21.9 pp
F · Rare high-lethalityAbscess + isolated mitral + age≥651050.0%50.0%0.0 pp−4.1 pp
Pooled (distinct · E⊆B, F=C counted once)4450.0%46.0%−4.0 ppMAD ESS 3.4 pp vs RISK-E 15.7 pp
Demographics
Age (years)
Female
Cardiac status
LVEF (%)
PAPs >50 mmHg
Prosthetic valve
Heart failure
IE characteristics
Aortic valve
Mitral valve
Isolated mitral
Multivalve
Abscess
Periannular ext.
Comorbidities
COPD
CKD stage ≥3
On dialysis
Microbiology
S. aureus
Culture-negative
Fungal
Pre-op critical state
Cardiogenic shock
Pre-op intubation
Pre-op IABP
Active endocarditis
Worked examples (Fig. 4)
Phenotype detection
A
Multivalve
S.aureus+CKD
B
Age≥80 NVE
COPD/EF<50
C
Abscess iso-mit
age≥65
D
PVE S.aureus
AO+mitral
E
NVE age≥80
LVEF<50%
F
Abscess iso-mit
age≥65
30-day mortality risk
For research use only · not a substitute for clinical judgment
ESS v3 · GIROC multicenter registry · n=5,215 · doi:10.5281/zenodo.20327974

EndoSysScore (ESS) is a stacking ensemble developed on the GIROC Italian multicentre registry (2000–2023; development cohort n=4,285; 23 centres; 30-day mortality 10.6%). Rare high-lethality phenotypes were identified in the derivation cohort by systematic multivariable pattern analysis and used to guide conditional synthetic data augmentation (CTGAN + TIMA clinician-supervised plausibility filtering). The final model combines three base learners (XGBoost shallow, XGBoost moderate, logistic regression) through a LightGBM meta-learner, with isotonic and beta calibration applied on real development data only.

Model performance was evaluated in a held-out internal validation cohort of 930 patients (109 deaths, 11.7% mortality). Development (n=4,285) and validation (n=930) cohorts were drawn from the same 23-centre GIROC registry via a stratified random patient-level split (≈80/20 within each centre); all 23 centres contributed to both cohorts, with no patient overlap. No validation-cohort patient was used for model training, synthetic augmentation, calibration, or phenotype optimisation. Fully external validation (leave-one-centre-out and prospective inter-registry evaluation) is the planned next step.

Phenotype transparency note: during phenotype definition, candidate phenotypes were required to demonstrate non-zero representation in the validation cohort to confirm applicability to rare IE presentations. This feasibility check verified representation only; phenotype definitions were fixed before any outcome data from the validation cohort were examined. Full methodology, reproducible code, and de-identified data are openly deposited on Zenodo and GitHub.

References
Gelsomino S, Moula A, Actis Dato GM, Parise G, Parise O, Lorusso R, Di Mauro M, on behalf of the GIROC investigators. EndoSysScore (ESS): Synthetic Data-Augmented Risk Prediction for 30-Day Mortality after Cardiac Surgery for Infective Endocarditis. Zenodo, 2026. doi:10.5281/zenodo.20327974
GIROC Study Group. Italian multicenter registry for outcomes in cardiac surgery for infective endocarditis. 23 centres, 2000–2023.
🗄
Zenodo deposit
Complete open deposit: de-identified development and validation datasets, trained model artefacts, multi-seed results, ablation study, and manuscript draft.
doi:10.5281/zenodo.20327974 →
GitHub repository
Full reproducible pipeline: TIMA data cleaning, CTGAN synthetic augmentation, model training, calibration, and validation. MIT licence.
sandro2462/sys-score →
Authors & attribution ✓ MIT licence
Sandro Gelsomino · Amalia Moula · Guglielmo Mario Actis Dato · Gianmarco Parise · Orlando Parise · Roberto Lorusso · Michele Di Mauro
Maastricht University Medical Centre · CARIM — Cardiovascular Research Institute Maastricht · The Netherlands
The EndoSysScore methodology, clinical phenotype framework, and research concept were conceived and developed by Sandro Gelsomino MD PhD in collaboration with the GIROC Study Group. Software engineering and statistical pipeline implementation were carried out with AI-assisted development. The scientific content, clinical decisions, and intellectual authorship are entirely those of the named authors.
Cite as
Gelsomino S, Moula A, Actis Dato GM, Parise G, Parise O, Lorusso R, Di Mauro M, on behalf of the GIROC investigators. EndoSysScore (ESS): Synthetic Data-Augmented Risk Prediction for 30-Day Mortality after Cardiac Surgery for Infective Endocarditis. Zenodo, 2026. doi:10.5281/zenodo.20327974