Dataset Information

Machine-Learning-Based Approach to Differential Diagnosis in Tuberculous and Viral Meningitis.

ABSTRACT:

Background

Tuberculous meningitis (TBM) is the most severe form of tuberculosis, but differentiating between the diagnosis of TBM and viral meningitis (VM) is difficult. Thus, we have developed machine-learning modules for differentiating TBM from VM.

Material and methods

For the training data, confirmed or probable TBM and confirmed VM cases were retrospectively collected from five teaching hospitals in Korea between January 2000 - July 2018. Various machine-learning algorithms were used for training. The machine-learning algorithms were tested by the leave-one-out cross-validation. Four residents and two infectious disease specialists were tested using the summarized medical information.

Results

The training study comprised data from 60 patients with confirmed or probable TBM and 143 patients with confirmed VM. Older age, longer symptom duration before the visit, lower serum sodium, lower cerebrospinal fluid (CSF) glucose, higher CSF protein, and CSF adenosine deaminase were found in the TBM patients. Among the various machine-learning algorithms, the area under the curve (AUC) of the receiver operating characteristics of artificial neural network (ANN) with ImperativeImputer for matrix completion (0.85; 95% confidence interval 0.79 - 0.89) was found to be the highest. The AUC of the ANN model was statistically higher than those of all the residents (range 0.67 - 0.72, P <0.001) and an infectious disease specialist (AUC 0.76; P = 0.03).

Conclusion

The machine-learning techniques may play a role in differentiating between TBM and VM. Specifically, the ANN model seems to have better diagnostic performance than the non-expert clinician.

SUBMITTER: Jeong YS

PROVIDER: S-EPMC8032912 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Diabetic nephropathy, hypertension, and glomerulonephritis are the most common causes of chronic kidney diseases (CKD). Since CKD of various origins may not become apparent until kidney function is significantly impaired, a differential diagnosis and an appropriate treatment are needed at the very early stages. Conventional biomarkers may not have sufficient separation capabilities, while a full-proteomic approach may be used for these purposes. In the current study, several machine learning algorithms were examined for the differential diagnosis of CKD of three origins. The tested dataset was based on whole proteomic data obtained after the mass spectrometric analysis of plasma and urine samples of 34 CKD patients and the use of label-free quantification approach. The k-nearest-neighbors algorithm showed the possibility of separation of a healthy group from renal patients in general by proteomics data of plasma with high confidence (97.8%). This algorithm has also be proven to be the best of the three tested for distinguishing the groups of patients with diabetic nephropathy and glomerulonephritis according to proteomics data of plasma (96.3% of correct decisions). The group of hypertensive nephropathy could not be reliably separated according to plasma data, whereas analysis of entire proteomics data of urine did not allow differentiating the three diseases. Nevertheless, the group of hypertensive nephropathy was reliably separated from all other renal patients using the k-nearest-neighbors classifier "one against all" with 100% of accuracy by urine proteome data. The tested algorithms show good abilities to differentiate the various groups across proteomic data sets, which may help to avoid invasive intervention for the verification of the glomerulonephritis subtypes, as well as to differentiate hypertensive and diabetic nephropathy in the early stages based not on individual biomarkers, but on the whole proteomic composition of urine and blood.