Development of a Risk Prediction Nomogram Model of Pyrotinib-induced Severe Diarrhea

Background: To identify the factors influencing pyrotinib-induced severe diarrhea and to establish a risk prediction nomogram model.

Methods: The clinical data of 226 patients received pyrotinib from two medical institutions from January 2019 to December 2023 were analysed retrospectively. A training set was made up of 167 patients from Hainan Cancer Hospital, and the external validation set was made up of 59 patients from Hainan West Central Hospital. Univariate and multivariate logistic regression analysis were used to identify independent factors influencing pyrotinib-induced severe diarrhea, and a risk prediction nomogram model was constructed, which was verified on patients in the external validation set.

Results: History of adverse reactions (ADRs), initial dose of pyrotinib, combination with capecitabine, thrombocytopenia, aspartate transaminase (AST), and use of probiotics or other drugs that regulate the gut microbiota were identified as independent influencing factors for pyrotinib-induced severe diarrhea (all P < 0.05). Based on these, a risk prediction nomogram model of pyrotinib-induced severe diarrhea was established. The area under the receiver operating characteristic curve was 0.794 and 0.863 in the training set and the external validation set, respectively. The calibration curve of the prediction model displayed good consistency both the two sets, which indicated that the model could have favourable predictive ability.

Conclusion: The risk prediction nomogram model of pyrotinib-induced severe diarrhea constructed in this study may identify high risk populations earlier so that clinicians can make appropriate decisions in time.