Establishment and verification of a surgical prognostic model for cervical spinal cord injury without radiological abnormality

Machine Learning-based Prediction of Prolonged Intensive Care Unit Stay for Critical Patients with Spinal Cord Injury

STUDY DESIGN

A retrospective cohort study.

OBJECTIVE

The objective of the study was to develop machine-learning (ML) classifiers for predicting prolonged intensive care unit (ICU)-stay and prolonged hospital-stay for critical patients with spinal cord injury (SCI).

SUMMARY OF BACKGROUND DATA

Critical patients with SCI in ICU need more attention. SCI patients with prolonged stay in ICU usually occupy vast medical resources and hinder the rehabilitation deployment.

METHODS

A total of 1599 critical patients with SCI were included in the study and labeled with prolonged stay or normal stay. All data were extracted from the eICU Collaborative Research Database and the Medical Information Mart for Intensive Care III-IV Database. The extracted data were randomly divided into training, validation and testing (6:2:2) subdatasets. A total of 91 initial ML classifiers were developed, and the top three initial classifiers with the best performance were further stacked into an ensemble classifier with logistic regressor. The area under the curve (AUC) was the main indicator to assess the prediction performance of all classifiers. The primary predicting outcome was prolonged ICU-stay, while the secondary predicting outcome was prolonged hospital-stay.

RESULTS

In predicting prolonged ICU-stay, the AUC of the ensemble classifier was 0.864 ± 0.021 in the three-time five-fold cross-validation and 0.802 in the independent testing. In predicting prolonged hospital-stay, the AUC of the ensemble classifier was 0.815 ± 0.037 in the three-time five-fold cross-validation and 0.799 in the independent testing. Decision curve analysis showed the merits of the ensemble classifiers, as the curves of the top three initial classifiers varied a lot in either predicting prolonged ICU-stay or discriminating prolonged hospital-stay.

CONCLUSION

The ensemble classifiers successfully predict the prolonged ICU-stay and the prolonged hospital-stay, which showed a high potential of assisting physicians in managing SCI patients in ICU and make full use of medical resources.Level of Evidence: 3.

Prognostic Value of mRNAsi/Corrected mRNAsi Calculated by the One-Class Logistic Regression Machine-Learning Algorithm in Glioblastoma Within Multiple Datasets

Glioblastoma (GBM) is the most common glial tumour and has extremely poor prognosis. GBM stem-like cells drive tumorigenesis and progression. However, a systematic assessment of stemness indices and their association with immunological properties in GBM is lacking. We collected 874 GBM samples from four GBM cohorts (TCGA, CGGA, GSE4412, and GSE13041) and calculated the mRNA expression-based stemness indices (mRNAsi) and corrected mRNAsi (c_mRNAsi, mRNAsi/tumour purity) with OCLR algorithm. Then, mRNAsi/c_mRNAsi were used to quantify the stemness traits that correlated significantly with prognosis. Additionally, confounding variables were identified. We used discrimination, calibration, and model improvement capability to evaluate the established models. Finally, the CIBERSORTx algorithm and ssGSEA were implemented for functional analysis. Patients with high mRNAsi/c_mRNAsi GBM showed better prognosis among the four GBM cohorts. After identifying the confounding variables, c_mRNAsi still maintained its prognostic value. Model evaluation showed that the c_mRNAsi-based model performed well. Patients with high c_mRNAsi exhibited significant immune suppression. Moreover, c_mRNAsi correlated negatively with infiltrating levels of immune-related cells. In addition, ssGSEA revealed that immune-related pathways were generally activated in patients with high c_mRNAsi. We comprehensively evaluated GBM stemness indices based on large cohorts and established a c_mRNAsi-based classifier for prognosis prediction.

Changes in neurological and pathological outcomes in a modified rat spinal cord injury model with closed canal

Most animal spinal cord injury models involve a laminectomy, such as the weight drop model or the transection model. However, in clinical practice, many patients undergo spinal cord injury while maintaining a relatively complete spinal canal. Thus, open spinal cord injury models often do not simulate real injuries, and few previous studies have investigated whether having a closed spinal canal after a primary spinal cord injury may influence secondary processes. Therefore, we aimed to assess the differences in neurological dysfunction and pathological changes between rat spinal cord injury models with closed and open spinal canals. Sprague-Dawley rats were randomly divided into three groups. In the sham group, the tunnel was expanded only, without inserting a screw into the spinal canal. In the spinal cord injury with open canal group, a screw was inserted into the spinal canal to cause spinal cord injury for 5 minutes, and then the screw was pulled out, leaving a hole in the vertebral plate. In the spinal cord injury with closed canal group, after inserting a screw into the spinal canal for 5 minutes, the screw was pulled out by approximately 1.5 mm and the flat end of the screw remained in the hole in the vertebral plate so that the spinal canal remained closed; this group was the modified model, which used a screw both to compress the spinal cord and to seal the spinal canal. At 7 days post-operation, the Basso-Beattie-Bresnahan scale was used to measure changes in neurological outcomes. Hematoxylin-eosin staining was used to assess histopathology. To evaluate the degree of local secondary hypoxia, immunohistochemical staining and western blot assays were applied to detect the expression of hypoxia-inducible factor 1α (HIF-1α) and vascular endothelial growth factor (VEGF). Compared with the spinal cord injury with open canal group, in the closed canal group the Basso-Beattie-Bresnahan scores were lower, cell morphology was more irregular, the percentage of morphologically normal neurons was lower, the percentages of HIF-1α- and VEGF-immunoreactive cells were higher, and HIF-1α and VEGF protein expression was also higher. In conclusion, we successfully established a rat spinal cord injury model with closed canal. This model could result in more serious neurological dysfunction and histopathological changes than in open canal models. All experimental procedures were approved by the Institutional Animal Care Committee of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, China (approval No. HKDL201810) on January 30, 2018.