A novel scoring system to predict survival in cirrhotic patients undergoing isolated lung transplantation: The PENS-CEPT score.
Eric J HyznyErnest G ChanChadi A HageVikrant RachakondaPablo G SanchezShahid M MalikPublished in: Clinical transplantation (2023)
Cirrhosis is usually regarded as a contraindication to isolated lung transplantation (ILT). We sought to determine which patients with cirrhosis could safely undergo ILT. Based on a retrospective analysis of patients with cirrhosis who underwent ILT at our center between 2007 and 2020, we developed an exclusionary algorithm (PENS-CEPT: Pittsburgh ExclusioN Score in Cirrhotics Evaluated for Pulmonary Transplant) to help determine which patients can undergo ILT with minimal incurred risk from their underlying liver disease. The score utilizes a combination of readily available clinical data and the presence (or absence) of spontaneous portosystemic shunts on preoperative cross-sectional imaging. Sixteen patients underwent ILT with a diagnosis of cirrhosis: nine with cystic fibrosis. On univariate analysis, only our model was able to predict 1 year survival. Of the nine patients that would have been approved using our model, there was only one short term death. Of the seven patients that would have been rejected by the model, all but one died within the first year with six dying of complications from liver failure. We are proposing a simple score utilizing routine clinical parameters and pre-operative imaging to determine the safety of ILT in cirrhotic patients. Further studies are required to validate this scoring system with the goal of safely increasing the opportunity for cirrhotic patients who would otherwise be rejected for ILT.
Keyphrases
- end stage renal disease
- chronic kidney disease
- ejection fraction
- newly diagnosed
- patients undergoing
- cross sectional
- peritoneal dialysis
- high resolution
- hepatitis b virus
- machine learning
- pulmonary hypertension
- risk factors
- liver failure
- patient reported outcomes
- extracorporeal membrane oxygenation
- deep learning
- electronic health record
- data analysis
- case control